174 | P a g e

DETERMINANTS OF BSE SENSEX: A

FACTOR ANALYSIS APPROACH




Arijit Ghosh
Assistant Professor, St. Xaviers College,
Kolkata India
writetoarijit@gmail.com

Samrat Roy
Assistant Professor, St. Xaviers College,
Kolkata India
samratsxc@gmail.com

Gautam Bandyopadhyay
Associate Professor, NIT,
Durgapur, India
math_gb@yahoo.co.in

Kripasindhu Choudhuri
Emeritus Fellow, Department Of Mathematics, Jadavpur University
India
chaudhuriks@gmail.com



175 | P a g e

Abstract.

The following paper tries to examine the primary factors responsible for affecting Bombay Stock
Exchange (BSE) in India. Further this paper attempts to investigate the relative influence of the
factors affecting BSE and thereby categorizing them. We considered the following determinants
Oil prices, Gold price, Cash Reserve Ratio, Food price inflation, Call money rate, Dollar price,
F D I, Foreign Portfolio Investment and Foreign Exchange Reserve (Forex). With the help of
multiple regression model and applying Factor analysis the primary factors are traced out.

In our research, we have estimated the relationship between BSE Sensex and some other
important economical factors and got some interesting results related to this. We have taken into
consideration the Multicollinearity problem among different independent variables and
attempted to eliminate it. We have used statistical methods to do the analysis based on monthly
basis database of different economical factors. Finally we got some relationships of those factors
with BSE Sensex. In our analysis we found that dollar price along with Factor 1i.e; External
Reserve and Factor score 2i.e; Inflation inertia are significantly affecting BSE Sensex. In
the context of Indian economy the appreciation of Dollar will bring in more foreign exchange
reserve which will act as stimulant to foster growth and in this process the injection of capital
flows will affect Sensex.
The fluctuations in Sensex due to Oil and CRR are significant. Any rise in Oil price will create
inflation inertia which will generate stochasticity in Sensex. The External reserves taken together
will act as resource generating Factor in attracting Foreign Capital inflows, which will make
Sensex more sensitive.


Key words:
Sensex, Economical determinants, Multiple Regression analysis, Multicollinearity, Factor Analysis.


1.0 INTRODUCTION:
Stock market is one of the major economic reflectors. Indian economy is currently emerging as a global
super power. Due to low labor cost and skillful manpower sectors like textile, garments, manufacturing,
banking and insurance has made a significant contribution to foster the growth potentials of the economy.
The Structural Adjustment Program adopted in 1991 had focused on stabilization and structural reforms
in this respect the changeover from inward orientation to outward strategies has generated euphoria in the
stock market. Hence the opening up of the economy has been successful in spreading its tentacles over
the economy. There are several factors which are directly or indirectly related to stock prices. Here while
observing stock market behavior we have taken into consideration Bombay Stock Exchange sensitive
index (BSE) in our database.

176 | P a g e

Due to the asymmetric information community, most of the people who are involved in share Market are
not aware of the information for Trading. The choice under uncertainty has led to the issue of moral
hazards in the capital markets. This paper tries to examine the interrelationship between different
determinants affecting Bombay Stock Exchange (BSE) in India. In this paper we considered the following
determinants Oil prices, Gold price, Cash Reserve Ratio, Food price inflation, Call money rate, Dollar
price, F D I, Foreign Portfolio Investment and Foreign Exchange Reserve (Forex).

The paper is divided into two sections. In section one an attempt has been made to discuss the scope, the
importance of selecting the determinants affecting the stock market and interrelationship between them.
In section two the methodology has been applied to examine the influence of those determinants and
grouping them using Factor analysis.

2.0 SCOPE AND OBJECTIVE

It is a well known fact that Dollar price or money exchange rate has a great influence on BSE Sensex. Our
research identifies the level of influence of dollar price on BSE Sensex. The oil price of India is
dependent on International Oil market. Any developing economy like India is dependent on Oil price, so
we tried to find out if oil price influences the BSE Sensex. The strength and stability of the host countrys
currency is measured by the level and volatility of call money rates. Gold price is included in the model as
an additional variable, to examine whether gold price contain any additional significant relation with
share price movements. Since gold has an asset value it works as an important savings material. The
technology transfer in terms of innovation and inventions to the developing country like India in phase of
Globalization in terms of FDI and Foreign Portfolio Investment has an impact on the volatility in the BSE
Sensex. The macroeconomic stability of any developing economy is highly dependent on food price
inflation. Its impact on BSE Sensex can be analyzed. The potentials of the economy i s strengthened by
the foreign exchange reserve which has some impact on Stock market.




3.0 LITERATURE REVIEW:

In the past decades, many researchers attempted to use different methods in order to predict decision
regarding share markets. Here we have described some previous research works related to our analysis
sectors and tried to find out their limitations to improve the analysis process better than before.

Firstly we reviewed on those papers/ articles where relation between stock prices and call money rate
were analyzed. In the paper of Kenneth E. Homa & Dwight M. Jaffee (1971) they have used stock price
as a dependent variable and supply of money as an independent variable. As he said, the nature of the
relationship between the money supply and common stock prices can be most easily described if a share
of common stock is viewed as an asset that yields its return to the investor over time. In this study the
177 | P a g e

relationship between the money supply and the stock market is estimated using the techniques of
regression analysis.

One attempt at estimating such relationship for use in simulating monetary effects within a
macroeconomic model is described in the research of Robert H. Rasche and Harold T. Shapiro (1968). A
fuller discussion of this common stock valuation formula can be found in Burton G. Malkeil (1963),
Martin Feldstein (1980) discussed a crucial cause of the failure of share prices to rise during a decade of
substantial inflation. The analysis here indicates that this inverse relation between higher inflation and
lower share prices during the past decade was not due to chance or to other unrelated economic events.
One of the analyses by Franco Modigliani and Richard Cohn (1979) also shows that it is unnecessary to
invoke a theory of systematic error of the type.

According to Kenneth E. Homa & Dwight M. Jaffee (1971), while forecasting share prices, further
realism could be introduced. In particular, short sales of stock, the tax treatment of short term and long
term capital gains, bills perhaps other assets, all could be introduced into the simulation. Finally, more
practical use might be made of timing implications of the model if forecasts were generated on a monthly
or weekly basis.

So we have tried with some other assets and economical factors to make the forecasting more realistic
according to this papers authors suggestion. According to Martin Feldstein (1980), full understanding of
the equilibrium relation between share prices and inflation requires extending the current analysis in a
number of ways. A more explicit portfolio model could derive asset demand equations from expected
utility maximization and could recognize that some institutional holdings are really indirect ways for
individuals to hold assets in a tax-favored way. For that purpose we have included Gold price as a risk
less assets in a tax-favored way. Martin Feldstein (1980) also concluded his study saying that the
simplification that the capital stock remains constant should be replaced by a more dynamic model that
recognizes the effect of inflation on capital accumulation.

We have done our analysis using some more dynamic model like correlation analysis, regression analysis
.The lacking of Martin Feldstein (1980) analysis was, a complete analysis of the effect of inflation on
share prices requires considering a wider range of alternative investments and incorporating the
possibility that perceived risk varies with inflation which was absent here. So we tried to consider a wide
range of alternative investments on monthly basis data to make the analysis more efficient.

The Global Economic Slowdown had a recessionary impact on the financial market leading to decline in
share prices and indices. In the recovery phase the economy has adopted expansionary fiscal policy to
accelerate aggregate demand. In this respect RBI has adopted a contractionary monetary policy overcome
the crisis which led to higher interest rate on bank deposits. Then we have tried to review on those papers/
articles which have done analysis to check whether stock prices and oil price are related to each other or
not. John Mauldin (2003) reported in an article of Swiss America Trading Corporation on the relationship
between oil prices and stock prices, where he found strong evidence that changes in oil prices forecast
stock returns. This predictability is especially strong in the developed countries markets. Among his (John
Mauldin (2003)) chosen 12 of the 18 countries, changes in oil prices significantly predict future market
returns on a lagging monthly basis. Not surprisingly, a rise on oil price suggests a lower stock market and
178 | P a g e

a drop in oil price infers a rise in stock prices. The magnitude of the oil price shift is also carried over into
the magnitude of the expected increase/decrease in stock prices.

John Mauldin (2003) used such a method, while adding alpha or excess returns over buy and hold, is still
volatile as heck (that's a technical term) and is wrong over 40% of the time in most countries. It is just
that when it is right, the returns are excessive. This also means that there could be certain random
entry/trend following variables. Though the study clearly showed that oil prices and stocks, especially if
there are big moves in oil, tend to go in opposite directions. Here we saw that though John Mauldin (6)
found a relationship between share price and oil price, but it is wrong over 40% of the time in most
countries. So we have tried to find this relation using different method to see whether the significance
level could be high.

The issue of whether stock prices and exchange rates are related or not have received considerable
attention after the East Asian crises. During the crises the countries affected saw turmoil in both currency
and stock markets. If stock prices and exchange rates are related and the causation runs from exchange
rates to stock prices then crises in the stock markets can be prevented by controlling the exchange rates.
Moreover, developing countries can exploit such a link to attract/stimulate foreign portfolio investment in
their own countries. Similarly, if the causation runs from stock prices to exchange rates then authorities
can focus on domestic economic policies to stabilize the stock market. If the two markets/prices are
related then investors can use this information to predict the behavior of one market using the information
on other market.
This article of R. Smyth and M. Nandha (2003) examines the relationship between exchange rates and
stock prices in Bangladesh, India, Pakistan and Sri Lanka using daily data over a six-year period from
1995 to 2001. Both the Engle-Granger two-step and Johansen cointegration methods suggest that there is
no long-run equilibrium relationship between these two financial variables in any of the four countries.
Granger causality tests find that there is unit-directional causality running from exchange rates to stock
prices in India and Sri Lanka, but in Bangladesh and Pakistan exchange rates and stock prices are
independent. Most of the empirical literature that has examined the stock prices-exchange rate
relationship has focused on examining this relationship for the developed countries with very little
attention on the developing countries. The results of these studies are, however, inconclusive. Some
studies have found a significant positive relationship between stock prices and exchange rates (for
instance Smith (1992), Solnik (1987), and Aggarwal (1981)) while others have reported a significant
negative relationship between the two (e.g., Soenen and Hennigar (1988)). On the other hand, there are
some studies that have found very weak or no association between stock prices and exchange rates (for
instance, Franck and Young (1972), Eli Bartov and Gordon M.Bodnor (1994)).

The study by Naeem Muhammad and Abdul Rasheed (2001) uses monthly data on four South Asian
countries, including Pakistan, India, Bangladesh and Sri- Lanka, for the period January 1994 to December
2000. They employed cointegration, vector error correction modeling technique and standard Granger
causality tests to examine the long-run and short-run association between stock prices and exchange rates.
The results of this study show no short-run association between the said variables for all four countries.
There is no long-run relationship between stock prices and exchange rates for Pakistan and India as well.
However, for Bangladesh and Sri Lanka there appear to be a bi-directional causality between these two
financial variables.
179 | P a g e


There is no theoretical consensus on the relationship between stock prices and exchange rates either. For
instance, portfolio balance models of exchange rate determination postulate a negative relationship
between stock prices and exchange rates. Franck and Young (1972) was the first study that examined the
relationship between stock prices and exchange rates. They use six different exchange rates and found no
relationship between these two financial variables. Aggarwal (1981) explored the relationship between
changes in the dollar exchange rates and change in indices of stock prices. He uses monthly U.S. stock
price data and the effective exchange rate for the period 1974-1978. His results, which were based on
simple regressions, showed that stock prices and the value of the U.S. dollar is positively related and this
relationship is stronger in the short run than in the long run. Solnik (1987) examined the impact of several
variables (exchange rates, interest rates and changes in inflationary expectation) on stock prices. He uses
monthly data from nine western markets (U.S., Japan, Germany, U.K., France, Canada, Netherlands,
Switzerland, and Belgium). He found depreciation to have a positive but insignificant influence on the
U.S. stock market compared to change in inflationary expectation and interest rates. Soenen and Hanniger
(1988) employed monthly data on stock prices and effective exchange rates for the period 1980-1986.
They discover a strong negative relationship between the value of the U.S. dollar and the change in stock
prices. However, when they analyzed the above relationship for a different period, they found a statistical
significant negative impact of revaluation on stock prices. Amare and Mohsin (2000) examine the long-
run association between stock prices and exchange rates for nine Asian countries (Japan, Hong Kong,
Taiwan, Singapore, Thailand, Malaysia, Korea, Indonesia, and Philippines). They use monthly data from
January 1980 to June 1998 and employed cointegration technique. The long-run relationship between
stock prices and exchange rates was found only for Singapore and Philippines. They attributed this lack of
cointegration between the said variables to the bias created by the omission of important variables.
When interest rate variable was included in their cointegrating equation they found cointegration between
stock prices, exchange rates and interest rate for six of the nine countries.

To examine the long-run relationship between stock prices and exchange rates Naeem Muhammad and
Abdul Rasheed (2001) employ the standard technique of cointegration. In another work Tamnun E
Mursalin, Ahmed Tanvir and Dr. Md. Jahangir Alam (2006) analyzed impact of various economical and
social determinants to predict decision of investors.



4.0 RESEARCH MODEL

We fit a Multiple Regression Model to the data set and carryout the analysis to examine the impact of the
determinants affecting Sensex and at the same time computing the degree of association among the
determinants. Further Factor Analysis is carried out to categorize the determinants into groups. Eventually
the crucial factors among Oil prices, Gold price, Cash Reserve Ratio, Food price inflation, Call money
rate, Dollar price, F D I, Foreign Portfolio Investment and Foreign Exchange Reserve (Forex) are traced
out using the above methodology. We have collected each sectors three years database on monthly basis
from specific sources. There are some holidays which didnt match in both share market & bank or
exchange office. So we had to take previous working days closing price as those days share price.
180 | P a g e







5.0 BACKGROUND OF EMPIRICAL STUDIES:

Generally the efficient market hypothesis states that markets are efficient if prices fully reflect all
available factors. Our analysis was based on some strong statistical methods like correlation analysis,
regression analysis, multiple regression analysis and Factor Analysis.



5.1 Regression Analysis:

Then we have done the regression analysis to justify the relations strength or weakness like as
we have found that there is a significant relation between share price and other specific factors
but we dont know how strong the relationship are. So by this analysis we have tried to figure it
out. Regression analysis is a method of modeling the relationships among two or more variables.
It is used to predict the value of one variable given the values of the others. We first review the
basic theory on crisp multiple linear regression. Suppose we have some data (x
1j
, x
2j
, , x
mj
,
y
j
), j = 1, 2, .., n, on (m + 1) variables x
1
, x
2
, ,x
m
and Y. The values of x
i
( i = 1, 2, . m)
are known in advance and Y is a random variable. We assume that there is no uncertainty in the
data for x
i
. We cannot predict a future value of Y with certainty. So we decide to focus on the
mean of Y, E(Y). Here we assume that E(Y) is a linear function of x
i
s, say
E(Y) = A
0
+ A
1
x
1
+ + A
m
x
m
.
Our model is Y
j
= A
0
+ A
1
x
1j
+ + A
m
x
mj
+ c
j
, j = 1, 2, . n.
The basic regression equation for the mean of Y is y = A
0
+ A
1
x
1
+ + A
m
x
m
, and we wish
to estimate the values of A
i
(i = 0,1, 2, . m).
Here we need the (1 - |)100% confidence interval for A
i
(i = 0, 1, 2, . m), where | is the level
of significance. First we require the crisp point estimators of A
i
(i = 0, 1, 2, . m) and o
2
, where
o
2
is the unknown population variance. We use the matrix notation to describe the estimators and
their confidence intervals.
Let us define the vectors as
181 | P a g e

A = [A
0
, A
1
, A
m
], c = [c
1
, c
2
, c
n
], y = [y
1
, y
2
, , y
n
] and the m n matrix X as
X =
|
|
|
|
|
.
|

\
|
mn n
m
m
x x
x x
x x
... 1
... ... ... ...
... 1
... 1
1
2 12
1 11

We also define A

= ]

, .... ,

[
1 0 m
A A A as the vector of point estimate of A. Then
t
A

= (X
t
X)
-1
Xy
t
,
which gives
m
A A A

, .... ,

1 0
.

5.2 Multiple regression analysis:

We have done multiple regression analysis also to check multiple variables effect on share price changes.
Multiple regression is used to account for (predict) the variance in an interval dependent, based on linear
combinations of interval, dichotomous, or dummy independent variables. Multiple regression can
establish that a set of independent variables explains a proportion of the variance in a dependent variable
at a significant level (through a significance test of R
2
), and can establish the relative predictive
importance of the independent variables (by comparing beta weights). Power terms can be added as
independent variables to explore curvilinear effects. Cross-product terms can be added as independent
variables to explore interaction effects. One can test the significance of difference of two R
2
's to
determine if adding an independent variable to the model helps significantly. (G. David Garson (2003)).

In general, multiple regression procedures will estimate a linear equation of the form:
X b X b X b a Y n n
* ... * *
2 2 1 1
+ + + + =
In the above the regression coefficients (or B coefficients) represent the independent contributions of each
independent variable to the prediction of the dependent variable. Here
Y
= BSE sensex,
X1
= dollar
price,
X 2
= oil price,
X3
= food price inflation and so on according to our analysis.

5.3 Multicollinearity:

Multicollinearity is one of the important problems in multiple regression analysis. It is usually
regarded as a problem arising out of the violation of the assumption that explanatory variations
are linearly independent. However, the mere satisfaction of this assumption does not preclude the
possibility of an approximate linear dependence among the explanatory variables and hence the
problem of multicollinearity. In other words, generally we do not speak in terms of the presence
of the absence of multicollinearity, but in terms of its degree.
A review of the treatment of multicollinearity in literature reveals that the hypothesis about the
consequences of multicollinearity is that a high degree of multicollinearity leads to high standard
errors of the estimates. What is implied here is that had the interdependence among the
explanatory variables been low, the estimated coefficients would have been statistically more
182 | P a g e

significant. Another consequence of multicollinearity as argued currently is that the inclusion of
a strongly correlated variable will bring only a marginal increase in . That is, if Y is regressed
on X2 and if X3 is highly correlated with X2 then the inclusion of X3 as another explanatory
variable in the equation will result in only a slight increase in , if at all any. The above two
consequences of multicollinearity are considered to be valid under all conditions. In fact, these
consequences are the direct results of the major assumption that the estimated variance of the
true error term is not being affected by the degree of multicollinearity. It is our contention that
the estimated variance of the true error term need not be invariant to the degree of
multicollinearity. It is argued here that a high degree of multicollinearity can result in a lower
estimate of the variance of the true error term. In fact under certain condition, a high degree of
multicollinearity will always lead to a lower estimated value for the variance of the error term.
Then it follows that firstly, a high degree of multicollinearity need not always result in high
values of the standard errors of the estimates. Secondly, inclusion of a highly correlated variable
(correlated with the existent explanatory variables) can some-times inflate the value of . Thus,
if data exhibit some degree of multicollinearity it need not always be reflected in the standard
errors being large. In such a case we do not have any built in safeguard against any rash
interpretation of predicted value. Moreover, the high or and t values in the regression could
have resulted from the high degree of multicollinearity.

5.4 Factor analysis:

Factor analysis is a collection of methods used to examine how underlying factors or
determinants influences the responses on a number of measured variables. There are basically
two types of factor analysis: exploratory and confirmatory. Exploratory factor analysis (EFA)
attempts to discover the nature of the constructs influencing a set of responses. Confirmatory
factor analysis (CFA) tests whether a specified set of constructs is influencing responses in a
predicted way.
Factor analyses are performed by examining the pattern of correlations (or co variances) between
the observed measures. Measures that are highly correlated (either positively or negatively) are
likely influenced by the same factors, while those that are relatively uncorrelated are likely
influenced by different factors.

6.0 EMPIRICAL STUDIES AND RESULTS

6.1 Data
The sample includes data of BSE Sensex, Oil prices, Gold price, Cash Reserve Ratio, Food price
inflation, Call money rate, Dollar price, F D I, Foreign Portfolio Investment and Foreign Exchange
Reserve (Forex) on monthly basis for the period from Jan07 to Mar10. There are some holidays which
didnt match in both share market & bank or exchange office. So we had to take previous days closing
price as those days parameters. To verify accurate gold price, we had to convert international price rate
(which was in dollar) into IRS.

6.2 Empirical results and discussions
The empirical results are presented here in the different subsections.
183 | P a g e


6.2.1 Correlation analysis
Here we used Pearson correlation analysis. It assumes that the two variables are measured on at least
interval scales, and it determines the extent to which values of the two variables are "proportional" to each
other.

Table 1: Correlation analysis of different determinants in comparison to BSE Sensex.
Pearson Correlation Method Correlations
Share Price 1.000
Dollar Price -0.705820539
Oil Price 0.530762551
Food Price Inflation -0.032220835
Gold Price 0.213896051
Call Money Rate 0.199697234
CRR 0.38248643
FDI 0.115076
FPI -0.1333682


6.2.2 Regression & Factor Output Analysis: -


Variables Enter ed/Removed
b
FOREX,
FDI, CMR,
FPI, CRR,
Foreign,
Gold, OIL,
Doller
a
. Enter
Model
1
Variables
Entered
Variables
Removed Method
All requested variables entered.
a.
Dependent Variable: Sensex
b.
Model Summary
.889
a
.791 .726 1557.25103
Model
1
R R Square
Adjusted
R Square
Std. Error of
the Estimate
Predictors: (Constant), FOREX, FDI, CMR, FPI, CRR,
Foreign, Gold, OIL, Doller
a.
184 | P a g e





Coefficients
a
60401.394 12242.893 4.934 .000
7.812 4.294 .386 1.819 .079 .160 6.258
167.114 160.660 .186 1.040 .307 .226 4.431
85.579 90.240 .126 .948 .351 .409 2.448
-78.145 441.748 -.031 -.177 .861 .237 4.223
-72.427 43.770 -.532 -1.655 .109 .070 14.334
-1169.062 243.581 -1.607 -4.799 .000 .064 15.551
.107 .144 .113 .743 .464 .314 3.184
.038 .213 .029 .180 .858 .275 3.641
.061 .036 .494 1.679 .104 .083 11.994
(Constant)
Gold
FPI
CMR
CRR
OIL
Doller
FDI
Foreign
FOREX
Model
1
B Std. Error
Unstandardized
Coef f icients
Beta
Standardized
Coef f icients
t Sig. Tolerance VIF
Collinearity Statistics
Dependent Variable: Sensex
a.
Cor relation Matrix
a
1.000 .760 -.213 .188 .353 .639 .707
.760 1.000 -.374 -.130 .543 .745 .618
-.213 -.374 1.000 .724 -.627 -.179 -.358
.188 -.130 .724 1.000 -.589 -.153 -.103
.353 .543 -.627 -.589 1.000 .578 .787
.639 .745 -.179 -.153 .578 1.000 .675
.707 .618 -.358 -.103 .787 .675 1.000
Gold
FPI
CRR
OIL
Doller
Foreign
FOREX
Correlation
Gold FPI CRR OIL Doller Foreign FOREX
Determinant = .001
a.
KMO and Bartlett's Test
.620
231.726
21
.000
Kaiser-Meyer-Olkin Measure of Sampling
Adequacy.
Approx. Chi-Square
df
Sig.
Bartlett's Test of
Sphericity
185 | P a g e



Total Variance Explained
3.896 55.658 55.658 3.896 55.658 55.658 3.379 48.267 48.267
1.780 25.434 81.091 1.780 25.434 81.091 2.298 32.824 81.091
.531 7.583 88.674
.474 6.771 95.445
.174 2.492 97.937
.103 1.476 99.414
.041 .586 100.000
Component
1
2
3
4
5
6
7
Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative %
Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings
Extraction Method: Principal Component Analysis.
Component Number
7 6 5 4 3 2 1
E
i
g
e
n
v
a
l
u
e
4
3
2
1
0
Scree Plot
186 | P a g e





Component Matrix
a
.868
.854
.842
.809
.734 .543
.868
-.602 .652
FOREX
Doller
FPI
Foreign
Gold
OIL
CRR
1 2
Component
Extraction Method: Principal Component Analysis.
2 components extracted.
a.
Rotated Component Matrix
a
.907
.861
.849
.846
.946
.864
.573 -.720
Gold
FPI
FOREX
Foreign
OIL
CRR
Doller
1 2
Component
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax w ith Kaiser Normalization.
Rotation converged in 3 iterations.
a.
Variables Enter ed/Removed
b
Foreign,
Gold,
FOREX,
FPI
a
. Enter
Model
1
Variables
Entered
Variables
Removed Method
All requested variables entered.
a.
Dependent Variable: REGR f actor score
1 f or analysis 1
b.
187 | P a g e







Model Summary

Model Summary
.991
a
.982 .980 .14214108
Model
1
R R Square
Adjusted
R Square
Std. Error of
the Estimate
Predictors: (Constant), Foreign, Gold, FOREX, FPI
a.
Coefficients
a
-4.002 .163 -24.581 .000
.003 .000 .403 10.051 .000
.058 .013 .193 4.637 .000
1.11E-005 .000 .266 7.431 .000
.000 .000 .265 7.001 .000
(Constant)
Gold
FPI
FOREX
Foreign
Model
1
B Std. Error
Unstandardized
Coef f icients
Beta
Standardized
Coef f icients
t Sig.
Dependent Variable: REGR f actor score 1 f or analysis 1
a.
Variables Enter ed/Removed
b
CRR, OIL
a
. Enter
Model
1
Variables
Entered
Variables
Removed Method
All requested variables entered.
a.
Dependent Variable: REGR f actor score
2 f or analysis 1
b.
Model Summary
.981
a
.963 .961 .19834199
Model
1
R R Square
Adjusted
R Square
Std. Error of
the Estimate
Predictors: (Constant), CRR, OIL
a.
Coefficients
a
-4.362 .179 -24.433 .000
.031 .002 .674 14.453 .000
.320 .040 .376 8.054 .000
(Constant)
OIL
CRR
Model
1
B Std. Error
Unstandardized
Coef f icients
Beta
Standardized
Coef f icients
t Sig.
Dependent Variable: REGR f actor score 2 f or analysis 1
a.
188 | P a g e

Mod
el R R Square
Adjusted
R Square
Std. Error
of the
Estimate
1
.874(a) .764 .744
1505.7392
3
a Predictors: (Constant), Doller, REGR factor score 1 for analysis 1, REGR
factor score 2 for analysis 1

Coefficients(a)

Mode
l
Unstandardized
Coefficients t
Standard
ized
Coefficie
nts Sig.
95% Confidence Interval
for B
B
Std.
Error Beta
Lower
Bound
Upper
Bound B Std. Error
1 (Constant) 57423.0
40
5027.17
9
11.423 .000 47217.324 67628.756
REGR factor
score 1 for
analysis 1
2206.55
5
360.209 .742 6.126 .000 1475.292 2937.817
REGR factor
score 2 for
analysis 1
-848.169 374.423 -.285 -2.265 .030 -1608.288 -88.051
Doller -953.222 112.123 -1.310 -8.502 .000 -1180.844 -725.600
a Dependent Variable: Sensex



The impact of the determinants affecting Sensex has been captured statistically by the multiple regression
models. The regression output reveals that 79% of the variation in Sensex can be explained by the
determinants taken under study. This is confirmed by the F statistic at 5% level. The remaining 21% is
left unexplained. The model has a better goodness of fit. However when we look at the individual
significance of the determinants it provides a different picture. All the independent variables except
Dollar price are statistically insignificant at 5% level. So in spite of high R
2
, the OLS estimates may have
large variance and covariance. This reflects the association among the independent variables. The
presence of Multicollinearity problem within the data set is evident from the above result. This calls for
diagnostic tests to affirm the presence of the problem. There by we go for correlation matrix.


The matrix provides an introspective view regarding the inter relationship among the variables. That
means the cell other than the shaded region explains strong correlations. Intuitively when the values of r
are less than -0.7 or more than 0.7 then the independent variables are highly correlated to each other. So
there is no way of disentangling the separate influence of the variables. The determinant value close to 0
189 | P a g e

confirms high correlation between independent variables. Further, there are certain diagnostic tests that
affirm multicollinearity. The KMO Bartlett Test Statistic is equal to 0.620 which exceeds 0.5 and hence,
the Null hypothesis of spherical matrix is rejected. This conclusion is further supported by Bartlett Test of
sphericity where the
2
statistic is significant at 5% level. Hence non spherical correlation matrix confirms
the presence of multicollinearity.
From the correlation matrix it is evident that CMR and FDI do not have any association with other
independent variables. Hence they do not have multicollinearity problem. Further analysis is carried out
except this two variables.

So to reduce the severity of the problem and to eliminate it we have to go for data reduction with the help
of factor analysis. There by using the extraction method via Principal component analysis the
communalities are computed. According to total variance explained matrix the first two components
explains 81% of the change in the independent variables. Further the graphical representation in terms of
scree plot reflects that there exists closeness among different eigen values of different component
numbers. Since there exists a combination among independent variables within this data set as confirmed
by the previous tests, data reduction is necessary. In this regard the original data set is converted to groups
on the basis of principal components. This is done by extraction method under factor analysis.

Our component matrix table shows Dollar, Forex, FPI, Foreign Portfolio Investment and CMR will
clearly be included in Component 1 and Gold, Oil and CRR in component 2. Since Gold and CRR belong
to both the components with different Loadings we have done rotation.

After rotation, loading of factors corresponding to different variables changes their corresponding values.
The corresponding matrix represents two components which includes the variables on the basis of
Varimax rotation method. According to this matrix, Gold, Forex, FPI, Foreign Portfolio Investment and
Dollar belongs to Component1 and Oil, CRR and Dollar belongs to component 2. Dollar is included in
both components. Further regression analysis is conducted between regression factor score1 with respect
to corresponding variables i.e; Gold, Forex, Foreign Portfolio Investment and FPI. While considering
Factor score1, the model yields better goodness of fit as measured by R
2
when we consider Dollar as
separate variable. The model summary shows that it is able to explain 98% of the variation in Factor score
1 due to the above mentioned variables. All the variables are statistically significant at 5% level.

Further regression is carried out with respect to factor score 2 corresponding to variables CRR and Oil.
The model summary reflects that 96% is explained by factor score 2, all the variables are statistically
significant at 5% level of significance.

Final regression is carried out between Dollar, factor score 1 and Factor score 2. In this case, the overall
model is statistically significant. It explains 76% of the variation in the dependent variables by the
independent variables. Further all the coefficients are statistically significant at 5% level.

Factor 1 can be named as External Reserve and Factor score 2 can be named as Inflation induced
variables.


190 | P a g e


7.0 CONCLUSION

In our research, we tried to find out the relationship between BSE Sensex and some other important
economical factors and got some interesting results related to this. We have used statistical methods to do
the analysis based on monthly basis database of different economical factors. Finally we got some
relationships of those factors with BSE Sensex. In our analysis we found that dollar price along with
Factor 1i.e; External Reserve and Factor score 2i.e; Inflation induced variables are significantly
affecting BSE Sensex. In the context of Indian economy the appreciation of Dollar will bring in more
foreign exchange reserve which will act as stimulant to foster growth and in this process the injection of
capital flows will affect Sensex. It is evident from the coefficient table that the rise in the price of Dollar
has a positive impact on Sensex.

Several other factors like Government Policies, political turbulence and social variables affects
fluctuations in BSE Sensex which can be analyzed statistically in future studies. Due to the constraint on
data base we have not considered the impact of political factors and turbulence on BSE Sensex. Here we
have analyzed nine independent variables but more quantitative factors can be included in further
research. The nonlinear dimension is yet to be explored.

8.0 REFERENCES:

Arijit Ghosh, Samrat Roy, Gautam Bandyopadhyay and Kripasindhu Choudhuri 28-30
th
Oct2010, Share
Market Analysis Using Various Economical Determinants, Online journal of American Institute of
Physics ISBN978-0-7354-0854-8 Vol.1

Aggarwal. R., 1981, Exchange Rates and Stock Prices: A Study of U.S. Capital Market under Floating
Exchange Rates, Akron Business and Economic Review, 7-12.

Bartov, E. and G. M. Bodnar, 1994, Firm Valuation, Earnings Expectations, and the Exchange-Rate
Exposure Effect, Journal of Finance, 49, 1755-1785.

Burton G. Malkeil, 1963, Equity yields, growth, and the structure of share prices, American Economic
review, LIII, 5, 1004-1031.

Franck. P. and Young, A., 1972, Stock price Reaction of Multinational Firms to Exchange Realignments,
Financial Management, 1, 66-73.

Franco Modigliani and Richard Cohn, 1979, Inflation, Rational valuation, and the market, Finan. Analysis
J., 35, 3-23.

John Mauldin, 2003, The Connection between Oil and Stock Prices, FLT Dec. Edition.

Kenneth E. Homa & Dwight M. Jaffee, 1971, The supply of money and common stock prices, Journal of
finance, 26, 1045-1066.
191 | P a g e


Martin Feldstein, 1980, Inflation and the stock market, The American Economic Review Vol 70, No. 5,
839-847.

Naeem Muhammad and Abdul Rasheed, 2001, Stock Prices and Exchange Rates: Are they Related?
Evidence from South Asian Countries, Applied Economics Research Center, Karachi-Pakistan.

Robert H. Rasche and Harold T. Shapiro, 1968, The F.R.B.-M.I.T. Economic model: its special features,
American Economic review, LVIII, 2, 136-137.

R. Smyth and M. Nandha, 2003, Bivariate causality between exchange rates and stock prices in South
Asia, Taylor and Francis Journals, 10, 699-704.

Smith, C., 1992, Stock Market and the Exchange Rate: A Multi-country Approach, Journal of
Macroeconomics, 14, 607-629.

Solnik. B., 1987, Using Financial Prices to Test Exchange Rate Models: A Note, Journal of Finance, 42,
141-149.

Soenen, L.A. and E.S. Hennigar, 1988, An Analysis of Exchange Rates and Stock Prices: the
U.S. Experience between 1980 and 1986, Akron Business and Economic Review, 7-16.

Tamnun E Mursalin, Ahmed Tanvir,Dr. Md. Jahangir Alam, Share Market Analysis Using Various
Economical and Social Determinants to Predict Decision for Investors(2006).

http://www.bseindia.com

http://ibef.org/download/fdi_12April10.htm

http://www.indiastat.com

http://www.kitco.com

http://www.moneycontrol.com


192 | P a g e






THE SUPPORT FROM THE
ORGANIZATIONAL STRUCTURE FOR
TACIT KNOWLEDGE IN PROJECT TEAM
7





Wang Lianjuan
Beijing University of Posts and Telecommunication
Beijing, China
lianjuan2003@126.com

Steven Si
Bloomsburg University of Pennsylvania
Bloomsburg, PA
ssi@bloomu.edu



7
This articls is the periodical achievmentof the program supported by ministry of education Knowledge
Management in Team: From the Perspective of Social Network Analysis (No. 10YJA630151).
193 | P a g e

Key words: Project team; Tacit knowledge;Knowledge management.
Abstract: Tacit knowledge is difficult to be shared because of its chracteristics. Project
team provide a benefit enviroment for tacit knowledge to be transferred and shared. Tacit
knowledge in team can be devided into two kinds. The first is outcome knowledge,
including the final product knowledge and know-who knowledge. The second is process
knowledge, including process of knowledge creating and process of team management.
Effective team meetings, communication tactic, standardization of team achievement are
all needed to promote tacit knowledge management in team.

Team, as the place where knowledge workers cooperate, has become a popular form adopted by
enterprises. Drucker(1996) thought that When a team find the right niche and can function, it is actually
the most effective organizational form. Senge(1990) considered that The basic learning unit of modern
organization is team, not the individual, so group learning is extremely important. Here the relationship
between them is that with the skin gone, to what can the hair attach itself . No team learning, no
organization learning.
Turners (1999) study indicated that "knowledge employees" is the main portion in project team,
they complete project tasks by communicating and cooperating with each other. Nonaka (1995) thought
that, in the process of cooperation, team members share knowledge in depth, learn from each other with
close contact, and create knowledge in the process. Team not only provides the basis for the individual
learning, and as the organizational bridge between employees and enterprise, it also gradually makes
explicited the knowledge which employees gain in the team, then transfer the knowledge to the company
level, so that the knowledge can be shared by other employees.
TACIT KNOWLEDGE IN TEAM
With the expansion of enterprise scale, the quantity of departments of organization is increasing, and
knowledge professional differentiation between departments is becoming more and more obvious. It
reduces the frequency and efficiency of communication between different departments, and different
departments can not share knowledge effectively, even they do not understand each other's business and
field. Highly specialization of knowledge means that it is needed to combine different professionals and
different background for the knowledge innovation. This kind of combination decides an important
working way for knowledge workers: workers with different professional and different backgrounds work
in the form of team.
The project team can make the members of different functional departments and even different
managerial hierarchies togethe. Its an effective method to solve these problems that the knowledge,
especially tacit knowledge, can not be shared and that employees can not communicate easily because of
the barrier between functional departments and between managerial hierarchies.
1. tacit knowledge
The concept of tacit knowledge was first put forward by philosopher Polanyi in The Study of Man in
1958.He believed that human have two kinds of knowledge, one kind of knowledge is usually expressed
by writing, charts, or mathematical formula. This is just one form of knowledge, and there is another kind
of knowledge which can not be systemically expressed, such as the knowledge about our behaviors. The
former can be called as explicit knowledge, and the latter tacit knowledge, which individual possesses but
is difficult to be expressed and conveyed clearly.
Hayek proposed the difference between articulated rules and non-articulated rules from the
perspective of jurisprudence and economics. He thought that " The non-rational factors of the basis of our
action is composed of our habits and technology, our preferences and attitude, our tools and our
institutions, these kinds of knowledge is the so-called tacit knowledge.
194 | P a g e

Sternberg has discussed the relationship among knowledge, human thinking and psychological
process from the perspective of psychology. He thought that the so-called tacit knowledge refers to the
knowledge action-oriented, which is procedural, and is gained with no assist from other people, and it can
urge peopleto realize their pursuit of the value goal. It is very important for real life to acquire and use
this kind of knowledge. In addition, tacit knowledge reflects individual ability learned from experience
and the ability of using knowledge in pursuing and realizing the individuals value goal.
Some scholars discussed tacit knowledge from the perspective of management and organizational
behavior. Drucker (1988) thought that Tacit knowledge, like some skills, can not be explained by
language, it can only be demo proved that it is there. The only way to learn this kind of skill is to
understand and practice. Nonaka (2000) considered that tacit knowledge is highly personalized, it is
difficult to be standardized, formalized, communicated, and shared, then he gave full attention to tacit
knowledge from the perspective of cognitive science of east, he thought that tacit knowledge includes
mental model and skills.
This view above about tacit knowledge mainly concerns individual attributes. Some researchers of
organizational theory further expand the connotation of tacit knowledge to the organizational level.
Nelson and Winter(1992), Spender (1993), and Alice (2000) thought that organizational tacit
knowledge exists in the internal of enterprise. Nelson and Winter(1982) introduced the organizational
routines mentioned by the research of knowledge economy of West today, which is equivalent to the
skills on the individual level. In most organizations, "How to do" is stored in these routines, and
individuals in organizations only can master these knowledge through learning by doing. Further
Spender (1993) divided organizational tacit knowledge into consciousness, automatic and shared tacit
knowledge, and each kind of tacit knowledge plays its distinctive role in strategy.
2Classification of tacit knowledge in project team
Davenport2000 wrote that knowledge is so elusive that we should explain the important part of
it.
There are many forms of tacit knowledge in project team, which exists in different ways.
Tacit knowledge can be divided into three hierarchies according to the difference of objective. The
first kind of knowledge is tacit knowledge on individual level, the second kind of knowledge is tacit
knowledge on team level,and the third kind of knowledge is tacit knowledge on enterprise level.
According to the process of the project management, tacit knowledge can be divided into the
following kinds: knowledg in the early phase, konwledge in the plan phase, konwledge in the implement
phase, konwledge in the acceptance phase, and knowledge in the daily detail management.
According to the technological characteristics of knowledge, tacit knowledge can be divided into
four classiffications: firstly, tacit technology knowledge, including informal ability, skills and knack that
is difficult to express; secondly, tacit management knowledge, including the knowledge of general
management and peculiar knowledge of project management; thirdly, tacit social relationship knowledge,
including the experience and skill of team work; Forthly, tacit cognition knowledge, including mental
models ,belief and values.
3. Classification of tacit knowledge in project team
We classify the tacit knowledge in project team according to the tacit levels. Tacit level is not
constant, and will develop constantly. As time goes on and with the activation of specific event,
communication between members will be more clearly and explicit. Some parts of the knowledge
generated during the operation of project can be gradually clarified and finally be changed into result
knowledge; other parts of knowledge exists in tacit form, may be understood by some members, but cant
be shared by other members, and maybe even disappear when the group dismiss. These kinds of
knowledge exist in the process of the project , such as process knowledge and team management process
knowledge. Tacit knowledge in Project team is shown in the below figure:


195 | P a g e





Fig 1: Knowledge in project team

The first, result knowledge
Although the result knowledge is more clear, this kind of knowledge is not explicit at first, it is
gradually clarified in the process of the project. Result knowledge can be divided into two categories as
shown in the above figure:
One is final product knowledge. The final product can be a product, a kind of service or a report, etc.
Whatever the product is , it always has knowledge in it. It is also a gradually clarified knowledge product
which comes from the communication and corporation of team members.
Second is social knowledge. The success of the team manifests not only in accomplishment of
mission, that is to say mission success , but also in relationship between team members, familiar with the
others domain and in accumulating the model to get along with orther team members, which can be
called social success. This relationship is very important to the partnership of members and the sharing of
knowledge.
The Second, process knowledge
Process knowledge comes from the operation of the project. Compared to the result knowledge, this
kind of knowledge is far more tacit, usually ignored by people. But this kind of knowledge plays a very
important role in accumulation, sharing and creation of knowledge in the team and enterprise. As time
goes on, the knowledge may disappear, unless we trace and summarize it in time.
Process knowledge mainly includes process knowledge from knowledge flow. Every team has its
own business process. People use and manage knowledge in the business activity, and the knowledge
itself forms process too and come into being knowledge flow.Knowledge flow describe the production
and development of knowledge, which can be a reference to other teams. Knowledge flow is relatively
difficult to be grasped, but with the development of the information technology, the trace of it will be
more available. The second is team management process knowledge. Team management is an art, we can
get a lot of learning experiences of management from the process. It is very necessary to summarize
experiences timely in order to improve the efficiency of team knowledge management.
ORGANIZATION DESIGN AND KNOWLEDGE PRODUCTION
Organizational structure refers to how to divide the work, how to group, and how to coordinate. The
structure of an organization is influenced by individual behavior and organizational environment in a
great extent. The individual mainly refers to the organizational manager who may influence the
organizational structure, on the other hand, organizational structure largely determines the way of
members in organization to contact with others, and then affect the behaviors of members.
The forms of traditional organizational structure mainly includes: the rank type, the division type,
and the matrix type. With the enterprise management form develops more flexibly, enterprise
organization form is becoming more and more diverse, and some new forms appearssuch as network
organization, team organization.
1. The relationship between organization design and knowledge production
Knowledge of
team management
Process knowledge
Social relationship
Process knowledge

Result knowledge
Final products
Knowledge
in project team
196 | P a g e

Organization design will affect the level of enterprise knowledge production. In order to improve the
level of enterprise knowledge management, the way of organization design should be considered.
Whether organization desig helps enterprise improve the level of knowledge production .
There are two aspects about the relationship between organization design and knowledge production.
One is the relationship between organization relaxation degree and knowledge production. Nonaka
(1995) proposed that the relationship between organization relaxation degree and knowledge production
should be considered when we design the organization. Relaxation degree is the allocation of resources to
deal with the uncertain situation, such as the diversity of internal knowledge of department , and the
reserve of members of project team. Relaxation resources can produce more motivation of innovation.
But as the diversity increasing, the enterprise may be have two risk: one is to choose the knowledge
production project which shouldn't be started; second is to stop the project which should be continued. So
the relationship between the diversity and the efficiency of knowledge production has an U-shaped-
against curve relationship, at the begining. With the increase of organization relaxation degree, the
efficiency of knowledge production grows constantly; however, when the organization relaxation degree
increases to a certain degree, the efficiency of knowledge production reach the maximum, and with the
further increase of organization relaxation degree, the efficiency of knowledge production will reduce.
Second is the similarity and diversity of knowledge production. Enterprise should consider the
similarity and diversity of knowledge production when setting up organization department. The
department with similar knowledge has a high efficiency, the controlability is stronger, but the diversity
of behavior is lowe. The behavior of department with diverse knowledge is totally different, it is difficult
to be controlled, but is benefit for individual to show their potential. Knowledge production needs
necessary diversity to create a sustainable production efficiency. Cummings(2004) research showed that
the performance of the task group with diverse knowledge is better than the group with single knowledge,
but if using a decentralized organization structure, the group with single knowledge can get the same
performance as the group of diverse knowledge get. Therefore, it is benefit for the production of
knowledge when using cross-functional project team or other organization form which makes the
members of different profession and department cooperate.

2. The optimal hierarchy of decentralization
Druker (1988) thought that almost all the knowledge workers are managers in knowledge-based
enterprise, they all need to make decisions in the process of their knowledge labor: judging which tools
should be taken to processing and screening information; and which method should be adopted to diffuse
and transfer relevant information to which kind of people.
The organization structure of enterprise adapted to knowledge management needs to be decentralized
as far as possible, and more rights of fieldwork disposition should be authorized to the grass-roots
knowledge employees. In this way, it not only improves the speed and efficiency of making decisions, but
can also better encourage employees to show their enthusiasm of innovation. Each employee, and each
grass-roots group is the center of innovation and profit creation, and is also the center of coordination and
that of self control.
Myers (1996) proposed the optimal hierarchy of decentralization, he thought that the total cost of
organization is the sum of the cost caused by poor quality information and that of incoherentmanagement
target. The farther distance from the top department (CEO office) is, the lower cost caused by poor
quality information will be, because decisions are made by the agent which is related to decision-making
and has more specialized knowledge, so this kind of cost will be reduced; But the farther distance from
the top department is, the higher cost caused by incoherent target will be. If the dispersion degree of
organization is higher, because of lacking of unified target, each member and each team in organization
will produce different knowledge and different targets, however, the cost will be increased when these
different kinds of knowledge are arranged and transferred to the top department. The best position of
decision-making is that when the reduction of the cost caused by poor quality information just covers the
increase of the cost caused by incoherent target.
197 | P a g e

3. The routine, rules and system of organization
Levit and March (1988) thought that organizations are learning through merging the
inference of history into the routine of behavior guidance. The routine includes situation, rules,
procedure, practices, strategies, and techniques used to build and operate organization, and it also
includes beliefs structure, frameworks, paradigms, norms, culture and knowledge which support,
interpret and boycott the normal routine. These routines are considered to be independent of individual,
even if members are changed, they will still exist. And these routines have more vitality than standard
operation procedures, they are helpful for basic knowledge of organization to be conceptualized, which is
the result of learning direct experience, interpretation (such as history, paradigm, reference frame and
culture) and others' experiences.
The routines and rules if formed firstly by standardizing the experience in operation. However, the
way that routines are formed by standardization maybe not be adopted, it should be determined by the
different stages of the development of enterprises. Mintzberg put forward five kinds of cooperation
mechanism of organization structure: (1) mutual adjustment; (2) direct supervision; (3) working process
standardization; (4) standardization of the results; (5) standardization of the skills. He pointed out that the
emergence of these five kinds of cooperation mechanism is a process. When the size of enterprise is
small, only having several people, members can take the method of mutual adjustment to coordinate with
each other. With the enlargement of enterprise scale and the increase of the complexity of organization, a
coordinate superior is needed, so the way of direct supervision appears. With the further enlargement of
the size and the further increase of complexity, it is necessary to take the way of standardization. General
work can adopt process standardization; But for some complex work, results standardization should be
adopted; if the work is more complex, not only the working process can not be standardized, but also the
results of the work are difficult to be formed to a standard, so it can only implement skills standardization;
While the enterprise is further complicated, such as high-tech enterprise, the way of standardization
coordination is also difficult to be implemented, then we should go back to adopt the way of mutual
adjustment.







4. Chain of communication
Command chain is an uninterrupted route of right , expanding from the top of organization to
the lowest level of the organization, and specifying who will report the work to whom. Command
chain is not only a kind of rights reflection, but also a kind of instruction of communication route.
Generally speaking, the longer communication route is, the more link of transmission in communication
route will be, the lower efficiency of communication wil l be, and the more serious distortion of
communication is.
Druker (1988) thought that the exchange and transmission of thought will not follow a linear
route through hierarchy, linear organization structure already can not adapt to the needs of
organization in knowledge-based background.
Alice (2000) also mentioned the relationship between enterprise organization structure and
tacit knowledge management. She thought that organization structure based on tacit knowledge
generally will appear a decentralized structure, and will use informal cooperation mechanism.
mutual adjustment
skills standardization
results standardization
process standardization
mutual adjustment
direct supervision
Fig 2: Five kinds of coordination mechanism of organization structure


198 | P a g e

Organization structure of linear function type involves many levels from the toppest of the
management level to the lowest level. and the communication chain involves many links. So that this kind
of organization structure is not benefit for sharing, innovating and accumulating knowledge. The
organization structure in division type and the matrix type can provide smoother convenience of
communication for members.
Of course, with the development of information technology and the impact of the tide of
decentralization, now, a staff can get information in a few seconds that only top managers could get 20
years ago. Staffs in any position can communicate with anyone, and don't need to take the formal way.
The influence on the enterprise knowledge management by communication command chain tends to be
weaker than before.
TEAM ORGANIZATIONAL FORMS AND RULES WHICH ARE BENEFIT FOR TACIT
KNOWLEDGE MANAGEMENT
1. Project team
The project team generally adopts matrix-type organization structure. Team members come from
different departments, work in different teams. When the project team dismiss, members will go back to
respective departments. In this matrix-type organization structure, members have two leaders at a certain
time, one is the project manager, another is the functional department manager. In such a situation, a
series of problems will appear, such as the contradicted order between project manager and functional
manager striving, and the problem of performance appraisal of members, etc.
In order to make up the shortage of matrix-type organization structure, some enterprises adopt task
force to accomplish the task, which means team members do their best to accomplish one a project until
the project ends. There are several advantages in this pattern, one of which is that from beginning to end,
all the participants in project need to report to coordinator or project manager of task force. This pattern
ensures the uniqueness of the command. Staffs will focus on one project, not many projects. When the
project ends, the task eliminates, and the members report to their function managers, then they will
participate in other new projects, or be sent to another task force.
Using group work is benefit for transferring and sharing tacit knowledge. Nowadays the working group
and the matrix-type organization are widespread forms. But the professional career of members will be
ruined if they are engaged in group work for a long term. The matrix-type organization goes against the
management principle that an inferior should not have two superiors. So Nonaka (1995) put forward the
lean organization that the performance of individuals is fully evaluated by group leaders when they are in
the group, but after the project ends, the performance of members who are going back to functional
departments is fully evaluated by department leaders. The mission of groups is to undertake projects,
requiring individuals to use their knowledge to create new knowledge. However, the mission of functional
departments is to train the employee. After the employees return back to functional departments, they will
be trained to improve. So lean organization forms a virtuous circle of "learning - application - learning".
2. Communication strategy of project team
Effective communication in project team is important for completing project tasks through promoting
the sharing and innovating knowledge. But in the process of communication, the influence among each
other team members, the difficult expression of tacit knowledge, and the means of communication will
inevitably affect the quality of communication.
Team members are eager to be accepted, so they have a tendency to workaccording to the rules f the
team. Numerous facts showed that team can give its members pressure, and makes them change their
attitudes and behaviors to be consistent with the team standards. This is the group psychology of team
members. Group psychology makes members have pressures when their personal view isnt consistent
with other members'. Then this psychology will drive members to be consistent with others. During the
199 | P a g e

team communication, the opinions of a person on a dominant position tend to affect other members'
views, and knowledge can not be expressed effectively because of being repressed.
When a group participates in activities that the rights difference plays a role, groupware support
system (a wide covered media) can solve the problem of rights difference by facilitating the team to use
the text signal instead of words signal. Anonymity and synchronization are the characteristics of
groupware support system, they can erase the effects of status difference. Anonymity can let members in
low position go against high-ranking person without fearing being found out. Synchronization doesn't
allow high-ranking person to dominate group discussion. In order to promote information exchange, it is
necessary to reduce constrained behaviors. Reducing constrained behaviors and exchanging ideas on
different levels can produce different interpretation about information, and it is helpful for organization
learning. Brainstorming, anonymous group creation, and electronic meeting are the effective methods to
reduce the problems of traditional team interaction.
The media adopted by team communication will also affect the effect of communication. The media
communication means mainly include: face-to-face exchange, video conference, telephone, email, fax,
official text, and etc. The influence on team communication by all the kinds of media is different. Daft
and Lengel (1983) introduced the concept of media wealthy. Media wealthy is an ability the media has to
change human understanding through clarifying the problems of enterprise. Wealthy features of one kind
of media include: the ability of providing rapid feedback; the ability of spreading various hints; the ability
of conveying personal sensory hint; and the ability of using natural languages. Swanson (1987) and
Markus (1994) put forward the concept of media coverage domain. Media coverage domain involves two
additional features: storage characteristic, namely the ability of storing information; and coverage breadth,
namely the ability of communicating with many people at the same time.












Fig 3: media wealthy degree and media coverage domain of media communication means

As can be seen from the figure, different media plays different roles in communication. The media
like face-to-face communication, with a higher wealthy degree, is helpful for promoting the establishment
of close and trust relationship between people, therefore, at the early stage of a team, in order to deepen
the emotions and trusts between members, most teams take some methods to let members meet. Besides,
deepening mutual understanding is an indispensable ring for building close and trust relationship. The
media, like E-mail and official text, is helpful for promoting fast and convenient communication because
of its characteristic of rapid spread. Communication media should be chosen according to certaine
situation.
3. Meetings of project team
The project manager needs to hold different kinds of meetings in the process of project. These
meetings can be classified into regular meeting and special meeting.
email
video conference
telephone
face-to-face exchange
media wealthy degree
high
high
low
low
official text
200 | P a g e

Regular meeting is held regularly, for instance, the first Monday of each month, or the end stage of
the development of electrical goods or of the whole project.
One kind of regular meeting is that of checking and commenting on project. The meeting is held in
each period of specific milestone, such as the time that the team completes a certain kind of electrical
goods. Participants include project managers, customer representatives and senior managers. They get
together to discuss the main event which has happened, and decide whether to continue, rework, even end
the project.
Another kind of regular meeting is that of evaluating project progress. These meetings are held at a
particular time interval, such as once a week or once every two weeks. The aim of these meetings is to
achieve the progress information of project schedule, budget and quality. Participants include project
manager, team leader and customer representatives. Then the project manager can make comments on
project progress after getting enough information.
The third kind of regular meeting is staff meeting. This meeting generally is held once a week. Only
team members attend it. Customers and senior management dont attend. This kind of meeting generally
lasts a short time, and members discuss only a few issues during the meeting. Team members get together
to discuss the success and failure of work, and exchange important information with each other.
The fourth kind of regular meeting is that of project review. This meeting is held after the end of
project. Members of team summarize the experiences and lessons which can be learned from the whole
project progress in the meeting,. And this kind of meeting is often ignored by enterprises, but it is
important for members to understand the knowledge which are created by themselves.
The project manager needs to hold not only regular meeting, but also the special meeting. Special
meeting is held in needs. One kind of special meeting is that project manager talks alone with another
member of project in a short time. They get together to discuss the immediate and the long-term task of
team members, the achievements, and the difficulties. These meetings must be held on the basis of fully
trust. If the project manager slightly shake the foundation of trust, the relationship between team members
and project manager will worsen.
The aim of dialogue meeting is to create a shared environment. Everyone can ask questions in this
kind of meeting. Senge (1998) called it a "probe field" which is a kind of group activities. Container can
be understood as a thought environment, people in this environment build the reality views to be a sum of
collective constructing, shared intention and faith of a team. Members can gradually recognize the
situation and problems through dialogue. Everyone adjusts their views themselves, and gradually reach a
consensus. This method focuses on synthesizing and improving the knowledge system. However, letting
people speak freely and organize by themselves can also bring some other effects, such as the spread and
creation of new knowledge.
4. Results standardization of project team
Results standardization of project team comes into being the routines, rules of enterprise. It is
helpful for enterprise to have rules to follow. There are many ways of results standardization:
One is chart and list. Through storing the results of project team in the way of charts and lists, it can
provide important reference for other project teams. But charts and lists provide information entries in an
isolated and separate way because of the lack of background materials. People need turn to memory to
compensate for the lack when storing information only in the form of charts and lists. In this situation,
descriptive explanation can reproduce the process of knowledge innovation with a great degree.
Another is Learning experiences description. Learning experiences is a tool to describe the case
occurring in the process of learning and changing. It has recorded the difficulties and events in learning.
Its key is the ideas of the events happening in the process of learning, the views of their actions, and the
difference between people about their various views. Learning experiences let the readers experience the
201 | P a g e

events happening in the process of learning just like on the spot, and help them make smarter decisions.
Its purpose is to provide background and information for others to learn. It is a process for organization to
gain, evaluate and spread the information of pilot projects.
According to the theory of five kinds of cooperation mechanism of organization structure proposed by
Mintzberg, in knowledge-based enterprise, standardization as a cooperation mechanism is just one of the
methods to save routines and rules, enterprise also needs to take the adjusting way to create a good
atmosphere of cooperation for team members to promote the sharing and innovation of knowledge
because of the complexity of the work and the limitation of the quantity of team members.
5. The knowledge departments and knowledge practitioners which are beneficial to knowledge
management
In order to promote knowledge management, enterprises also need to appoint special knowledge
practitioners. knowledge practitioners are the employees on the ground and the department managers
which are responsible for "accumulating knowledge, producing knowledge and updating fuzzy and
clarifying knowledge". The diathesis requirements for them include high intellectual standard, dedicated
consciousness of transforming the world by their own views, the rich experiences inside and outside the
company, the skills of communicating with customers and colleagues, and the broad bosom when
discussing or debating with others on public.
Knowledge practitioners in different enterprises are different. Enterprises need to consider the
actual situation of enterprise to appoint knowledge practitioners. According to the needs of knowledge
management, the major knowledge practitioners can be knowledge project managers, knowledge
department managers, etc.
The responsibility of knowledge project managers is planning a feasibility study, organizing team,
arranging plan, dispatching resources, supervising the progress of work, solving some temporary
problems, and summarizing. Because knowledge project is different from general projects, it concerns
more on person less on objects.So the knowledge project manager should be considered about hie oe her
ability of dealing with people.
For knowledge-intensive organization, sometimes it needs to appoint a knowledge department
manager.The requirements of him or her are similar to that of CKO. The difference is a knowledge
department manager is not as good as CKO on holding all the cards, but need more professional quanlities
than CKO.
In order to promote the communication and exchange between different project teams, especially
when these project teams have a great barrier in space, it needs to install a new position of knowledge
management practitioners. Bresnen proposed the method to solve the problem of knowledge transfer for
construction project team. Because construction teams are in different regions, it is difficult for them to
communicate with each other, and because the form and dissolution sometimes happen, so members
unavoidably change frequently. In such cases, a team always reduplicate invention. Thus Bresnen
proposed a position called regional engineering manager (REM). In one region team sets up one fixed
position, REM. REM in charge of improving the communication between each construction team,
assisting the training and development of each construction team engineer in each region. REMs
generally meet on the engineering meeting every three months. At the meeting, they discuss many
problems in many fields, and they build and strengthen the social networks through the meeting. REM
also holds the annual convention for engineers in each construction site. At the meeting, members discuss
the lessons and successful experiences of project, the work doing at hand, some problems existed at
present, and the dissatisfaction of job, etc.


Conferences
1 Alice Lam(2000), Tacit Knowledge, Organizational Learning and Societal Institutions: An
202 | P a g e

Integrated Framework, Organization Studies, 23(3): 487-513.
2 Barbara L., & March, J.G(1988)., Organizational Learning, Annual Review of Sociology, 14:
319340.
3 Cummings, J. N.(2004), Work Groups, Structural Diversity, and Knowledge Sharing in a Global
Organization, Management Science, 50(3), 352-364.
4 Daft, R.L. & Lengel, R.H.(1996), Information richness: a new approach to managerial behavior and
organization design. In Ethel Auster & Chun Wei Choo,. (Eds.) Managing information for the
competitive edge. (pp.167-216). New York, NY: Neal-Schuman.
5 Davenport, T. H.(2000), Working Knowledge: How Organizations Manage What they Know,
Harvard Business School Press.
6 Drucker, P. F.(1988), The Coming of the New Organization, Harvard Business Review, 66 (1),45
53.
7 Hayek, F. A. (1978), New Studies in Philosophy, Politics, Economics and the History of Ideas,
Chicago: The University of Chicago Press.
8 Markus, M. L.(1994), Electronic mail as the medium of managirial choice, Organization Science,
5:502-527.
9 Mike, B., & Linda, E.(2003), Social Practices and the Management of Knowledge in Project
Environments, International Journal of Project Management, 21(3): .157-166.
10 Myers, P. S.(1996), Knowledge management and organization design, Boston: Butterworth-
Heinemann.
11 Nelson, R.R. & Winter, S.G.(1982), An Evolutionary Theory of Economic Change, Belknap Press.
12 Nonaka I., & Takeuchi H.(1995), The Knowledge Creating Company, Oxford University Press.
13 Nonaka , I., Toyama R., & Konno N.(2000), SECI,Ba and Leadership: a Unified Model of Dynamic
Knowledge Creation, Long Range Planning, 33(1) :5 - 34.
14 Polanyi, M.(1958), Personality knowledge. Chi- cago: University of Chicago Press.
15 Senge, P. M.(1990), The Fifth Discipline:The Art and Practice of the Learning Organization,
Bantam Doubleday.
16 Spender, J.C.(1993), Competitive Advantage From Tacit Knowledge, Academy of Management
Proceedings, 9(2), 37-40.
17 Sternberg, R. J.(1985), Beyond IQ: A triarchic theory of human intelligence. New York: Cambridge
University Press.
18 Swanson, B. E.(1987), Information channel disposition and use, Decision Science, 18: 131-145.
19 Turner J. R.(1999), The Handbook of Project-based Management, McGraw-Hill Publishing
Company.




203 | P a g e













IN SEARCH OF VARIOUS STRATEGIC
FACETS OF BEHAVIORAL CHANNEL
MANAGEMENT: AN EMPIRICAL STUDY
ON INDIAN DISTRIBUTION


Kaushik Mandal
Department of Management Studies,
National Institute of Technology,
Durgapur, India
kaushikmandal.nit@gmail.com

Goutam Bandyopadhaya
Department of Management Studies,
National Institute of Technology,
204 | P a g e

Durgapur, India
math_gb@yahoo.co.in

Koushick Roy
Department of Management Studies,
National Institute of Technology,
Durgapur, India
koushickroy@gmail.com
205 | P a g e

Key Words: Influence Strategies, Power Strategies, Exploratory factor analysis, Confirmatory
factor analysis, Reliability & Validity.

ABSTRACT:
India has been witnessing a sea change in terms of competition since the decade of nineties.
Indian players, especially players of FMCG, Cement & Durable, are compelled to change their
functional operations under this high degree of competitions. Therefore, functional operation
like distribution management is considered important in terms of effectiveness & efficiency. For
the purpose of effective & efficient distribution, marketers have to take not only the care of
logistics but also they have to take care of behavioral management of channel partners. Globally
accepted literatures have already proposed for strategies of power & influence in this regard.
But some of the empirical researches have raised questions on applicability of those strategies in
relation to developing nations. Considering this question valid and relevant we have identified
our research plan vis--vis Indian distribution.

For this study we have selected three objectives.
i) Development of scale for identification of channel control strategies of marketer.
ii) Identification of strategies for controlling channel partner.
iii) Identify the importance of each specified strategies.
For the empirical purpose of present research, we have executed a survey in southern part of
west Bengal, a state of India, with randomly selected 166 respondents who are distribution
channel partners in profession. We have selected southern part of the state of West Bengal
because this region is having representative character of India.
For the purpose of attainment of first objective stated above, we have developed a valid and
reliable construct by combining items from the globally recognized scales of power & influence
strategies. We have reached to second & third objectives by using appropriate statistical tools
used for Exploratory & Confirmatory factor analysis.
Finally, we are able to identify four important strategic-facets namely i) Intelligence ii) Big
Bossing iii) Contract Orientation iv) Expert for Indian distributions. Result indicates deviation
from the global research outcome. Thus, the present research is very much contributory to
academic & management world, especially in reference to developing nations.

INTRODUCTORY NOTE
Modern marketing practices across the globe talk about efficiency of distribution management.
Objective of timely placement of the right product to right customer group brings challenges for
attainment of efficiency in the system of distribution. In fact, conflicts in the distribution system
are less physical, more human, in nature. Mostly behavioral reasons are the causes behind
channel conflict between marketers and channel partners. Therefore behavioral management of
distribution system is highly required to handle channel conflict between marketers and channel
partners.
Indian environment is same as in global context. Moreover, India has been witnessing a sea
change in their economic environment from the inception of economic reforms of nineties. Since
then global players have shown their foot fallings in Indian drawing rooms. Thus Indian
customers are enjoying wider choices among competitive and equally likely brands. Big
206 | P a g e

shopping malls to village hut have been experiencing a range of competing brands. As a result,
they become brand switcher, if not split loyal. This scenario of competition is maximum in
India in respect of three industries namely, FMCG (Fast moving consumer goods), Cement &
Consumer durable. These three industries are unique in characteristics in India having market
dominated by middle class, middle income customer with moderate but not with high capacity of
purchasing. Hence customers are interested for cost efficient / economy brand which is also
having value in functional as well as non functional dimensions.
Therefore, we have learned two important facets of Indian competition. First one, to get &
sustain the competitive position in the mind of the customer, availability of the product in the
shelf at the point of purchase is minimal or primary condition. Second facet talks about getting a
distinct position in the mind of the customer marketer who has to generate functional & non
functional values. Where cost efficiency of the product is definitely a good and important
functional value for middle class, middle income customers with moderate power of purchasing.
In this backdrop it is wise to infer that timely supply of the product with efficient
mechanism is the key to achieve success in keen competition. Thus it calls for applications of
proper behavioral management of the marketers vis--vis channel partners. Good behavioral
management does simultaneous reduction of channel conflict and increase of motivation among
channel partners. Thus motivated channel partners have to follow the guidelines of the marketer
to create edge in competition.


REVIEW OF LITERATURES
We can classify the researches in the field of behavioral management of channel easily with
three groups. First group of literatures talks about power sources and dependence of channel
members. Second group of literature focuses on influence strategy of the channel partners over
each other. Third group of researches considers power, power sources & influence strategies are
having same perspective and that is compliance of channel partners in the context of channel
management practices of the channel leader (marketer). First group of literatures concentrates on
usage of social power in behavioral management of channel. Social power is the concept that has
been mostly used by researchers in the field of social psychology. Also in the literatures of
politics power bases are defined, well before management literature. For example Simon
(1953) has talked about measurement of power. French & Raven (1959) & Cartwright (1959)
have employed this concept of social power for the first time in the literature of management.
French & Raven (1959) have identified the five social power bases and these are reward,
coercive, expert, referent & legitimate. Later Raven (1965) has added one more power base and
that is information exchange. Operationalisation of these social power construct was earnestly
required for management practices. Swasy (1979) has generated a reliable & valid scale to bridge
the gap between concept and practices in the context of channel management. Ansary & Stern
(1972) are pioneer in measuring power in the context of a behavioral management of a
distribution channel. For this purpose they were not taking considering ideas promoted by French
& Raven (1959). In fact they have proposed a model, where they have tried to estimate power as
a function of weaker channel members dependence on stronger channel member and sources of
power perceived by relatively weaker channel partner over relatively stronger channel partner.
The effort they made taken was in vain as they failed to establish any good result. But this result
probably would have prompted Hunt & Nevin (1974) to carry out research in the same direction.
They have unified the concept proposed by Simon (1953), French & Raven (1959) and the
207 | P a g e

theory that Ansary & Stern (1972) was interested to promote. Hunt & Nevin (1974) have
modified the power model by classifying power into two sources that are coercive sources &
non-coercive sources. Further under non-coercive sources they have mentioned reward,
expertise, legitimacy & referent power. With the help of a technique called multiple
classification analysis they have empirically tested the relationship between power of a channel
member and available sources of power. Furthermore, they have identified, in case of a
franchisor-franchisee form of distribution channel , that franchisor depends on coercive form of
power to control franchisee but franchisee is more satisfied when franchisee is using non
coercive form of powers. Completely different approach has also been seen. In the line consistent
with Bacharach & Lawler (1980), Frazier (1983) has critically appraised French & Raven (1959)
way of viewing powerbases. According to them power bases mentioned by French & Raven
(1959) is more of instrument of executing power, less of sources of power. In fact they have
assumed for dyadic relationship between two channel partners. Thus they believe role
performances of each channel partner are the key to create dependence between them.
Some of the researches are very focused on the application of the social power theory in
practice of channel management. Improper control over channel partners may results in to
channel conflict. Hunt & Nevin (1974) have introduced the issue by highlighting satisfaction of
channel partners. Actually dissatisfaction of channel partner causes conflict. First time Lusch
(1976) has measured impact of power sources on conflict within channel. In the same line Gaski
(1984) has also opined misuse of power would lead to conflict. But he expressed his non
acceptance of viewing power and sources of power in isolation. He has opined power
manifestation is highly related to exercise of power. Therefore, he has found role of influencing
channel partner is important. Like Gaski (1984), several researchers like Frazier & Summer
(1986), Kale (1986) & Frazier et. al (1989) are in favor of the school of thought that is based on
interface among influence strategies , power & sources of power.
Frazier & Summer (1986) also have advocated for interrelation between power and
influence strategies. Empirically they have shown influence strategies are different ways for
applying power, either in form of threat or in the form of recommendation. Furthermore, Frazier
& Summer (1986) have found negative association between coercive influencing and dealers
perceived satisfaction. According to theory, high coercion of channel leader over channel partner
would lead to low satisfaction of the dealers. Almost similar line of argument has been presented
by Kale (1986). Moreover, Kale (1986) has recommended for carrying out research in
developing nations with the employment of power & influence scale that are already established
on developed countries. He has executed his research on India and found a strong relation
between power and degree of pressure involved in exercising influencing strategies to change
channel partners behavior. Moreover, he has found positive associations among various
influence mechanisms and between marketers power and each of influence mechanism. In his
study Kale (1986) calls for carrying out study across several industries in developing nations like
India. Frazier et. al. (1989) have also opined that business environment in which channel is
operating should be considered for setting channel theory. They have also expressed the need for
executing study across several industries.
Third group of literature covers the influence strategies & its usage on management of channel
partners behavior. Boyle et. al. (1992), Frazier & Sheth (1985), Payan & Mcfarland (2005) are
prominent among the list of this group of researchers. Boyle et. al. (1992) have studied the
influence strategies as a strategic move by which marketer instruct the channel partner to alter
their behavior in favor of marketer. This work is highly contributory to the literature because of
208 | P a g e

they have developed the reliable & valid construct of influence. Furthermore, they have
established some of the influencing strategies which are having negative impact on relationship
between marketer & channel partner. Keith et. al. (1990) have also found almost the same types
of results. According to them channel partners attitude is dependent on channel partners
dependence on marketer and influence used by marketer. Though they have used the term
influence, they have actually employed power bases as a construct for influence. In an
alternative version Frazier & Sheth (1985) have conceptually shown attitude & behavior about
implementation of channel programme designed by marketer and that is dependent on various
form of influencing strategies. Payan & Mcfarland (2005) in recent study have been theorizing
the effect of influence strategies on compliance of channel partner. They are in favor of applying
non-coercive influence strategies on channel partner when adequate level of quality
advertisement is assured. According to the theory, coercive influence acts only in case of high
dependence of channel partners over their channel leader.
The commonality find in all three types of literature is in relation to purpose of behavioral
management of channel partners by their channel leader (marketer). These three groups of
literature are three different ways to meet the purpose. All these theories have been developed in
the context of advanced developed nations. Few authors have carried out research in developing
nations. Result found is interesting and calls for further research in the channel management of
developing nations. Practical knowledge based on exposure of channel behavior of Indian
distribution channel members of FMCG, Cement & Consumer durables helps us to understand
that Indian channel leader (marketer) combines the paths to reach the purpose of effective &
efficient behavioral management. Reason behind this combined approach is probably because
that both power & influence are used for same perspective and that is in compliance of channel
partner. Therefore, we believe that empirically we should test how power & influence are
positioned in the mind of channel partners. Kale & other authors like Frazier et. al. (1989) are
advocating for empirical test results across industries of developing nations. This argument helps
us to decide that we should carry out empirical test in our country India. Literature review also
suggests to identify two important scales one for influence and another for power. From the
review we also find that scales, used by Boyle et. al. (1992) & Swasy (1979) respectively for
influence & power , are two good scales in terms of reliability & validity. Thus, if we combine
these two scales we would have sixty items in hand. Methodology based literatures suggest, it is
not possible for respondents to operate with large numbers of items (Allen & Rao, 2010). Hence,
methodological literatures suggest reducing the items by proper item analysis.

RESEARCH OBJECTIVES & PROCESS UNDERTAKEN
Research Objectives:
Research objectives are three dimensional in nature. First of all we have to develop an instrument
(scale) suitable in the Indian context of distribution to identify & measure channel management
practices in India. Secondly using this scale we have to identify various strategies of Indian
distribution channel management. Thirdly, we have to measure importance of each of the
strategies identified from Indian channel management practices.

Research Process Undertaken:
209 | P a g e

In the previous section we have reviewed three groups of literatures. All these literatures
converge to single perspective that how to direct channel partners by the channel leader i.e.
marketer. Naturally a research query comes into our mind that can we develop a single scale
which combines scale used for measurement of influence strategies & scale deployed for
measurement of perception in relation to power sources possessed by marketer?
For the purpose of answering first research objective we have to mix all the items used by both
schools of literatures. In the present study we have combined 31 items from the scale used by
(Swasy 1979) & 29 items from the scale used by (Boyle et.al 1992). Our next task is to filter the
scale with respect to Indian business environment. It is justified because India is having different
cultural identities which resulted in highly inter connected social relationship among populations.
Thus we have to execute a survey using questionnaire covering all these 60 items. For the
purpose of filtration we have to use multiple correlations among all these 60 items. Next we have
to identify all those items each of which has high (more than .60 correlation) and significant (less
than .05 probability value) correlation with at least one of the 59 items. By this we can able to
handle multi co-linearity problem in advance. This filtration process would generate xx no of
items out of initially gathered 60 items. For details of sampling, see Table: 1.
Table: 1
Sampling at a Glance
Industries Population
Randomly
Selected
Responded
Randomly
selected Hold
out Sample
Water purifier 53 53 53
56
Cement 117 80 56
FMCG 121 80 57
Total 291 213 166

In the next stage with these xx items we would like to carry out another survey among the
specified respondents with an objective to identify various facets. It is ideal to carry out two
surveys one is pilot kind of survey for the purpose of filtration & next survey for identification of
various latent facets behind it. However, for minimizing cost & time, researchers may use one
survey in place of two mentioned above. For the purpose of survey among the respondents who
are channel participants for the present work, we have decided to execute on south Bengal region
of West Bengal. West Bengal is one state which represents cosmopolitan, rural encircled urban
area & urban middle class dominated characters of India.
South Bengal region is very much representative of the features stated above. Further we have
collected name & addresses of channel participants from three industries FMCG, Cement &
Water purifier located in this region from various sources like i) yellow pages ii)respective trade
associations directory & iii) executives of major companies of these three industries operating in
this region. We have prepared three lists of channel partners of FMCG, Cement & water purifier
industries. For water purifier group of channel partner all the 53 respondents have been selected.
We have visited personally to all those 53 channel participants and interviewed with structured
questionnaire using construct developed for the study. Likewise for cement, 80 respondents have
been selected randomly out of which 56 have been responded positively at the time of visit to
their place for the purpose of interview. In the similar way, 80 respondents have been selected
randomly but only 57 responded positively in time of visit to their place for interview with
210 | P a g e

structured questionnaire having pre-selected construct. In the similar way we have selected 56
samples randomly for the purpose of validity checking.
It is here not out of place to mention that exploratory factor analysis with suitable statistical
analysis is most useful for the purpose of identification of latent factors (facets) behind. But
further we have to confirm the extracted facets (factors) by using confirmatory factor analysis. In
between we have to calculate the importance of each of the identified facets by using the
philosophy of Principal Component Analysis (Johnson & Whichern, 2008). According to this
philosophy proportion of variance of each principal component is considered to be weight
(importance) for the same. We can refer Guha & Chakraborty (2002) for similar usage of PCA.
It is highly required to check the reliability and validity of the construct developed in this
present work. Exploratory analysis helped us to identify four important facets namely
i) Intelligence ii) Big Bossing iii) Contract Orientation iv) Expert. Each of these facets is having
some items. Now it is necessary that items within facet should have sufficient inter-correlation
among them. It is because; these inter-correlations are denoting internal consistency among the
items measuring that factor. In fact analysis in relation to internal consistencies of the items
within a factor is known as reliability analysis. Cronbach Alpha is a statistical measure which
examines the average internal consistency among the items. More than .7 value of cronbach
alpha is considered as acceptable reliability of the factor. Reliability is the boundary condition
for validity. In spite of good reliability, possibilities are there that the factor is not having
acceptable validity. Actually, acceptable validity of a construct means the construct is able to
measure successfully the concept it is used to measure. For this purpose, we are using Campbell
& Fiske( 1959) criteria for checking the validity. As per the rule we have to take one more set of
responses called hold out sample. Campbell & Fiske (1959) have proposed for measurement of
validity by two ways namely convergent validity & discriminant validity. Convergent validity is
agreement among several attempts to measure same concept (construct). Discriminant validity is
the dis-agreement among several attempts to measure various concepts (construct). Severability
of attempts means attempts by different samples or attempts by different methods. In the case of
present research we use two groups of samples. One of which is used as sample for research and
another is small group of sample used for the purpose of measuring validity. This is also known
as hold out sample. As we have done reliability test for factor construct correlation between any
two item within a factor which is always positive and significant. In the similar way correlation
between any two items between factors are insignificant. Now it is to check whether this reliable
result is stable for any other available sample. If result is positive as per the criteria, we consider
it as a validity of the model (construct) proposed for. To satisfy convergent validity for a
particular item we have to check its correlation with any item belonging to the same factor but
that responses can be drawn from hold out samples. Significant and positive correlation indicates
satisfaction of convergent validity. To satisfy discriminant validity for an item we have to check
its correlation with any item belonging to different factors but we have to generate responses for
this item from hold out samples. Insignificant or nearer to zero correlation indicates satisfaction
of discriminant validity. But the proposed validity of the model is not free from limitations.
Confirmatory Factor Analysis can overcome this situation. Diagram: 1 shows the research
process at a glance.



211 | P a g e



212 | P a g e

Diagram: 1
Process Undertaken In Present Research













































Marketer

Channel Partner
Perception of
Sources of Power
marketer has

Influence
Strategies

Instrument of
measuring
Perception of power


Instrument of
measuring
Influence Strategies

Combined Scale
of 60 Items

Filtered by
criteria
-- High (>.60)
& significant
(<.05)
Correlation

Research
objectives

Survey to
Channel Partner
Survey to
Channel Partner
after random
selection
Exploratory
Factor Analysis
(EFA)

May be
conduct
ed
simultan
eously

Derived xx
items scales

Factor Extracted

Importance of
Factors
determined

Confirmed the
factors extracted
by Confirmatory
Factor Analysis
(CFA)

1
2
3
Reliability &
Validity
Analysis
213 | P a g e

Primarily we have considered all those items which have been used for constructing Power and
Influence scales. Next to that we have identified only those items which have high (more than
.60 correlation) and significant (less than .05 probability value) correlation with at least one of
the 59 items. By these filtration process 12 items (See Table: 2) have been selected for further
analysis.
Table: 2
Items Considered
Symbols of
Items
Description of items
POWER 6
The information provided by company about this situation
makes sense.
POWER 8 Being similar to company is good
POWER 9 The information which is provided by company is logical.
POWER 11
Company's expertise makes other channel member more likely
to be right.
POWER 14 I want to be similar to company
POWER 26 Company's knowledge makes other channel member right.
INFLUENCE 5
Company refers to portion of our franchise agreement which
favor their position to gain our compliance on a particular
demand.
INFLUENCE
11
Company makes a point to refer to any legal agreements we
have when attempting to influence our actions.
INFLUENCE
23
Company uses section of our sales agreements as a ' Tool ' to
get us to agree to their demands.
INFLUENCE
24
Company communicates their ability to Make Thing
Difficult for our business if specific demands are not met.
INFLUENCE
27
Company states that specific services will be discontinued for
not complying to requests.
INFLUENCE
29
Company threatens to reduce the amount of business they will
do with our firm, should their demand not be met.

Exploratory Factor Analysis:
We measure the useful statistics Kaiser-Meyer-Olkin (KMO) to check the appropriateness of
sample adequacy. Generally KMO measure is an index used to evaluate the sample adequacy of a factor
analysis. A high value (between .5 & 1.0) indicates that factor analysis is adequate in terms of sample
(Malhotra, 2009). Value of KMO measure of sample adequacy in case of present research is .701 which
signifies the purpose of adequacy. Bartletts test of sphericity is used to check whether the appropriate
inter correlation exist for running factor analysis or not. The greater the value of test statistic, factor
model becomes more appropriate. The approximate chi-square value is 762.527 with 66 degree of (
12
C
2
)
freedom, which is significant at .05 level. Thus considering all the above facts, we are eligible to use
214 | P a g e

factor analysis to identify the strategies for management of behavior of channel partner in this present
analysis.

Table: 3
KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling
Adequacy. .701
Bartlett's Test of
Sphericity
Approx. Chi-Square 762.527

Degree of freedom
66

Significance
.000

We have used principal component analysis to determine the minimum number of factors that
will account for maximum variance in the data (for all those eigenvalues scores more than 1). This initial
solution is then rotated by using varimax method with Kaizer Normalization to select the variables with
high loading on a particular factor. The principal component analysis along with varimax rotation reduces
the 12 variables in to four factors (See Table No.: 4). All those factors are having corresponding
eigenvalue greater than 1. Each factor is composed of variable which is having factor loading of more
than 0.50 (Hair et. al 2009). Four factors are extracted with 72.32 % cumulative variance. Table: 4 depict
the factors after rotation. Percentage variance explained by cumulative total variance explained for each
factor is considered to be the importance of the said extracted factor. Factor 1 represents 37.09% (i.e.
26.83/72.33) of extracted factors variance explained. Likewise factor 2, 3 & 4 each represents 29.27%,
18.95% & 14.67% respectively.












215 | P a g e

Table: 4
Result of Factor Analysis & Regression Analysis


This analysis yielded four interpretable factors contained four, three, three and two items respectively
(See Table no.: 4).We have used naming for the purpose of unique representation of the latent factor
behind items grouped under the particular factor.
Latent factor behind factor 1 represents power of Intelligence of the channel principal over channel
subordinate. This indicates channel principals intelligence power in respect of leadership & knowledge
of the business which is being used to control the channel participant. So, we have named the latent factor
as Intelligence.
Similarly items under factor 2 describe the latent fact of big brother attitude of the channel principal over
channel subordinate. This indicates role of a channel principal as big brother in a family to control the
channel participants. As a result, we have named the latent factor as Big Bossing.
Items under factor 3 covers channel principals instructive behavior. It represents influence strategy based
on legal contract between channel principal and channel subordinates. Hence, we have named the factor
as Contract Orientation.
Factor 4 expresses knowledge & skill of the expertise which is valuable for smooth running of the
business. That is why we have named this factor as Expert.
Items Intelligence Big Bossing
Contract
Orientation
Expert
Reliability
(Cronbach's Alpha)
Variance
Explained
Importance
of the Factor**
R
2
POWER 9 0.82
POWER 14 0.79
POWER 6 0.77
POWER 8 0.76
INFLUENCE 24 0.86
INFLUENCE 29 0.83
INFLUENCE 27 0.82
INFLUENCE 11 0.88
INFLUENCE 5 0.85
INFLUENCE 23 0.77
POWER 26 0.92
POWER 11 0.87
** (% Variation Explained by the Factor/Total Cumulative Variance)
98.5%
96.3%
96.2%
97%
37.09%
29.27%
18.95%
14.67%
26.83%
21.17%
13.71%
10.61%
Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.
Rotation Method: Varimax with Kaiser Normalization
Rotation converged in 5 iterations.
0.799
0.815
0.815
0.796
216 | P a g e

Exploratory results shows each of the extracted factors are either generated by the items of power scale or
developed by the items of influence scale. Therefore, we can infer safely that power & influence are
positioned distinctly in the mind of channel participants. Further statistical analysis makes the conclusion
more concrete.

Analysis of Reliability and Validity of the Construct:
Cronbachs alpha for all four factors are more than .70 (See Table No.:4). Thus the factor reliabilities
have satisfied the criteria. We have to check the validity for 12 items. So, responses of actual sample on
these 12 items have to be correlated with response of hold out sample for 12 items. Thus we need 12 x
12= 144 correlation results to check the validity of the 12 items. Those who are interested can check these
detailed results in appendix. In summary we can safely claim that all the construct are valid because in
almost all the 90-95% cases the criteria holds strongly.

Statistical Explanation:
We use regression technique to understand the explanatory power of independent variable which is
unobserved (factor score) by dependent variables which is observed score on item comes under said
factors. We get R
2
value of four factors are 98.5%, 96.3%, 96.2% & 97% (See Table No.:2) which
replicates a better predicts of the different power used by channel principal. It proves how observe
variables perfectly converge to unobserved one. We have also examined the t-test which measures the
significance of the partial correlation of variables reflected in each factor. The detailed result of regression
analysis of each extracted source of Power has been analyzed in this regard.
In Every analysis, we have found the model with a large regression sum of squares in comparison to the
residual sum of squares which indicates that the model accounts for most explained of variation in the
dependent variable. So, it is obvious, low residual sum of square indicates that the model fits to explain a
lot of the variation in the dependent variable. The mean square is the sum of squares divided by the
degrees of freedom. The F statistic is the regression mean square (MSR) divided by the residual mean
square (MSE).The regression degrees of freedom is the numerator and the residual degrees of freedom is
the denominator degrees of freedom for the F statistic. The total number of degrees of freedom is the
number of cases minus 1. If the significance value of the F statistic is small (smaller than say 0.05) then
the independent variables do a good job by explaining the variation in the dependent variable. If the
significance value of F is larger than say 0.05 then the independent variables do not explain the variation
in the dependent variable. The independent variables involved with every dependent variblaes (i.e.four
extracted factors) are not measured in different units, so the unstandardized coefficients or betas are an
attempt to make the regression coefficients more comparable. The t-statistics can help us to determine the
relative importance of each variable over the factor under which that items are considered.
217 | P a g e

Confirmatory Factor Analysis:
We have conducted Confirmatory factor analysis (CFA) on the result of exploratory factor analysis by
using software called analysis of moment structure (AMOS version 7). Generally we use CFA to confirm
the exploratory factor model. CFA is same as a structural equation modeling (SEM) technique. We use
CFA to determine the goodness of fit between hypothesized model & sample data. We can add a path to
a factor model based on the combined understanding of theoretical, logical and empirical knowledge.
Modification indices guide us for effective path addition to the model. Using AMOS we get idea for
addition of path to improve the goodness fit of the proposed factor model.

Figure: 1
Initial Model Derived From Exploratory Factor Analysis


218 | P a g e




Four factor solutions are derived from exploratory factor analysis (EFA). Initial EFA result a poor model
fit (Chisquare:106.355, df:48; P:.000; GFI:0.901; AGFI:0.839; PGFI:0.554; RMR:0.156; NFI:0.865;
CFI:0.919; TLI:0.888; PNFI:0.629; FMIN:0.645; RMSEA:0.086; AIC:166.355; ECVI:1.008). Based on
the modification indices (MI), covariance & regression path we have modified the result (See Figure:
2).Final model shows an excellent fit in data, where Chi-square: 31.302, df: 44; P: 0.925; GFI: 0.97;
AGFI: 0.947; PGFI: 0.547;RMR: 0.093; NFI: 0.960; CFI: 1.00; TLI: 1.026; PNFI: 0.000;FMIN:.190;
RMSEA: 0.000;AIC:99.302;ECVI: 0.602.

POWER8
POWER9
POWER14
INFLUENCE24
INFLUENCE29
INFLUENCE27
INFLUENCE5
INFLUENCE11
INFLUENCE23
POWER26
POWER6
POWER11
E1
E2
E3
E4
E5
E6
E7
E8
E9
E10
E11
E12
Intelligence
Big Bossing
Contract
Orientation
Expert
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
219 | P a g e

Figure: 2
Final Model Confirmed by Confirmatory Factor analysis


There are good number of rules for evaluating goodness of model fit between the hypothesized
model and sample data. We have used the following quasi statistical rules to judge the goodness of model
fit:
CMIN (Minimum Discrepancy): CMIN is a measure by which we try to estimate whether any
discrepancy exit between observed & estimated matrices or not. Actually it measures the difference
between the unrestricted sample covariance matrix and the restricted covariance matrix. It also measures
the likelihood of generating a
2
value which is more than
2
value in case of acceptance of null
hypothesis. However, this measure is very sensitive to sample size. Secondly, we may not consider it as a
correct model fit even of the situation when observed & estimated matrices differ significantly. In spite of
these limitations CMIN is one of the basic measures in reference to CFA (Byrne, 2010). For the purpose
POWER8
POWER9
POWER14
INFLUENCE24
INFLUENCE29
INFLUENCE27
INFLUENCE5
INFLUENCE11
INFLUENCE23
POWER26
POWER6
POWER11
E1
E2
E3
E4
E5
E6
E7
E8
E9
E10
E11
E12
Intelligence
Big Bossing
Contract
Orientation
Expert
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
220 | P a g e

of present research we have observed relatively smaller chi-square value that is 30.1 at 44 degrees of
freedom. The corresponding p value is .945. Both indicate a goodness of model fit.
Root Mean Square Residual (RMSR (Also known as RMR)): RMSR is another residual based
measure. Actually it is a mean value of the residuals derived after comparing observed & estimated
matrices. These matrices either may be of correlation matrix or may be of covariance matrices. Naturally
average value of residual nearer to zero is desirable for the model fit. For the present work we have
received the value 0.093, which is fair to accept the goodness of model fit.
Goodness of Fit Index (GFI): Another non statistical measure depicts overall degree of model fit. Value
nearer or equal to 1 indicates good fit and value nearer or equal to zero reflects poor fit. It is basically
squared residual between actual & predicted data. Thus it is insensitive to sample size. For the present
research GFI is 0.97 reflects good fit.
Adjusted Goodness of Fit Index (AGFI): Adjusted goodness fit index is an extension of GFI. By
adjusting GFI with ratio between degrees of freedom for the model proposed for and the degrees of
freedom for the model which is null, we get AGFI. More than 0.9 value is appropriate for this measure. In
the present analysis we have seen AGFI value is .946. Hence present model is well accepted.
Parsimony Goodness of Fit Index (PGFI): PGFI adjusts GFI with some parsimony measure. Higher
value is indicating a good result. In our analysis it is 0.547 that is excess of 0.5, shows good fit.
Normed Fit Index (NFI): NFI is a relative measure to examine how proposed chi-square value exceeds
from null chi-square value. Excess is equal to null
2
value is considered as perfect fit (NFI is equal to 1).
However NFI is equal to 0.9 is acceptable level for acceptance. Our analysis i shows NFI as 0.96. Thus
we accept the model easily.
Comparative Fit Index (CFI): It is a modified measure of NFI taking into consideration of sample size.
In our work CFI value is 1 which explains good fit of the model.
Tucker Lewis Index (TLI): TLI is a kind of incremental method combines with parsimony in measure.
In the same line with NFI, it measures how proposed
2
value per degree of freedom deviates from null
2

value per degree of freedom. Hence, also we consider that deviation equal to null
2
per degree of
freedom as perfect fit (TLI value is 1). However, TLI value 0.9 is also acceptable. In the present case TLI
value is 1.06 indicates a very perfect fit.

Parsimony Normed Fit Index (PNFI): It is modified measure of normed fit index. By using this
measure, we modify the model in terms of avoidance of over-fitting of data. NFI is adjusted with a ratio
221 | P a g e

between degree of freedom of the null model. Simply it means NFI per unit ratio of degrees of freedom
between proposed and null models. By this measure it is easy to compare between models and infer
whether right fitting of the data has taken place or not. PNFI value in the present case is 0.64 expresses
better fit of the model.
Minimum Discrepancy Fit Function (FMIN): (N-1) FMIN=CMIN. From this relationship it is clearly
indicating that FMIN is related to CMIN measure but it is dependent on sample size. In case of large
sample, model is not acceptable if the value of FMIN is substantially away from zero. For our work, we
have FMIN value 0.190 and which is nearer to zero, help us to infer for a acceptance of the model.
Root Mean Square Error of Approximation (RMSEA): RMSEA is best in use to confirm the model
with large sample. Like RMSR, RMSEA is to measure the difference per degree of freedom. However,
not like RMSR, RMSEA is measured on population. RMSEA considers the error of approximation in
terms of population per degree of freedom. Value less than 0.50 indicates good fit. In the present research
RMSEA is calculated as 0.000. As a result acceptability is very high.

Closeness of Fit (PCLOSE): PCLOSE is a test for closeness of fit. Actually, it tests whether RMSEA is
good in the population or not. P value less than 0.05 in non acceptable value for this test. Some
researchers have suggested that p value for this test should be more than .5. In our case p value is 0.998.
Therefore, we can conclude safely that initial model fits very well with the available data.

Akaikes Information Criterion (AIC): This is a measure by which we compare among models, where
models are varied in terms of construct numbers. Objectives of these criteria are to check the parsimony
in the model. Simply this measure estimates whether we have fit the model fit by using too many
coefficients. However, finally it is a comparative measure. We consider result which is lesser than all
comparables, as a good result. For the present case AIC & other related results are lower than the other
independent & saturated models. Hence, it is accepted as a good fit.
Expected Cross-validation Index (ECVI): We use ECVI for the purpose of checking validity of the
proposed model. It is based on the philosophy that fitted co-variance matrix is likely to respect in case of
a new sample of similar size but drawn from a single population. Thus ECVI expects minimum difference
between fitted covariance matrix of actual sample and expected covariance matrix of hold out sample for
best fit result. In case of the present research we have seen that ECVI value is 0.602 that is less than the
ECVI value of independent & saturated model. Hence, it indicates a good fit.
222 | P a g e

Hoelter Model: Hoelter Critical N (or CN) is estimating whether size of the sample is adequate enough
to produce a model which is sufficient for
2
test. According to this model any value more than 200,
indicates good result. For our case both the results are more than 300 and thus satisfy the requirement
successfully.
We present the fit indices of initial model & final model in Table 5.





Table: 5
Comparative Statement of Initial & Final Model
Models
Chi-
Square
df P RMR GFI AGFI PGFI NFI CFI TLI
Initial Model
(EFA)
106.355 48 0.000 0.156 0.901 0.839 0.554 0.865 0.919 0.888
Final Model
(CFA)
31.302 44 0.925 0.093 0.970 0.947 0.547 0.960 1.000 1.026

Table: 5 (continued)
Comparative Statement of Initial & Final Model
Models PNFI FMIN F0 RMSEA AIC ECVI
Hoelter
(.05)
Initial Model (EFA) 0.629 0.645 0.354 0.086 166.355 1.008 102
Final Model (CFA) 0.000 0.190 0.000 0.000 99.302 0.602 319

223 | P a g e

Finally we can conclude by stating that our proposed model fit in terms of i) absolute fit measures like
CMIN, FMIN, RMSR, RMSEA & ECVI ii) Parsimonious fit measures like PGFI, PNFI & AIC and iii)
Incremental fit measures like TLI, NFI. AGFI & CFI (Hair et. al., 2006). Hence, model fits perfectly.

MANAGERIAL IMPLICATION:
We have extracted three important implications vis--vis the present research work. First of all
this research helps us to develop a valid & reliable scale (instrument) that would able to measure
the mechanism of channel control by the channel leaders for various distribution environment
of developing nations similar to Indian environment. Second implication is related with
identification of strategies and their interrelation, if any, for the distribution channel over which
survey has taken place. Four important strategies are identified for creating behavioral control
over channel participants with an object to smooth functioning of the channel. Most important is
these four strategies are sharply different from proposed strategies of the existing literatures
based on developed nations. We have also experienced that two strategies are developed by
items drawn from influence strategy scale and two strategies are generated from items chosen
from power scale. It shows channel participants of developing nations also view power &
influence non-identical. Furthermore, we have seen partial interrelation exist between
intelligence and contract orientation and also between expert & contract orientation. This result
infers effect of expert on contract formation and impact of business intelligence in contract
execution.
Third implication is related to measurement of importance of each strategy to the concerned
distribution channel. We have seen importance of intelligence is highest 37%. So it is most
important channel control mechanism. Big Bossing is next important channel control mechanism
having importance almost 30%. Most interesting observation is that importance of expert power
to control channel is lowest. That may be because of high competition. All competitors are
having that power common.
Therefore, this work will provide a guideline to the managers of developing nations for the
purpose of behavioral management of their esteem channel partners. Last but not the least this
paper contributes academics of channel management by providing a customized version of
theory of behavioral management after empirical testing of the globally accepted model.














224 | P a g e














REFERENCES

Allen, Derek. R. & Rao, Tanniru. R.(2010), Analysis of Customer
Satisgfaction Data, New Age International Publisher, New Delhi , India, pp:
34-35.
Bacharach, Samuel and Edward Lawler (1980), Power and politics in
Organizations, San Francisco: Jossey-Bass.
Boyle, Dwyer (1992) Influence Strategies In Marketing Channel: Measures
& Use In Diff. Relation Ship Structures Journal Of Marketing Research,
29:462-73.
Byrne Barbara M (2010) Structural Equation with AMOS 2nd Edition,
Routledge
Campbell, D.T. & Fiske, D.W (1959), Convergent and discriminant
validation by the multitrait-multimethod matrix, Psychological Bulletin ,
56:81-105.
Darwin Cartwright (1959) A Field Theoretical Conception of Power. In D.
Cartwright (ed.) Studies in Social Power (Ann Arbor: Institute for Social
Research).
El-Ansary , Adel I, and Louis W.Stern (1972), Power Management In The
Distribution Channel Journal Of Marketing Research , 9:47-52.
Fraizer & Sheth (1985) An Attitude Behavior Framework For Distribution
Channel Management Journal Of Marketing, 49:38-49.
225 | P a g e

Fraizer, James Gill, and Sudhir Kale (1989), Dealer Dependence Levels and
Reciprocal Actions in a Channel of Distribution in a Developing Country.
Journal of Marketing , 53:50-69.
Frazier (1983) On the Measurement of Interfirm Power in Channel of
Distribution Journal of Marketing Research, 20:158-66.
Frazier and John O. Summers (1986), Perceptions of Interfirm Power and its
use within a franchise channel of distribution. Journal of Marketing
Research, 23:169-76.
French. John R. and Bertram Raven. (1959).The basis of social power. In
D. Cartwright (Ed). Studies in social power. Ann Arbor. MI: Institute of
Social Research.150-167
Ghaski J F (1984) The Theory of Power and Conflict in Channel of
Distribution, Journal of Marketing ,48:9-29.
Guha A , Chakraborty D (2002) Relative position of HDI across Indian
states: some exploratory research Artha Beekshan, 3:170.
Hair J F, Black W C, Babin B J, Anderson R E, Tatham R L (2006),
Multivariate Data Analysis 5th Ed., Pearson Education, New Delhi, India
Chapter 3,11 & appendix 11B.
Hunt, Shelby D. and John R. Nevin (1974), Power in a channel of
distribution: Sources and consequences, Journal Of Marketing Research ,
11:186-93.
Johnson Richard A. and Whichern Dean W(2008),Applied Multivariate
Statistical Analysis Sixth edition, Pearson Education , New Delhi, India,
pp:510-11.
Kale, Sudhir H.(1986) Dealer perceptions of manufacturer power and
Influence Strategies in a developing country Journal of Marketing Research,
23:387-93.
Keith J.E, Tackson D.W, JR., Crosby L.A.(1990) Effects of alternative types
of Influence Strategies under different channel dependence structures
Journal of Marketing, 54:30-41.
Lusch, Robert F.(1976) Sources of Power: Their Impact on Intrachannel
Conflict Journal of Marketing Research, 13:382-90.
Malhotra N K (2008), Marketing research- An Applied Orientation, 5 th
Ed. , Pearson Education.
226 | P a g e

Payan, Richard G Mc Farland (2005) Decomposing Influence Strategies:
Argument Structure And Dependence As Determinants O The Effectiveness
Of Influence Strategies In Gaining Chanel Member Compliance.Journal of
Marketing, 69:.66-79.
Raven, Bertram H. (1965) Social Influence and Power, In Ivan D. Steiner
and Martin Fishbein, eds., Current Studies in Social psychology. New York:
Holt, Rinchart and Winston,371-82.
Simon, Herbert(1953) Notes on the observation and measurement of
Power, Journal of politics, 15:500-18.
Swasy Jhon L (1979), Measuring The Bases Of Social Power. In William L
Willie (Ed).Advances In Consumer Research ,6:340-346, Ann Arbor MI















Appendix:

Output of AMOS

Sample size = 166
Variable Summary
Observed, endogenous variables

POWER6
POWER8
POWER9
POWER14
INFLUENCE24
INFLUENCE29
227 | P a g e

INFLUENCE27
INFLUENCE5
INFLUENCE11
INFLUENCE23
POWER26
POWER11

Unobserved, exogenous variables
E1
E2
E3
E4
E5
E6
E7
E8
E9
E10
E11
E12
Intelligence
Big Bossing
Contract_Orientation
Expert
Variable counts (Group number 1)
Number of variables in your model: 28
Number of observed variables: 12
Number of unobserved variables: 16
Number of exogenous variables: 16
Number of endogenous variables: 12
Parameter summary (Group number 1)
Weights Covariances Variances Means Intercepts Total
Fixed 16 0 0 0 0 16
Labeled 0 0 0 0 0 0
Unlabeled 10 8 16 0 0 34
Total 26 8 16 0 0 50
Models
Computation of degrees of freedom (Default model)
Number of distinct sample moments: 78
228 | P a g e

Number of distinct parameters to be estimated: 34
Degrees of freedom (78 - 34): 44
Result (Default model)
Minimum was achieved
Chi-square = 31.302
Degrees of freedom = 44
Probability level = .925

Maximum Likelihood Estimates









Regression Weights: (Group number 1 - Default model)
Estimate S.E. C.R. P Label
POWER8 <--- Intelligence .675 .117 5.754 ***
POWER26 <--- Expert 1.000
POWER6 <--- Intelligence 1.000
POWER9 <--- Intelligence 1.103 .139 7.929 ***
POWER14 <--- Intelligence .803 .122 6.571 ***
INFLUENCE24 <--- Big Bossing 1.000
INFLUENCE29 <--- Big Bossing .811 .091 8.867 ***
INFLUENCE27 <--- Big Bossing .855 .096 8.868 ***
INFLUENCE5 <--- Contract_Orientation 1.000
INFLUENCE11 <--- Contract_Orientation 1.110 .120 9.250 ***
INFLUENCE23 <--- Contract_Orientation .965 .107 9.004 ***
POWER11 <--- Expert 3.673 2.552 1.440 .150
INFLUENCE23 <--- POWER8 -.257 .069 -3.699 ***
INFLUENCE23 <--- POWER26 -.206 .067 -3.088 .002

Covariances: (Group number 1 - Default model)
Estimate S.E. C.R. P Label
Intelligence <--> Big Bossing -.187 .119 -1.571 .116
Intelligence <--> Contract_Orientation -.048 .132 -.361 .718
229 | P a g e

Estimate S.E. C.R. P Label
Intelligence <--> Expert .131 .111 1.183 .237
Big Bossing <--> Contract_Orientation .679 .173 3.912 ***
Big Bossing <--> Expert .057 .059 .955 .340
Contract_Orientation <--> Expert -.044 .056 -.795 .427
E4 <--> E2 .787 .146 5.379 ***
E1 <--> E12 -.227 .091 -2.499 .012





Total Effects (Group number 1 - Default model)
Expert Contract_Orientation Big Bossing Intelligence POWER26 POWER8
POWER26 1.000 .000 .000 .000 .000 .000
POWER8 .000 .000 .000 .675 .000 .000
POWER11 3.673 .000 .000 .000 .000 .000
INFLUENCE23 -.206 .965 .000 -.173 -.206 -.257
INFLUENCE11 .000 1.110 .000 .000 .000 .000
INFLUENCE5 .000 1.000 .000 .000 .000 .000
INFLUENCE27 .000 .000 .855 .000 .000 .000
INFLUENCE29 .000 .000 .811 .000 .000 .000
INFLUENCE24 .000 .000 1.000 .000 .000 .000
POWER14 .000 .000 .000 .803 .000 .000
POWER9 .000 .000 .000 1.103 .000 .000
POWER6 .000 .000 .000 1.000 .000 .000
Direct Effects (Group number 1 - Default model)
Expert Contract_Orientation Big Bossing Intelligence POWER26 POWER8
POWER26 1.000 .000 .000 .000 .000 .000
POWER8 .000 .000 .000 .675 .000 .000
POWER11 3.673 .000 .000 .000 .000 .000
INFLUENCE23 .000 .965 .000 .000 -.206 -.257
INFLUENCE11 .000 1.110 .000 .000 .000 .000
INFLUENCE5 .000 1.000 .000 .000 .000 .000
INFLUENCE27 .000 .000 .855 .000 .000 .000
INFLUENCE29 .000 .000 .811 .000 .000 .000
INFLUENCE24 .000 .000 1.000 .000 .000 .000
POWER14 .000 .000 .000 .803 .000 .000
POWER9 .000 .000 .000 1.103 .000 .000
POWER6 .000 .000 .000 1.000 .000 .000
230 | P a g e


Indirect Effects (Group number 1 - Default model)
Expert Contract_Orientation Big Bossing Intelligence POWER26 POWER8
POWER26 .000 .000 .000 .000 .000 .000
POWER8 .000 .000 .000 .000 .000 .000
POWER11 .000 .000 .000 .000 .000 .000
INFLUENCE23 -.206 .000 .000 -.173 .000 .000
INFLUENCE11 .000 .000 .000 .000 .000 .000
INFLUENCE5 .000 .000 .000 .000 .000 .000
INFLUENCE27 .000 .000 .000 .000 .000 .000
INFLUENCE29 .000 .000 .000 .000 .000 .000
INFLUENCE24 .000 .000 .000 .000 .000 .000
POWER14 .000 .000 .000 .000 .000 .000
POWER9 .000 .000 .000 .000 .000 .000
POWER6 .000 .000 .000 .000 .000 .000
Model Fit Summary
CMIN
Model NPAR CMIN DF P CMIN/DF
Default model 34 31.302 44 .925 .711
Saturated model 78 .000 0

Independence model 12 785.538 66 .000 11.902


RMR, GFI
Model RMR GFI AGFI PGFI
Default model .093 .970 .947 .547
Saturated model .000 1.000
Independence model .623 .527 .441 .446
Baseline Comparisons
Model
NFI
Delta1
RFI
rho1
IFI
Delta2
TLI
rho2
CFI
Default model .960 .940 1.017 1.026 1.000
Saturated model 1.000 1.000 1.000
Independence model .000 .000 .000 .000 .000
Parsimony-Adjusted Measures
231 | P a g e

Model PRATIO PNFI PCFI
Default model .667 .640 .667
Saturated model .000 .000 .000
Independence model 1.000 .000 .000
FMIN
Model FMIN F0 LO 90 HI 90
Default model .190 .000 .000 .013
Saturated model .000 .000 .000 .000
Independence model 4.761 4.361 3.836 4.931
RMSEA
Model RMSEA LO 90 HI 90 PCLOSE
Default model .000 .000 .017 .998
Independence model .257 .241 .273 .000



AIC
Model AIC BCC BIC CAIC
Default model 99.302 105.118 205.109 239.109
Saturated model 156.000 169.342 398.735 476.735
Independence model 809.538 811.591 846.882 858.882
ECVI
Model ECVI LO 90 HI 90 MECVI
Default model .602 .679 .692 .637
Saturated model .945 .945 .945 1.026
Independence model 4.906 4.382 5.476 4.919
HOELTER
Model
HOELTER
.05
HOELTER
.01
Default model 319 363
Independence model 19 21

232 | P a g e











Validity Analysis:

233 | P a g e


234 | P a g e






DEMOGRAPHIC PREDICTORS OF
CONSUMER CONCERN ABOUT AIR
POLLUTION, ORIENTATION TOWARDS
MATERIALISM, AND WILLINGNESS TO
PURCHASE INNOVATIVE NEW
PRODUCTS





William T. Neese
Bloomsburg University of Pennsylvania
wneese@bloomu.edu


235 | P a g e

Keywords: Consumers; Demographics; Environmentalism; Materialism; Innovativeness.

ABSTRACT

Research studying the impact of environmentalism on consumer purchase decisions has
increased in recent years, yet the resulting body of knowledge has produced inconsistent results.
Interactions between consumer demographics and environmentalism have been reported in
several studies, but these studies sporadically treat demographics as moderating effects,
segmentation variables, or direct predictors of a wide array of dependent variables thought to
impact environmentally-conscious consumption decisions. This study analyzes the extent to
which consumer demographic variables predict concern about air pollution, attitude toward
materialism, and willingness to try innovative new products in a purchase decision for a hybrid
automobile. Subjects were intercepted at malls physically located in the geographic market for
the dealership featured in test advertising, and exposed to one test advertisement similar to an
actual newspaper ad used by the sponsor. The eight test advertisements, consumer involvement
with automobiles, their attitude toward the advertisement viewed and the featured brand, brand
beliefs, and purchase intentions for that brand were measured and included in this analysis as
covariates to produce a cleaner test. This methodology was not reported in the literature
reviewed for this manuscript, and is unique to this study. Five of seven demographic variables
tested in a series of Multivariate Analysis of Covariance (MANCOVA) runs significantly
influenced one or more of the twelve dependent variables studied. A literature review, detailed
statistical analysis, and discussion of results are presented.
INTRODUCTION

Interest in what can be termed environmentally friendly marketing has grown over the past half century
in the United States and worldwide. Citizens of many nations are increasingly concerned that
nonrenewable natural resources such as petroleum made from fossil fuels will soon be depleted, leaving
nothing but an atmosphere heavily polluted by the likes of automobile emissions. According to Pepper,
Jackson and Uzzell (2009), research into sustainable consumption has become a burgeoning field across
many perspectives, including in economics, anthropology, psychology, sociology, human geography, and
marketing (p.126). Conspicuous consumption of certain products (e.g., luxury automobiles) has long
been known to serve as an achievement or status badge used by consumers to differentiate themselves
from others (Ivanic and Nunes 2009). Rindfleisch, Freeman and Burroughs (2000) reported that
materialistic individuals prefer an automobile high in status appeal (p.36). When consumers driven
more by an outward-focused environmental orientation are forced to reconcile with self-oriented status-
driven consumption, incongruent attitudes can materialize. Kilbourne and Pickett (2008) reported that
environmental beliefs are negatively correlated with materialism, as did Banerjee and McKeage (1994).

One comprehensive model of the human value system that effectively incorporates this conflict in
consumption goals was proposed by Schwartz (1994), who sets forth a circumplex (p.24) illustrating
how ten motivational value types are both compatible with and conflict with other values. For example,
the set of values described as Openness to Change is directly opposed to another set characterized by
conformity and tradition. The set of values that relates to hedonism and achievement for oneself are in
direct conflict with those universal values that transcend oneself, such as benevolence towards others or
caring about the environment for the collective good. Unfortunately, Schwartzs (1994) circumplex is
complicated and not specifically designed for use in marketing studies, plus the Schwartz Value Survey
(SVS) requires the use of a large array of variables to implement the questionnaire. None of the measures
included in the SVS contain hierarchy-of-effects measures needed here to capture the impact of consumer
decision-making. However, using this model to establish a general parameter, it is possible to identify a
236 | P a g e

more parsimonious set of variables commonly found in the consumer decision-making literature to serve
as dependent variables capturing the conflict inherent in purchase decisions related to the environment.

Three separate but related constructs were selected from the literature to be dependent variables in this
analysis: (1) Pollution Concern; (2) Materialism, and (3) Innovativeness. They are illustrated as a Venn
diagram in Exhibit One and verbally detailed in Exhibit Two. A summary of the Pearson correlations for
all 12 items is also presented in Exhibit One, which reveals that Pollution Concern and Innovativeness are
the most correlated with 60.0 percent of the 15 possible relationships significant, whereas Pollution
Concern and Materialism are least correlated with only 10.0 percent of the 20 possible relationships
significant. Materialism and Innovativeness are not strongly connected, either, with only 16.7 percent of
the possible 12 relationships between sets of items significant. Materialism and Innovativeness exhibits
the fewest negative correlations (41.7%), with Pollution Concern and Materialism demonstrating the most
negative correlations (90.0%). In all, this correlation analysis is consistent with the literature cited, and
provides a reasonable degree of validity for the dependent model.

EXHIBIT ONE: Dependent Constructs


Pearson Correlation Summary:
Pollution Concern + Innovativeness
Total Correlations Possible = 15
# Significant at .01 = 7 = 46.7%
# Significant at .05 = 2 = 13.3%
Negative Correlations = 9 = 60.0%
Pollution Concern + Materialism
Total Correlations Possible = 20
# Significant at .01 = 0 = 00.0%
# Significant at .05 = 2 = 10.0%
Negative Correlations = 18 = 90.0%
Materialism + Innovativeness
Total Correlations Possible = 12
# Significant at .01 = 0 = 00.0%
# Significant at .05 = 2 = 16.7%
Negative Correlations = 5 = 41.7%

EXHIBIT TWO: Dependent Variables
Pollution Concern Materialism Innovativeness
Antil (1985) Richins (1987) Leavitt and Walton (1975)
(1) Air pollution from
automobiles is a critical
problem today.
(2) Not enough is being
done to save scarce
petroleum resources.
(3) Id be willing to walk
or ride a bicycle to reduce
air pollution.
(4) More fuss is being
made about air pollution
than is really justified.
(5) I rarely ever worry
about the effects of air
(6) It is important for me to
have really nice things.
(7) People place too much
emphasis on material
things.
(8) Money cant buy you
happiness.
(9) The things people own
say a lot about how
successful they are.
(10) I might try a new and
different type of
automobile.
(11) I would rather be safe
than sorry buying a new
automobile.
(12) Hybrid automobiles
are probably just a
marketing gimmick.

Scale: 7-Point Likert
(1) Strongly Disagree
(2) Disagree
(3) Slightly Disagree
Environmentalism
(Pollution Concern)
Innovativeness Materialism
237 | P a g e

pollution.




(4) Neutral
(5) Slightly Agree
(6) Agree
(7) Strongly Agree
LITERATURE REVIEW & METHODOLOGY

A wealth of marketing literature connects the three variables in Exhibit One. One of the earliest studies
found for this review was reported by Kassarjian (1971), who uncovered some of the marketing
correlates of attitudes toward air pollution (p.61). More recently, Phillips, Englis, and Solomon (2010)
studied the link between environmentally conscious consumer behavior and innovativeness, and found
strong support for their model. According to these authors, While the issue of environmentalism is not
new, for many consumers the specific actions available to them should they wish to ameliorate ecological
problems are environmentally responsible behaviors and green products are new (p.879).

One of the most prolific contributors to topics of interest in this analysis seems to be Russell Belk, who
has published a multitude of research articles related to materialism (e.g., Belk 1985), gift-giving, and
sharing (e.g., Belk 2010). At the macro level, Huang and Rust (2011) recently report the results of a
study they conducted in a wealthy country, saying this: Taken as a whole, the findings imply that
societal consumption patterns should be sensitive to aspects of environmental impact and social justice,
even if altruistic motivations are absent (p.40). What they mean in terms of materialism at the national
level is this: the poorer the poor countries are, the less the rich countries should consume (p.40). At the
individual consumer (i.e., micro) level, Park, Burns, and Rabolt (2007) found fashion innovativeness and
materialism to be positively related to attitude toward purchasing foreign fashion goods online.

Demographic Independent Variables. Kassarjian (1971) reported the following: In regard to
marketing correlates of air pollution, it is apparent from this study that there is no simple segmentation
variable other than the attitude itself. Demographic variables such as age, sex, socioeconomic status, and
political party membership do not seem to be relevant (p.65). Since that first study, many researchers
have reported significant results concerning the variables included in Exhibit One and a variety of
demographic descriptors, yet none have modeled them together. Chan-Halbrendt, Fang, and Yang (2009)
explored the propensity of Chinese consumers to substitute plastic bags with more environmentally
friendly alternatives, and found: Based on the latent class and socio-demographic segmentation results,
there are preference distinctions among age groups (p.179). Ponchio and Aranha (2008) analyzed
materialism as a predictor of purchase behavior for low income consumers; Clark et al. (1999) discussed
the backlash against materialism among teenagers as a group, and Wei (2005) even filtered demographic
variables through cognitive age, a composite psychological variable with profound meanings (p.638).
According to Wei (2005), Results of this study suggest the mediational role of cognitive age between
consumers demographic characteristics and their innovativeness (p.638).

La Barbera and Gurhan (1997) did use several demographic variables as predictors of a dependent
variable they call Subjective Well-Being (SWB), and also modeled materialism and religiosity as
independent variables. They found partial support for the hypotheses proposed including several
demographics on SWB, but not for marital status and education. Interactions between the demographic
variables and materialism were discussed, as well as problems associated with multicollinearity caused by
the demographics. The mixed results reported in the LaBarbera Gurhan study are consistent with those
reported in several other analyses. According to Tellis, Yin and Bell (2009), Our literature review
suggests no clear consensus about the relationship between demographics and consumer innovativeness
(p.6). They ascribe these equivocal results to inconsistent definitions used across studies, a problem not
uncommon in other fields of research. Finally, these authors list six demographic variables which they
238 | P a g e

found have often been used in market segmentation studies: age, education, gender, income, mobility,
and number of children.

Hierarchy of Effects Covariates. In order to produce a more sensitive test of the true effects of age on
the dependent variables previously detailed, covariates must be included in the quantitative analysis (Hair
et al. 2010). Since this is an advertising effectiveness study, the most appropriate set of variables to
consider as covariance must include the hierarchy of effects as first established in the marketing literature
by Lavidge and Steiner (1961). Iterations of the hierarchy of effects theory extend back in time well over
one hundred years (Barry 1987), and merits of the concept have been rigorously debated for decades
(Barry 2002; Weilbacher 2001). Respondent attitudes about the advertisement and the featured brand,
their beliefs about the brand, and their purchase intentions are likely to moderate any impact demographic
characteristics will have on the dependent variables being analyzed here, so these effects will be
statistically considered along with the impact of the actual test advertisements themselves.

Consumer Involvement Covariate. Level of involvement has long been known to exert a significant
impact on consumer decision making (Zaichkowsky 1985; 1986). Generally speaking, involvement is
interest in or importance of some object, event, or idea to a person. It ranges from apathy (low
involvement) to fanaticism (high involvement), and tends to increase when the financial, social, or
physical risks of making the wrong decision are greater. Several early consumer involvement studies
compared involvement levels across product categories and found automobiles to be in the high range
(Richins and Bloch 1991; Vaughn 1986; Zaichkowsky 1987). High involvement products such as
automobiles are visibly related to social identity and status (i.e., they have badge appeal), resulting in
that purchase decision potentially having importance in a social context (Jansen and Jager, 2002). Since
an automobile is featured in the test copy used here, involvement with automobiles is also statistically
considered in the current study.

RESULTS

A mall intercept study was conducted at two different locations in the geographic area from which a
majority of customers is drawn by the actual automobile dealership featured in the test advertisements.
The demographic profile of the resulting sample is displayed in Table One.

TABLE ONE: Demographic Statistics
Label Category Frequency Percent Mean S.Dev.
Marital Status (1) Never Married
(2) Married
(3) Separated
(4) Divorced
(5) Widowed
134
119
5
16
6
47.9
42.5
1.8
5.7
2.1
1.72 .917
Household
Income
(1) Under $19,999
(2) $20-29,999
(3) $30-39,999
(4) $40-49,999
(5) $50,59,999
(6) $60-69,999
(7) Over $70,000
38
24
26
33
52
43
64
13.6
8.6
9.3
11.8
18.6
15.4
22.9
4.51 2.067
Gender (1) Male
(2) Female
136
144
48.6
51.4
1.51 .501
Age (1) Under 25Years
(2) 25-34 Years
99
52
35.4
18.6
2.59 1.549
239 | P a g e

(3) 35-44 Years
(4) 45-54 Years
(5) 55-64 Years
(6) 65-74 Years
(7) Over 74 Years
44
48
28
5
4
15.7
17.1
10.0
1.8
1.4
Education (1) Less Than H.S.
(2) High School
(3) Some College
(4) Undergraduate
(5) Graduate
8
97
100
60
15
2.9
34.6
35.7
21.4
5.4
2.92 .941
Race/
Ethnicity
(1) African Amer.
(2) Asian
(3) Caucasian
(4) Hispanic
(5) Native Amer.
4
2
271
1
2
1.4
.7
96.8
.4
.7
2.99 .364
Head of
Household
Occupation
(1) Executive/Mgt.
(2) Professional
(3) Marketing/Sales
(4) Secretary/Cleric.
(5) Agri./Forestry/Fish.
(6) Construction
(7) Manufacturing
(8) Transportation
(9) Laborer
(10) Other
(11) Retired
(12) Student
44
53
22
14
13
19
12
4
22
33
16
28
15.7
18.9
7.9
5.0
4.6
6.8
4.3
1.4
7.9
11.8
5.7
10.0
5.74 3.952
The sample includes 280 respondents who processed one of eight test advertisements, then completed a
questionnaire that captured the hierarchy of effects results presented in Table Two. Reliability ratings for
these multi-item scales all meet or exceed minimum acceptable levels (Peterson 1994). In addition, Table
Two lists the demographic variables that were significantly (<.05) adjusted by the covariates included in
this analysis. Only attitude towards the advertisement and the treatments themselves failed to covary any
independent variables. Virtually all the predictors were significantly adjusted by almost all the covariates,
a fact not previously reported in the literature reviewed.

TABLE TWO: Covariate Reliability and Significance (<.05)
Multi-Item Scale Label #
Items
Coeff.
Alpha
Significant
Covariate For
Involvement with Automobiles 15 .87 Marital Status; Income;
Gender; Age; Education;
Race/Ethnicity;
Occupation
Attitude Toward the Advertisement 9 .81 None
Attitude Toward the Brand 9 .88 Income; Gender;
Education;
Race/Ethnicity;
Occupation
Brand Beliefs 3 .74 Marital Status; Income;
Gender; Age; Education;
Race/Ethnicity;
240 | P a g e

Occupation
Purchase Intentions 3 .84 Marital Status; Income;
Gender; Age; Education;
Race/Ethnicity;
Occupation

A series of seven Multivariate Analysis of Covariance (MANCOVA) runs were made, with
Environmentalism, Materialism, and Innovativeness entered as dependent variables, one demographic
characteristic entered as the independent variable, and involvement with automobiles, attitude toward the
advertisement, attitude toward the featured brand (an actual vehicle), brand beliefs, purchase intentions,
and eight test advertisements modeled after an actual newspaper ad used by the dealership. Results of
these MANCOVA runs are included in Table Three.

TABLE THREE: Multivariate Analysis of Covariance (MANCOVA)
Effect Quantitative
Test
Value F Hypo-
thesis df
Error df Sig.
of F
MARITAL
STATUS
Pillais Trace
Wilks Lambda
Hotellings Trace
Roys Largest Root
.242
.777
.262
.110
1.401
1.401
1.400
2.383
48.000
48.000
48.000
12.000
1044.000
995.882
1026.000
261.000
.039
.039
.039
.006
TABLE THREE (continued): Multivariate Analysis of Covariance (MANCOVA)
HOUSE-
HOLD
INCOME
Pillais Trace
Wilks Lambda
Hotellings Trace
Roys Largest Root
.342
.699
.374
.148
1.317
1.319
1.320
3.214
72.000
72.000
72.000
12.000
1566.000
1398.595
1526.000
261.000
.042
.041
.040
<.001
GENDER Pillais Trace
Wilks Lambda
Hotellings Trace
Roys Largest Root
.054
.946
.058
.058
1.252
1.252
1.252
1.252
12.000
12.000
12.000
12.000
261.000
261.000
261.000
261.000
.248
.248
.248
.248
AGE Pillais Trace
Wilks Lambda
Hotellings Trace
Roys Largest Root
.425
.635
.487
.239
1.657
1.690
1.720
5.191
72.000
72.000
72.000
12.000
1566.000
1398.595
1526.000
261.000
.001
<.001
<.001
<.001
EDU-
CATION
Pillais Trace
Wilks Lambda
Hotellings Trace
Roys Largest Root
.278
.748
.304
.128
1.622
1.624
1.624
2.790
48.000
48.000
48.000
12.000
1044.000
995.882
1026.000
261.000
.005
.005
.005
.001
RACE/
ETH-
NICITY
Pillais Trace
Wilks Lambda
Hotellings Trace
Roys Largest Root
.158
.850
.168
.073
.895
.896
.898
1.587
48.000
48.000
48.000
12.000
1044.000
995.882
1026.000
261.000
.676
.674
.672
.095
OCCU-
PATION
Pillais Trace
Wilks Lambda
Hotellings Trace
Roys Largest Root
.579
.539
.660
.246
1.209
1.229
1.247
5.357
132.000
132.000
132.000
12.000
2871.000
2071.377
2741.000
261.000
.056
.043
.032
<.001

Five demographics were significant predictors at the multivariate level. Therefore, analysis of
univariate results for the dependent items is warranted. Only results significant at <.05 are
included in Table Four, and two demographic variables that did not yield significant multivariate
results are excluded.
241 | P a g e


TABLE FOUR: Univariate Statistics for Significant Multivariate Tests (Sig. <.05)
Source Dependent Variable
(See Exhibit One)
Sum of
Sqrs.
df Mean
Square
F Sig.
of F
MARITAL
STATUS
(3) Pollution Concern 31.747 4 7.937 2.609 .036
HOUSEHOLD
INCOME
(10) Innovativeness
(12) Innovativeness
26.390
38.241
6
6
4.398
6.374
2.362
3.044
.031
.007
AGE (2) Pollution Concern
(5) Pollution Concern
(8) Materialism
(12) Innovativeness
22.880
62.17
36.973
30.595
6
6
6
6
3.813
10.363
6.162
5.099
2.271
3.919
2.204
2.403
.037
.001
.043
.028
EDUCATION (2) Pollution Concern
(5) Pollution Concern
17.341
52.237
4
4
4.335
13.059
2.569
4.907
.038
.001
TABLE FOUR (continued): Univariate Statistics for Multivariate Tests (Sig. <.05)
OCCUPATION (1) Pollution Concern
(3) Pollution Concern
(10) Innovativeness
52.745
77.255
41.297
11
11
11
4.795
7.023
3.754
3.045
2.381
2.039
.001
.008
.025

For Marital Status, only one dependent item resulted in differences among respondents at the <.05 level,
and that relates to the statement Id be willing to walk or ride a bicycle to reduce air pollution. Table
Five contains the mean pattern for this item, and indicates that Married individuals are less likely than
average (i.e., Total) or any other category for that matter to agree with this statement. Married individuals
comprise 43% of the group.

TABLE FIVE: Marital Status Descriptive Statistics (Sig. <.05)
Dependent Item # Variable
Categories
Mean Standard
Deviation
N
(3) Id be willing
to walk or ride a
bicycle to reduce
air pollution.
(Sig. of F = .036)
Never Married
Married
Separated
Divorced
Widowed
TOTAL
3.63
3.29
4.20
4.25
5.17
3.56
1.842
1.833
2.168
1.844
1.941
1.865
134
119
5
16
6
280

The significant impact of Household Income on dependent variables is concentrated on the
Innovativeness construct. $20-29,999 and $30-39,999 respondents tend to be more willing to try a new
type of automobile and disagree that hybrid technology is a marketing gimmick (Table Six). However,
means are fairly dispersed across categories.

TABLE SIX: Household Income Descriptive Statistics (Sig. <.05)
Dependent Item # Variable
Categories
Mean Standard
Deviation
N
(10) I might try a
new and different
type of
automobile.
(Sig. of F = .031)
Under $20,000
$20-$29,999
$30-$39,999
$40-$49,999
$50-$59,999
$60-$69,999
Over $70,000
2.63
3.67
2.96
2.76
2.88
2.84
2.64
1.051
1.834
1.865
1.480
1.231
1.271
1.314
38
24
26
33
52
43
64
242 | P a g e

TOTAL 2.85 1.405 280
(12) Hybrid
automobiles are
probably just a
marketing
gimmick.
(Sig. of F = .007)
Under $20,000
$20-$29,999
$30-$39,999
$40-$49,999
$50-$59,999
$60-$69,999
Over $70,000
TOTAL
5.45
4.42
4.12
5.00
4.73
5.07
5.23
4.94
1.483
1.472
1.862
1.323
1.359
1.404
1.466
1.502
38
24
26
33
52
43
64
280
Table Seven contains the results for the Age demographic, and includes four dependent items for
examination: Not enough is being done to save scarce petroleum resources;
I rarely ever worry about the effects of air pollution; Money cant buy you happiness; and Hybrid
automobiles are probably just a marketing gimmick. Age is the only demographic characteristic that
produced significant differences across respondents for all three dependent variables. The first two items
relate to pollution concern, and interestingly it is middle-aged consumers that tend to agree more with
those statements. One might even interpret these results as contradictory: on the one hand, they agree
more that not enough is being done to save petroleum resources, yet on the other hand, they rarely ever
think about it. The oldest respondent group agreed most that money cannot buy happiness (perhaps
wisdom from years), whereas younger consumers tended more to believe that hybrid automobiles are just
a marketing gimmick. Again, we see the inconsistency in demographic data: on the last item, the 65-74
cohort was more in agreement with the Under 25, 25-34, and 35-44 year olds than their older
contemporaries.

TABLE SEVEN: Age Descriptive Statistics (Sig. <.05)
Dependent Item # Variable
Categories
Mean Standard
Deviation
N
(2) Not enough is
being done to save
scarce petroleum
resources.
(Sig. of F = .037)
Under 25 Years
25-34 Years
35-44 Years
45-54 Years
55-64 Years
65-74 Years
Over 74 Years
TOTAL
5.43
5.46
5.45
5.65
6.29
5.20
4.75
5.55
1.162
1.306
1.562
1.604
.937
1.304
2.630
1.362
99
52
44
48
28
5
4
280
(5) I rarely ever
worry about the
effects of air
pollution.
(Sig. of F = .001)
Under 25 Years
25-34 Years
35-44 Years
45-54 Years
55-64 Years
65-74 Years
Over 74 Years
TOTAL
4.00
4.27
4.73
4.25
5.50
5.00
4.25
4.38
1.604
1.586
1.575
1.896
1.622
1.225
2.363
1.702
99
52
44
48
28
5
4
280
(8) Money cant
buy you
happiness.
(Sig. of F = .043)
Under 25 Years
25-34 Years
35-44 Years
45-54 Years
55-64 Years
65-74 Years
Over 74 Years
TOTAL
2.31
2.73
2.45
2.15
1.43
2.40
3.25
2.31
1.536
1.911
1.861
1.650
.997
1.517
2.630
1.681
99
52
44
48
28
5
4
280
243 | P a g e




TABLE SEVEN (continued): Age Descriptive Statistics (Sig. <.05)
(12) Hybrid
automobiles are
probably just a
marketing
gimmick.
(Sig. of F = .028)
Under 25 Years
25-34 Years
35-44 Years
45-54 Years
55-64 Years
65-74 Years
Over 74 Years
TOTAL
5.24
5.08
5.11
4.42
4.32
5.20
4.25
4.94
1.378
1.412
1.316
1.541
1.827
1.643
2.363
1.502
99
52
44
48
28
5
4
280

Education only had an impact on pollution concern (Table Eight). Respondents with higher levels of
education tended to agree more that not enough is being done to preserve scarce resources and (ironically)
they rarely worry about the effects of air pollution. This apparent inconsistency mirrors Age, and likely
explains why multicollinearity problems have been reported in studies that use demographic variables as
predictors in a multiple regression analysis (e.g., LaBarbera and Gurhan 1997). As with Education,
Occupation, and Income (i.e., three distinct components of Social Class), Age and Education might be
two different variables reflecting a common influence.

TABLE EIGHT: Education Descriptive Statistics (Sig. <.05)
Dependent Item # Variable
Categories
Mean Standard
Deviation
N
(2) Not enough is
being done to save
scarce petroleum
resources.
(Sig. of F = .038)
< High School
High School
Some College
Undergraduate
Graduate
TOTAL
4.50
5.38
5.81
5.43
5.93
5.55
1.927
1.475
1.061
1.489
1.163
1.362
8
97
100
60
15
280
(5) I rarely ever
worry about the
effects of air
pollution.
(Sig. of F = .001)
< High School
High School
Some College
Undergraduate
Graduate
TOTAL
2.13
4.48
4.26
4.53
5.07
4.38
.991
1.640
1.756
1.652
1.335
1.702
8
97
100
60
15
280

Finally, data for the Occupation MANCOVA run are presented in Table Nine, and most clearly indicate
that only retired individuals are willing to walk or ride a bicycle to reduce air pollution, whereas those
employed in agriculture/forestry and transportation/shipping are least willing. Since these occupations
generally required heavy equipment operations skills, these results are perhaps somewhat predictable. In
fact, the group least likely to believe that air pollution is a critical problem works in agriculture or forestry
jobs. Interestingly, many of the blue-collar fields (i.e., Secretarial/Clerical, Agriculture, Construction,
Labor) are more innovative toward the purchase of new or different automobiles, whereas the
Executive/Manager, Professional, and Marketing/Sales fields are less in agreement with that statement
and appear to be more traditional.
TABLE NINE: Occupation Descriptive Statistics (Sig. <.05)
Dependent Item
#
Variable Categories Mean Standard
Deviation
N
(1) Air pollution Executive/Manager 5.50 1.455 44
244 | P a g e

from automobiles
is a critical
problem today.
(Sig. of F = .001)
Professional
Marketing/Sales
Secretarial/Clerical
Agriculture/Forestry
Construction
Manufacturing
Transportation
Labor
Other
Retired
Student
TOTAL
5.58
5.59
6.21
3.85
5.89
5.92
5.25
6.05
5.24
6.00
5.75
5.59
1.134
1.221
.699
2.375
.875
.900
1.708
1.090
1.347
1.592
1.206
1.360
53
22
14
13
19
12
4
22
33
16
28
280
(3) Id be willing
to walk or ride a
bicycle to reduce
air pollution.
(Sig. of F = .008)
Executive/Manager
Professional
Marketing/Sales
Secretarial/Clerical
Agriculture/Forestry
Construction
Manufacturing
Transportation
Labor
Other
Retired
Student
TOTAL
3.64
3.43
4.14
4.00
2.23
3.26
3.92
2.25
3.77
2.82
5.12
3.71
3.56
1.480
1.927
1.583
1.710
1.589
1.695
2.234
.957
1.744
2.038
1.784
1.979
1.865
44
53
22
14
13
19
12
4
22
33
16
28
280
(10) I might try a
new and different
type of
automobile.
(Sig. of F = .025)
Executive/Manager
Professional
Marketing/Sales
Secretarial/Clerical
Agriculture/Forestry
Construction
Manufacturing
Transportation
Labor
Other
Retired
Student
TOTAL
2.75
2.57
2.36
3.79
3.15
3.32
2.83
2.75
3.36
2.36
3.19
2.96
2.85
1.332
1.185
1.093
1.718
1.519
1.493
1.528
.957
1.497
.962
2.228
1.319
1.405
44
53
22
14
13
19
12
4
22
33
16
28
280



CONCLUSION

Demographic variables clearly help determine consumer predispositions towards the environment
(pollution concern), materialism, and willingness to purchase innovative new products such as a hybrid
automobile. Unfortunately, this study demonstrated once again the difficulty in deriving a clear pattern
from the data. Although statistically significant differences did emerge for five of seven variables tested,
an examination of the distribution of means for those items failed to consistently yield clear information
describing the specific results. However, these results might well be different for other studies depending
on the research parameters. One thing is consistent: the body of literature and results from the current
analysis clearly demonstrate that personal demographic characteristics do significantly influence a
245 | P a g e

consumers orientation toward the environment, materialism, and innovation. Managers need to consider
whatever combination of demographic and other variables best fit their situation as antecedents to their
marketing campaigns, because that backdrop does have a significant impact on important decision
parameters outside of the direct influence the company can exert through the marketing mix.

REFERENCES

Antil, John A. (1985), Socially Responsible Consumer: Profile and Implications for Public Policy, Journal
of Macromarketing, Fall: 18-39.

Banerjee, Bobby and Kim McKeage (1994), How Green is My Value: Exploring the Relationship
Between Environmentalism and Materialism, in Advances in Consumer Research, C.T. Allen and D.R.
John (Eds.), 21: 147-152. Provo, UT: Association for Consumer Research.

Barry, Thomas E. (1987), The Development of the Hierarchy of Effects: An Historical Perspective, in
Current Issues & Research in Advertising, J.H. Leigh (Ed.), 10(2): 251-295.

__________ (2002), In Defense of the Hierarchy of Effects: A Rejoinder to Weilbacher, Journal of
Advertising Research, 42(3): 44-47.

Belk, Russell (1985), Materialism: Trait Aspects of Living in the Material World, Journal of Consumer
Research, 12 (December): 265-280.

__________ (2010), Sharing, Journal of Consumer Research, 36 (February): 715-734.

Chan-Halbrendt, Catherine, Di Fang, and Fang Yang (2009), Trade-Offs between Shopping Bags Made
of Non-Degradable Plastics and Other Materials, Using Latent
Class Analysis: The Case of Tianjin, China, International Food & Agribusiness Management Review,
12(4): 179-197.

Clark, Andrew, Shanda Deziel, Susan McClelland, and Susan Oh (1999), How Teens Got the Power,
Macleans, 112(12): 42-46.

Hair, Joseph F., Jr., William C. Black, Barry J. Babin, and Rolph E. Anderson (2010). Multivariate Data
Analysis (7
th
Edition), Upper Saddle River, NJ: Prentice Hall.

Huang, Ming-Hui and Roland Rust (2011), Sustainability and Consumption, Journal of the Academy of
Marketing Science, 39(1): 40-54.

Ivanic, Aarti S. and Joseph C. Nunes (2009), The Intrinsic Benefits of Status: The Effects of Evoking
Rank, in Advances in Consumer Research, A.L. McGill and S. Shavitt (Eds.), 36: 12-15. Duluth, MN:
Association for Consumer Research.

Janssen, Marco A. and Wander Jager (2002). Stimulating Diffusion of Green Products, Journal of
Evolutionary Economics, 12 (3): 283-306.

Kassarjian, Harold H. (1971), Incorporating Ecology into Marketing Strategy: The Case of Air Pollution,
Journal of Marketing, 35(3): 61-65.

Kilbourne, William and Gregory Pickett (2008), How Materialism Affects Environmental Beliefs,
Concern, and Environmentally Responsible Behavior, Journal of Business Research, 61(9): 885-893.
246 | P a g e


Lavidge, Robert J. and Gary A. Steiner (1961), A Model for Predictive Measurements of Advertising
Effectiveness, Journal of Marketing, 25(6): 59-62.

Leavitt, Clark and John Walton (1975), Development of a Scale for Innovativeness, in Advances in
Consumer Research, M.J. Schlinger (Ed.), 2: 545-554. Ann Arbor, MI: Association for Consumer
Research.

Park, Hye-Jung, Leslie D. Burns, and Nacy J. Rabolt (2007), Fashion Innovativeness, Materialism, and
Attitude toward Purchasing Foreign Fashion Goods Online across National Borders: The Moderating
Effect of Internet Innovativeness, Journal of Fashion Marketing & Management, 11(2): 201-214.

Pepper, Miriam, Tim Jackson, and David Uzzell (2009), An Examination of the Values that Motivate
Socially Conscious and Frugal Consumer Behaviours, International Journal of Consumer Studies, 33:
126-136.

Peterson, Robert A. 1994. A Meta-Analysis of Cronbachs Coefficient Alpha. Journal of Consumer
Research 21 (September): 381-91.

Phillips, Diane M., Basil G. Englis, and Michael R. Solomon (2010), The Role of Innovativeness in
Environmentally Conscious Consumer Behavior, in Advances in Consumer Research, M.C. Campbell, J.
Inman, and R. Pieters (Eds.), 37: 878-880. Duluth, MN: Association for Consumer Research.

Ponchio, Mateus C. and Francisco Aranha (2008), Materialism as a Predictor Variable of Low Income
Consumer Behavior When Entering Into Installment Plan Agreements, Jopurnal of Consumer Behavior,
7(1): 21-34.

Richins, Marsha L. (1987), Media, Materialism, and Human Happiness, in Advances in Consumer
Research, M. Wallendorf and P. Anderson (Eds.), 14: 352-356. Provo, UT: Association for Consumer
Research.

__________ and Peter H. Bloch (1991). Post-Purchase Product Satisfaction: Incorporating the Effects of
Involvement and Time, Journal of Business Research, 23 (2): 145-58.

Rindfleisch, Aric, Dan Freeman, and James E. Burroughs (2000), Nostalgia, Materialism, and Product
Preference: An Initial Inquiry, in Advances in Consumer Research, S.J. Hoch and R.J. Meyer (Eds.), 27:
36-41. Provo, UT: Association for Consumer Research.

Schwartz, Shalom H. (1994), Are There Universal Aspects in the Structure and Contents of Human
Values? Journal of Social Issues, 50: 19-45.

Tellis, Gerard J., Eden Yin, and Simon Bell (2009), Global Consumer Innovativeness: Cross-Country
Differences and Demographic Commonalities, Journal of International Marketing, 17(2): 1-22.

Vaughn, Richard (1986). How Advertising Works: A Planning Model Revisited, Journal of Advertising
Research, 26 (1): 57-66.

Wei, Sheng-Chung (2005), Consumers Demographic Characteristics, Cognitive Ages, and
Innovativeness, in Advances in Consumer Research, G. Menon and A.R. Rao (Eds.), 32: 633-640.
Duluth, MN: Association for Consumer Research.

247 | P a g e

Weilbacher, William M. (2001), Point of View: Does Advertising Cause a Hierarchy of Effects?
Journal of Advertising Research, 41(6): 19-26.

Zaichkowsky, Judith L. (1985). Measuring the Involvement Construct, Journal of Consumer Research,
12 (December): 341-52.

____________ (1986). Conceptualizing Involvement, Journal of Advertising, 15 (2): 4-14, 34.

____________ (1987). The Emotional Aspect of Product Involvement, in Advances in Consumer
Research, M. Wallendorf and P. Anderson (Eds.), 14 (1): 32-35. Provo, UT: Association for Consumer
Research.


248 | P a g e









SOCIAL DOMAIN THEORY AS A
FOUNDATION FOR ASSESSING STUDENT
VIEWS OF FACULTY SOCIAL
NETWORKING INTERACTIONS



Patricia L. Nemetz
Eastern Washington University
pnemetzmills@ewu.edu


249 | P a g e

Keywords: Social Networking, Higher Education, Social Domain Theory

ABSTRACT

As educators consider using social networking sites, like Facebook, for educational innovations, they
must be aware of possible vulnerabilities associated with the blurring of social and professional
boundaries associated with such use. This research uses social domain theory to examine how students
rate the appropriateness of various faculty postings, behaviors, and responses on Facebook when used for
educational purposes. Results were consistent with expectations described by social domain theory and
previous research. Principal Component Analysis found scenarios inter-correlated within three major
components, with the extracted components showing face validity with conventional, personal choice,
and moral domains. Students generally found faculty Facebook postings related to conventional issues
more appropriate than postings related to personal choice and moral issues. Privacy concerns are
somewhat more complex, with the introduction of a privacy setting an important consideration for some
scenarios. Older students and females were more likely to find some conventional postings and personal
choice intrusions by faculty more inappropriate than younger students and males. Furthermore, students
who disagreed that faculty should be more accessible on Facebook were more likely to find conventional
and personal choice scenarios inappropriate, suggesting they would prefer a firm boundary between
faculty use of Facebook and their own social use of Facebook.

INTRODUCTION

An explosion in the use of online social networking tools suggests that student life may be undergoing
social, cultural, and moral transformations that need objective analysis for developing guidelines of
acceptable online academic behavior (Fougler, Ewbank, Kay, Popp, & Carter, 2009; Manafy, 2010).
Social networking describes web-based services that allow individuals to (1) construct a public or semi-
public profile within a bounded system, (2) articulate a list of other users with whom they share a
connection, and (3) view and traverse their list of connections and those made by others within the
system (boyd and Ellison, 2007, p. 211). Current popular social networking services include Facebook,
MySpace, Instant Messaging, and Twitter, among others. As professors consider the option of using
online social networking platforms for educational innovations, they must weigh the benefits of such
innovations against the potential vulnerabilities inherent in the use of such technologies for educational
purposes.

Several educators argue that social networking platforms hold great promise for educational innovation
through their use of informal online interactions. Social and cultural enhancements embedded in
networking technologies can bypass learning by way of passive information-absorption and facilitate
learning through the social interactions of creating, editing, collaborating, and community learning (Bull,
Thompson, Searson, Garofalo, Park, Young, & Lee, 2008; Greenhow, Robelia, & Hughes; OReilly,
2009; Twu, 2009). Both faculty and students acknowledge that such tools can enhance learning (Jones,
Johnson-Yale, Millermaier, & Perez, 2008). Additional insights from Fabos (2008) suggest that
communicating through images, sounds, and digital media, with motives for social interaction, rather than
for academic literacy, provides students with as much, if not more, information about a particular topic
than a professor may have. Professors, however, can enhance these experiences by compelling more
insightful thinking, because students do not read information with the same critical eye as their teacher
(Fabos, 2008).

The use of online social networking platforms, however, can be a minefield of problems if university
faculty fail to understand the many ways that social networking activities transcend classroom walls,
muddy the boundaries between university and non-university jurisdictions, and lead to lack of clarity
250 | P a g e

about what constitutes legitimate professional authority and responsibility (Fougler, et al., 2009).
Several hindrances to the use of social networking sites for direct educational purposes may include
concerns about privacy (Gross & Acquisiti, 2005; Manafy, 2010), the erosion of personal and
professional boundaries (Lipka, 2007, Wandel, 2008; Young, 2009), and the personal preferences of
students (Twu, 2009). Jordan (2009) describes the proliferation of digital technologies as a world in
which people live increasingly hybrid lives where the physical and the digital, the real and the virtual,
interact. In this world, online and offline identities may overlap and interdigitate, erasing prior
boundaries in social, cultural, linguistic, political, and economic domains (Jordan, 2009, p. 184). The
blurring of boundaries and the fusion of the real and the virtual in hybrid settings (Jordan, 2009, p. 174)
has profound implications for social conventions, expectations, and life experiences. In an educational
setting, the sharing of too much information with unintended individuals may be extremely detrimental
(Lego Munoz & Towner, 2009). For example, there exist numerous instances of students being
reprimanded or suspended because of unprofessional or illegal activities displayed through personal
comments, profile information, and photographs posted on social networking sites (Brooks, 2007; Lipka,
2007; Person, 2006). Some individuals feel that social networking sites should be a social space free
from formality and authority figures (Charnigo & Barnet-Ellis, 2007; Selwyn, 2009), with the incursion
of authorities viewed as an inappropriate invasion into the social world of unbridled interaction (Lego
Munoz & Towner, 2009).

Both students and faculty are vulnerable to mistakes and violations of acceptable behavior if the
blurring of boundaries creates uncertainties about privacy, authority, and the utility of social networking
technologies (Jordan, 2009). One theory that may help in the exploration of issues involved with online
social networking is Turiels (1983; 2002) social domain theory, which provides a foundation for describing
and distinguishing moral and nonmoral domains in complex social issues. It has been used extensively in
studies of student views of teachers authority and its jurisdictional limitations (Smetana & Asquith, 1994;
Smetana & Bitz, 1996). The purpose of this exploratory research, therefore, is to use social domain theory
to ascertain students views of faculty online social networking interactions so that its findings might
inform educators decisions regarding the if and how of best integrating social networking in the
classroom.


LITERATURE REVIEW

Research on the use of social networking includes studies showing both benefits and pitfalls. Several studies
have examined how students use social networking. These studies indicate that students use social
networking sites to maintain or strengthen their offline social networks (boyd & Ellison, 2007), and that they
are a primary tool of communication and socialization among college students (Golder, Wilkinson, &
Huberman, 2008). A large study commissioned by the Educause Center for Applied Research (ECAR)
surveyed college students and found that 26.9% of undergraduate students spent 6-10 hours a week on social
networking sites and 55.8% spend 5 hours or less a week (Salaway, Caruso, & Mark), with Facebook being
the most popular, and Myspace and Bebo of secondary importance. Among college-age students, adoption
rates of Facebook are reported to be 85-96% (Salaway, Caruso, & Mark, 2008; Stutzman, 2008; Tufekci,
2008). Students use social networking sites mostly for maintaining a social connection to friends, but also
for exchanging information, learning about others, expressing opinions, dating, and using applications
(Agarwal & Mital, 2009; Lampe, Ellison & Steinfield, 2006; Madge et al., 2009; Raacke & Bonds-Raacke,
2008; Salaway, Caruso, & Mark, 2008). Research based on content analysis of students postings shows
highly intimate online communications, with rich thematic elements, such as family and social issues, risk-
behavior admissions, disclosure of personally identifiable information, frequent peer interaction, the use of
sexual and profane language, and intimate, candid discussions (Williams & Merten, 2008).

251 | P a g e

Research has found that education and education-related interactions are the subject of social
networking interaction across all student age groups. Students are better able to communicate with other
students about coursework, and such interaction increase class satisfaction and provide a sense of
community among students (Beaudoin, 2002; Chu & Meulemans, 2008; Dawson, 2006; Driver, 2002;
Greenhow & Robelia, 2009a, 2009b; Jones et al., 2008; Kaya, 2010; Madge et al., 2009; Salaway, Caruso,
& Mark, 2008). Another study found that students prefer Facebook to the use of course management
software, like Blackboard, for discussions (Chu & Meulemans, 2008. In addition, some empirical
evidence indicates that the use of social networking sites increases competencies in technology, creativity,
and communication (Greenhow & Robelia, 2009b).

Critics cite several concerns about combining education with social networking. Most frequently mentioned
as reasons not to use social networking in a classroom setting are issues related to safety, privacy, and
erosion of professional boundaries (Cain, 2008; Griffith & Liyanange, 2008; Hewitt & Forte, 2006). The
ECAR undergraduate survey indicates that about 55% of students are concerned about the misuse of
personal information, with females and older students being more concerned (Salaway, Caruso, & Mark,
2008). One in eight college women report having been stalked, so the possibility that social networking
might facilitate cyberstalking by miscreant faculty or fellow students is one concern (Kirkland, 2010). High-
profile college-campus incidents that result in disciplinary action or unintended consequences may also
hinder students from granting access to their online profiles beyond their intimate social circles. Past
campus incidents include athletic suspensions or dismissals for inappropriate verbal and photographic
postings (Armour, 2006; Brooks, 2007; Drew, 2010), academic suspensions for postings of inappropriate
photographs of selves, other students, or dignitaries (Gruss, 2007; Iyengar, 2006), and loss of prospective
jobs for posting risqu online persona (Finder, 2006).

Attempts to train future faculty about violations of safety and privacy found divergent views among
faculty-in-training about appropriate boundaries and moral impositions related to activities normally outside
the realm of faculty and student interaction. One study that used case studies based on Turiels (1983)
social domain theory as a training device met with considerable ambiguity from faculty-in-training that led
to a lack of clarity about what constitutes legitimate teacher conduct and authority on social networking
sites (Fougler, et al., 2009). This study lends support to the notion that blurring between the physical and
virtual realms of social institutions can lead to social transformations fraught with uncertainty (Jordan,
2009). Concerns about student-teacher online relationships have led several school districts at the K-12
level to regulate or ban teachers from using online social networks for educational purposes (Affleck, 2010;
Bowean & Mack; 2010; Garrow, 2010; Kieffer, 2010). Reports of university faculty gaffes or misuse are
fewer, but universities have issued guidelines on the use of social networking when representing the
university (EWU Board of Trustees; 2010).

Research findings on college students acceptance of direct faculty access to their social networking sites
are somewhat inconsistent and ambiguous. Roblyer, McDaniel, Webb, Herman, & Witty (2010) found that
students are much more likely than faculty to use Facebook and are more open to the use of social
networking sites to support classroom work. Other empirical data indicates that faculty presence on
Facebook, with profiles rich in self-disclosure, increase student motivation, affective learning, and teaching
credibility (Mazer, Murphy, and Simonds, 2007; 2009). In one survey (Hewitt and Forte, 2006), students
found faculty presence on Facebook acceptable, with males being more accepting than females. Students
are satisfied with faculty presence because of the availability of alternative means of communication, the
ability to know faculty better (Hewitt and Forte, 2006), and the satisfaction of having virtual office hours
available (Li & Pitts, 2009). But other research found that while students are highly favorable toward the
use of social networking sites for social purposes, they are more tentative about their use for utilitarian
purposes, such as work and education ((Nemetz, Aiken, Cooney, and Pascal, 2011). Furthermore, it
appears that the degree of social connectedness inherent in students self-schema (Wang and Mowen, 1997)
is a factor in their judgment of a faculty presence. More socially-connected, or less independent, students
252 | P a g e

are tentative about faculty presence in their networks for social purposes, but are favorable toward a faculty
presence for utilitarian purposes, like virtual office hours and improved accessibility (Nemetz, Aiken,
Cooney, and Pascal, 2011). Little research is available that assesses student judgment of specific faculty
behaviors on social networking platforms, hence, the intent of this research.


RESEARCH FOUNDATION

As a foundation for evaluating how students might assess faculty behaviors, requests, posts, and
responses when using social networking platforms for educational purposes, this research uses social
domain theory (Turiel, 1983). Social domain theory (Nucci, 2001; Smetana, 1995, 2002, Turiel, 1983,
1998) has been used to demonstrate that individuals have different types of social interactions and that
these varied interactions lead to the development of different domains of social knowledge. Specifically,
people think about moral matters, conventional matters, and personal matters in different ways (Davidson,
Turiel & Black, 1983; Nucci, 1981; Smetana, 1988; Smetana, 2006). Moral issues are acts that pertain to
others rights or welfare (such as notions of harm, fairness, and rights). In a university setting, a moral
issue might involve the decision of whether to cheat or not on a test. Conventional issues refer to the
arbitrary and agreed-upon uniformities in social behavior that are alterable and context-dependent (such
as social roles, institutional organization, and matters of social efficiency) (Nucci, 1996; Willard, 1997).
For example, it may be a norm at most universities for a student to find a seat and remain in that seat
during the entire class period. Furthermore, professors may prohibit late arrival to the classroom in order
to facilitate uninterrupted and heightened attention on the subject matter in a lecture. The expectations
here are context-dependent in that students give legitimacy to these issues in a classroom setting, but may
not be willing to respond to similar norms in a different setting, such as at a football game. Personal
issues have consequences only to the actor and are thus viewed as beyond societal regulation and moral
concern (such as control over the body, and preferences and choices regarding personal appearance,
friends, and hobbies) (Nucci, 1996, 2001). A personal issue might be the choice for a student to wear a
beard, long hair, shorts, and sandals to class.

Some issues involve domain overlap; these issues raise moral concerns as well as concerns about social
conventions or personal choice, and are known as multifaceted issues (Nucci, 1989). A multifaceted
issue in a university setting might involve rules against setting off fire alarms in dormitories. Here, there
is a moral issue of using emergency resources to respond to a prank when those resources might be
needed for a legitimate emergency elsewhere, as well as the conventional issue of disrupting student
residents who need to respond by leaving the dormitory. One additional subset domain involves issues
that are considered prudential, or prudentially advisable, and include decisions about safety, health, and
comfort. Examples in this category are decisions about smoking, alcohol, and unsafe driving (Smetana
and Asquith, 1994). Evaluations of these prudential issues might be considered overlapping in moral,
conventional, and personal domains, and their domain evaluation is somewhat age-dependent (Smetana &
Asquith, 1994).

Both the age of students and the jurisdictional boundaries of the authorities involved are relevant
when understanding student judgments of social domains. Research has examined the age-related
changes and the ways in which people reason about moral and nonmoral concerns and found patterns of
development and understanding consistent with limitations on both authority and jurisdiction as students
age (Nucci, 2001). Numerous studies have assessed reasoning about role-related authority, especially that
of the scope and limits of parent and teacher authority (Laupa, 1991; Laupa, 1995; Laupa & Turiel, 1993;
Smetana & Asquith, 1994; Smetana & Bitz, 1996; Smetana, Campione-Barr, and Daddis, C., 2004;
Smetana and Chuang, 2001; Smetana & Daddis, 2002). These studies, involving children and
adolescents, indicate that as students age, they are more likely to judge teachers legitimate authority as
253 | P a g e

limited to the boundaries of the school context. For example, among high-schoolers, issues such as drug
and alcohol use are viewed as personal/prudential decisions that are outside of school jurisdiction (Nucci,
Guerra, & Lee, 1991), unless the use occurs within the confines of the school (Smetana & Bitz, 1996).
Considering the more mature ages of college students, as well as the increased independence from the
parental domicile, such issues are even more likely to be viewed as personal/prudential decisions outside
the bounds of university authorities, unless the usage occurs in university or university-related contexts.
It is a reasonable expectation that college students would view university authority as extremely limited to
mostly conventional issues and a few moral issues that involve university matters.

The uncertainties created by the blurring of boundaries when real and virtual worlds merge for
educational purposes create additional ambiguities about the proper role of faculty on social networking
sites (Jordan, 2009). Technology leaders have even gone so far as to state that privacy is no longer a
social norm (Mark Zuckerberg, in Manafy, 2010), thus blurring any reasonable expectation of privacy in
many settings. Given that social networking sites are currently used by students primarily for social
purposes, and that those social interactions include content that might be considered risqu or
inappropriate in formal settings, educators might find themselves exposed to information about students
normally kept outside official university boundaries. Conversely, faculty who mix the use of social
networking sites for both social and educational purposes may find themselves exposing information that
the university expects them to keep private. For example, faculty in state-supported universities are often
prohibited from expressing religious and political preferences in the classroom as a matter of policy to
avoid legal entanglements (EWU Board of Trustees, 2010). If such information is posted in the profile of
the faculty member, the posting itself may be a violation of policy, and in addition, may impact
educational effectiveness by creating a response bias in students familiar with the preferences of faculty.
Even if privacy and group settings are used to create some boundary between social and educational use,
profile information is far more accessible online than in other contexts.

Besides jurisdictional ambiguity, some evidence exists to suggest that students and faculty view some
moral matters differently in technological contexts than in non-technological contexts. Willard (1997)
posits that actions taken in cyberspace are distanced from the resulting harm, and that distance may
impair a persons ability to discern an underlying moral value. For example, one study found that reading
someones personal diary was considered an invasion of privacy by a much larger number of participants
than was recording a live image video of a person in a public setting filmed without the persons
knowledge (Friedman, Kahn, Hagman, Severson, & Gill, 2006). Furthermore, activities such as illegal
file sharing, piracy, and hacking (Friedman, 1997) often take place extensively in university settings,
while similar, but more visible activities, such as shoplifting, theft, and destruction of property are viewed
as obvious moral violations. Friedmans (1997) research revealed that online violations are a
consequence of the perceived distance between the actor and potential victims, the indirect nature of the
actions, and the lack of established consequences for such behavior online.

In summary, a multitude of uncertainties, ambiguities, and contingencies must be considered when
examining how students assess faculty behaviors, requests, posts, and responses when using social
networking platforms for educational purposes. While hypothesizing exacting relationships about such
assessments is difficult for purposes of this research because social domain theory typically uses scenarios
and in-depth interviews for its methodologies (Nucci, 2001), several expectations about college-student
views might be suggested to guide this exploratory research. It is not unreasonable to expect that college
students would view a fairly narrow scope of online activities as legitimate concerns of faculty when
using social networking sites for educational purposes. In particular, those activities related to
conventional matters of organization and educational proficiency would be viewed with the most
legitimacy, while most moral issues (other than those related to education itself) and personal issues
would be viewed as beyond faculty purview most of the time, particularly if they are handled in a manner
that neglects concerns about privacy. Furthermore, there may be enough uncertainties that some students
254 | P a g e

may prefer to place their own boundary on their social networking by avoiding the use of social
networking sites for educational purposes altogether. In addition, there may be a few issues that are
viewed differently due to different life experiences, so age and gender may play a role in the assessments.


METHODS

Sample: A survey that included several social-networking-usage questions, demographic questions, and
brief scenarios was created, and then tested for ease of use by discussing items in-depth with several
students. After making a few minor corrections based on student feedback, the final survey was
uploaded to an online survey service and administered to two undergraduate business classes at a
medium-sized state university in the northwestern United States. The classes were chosen based on their
relatively large size and experience using a blended learning platform. Because the universitys
distance-learning program is fairly small, the blended leaning classes were chosen so that research
respondents had at least some familiarity with aspects of online learning. A large portion of the class was
taught using Blackboard course management software, which included video lectures, online reading
materials, embedded class activities, and online grading and feedback. In-person class attendance was
optional and covered the same lecture material as that in the online videos. Social networking sites were
not used for educational purposes for these classes, but students accessed the professor through e-mail on
a regular basis and vice versa.

The convenience sample included 110 students. The gender reported was 48% female and 52% male.
Respondents reported an average age of 27, and age ranged from 20 to 47. Reported GPA averaged 3.3,
which is representative of typical class GPA averages for this program, given that students must have a 2.75
GPA to be accepted into the program. Summary statistics for respondents are listed in Table 1.

Table 1

Summary Statistics and Frequencies for Respondent Sample

N Min Max Mean Std. Dev.
Age 108 20 47 26.94 7.102
Reported GPA 108 2.7 4.0 3.34 0.381

Measures: The questionnaire itself had three main parts. First, respondents read a short introduction
that guaranteed anonymity, conveyed the voluntary nature of completing the survey, and assured no
repercussions for choosing not to complete the survey. Second, subjects were asked their opinion about
the usage of Facebook, a popular social networking site, for educational purposes. Next, they were
presented with 9 brief scenarios that described fictional behaviors of faculty while using Facebook for
educational purposes. They were asked to rate the behaviors using a seven- point Likert-type rating scale,
with Extremely Inappropriate anchoring a score of 1 and Completely Appropriate anchoring a
score of 7. They were asked to rate each scenario in two different conditions if the behavior occurred
when Facebook was open to all contacts on their accounts, and if the behavior occurred when Facebook
was open only to those listed in a group set up for members of an online class. After rating the
scenarios, the respondents were again asked their opinion of faculty using Facebook. Third, the survey
inquired about the subjects age, gender, estimated GPA, and international-student status. The scenarios
are listed in Table 2.

Table 2

255 | P a g e

Facebook Scenarios Depicting Fictional Faculty Behaviors*


1. After a reading an article on affirmative action, your professor requires you to post your
thoughts on racism.
2. Your professor announces on the Wall that (s)he made a mistake in the assignment, and
problem number 12, not number 13, is due on Monday.
3. Your professor takes a poll to determine if more students prefer the test on Tuesday or on
Wednesday.
4. Your professor posts a wall message that says congratulations on being excellent students and
finishing a difficult class
5. You post a message that says End of quarter time to party! Your professor comments on
your Facebook, I could use a break from all the grading. Wheres the action?
6. You posted that your family and friends are coming to take you out for your 21
st
birthday party
next month. Your professor sees your Facebook profile picture, which shows you obviously
drinking alcohol and being pretty tipsy. Your professor comments that you should be reported
to the dean of students for underage-drinking disciplinary action.
7. You and several students have a thread of discussion going about what classes to take next
quarter and who to take them from. Your professor jumps in and tells you who has a bad
reputation as a professor.
8. Your professor posts a comment that your lip ring looks unprofessional.
9. Your professor posts the affirmation, Jesus Christ is my Lord and Savior, with the option for
you to like or comment

*Likert-scale ratings: 1=Extremely Inappropriate 7=Completely Appropriate
Scenarios were rated in conditions of 1. Openness to all contacts; and; 2. Openness to only class contacts


Analysis: No a priori hypotheses were presented for analysis, however, the statistical technique of
principal components analysis was used to find social domain commonality within the scenarios. With
social domain theory as the research foundation, it was expected that issues would be understood on the
basis of fit with specific social domains. Through a process of reasoning based on certain criteria (rule
contingency, rule alterability, rule generalizability, act generalizability, and act severity), researchers have
determined that people make a conceptual distinction among conventional, moral, and personal issues
(Nucci, 2001). Based on these distinctions, scenarios can be evaluated for common patterns of
conceptualization that match the domains. In essence, some commonality within scenarios elicit s
responses that result in a combination of those scenarios into the same social domain. A technique that is
used to analyze groups of correlated responses that represent one or more common domains is principal
components analysis (PCA) (Henriques, 2010). PCA is used to find optimal ways of combining
responses into a small number of subsets in order to explain a maximal amount of variance (Suhr, 2005).
This statistical technique is to be distinguished from factor analysis, which is more typically used in
psychometrics for purposes of finding an underlying construct. Social domain theory makes no claim of
an underlying construct; rather it suggests common conceptualizations of issues within each domain.
Principal component analysis also makes no claim of finding an underlying construct, but rather finds
groups which have the greatest amount of commonality explained by the response set (Henriques, 2010).
Principal components analysis (with varimax rotation and Kaiser normalization), therefore, was the
technique used to determine how the scenarios grouped into specific social domains.

In addition to principal components analysis, several correlations and t-tests were used to
determine if student assessment of faculty behaviors within specific domains correlated with gender, age,
and opinions of using Facebook for educational purposes.
256 | P a g e



RESULTS

Summary statistics for student responses to the scenarios are shown in Table 3 for two conditions:
when a students Facebook interactions are open to all the students contacts, and when the interactions
are open only to a class in a Facebook group.

Table 3

Means and Standard Deviations for Student Ratings of Each Facebook Scenario*

Open to All Open to Class
Only
Scenario Min Max Mean St.
Dev
Mean St. Dev
Racism Assignment
Assignment Change Announcement
Drinking Violation Comment
Party Information Request
Excellent Students Comment
Prof Reports Reputation of Other Profs
Test Preference Poll
Lip Ring Comment
Jesus Personal Savior Statement
1
1
1
1
1
1
1
1
1
7
7
7
7
7
7
7
6
7
2.74
4.09
1.94
2.56
4.78
2.09
4.67
1.70
2.78
1.896
2.245
1.835
1.940
2.405
1.754
2.162
1.277
2.159
3.93
4.83
1.98
2.69
5.39
2.13
5.20
1.94
2.63
2.278
2.299
1.798
2.143
2.216
1.686
2.086
1.503
2.199
* 1=Extremely Inappropriate; 7=Completely Appropriate

For the open-to-all-contacts condition, students rated the appropriateness of the scenarios in the following
order, from least appropriate to most appropriate: lip ring comment, drinking violation comment,
professor reports reputation of other professors, party information request, racism assignment, Jesus
statement, assignment change announcement, test preference poll, excellent students comment. The least
variation in response occurred for the lip ring comment (sd=1.277), and the most variation occurred for
the professor reports reputation of other professors comment (sd =2.405). The order changes for the
open-only-to-class condition, where the Jesus statement moves from being the 6
th
least appropriate
scenario in the open-to-all condition to the 4
th
least appropriate scenario. Except for the Jesus
statement, all scenarios became more appropriate when students answered in the open-to-class-only
condition than in the open-to-all condition. This result suggests that when a moderate privacy barrier is
created, students recognize it as a minor remedy for generally inappropriate online interactions. In both
conditions, all scenarios, except the three dealing with class information, are rated on the inappropriate
end of the scale.

The results for the principal components analysis are shown in Table 4. The table lists the
correlations of the ratings for each scenario with the components extracted. The principal components
(PCs) were named for the items most strongly correlated with them (shown underlined in the table). The
first principal component represents the conventional domain with high correlations for the ratings of the
following scenarios: assignment change announcement (r=0.782), excellent students comment
(r=0.799), and test preference poll (r=0.906). These activities can be considered typical organizational
or motivational strategies of faculty. The second PC, named the personal choice domain, is associated
with activities that normally involve personal choice by students, but the scenarios describe some type of
faculty intrusion into that choice. Its highest correlations are with ratings for the following scenarios:
party information request (r=0.744), professor reports the reputation of other professors (r=0.823),
257 | P a g e

and lip ring comment (r=0.742). The third PC is named for the moral domain and shows the strongest
correlations with ratings for scenarios about reporting drinking violations (r=-0.566) and exposing
ones religion as a professor (r=0.813). The racism assignment scenario does not have a correlation
above 0.5 with any domain, suggesting it might be a multifaceted issue evaluated by students as
belonging in the conventional and personal domains. The relatively even distribution of correlation
coefficients for the ratings of the drinking violation scenario in each of the three domains also suggests
it may be evaluated by students as belonging to conventional, personal, and moral domains. Interestingly,
the racism assignment has a mean on the inappropriate end of the scale (m=2.74), even in the
condition where interactions are exposed to class members only (m=3.93).

Table 4

Rotated Component* Matrix for Facebook Scenarios


Scenario
Dominant Component Domain
Conv Pers Moral
Racism Assignment
Assignment Change Announcement
Drinking Violation Comment
Party Information Request
Excellent Students Comment
Prof Reports Reputation of Other Profs
Test Preference Poll
Lip Ring Comment
Jesus Personal Savior Statement
.487
.782
.408
.202
.799
.128
.906
.024
.323
.429
.158
.436
.744
.192
.823
-.023
.742
.208
-.179
.122
-.566
-.023
-.053
.285
.204
-.069
.813
*Extraction Method: Principal Component Analysis.

To determine if the naming of the principal component domains had face validity, a brief survey
was administered to 10 additional students subsequent to administering the first survey. In this survey,
written explanations of conventional, personal, and moral domains were given to the students. The 9
scenarios for this research were then listed, and students were asked to indicate if they thought the
scenario involved mostly conventional, mostly personal, or mostly moral issues. If they thought it
involved several domains without one dominating, they were asked to list which domains were relevant to
that scenario. Students were nearly unanimous in their assessments of the scenarios, and those
assessments were nearly the same as the numerical results using PCA. Students also agreed that the
racism assignment and drinking violation scenarios were multi-faceted issues, with racism
assignment scenario belonging to the conventional and personal domains, and drinking violation
scenario belonging to all three domains. The only disagreement occurred with the Jesus statement.
Three students thought it was a personal issue, while 7 students thought it was a moral issue. Discussion
indicated that the 7 students who chose the moral domain were concerned that such a statement might
violate freedom of religion rights and were concerned that if they saw that another student liked the
comment, they might feel obligated to like the comment as well, or risk falling out of favor with the
professor. The three students who thought it was a personal issue indicated that they believed the
professor had a right to express their religion, which they viewed as a personal choice.

Of some interest are the average means for the scenario ratings that correlate most highly with each PC
domain. For the three scenario ratings that most highly correlate with the conventional PC domain, the
average of the means is 5.416 (falling on the appropriate end of the scale), while the averages of the
258 | P a g e

means for the personal choice PC domain and moral PC domain are 2.117 and 2.349, respectively (falling
on the inappropriate end of the scale). In general, these results suggest that faculty interactions that
involve behaviors that fall within the conventional domain are viewed as more appropriate than those that
fall in the personal choice and moral domains. These findings are in alignment with previous literature
(Laupa, 1991; Laupa, 1995; Laupa & Turiel, 1993; Smetana & Asquith, 1994; Smetana & Bitz, 1996;
Smetana, Campione-Barr, and Daddis, C., 2004; Smetana and Chuang, 2001; Smetana & Daddis, 2002),
as well as with the expectation of social domain theory that forms the research foundation.

The remaining statistical analyses were completed by using principal components scores, as
suggested by Suhr (2005). To determine if age and grade point average have any effect on
appropriateness ratings of scenarios, Pearson correlations were computed for age and GPA for each
PC score. Results are shown in Table 5. The only significant correlation is age with the personal
choice PC, and it is negative (r=-0.217; p<0.01). As age increases, ratings of the appropriateness of
interference with personal choice decreases, a result also in alignment with previous research (Nucci,
Guerra, & Lee, 1991; Smetana & Bitz, 1996). One additional correlation was computed to determine if
student agreement that faculty should be more accessible on Facebook related to their ratings of scenario
appropriateness in each domain. Results suggest that the less students agreed that faculty should be
accessible on Facebook, the more likely they were to rate scenarios in the conventional domains and
personal choice domains as inappropriate (r=0.346, p<0.01; r=0.412, p<0.01, respectively).





Table 5

Pearson Correlations for Age and GPA with Principal Component Scores

Age GPA More Faculty Access
Conventional PC
Personal Choice PC
Moral PC
-0.132
-0.217*
-0.050
-0.044
-0.021
0.049
0.346**
0.412**
-0.120
**p0.01; *p<0.05

To determine if gender had an effect on appropriateness ratings, t-tests were performed on each
PC, then on the ratings for each scenario. Results are shown in Tables 6 and 7, respectively. Results are
significant for the Conventional PC Score and Personal Choice PC Score. The t-test results for each
scenario rating show significant differences between men and women for the following scenarios: racism
assignment, drinking violation comment, professor reports reputation of other professors, test preference
poll, and lip ring comment. For all these scenarios, women rated the scenarios as significantly more
inappropriate than men.

Table 6

T-Test Results for Gender Differences of Each Principal Component Score

Gender
N Mean Std. Dev. F
Conv PC Score Female 50 14.181 6.668 4.820*
Male 56 15.823 5.576
259 | P a g e

Pers PC Score Female 50 7.669 3.456 3.766*
Male 56 9.815 4.800
Moral PC Score Female 50 2.621 2.504 0.244
Male 56 1.966 2.342
*p<0.05


In summary, results were consistent with expectations described by social domain theory and previous
research. Principal Component Analysis found scenarios inter-correlated within three major components,
with the extracted components showing face validity with conventional, personal choice, and moral
domains. Students generally found faculty Facebook postings related to conventional issues more
appropriate than postings related to personal choice and moral issues. Privacy concerns are somewhat
more complex, with the introduction of a privacy setting an important consideration for some scenarios,
such as the racism assignment. Older students and females were more likely to find some conventional
postings and personal choice intrusions by faculty more inappropriate than younger students and males.
Furthermore, students who disagreed that faculty should be more accessible on Facebook were more
likely to find conventional and personal choice scenarios inappropriate, suggesting they would prefer a
firm boundary between faculty use of Facebook and their own social use of Facebook.


Table 7

T-Test Results for Gender Differences of Each Facebook Scenario Rating

Gender
N Mean Std. Dev.
F
Racism Assignment Female 52 2.27 1.573 10.732**
Male 56 3.18 2.072
Assignment Change Female 50 4.28 2.365 1.627
Male 56 3.93 2.139
Drinking Violation Female 50 1.64 1.367 9.785**
Male 56 2.21 2.147
Prof to Party Request Female 52 2.42 1.923 0.000
Male 56 2.68 1.964
Excellent Student
Comment
Female 52 4.35 2.520 3.520
Male 56 5.18 2.241
Prof Reports
Reputation of Other Profs
Female 52 1.73 1.206 8.795**
Male 56 2.43 2.096
Test Preference Poll Female 52 4.42 2.396 6.044*
Male 56 4.89 1.913
Lip Ring Comment Female 52 1.42 .893 11.398**
Male 56 1.96 1.513
Jesus Comment Female 52 3.00 2.187 0.561
Male 56 2.57 2.131
*p<0.05; **p<0.01


DISCUSSION
260 | P a g e


As social networking technology evolves, and its functionality increases, faculty are left with the
decision to embrace a technology that seems increasingly relevant to students, or to avoid it in favor of a
more educationally-dedicated platform. For students, the familiarity and ease of use associated with
social networking sites may justify expanding the use of such sites for educational purposes. The
hesitations, concerns, and mistakes of such expansion, however, must also be be considered as the
blurring boundary created by the digital realm transforms cultural, social, and professional expectations.
This research was driven by such concerns and sought some guidance through the use of social domain
theory as an analytical tool for exploring student assessments of faculty postings, responses, and
behaviors when using Facebook for educational purposes.

Results suggest that social domain theory does offer some limited, but clarifying, guidance.
Through principal components analysis and the use of 9 fictional scenarios rated for appropriateness by
students, the research found high inter-correlations among issues designated as belonging to conventional,
personal choice, and moral domains. A subsequent test for face validity found strong indications that the
designations for the extracted components were in agreement with student designations for the domains
relevant to issues in the scenarios. Furthermore, as suggested by social domain theory, appropriateness
ratings for faculty involvement in conventional issues were found to be higher than those for personal
choice and moral issues, which were found to be highly inappropriate. Faculty involvement in
conventional issues tended to be designated as even more appropriate when the use of Facebook included
a privacy barrier in the form of a group setting allowing only class members to see postings. Older
students and females were more discriminating about appropriate behavior than younger students and
males, with several scenarios involving conventional issues and several involving personal choice rated as
more inappropriate by them. Such findings are all in alignment with expectations posited by social
domain theory and previous research (Laupa, 1991; Laupa, 1995; Laupa & Turiel, 1993; Smetana &
Asquith, 1994; Smetana & Bitz, 1996; Smetana, Campione-Barr, and Daddis, C., 2004; Smetana and
Chuang, 2001; Smetana & Daddis, 2002). In addition, students who disagreed that faculty should be
more accessible through Facebook were more likely to find conventional and personal choice scenarios
more inappropriate, suggesting that they wished to be free of a faculty presence on Facebook, regardless
of the facultys purpose.

Perhaps of more interest are findings related to the scenarios that did not fit as neatly into
designated domains. The racism assignment scenario might be representative of a situation into which
faculty might venture easily, unaware of potential controversy, and blinded by the normalcy of requiring
such an assignment in a classroom setting. In-depth discussions with students about that scenario
revealed their reluctance to post their opinions about controversial topics, both in plain view of all their
contacts and in plain view of just their classmates. Writing about a controversial topic to be viewed only
by the eyes of the professor was felt to be a much safer experience than posting their thoughts in
writing for all to see. Even if they were asked to participate in a classroom discussion on a controversial
topic, students thought it was different than posting their views because they believed the professor could
be more directive and control potential outbreaks and inflammatory arguments in a classroom setting,
moreso than online. They also believed that a classroom discussion usually involved voluntary
participation, which was more palatable than a required posting. In assessing this scenario, several
students acknowledged they were stymied by the fact that they saw the conventional aspect of a required
assignment as important to their success in class, while also believing their views of the topic and
willingness to discuss them in front of others were a personal choice, hence the inappropriateness rating.

The complexity of their assessments was also evident in the drinking violation and Jesus
statement scenarios. Several students were sympathetic to the conventional and moral issues present in
the drinking violation scenario and acknowledged that faculty would be in a tight spot if they were
required by their position to report illegal activities if they saw them. On the other hand, they believed a
261 | P a g e

posting from a setting outside the university, from the past, with no evident disruption to the immediately
surrounding community, in which a student near the legal drinking age is shown drinking alcohol, did not
rise to the level of needing to involve authorities. From the students perspective, it was seen as a personal
choice outside the jurisdiction of university personnel. Discussions indicated that nuanced thinking about
this issue, from both the faculty and student perspectives, led the secondary sample of students (those
responding for purposes of face validity) to designate the issues as involving all social domains. Still,
many of them believed faculty should use their better judgment, and viewed reports to the authorities as
inappropriate, just as the primary sample did. For the scenario involving a professor posting their
religious beliefs, nuanced thinking was also evident. For those students who viewed the scenario mostly
as a personal issue, they were influenced mostly by the public nature of most profile postings. They
believed the professor had a right to religious expression, particularly when open to the broader public.
When in the setting opened only to the class, they saw the posting as involving other domains because it
did not fit the professional boundary expected once a professor enters the educational realm. They
understood the concerns about such a statement eliciting a response bias from students.

As a whole, this research adds to the growing body of literature that suggests faculty must use
online sites for educational purposes with care. Simple, direct, and unaltered migration of in-class
activities into an online format may not be wise. The lack of social and nonverbal clues, combined with
blurred boundaries, may be a challenge for faculty and students alike when conducting open or closed
online discussions. Discussion closed to outsiders, and aligned with equally closed personal boundaries,
may be the safest route to transitioning to an online format involving social networking platforms. Such
a sterile approach might be unappealing to those who see the benefits of learning in an easygoing,
collaborative environment. For those who venture into this broader setting to achieve innovative and
exciting educational outcomes, caution is still advised. End-of-class assessments of appropriateness of
various assignments, discussions, and online behaviors would assist faculty in honing their online skills
and personalities so that vulnerabilities are minimized. Over time, social transformations associated
with the blurred boundaries of educational integration may render a clearer picture of what constitutes
appropriate activity, but in these embryonic stages, continued assessment is advised.

This research includes methodological innovations, as well as several limitations. First, the use
of principal components analysis allowed for the collection of large amounts of data through the survey
technique. This methodology was an efficient method for data collection and involved the use of
objective statistical analysis. The collection of additional data from a smaller sample to test for face
validity also allowed for a richer discussion of conceptualization about social domains. The combination
of techniques provided a firmer foundation on which to rest conclusions than the use of either method
alone, in isolation. Analysis of scenario data often involves in-depth discussions with smaller numbers of
respondents, along with arduous content analysis. The methodology for this research is a possible
alternative for other scenario studies. That said, future research on the same topic would benefit from
using a wider range of scenarios, along with the in-depth discussions and content analysis. In addition to
such methodologies, methods that compare faculty responses with student responses would highlight
potential vulnerabilities of online education by finding where differences between faculty and students
exist. What faculty view as appropriate may diverge significantly from what students view as
appropriate, and vice versa. Studies with larger and more varied samples would also lead to more
generalizable conclusions. This research study was limited to mostly students enrolled in business
disciplines with a narrow experience in online education; students in other disciplines and with more
online experience may think differently. Further study is warranted with broader samples. Last, data
collected from ongoing end-of-class assessment would offer an additional means of learning how to deal
with the nuanced concerns of appropriate faculty online behavior.

In conclusion, the use of new technologies for education opens great possibilities for adapting to
diverse learning populations. The possibilities and diversity of responses, however, also offer the
262 | P a g e

potential for missteps and mishaps along the way. Ongoing experimentation is ever important, but it also
begs the question of how to proceed in ways that best benefit our students, our faculties, and our
constituents interested in favorable educational outcomes. In understanding the educational outcomes, the
social domain must not be neglected, as truly educated citizens must understand how to function as
members of society as well as how to be readers of great books.

REFERENCES

Affleck, J. (2010). Online and out of line. Townsville Bulletin. Townsville, Md.: Dec 25, p. 83.
Agarwal, S., & Mital, M. (2009). An exploratory study of Indian university students' use of social networking
sites: Implications for the workplace. Business Communication Quarterly, 72(1), 105-110.
Armour, N. (2006). Athletes learn to regret overexposure on Web. Journal Gazette. Ft. Wayne, IN: May 30,
1B.
Beaudoin, M. (2002). Learning or lurking? Tracking the 'invisible' online student. Internet and Higher
Education, 5(2), 147-155.
Bowean, L. and Mack, K.(2010). Can your teacher be your 'friend' online?
Chicago Tribune. Chicago, Ill.:Aug 10, p. 1.1
boyd, d. m., & Ellison, N. B. (2007). Social network sites: Definition, history, and scholarship. Journal of
Computer-Mediated Communication, 13(1), 210-230.
Brooks, B. G. (2007). CU Internet postings led to dismissals: [FINAL Edition]. Rocky Mountain News.
Denver, CO: Jan 21, 2007, 14SPORTS.
Bull, G., Thompson, A., Searson, M., Garofalo, J., Park, J., Young, C., et al. (2008). Connecting informal and
formal learning: Experiences in the age of participatory media. Contemporary Issues in Technology and
Teacher Education, 8(2), 100-107.
Cain, J. (2008). Online social networking issues within academia and pharmacy education. American Journal
of Pharmaceutical Education, 72(1), Article 10.
http://www.pubmedcentral.nich.gov/articlerender.fci?artid=2254235.
Charnigo, L., & Barnett-Ellis, P. (2007). Checking out facebook.com: The impact of a digital trend on academic
libraries. Information Technology and Libraries, 26(1), 23-34.
Chu, M., & Meulemans, Y. N. (2008). The problems and potential of MySpace and Facebook usage in
academic libraries. Internet Reference Services Quarterly, 13(1), 69-85.
Davidson, P., Turiel, E., & Black, A. (1983). The effect of stimulus familiarity on the use of criteria and
justifications in children's social reasoning. British Journal of Developmental Psychology, 1(1), 49-65.
Dawson, S. (2006). A study of the relationship between student communication interaction and sense of
community. Internet and Higher Education, 9(3), 153-162.
263 | P a g e

Drew, J. (2010). BYU football notes: Coach allows social networking sites, but urges caution. The Salt
Lake Tribune, Salt Lake City, UT: Nov 10.

Driver, M. (2002). Exploring student perceptions of group interaction and class satisfaction in the web-
enhanced classroom. Internet and Higher Education, 5(1), 35-45.

EWU Board of Trustees (2010). Social Media (Draft). EWU Policy 203-04. Eastern Washington
University, Nov.
Fabos, B. (2008). The price of information: Critical literacy, education, and today's internet. In J. Coiro, M.
Knobel, C. Lankshear & D. J. Leu (Eds.), Handbook of research on new literacies (pp. 839-870). New York:
Erlbaum.
Farmer, J. (2010). Behavior adoption on social networks. 20bits by Jessie Farmer.
http://20bits.com/articles/behavior-adoption-on-social-networks/
Finder, A. (2006). When a risque online persona undermines a chance for a job. New York Times (Late Edition
(East Coast)). New York, NY: Jun 11, 2006, 1.1.
Fougler, T., Ewbank, A., Kay, A., Osborn Popp, S., & Carter, H. L. (2009). Moral spaces in MySpace:
Preservice teachers' perspectives about ethical issues in social networking.
Journal of Research on Technology.

Friedman, B. (1997). Social judgments and technological innovation: Adolescents' understanding of
property, privacy, and electronic information. Computers in Human Behavior, 13(3), 327-351.

Friedman, B., Kahn, P. H., Jr., Hagman, J., Severson, R. L., & Gill, B. (2006). The watcher and the
watched: Social judgments about privacy in a public place. Human-Computer Interaction, 21(2), 235-
272.

Garrow, H. B. (2010) Board proposes online restrictions from teachers to students. McClatchy - Tribune
Business News. Washington:Dec 27.
Golder, S., Wilkinson, D., & Huberman, B. (2008). Rhythms of social interaction: Messaging within a
massive online network. Palo Alto, CA: HP Labs.
Greenhow, C., & Robelia, E. (2009a). Informal learning experiences and identity formation in online social
networks. Learning, Media and Technology, 34(2), 119-140.
Greenhow, C., & Robelia, E. (2009b). Old communication, new literacies: Social network sites as social
learning resources. Journal of Computer-Mediated Communication, 14(4), 1130-1161.
Greenhow, C., Robelia, E., & Hughes, J. (2009). Web 2.0 and educational research: What path do we take now?
Educational Researcher, 39(4), 246-259.
Griffith, S., & Liyanage, L. (2008). An introduction to the potential of social networking sites in education.
Proceedings of the Emerging Technologies Conference.
264 | P a g e

Gross, R., & Acquisiti, A. (2008). Information revelation and privacy in online social networks. ACM
Workshop on Privacy in the Electronic Society, 71-80.
Gruss, M. (2007). For most of us, unflattering photos are our cross to bear. Virginian Pilot. Norfolk, VA: Oct
17, E1.
Henrique, J. B. (2010) Factor analysis versus principal components analysis (PCA). The Statistics Page,
University of Wisconsin, Madison. http://psych.wisc.edu/henriques/pca.html.

Hewitt, A., & Forte, A. (2006). Crossing boundaries: Identity management and student/faculty relationships on
the Facebook. Computer Supported Cooperative Work Conference, Banff, Alberta, Canada.
Iyengar, K. (2006). Fact facts. Tulsa World. Tulsa, OK: Jun 2, D1.
Jones, S., Johnson-Yale, C., Millermaier, S., & Perez, F. S. (2008). Academic work, the internet, and U. S.
college students. Internet and Higher Education, 11(3-4), 165-177.
Jordan, B. (2009). Blurring boundaries: The "real" and the "virtual" in hybrid spaces. Human Organization,
68(2), 181-193.
Kaya, T. (2010) New college networks, unlike Facebook, connect 'Social' to Studies. The Chronicle of
Higher Education. Washington:Oct 10.

Kieffer, C. (2011) State releases educators' ethics code. McClatchy - Tribune Business News.
Washington:Jan 23.
Kirkland, C. J. (2010). Stalkers pose a real threat. Hartford Courant. Hartford, CT: Apr 11, 2010, C1.
Lampe, C., Ellison, N., & Steinfield, C. (2006). A Face(book) in the crowd: Social searching vs. social
browsing. In Proceedings of ACM Special Interest Group on Computer-Supported Cooperative Work, ACM
Press (pp. 167-170). http://portal.acm.org/citation.cfm?id=1180901.
Laupa, M. (1991). Children's reasoning about three authority attributes: Adult status, knowledge,
and social position. Developmental Psychology, 27(2), 321-329.
Laupa, M. (1 995). Children's reasoning about authority in home and school contexts. Social
Development, 4(1), 1-16.
Laupa, M., &Turiel, E. (1993). Children's concepts of authority and social contexts. Journal of
Educational Psychology, 85(1), 191-197.
Li, L., & Pitts, P. J. (2009). Does it really matter? Using virtual office hours to enhance student-faculty
interaction. Journal of Information Systems Education, 20(2), 175-186.
Lipka, S. (2007, December). For professors, 'friending' can be fraught. The Chronicle of Higher Education,
54(15), A. 1.
265 | P a g e

Madge, C., Meek, J., Wellens, J., & Hooley, T. (2009). Facebook, social integration and informal learning at
university: 'it is more for socializing and talking to friends about work than for actually doing work.'. Learning,
Media and Technology, 34(2), 141-155.
Manafy, M. (2010). Social web etiquette. EContent, 33(3), 5.
Munoz, C. L., & Towner, T. L. (2009). Back to the Wall: How to use Facebook within the classroom.
Unpublished manuscript. Department of Marketing and Entrepreneurship, Silberman College of Business,
Fairleigh Dickinson University, Madison, NJ.
Nemetz, P., Aiken, K., Cooney, V., & Pascal, V. (2011) Student Separateness-Connectedness as a
Determinant of Acceptable Faculty Involvement in Students Social Networks. Internet and Higher
Education, In Press.

Nucci, L. (1981). Conceptions of personal issues: A domain distinct from moral or societal concepts.
Child Development, 52(1), 1 14-121.

Nucci, L. (1989). Challenging conventional wisdom about morality: The domain approach to values
education. In L. Nucci (Ed.), Moral development and character education (pp. 183-203). Berkeley, CA:
McCutchan Publishing.

Nucci, L. P. (2001). Education in the Moral Domain. Cambridge: Cambridge University Press.
Raacke, J., & Bonds-Raacke, J. (2008). Myspace and Facebook: Applying the uses and gratifications theory to
exploring friend-networking sites. Cyberpsychology & Behavior, 11(2), 169-174.
Roblyer, M.D, McDaniel, M., Webb, M., Herman, J., Witty, J. V. (2010). Findings on Facebook in higher
education: A comparison of college faculty and student uses and perceptions of social networking sites.
Internet and Higher Education, 13(3), 134-140.
Salaway, G., Caruso, J. B., & Mark, R. N. (2008). The ECAR study of undergraduate students and
information technology, 2008 (vol. 8). Boulder, CO: EDUCAUSE.
Selwyn, N. (2009). Faceworking: Exploring students' education-related use of Facebook. Learning, Media and
Technology, 34(2), 157-174.
Smetana, J. G. (1988). Adolescents' and parents' conceptions of parental authority. Child Development,
59(2), 321-335.

Smetana, J. G.(1995). Morality in Context: Abstractions, Ambiguities, and Applications. In R.
Vasta (ed.), Annals of Child Development (Vol. 10). London: Jessica Kingsley
Publishers.

Smetana, J. G. (2002). Culture, Autonomy, and Personal Jurisdiction in Adolescent-Parent
Relationships. In H. W. Reese and R. Kail (eds.), Advances in Child Development and
Behavior (Vol. 29). Orlando, Fla.: Academic Press.

Smetana, J. G. (2006). Social-cognitive domain theory: Consistencies and variations in children's moral
and social judgments. In M. Killen & J. G. Smetana (Eds.), Handbook of Moral Development (pp. 119-
153). Mahwah, NJ: Lawrence Erlbaum Associates Publishers.
266 | P a g e


Smetana, J. G, & Asquith, P. (1994). Adolescents' and parents' conceptions of parental authority and
personal autonomy. Child Development, 65(4), 1147-1162.

Smetana, J. G, & Bitz, B. (1996). Adolescents' conceptions of teachers' authority and their relations to
rule violations in school. Child Development, 67(5), 1153-1172.

Smetana, J. G., Campione-Barr, N., and Daddis, C. Longitudinal Development of Family Decision-
Making: Defining Healthy Behavioral Autonomy for Middle Class African American Adolescents.
Child Development, 2004, 75(5), 117.

Smetana, J. G., and Chuang, S. Middle-Class African American Parents Conceptions of Parenting in the
Transition to Adolescence. Journal of Research on Adolescence, 2001, 11(2), 177198.

Smetana, J. G., and Daddis, C. Domain-Specific Antecedents of Psychological Control and Parental
Monitoring: The Role of Parenting Beliefs and Practices. Child Development, 2002, 73(2), 563580.
Stutzman, F. (2008). Experience and privacy: Exploring the disclosure behaviors of established Facebook
users. ASIST Annual Meeting. Conference Poster.
http://fredstutzman.com/papers/ASIST2008P_Stutzman.pdf.
Suhr, D. D. (2005). Principal Component Analysis vs. Exploratory Factor Analysis. SUGI 30
Proceedings, Philadelphia, PA. http://www2.sas.com/proceedings/sugi30/203-30.pdf.
Tufekci, Z. (2008). Can you see me now? Audience and disclosure regulation in online social network sites.
Bulletin of Science, Technology & Society, 28(1), 20-36.
Turiel, E. (1983). The development of social knowledge: Morality and convention. Cambridge, MA:
Cambridge University Press.

Turiel, E.(1998). Moral Development. In W. Damon (ed.), Handbook of Child Psychology, Vol.
3. In N. Eisenberg (ed.), Social, Emotional, and Personality Development (5th ed.). New
York: Wiley.

Turiel, E. (2002). The culture of morality: Social devehpment, context, and conflict. New York:
Cambridge University Press.
Twu, H. (2009). Effective wiki strategies to support high-context culture learners. TechTrends: Washington,
53(5), 16-21.
Wandel, T. L. (2008). Colleges and universities want to be your friend: Communicating via online social
networking. Planning for Higher Education, 37(1), 35-48.
Williard, N. (1997) Moral development in the information age. Center for Advanced Technology in
Education. University of Oregon College of Education, Eugene, OR.
http://tigger.uic.edu/~lnucci/MoralEd/articles/willard.html.

Williams, A., & Merten, M. (2008). A review of online social networking profiles by adolescents: Implications
for future research and intervention. Adolescence, 43(170), 253-274.
267 | P a g e

Young, J. R. (2009, February 9). How not to lose face on Facebook, for professors. The Chronicle of Higher
Education, 55(22), A. 1.