Vous êtes sur la page 1sur 20

Journal of Vocational Behavior 58, 142161 (2001) doi:10.1006/jvbe.2000.1756, available online at http://www.idealibrary.

com on

Impression Management by Association: Construction and Validation of a Scale


Martha C. Andrews
West Virginia University

and K. Michele Kacmar


Florida State University This study describes the construction and initial validation of a scale that measures four indirect impression management tactics identied by Cialdini (1989). The scale measures the connection-focused tactics of Boasting, Blurring, Blaring, and Burying. The scale development consisted of the multiple stage process identied by Hinkin (1998). Stage 1 included holding three separate focus groups to generate the initial items. Stages 2 and 3 included two data collections to rst explore and then conrm the factor structure of the impression management by association scale. A nal data collection in Stage 4 assessed the convergent and discriminant validity of the impression management by association scale. Overall, the impression management by association scale satisfactorily represented the four factors conceptualized by Cialdini (1989), although more renement is needed to better distinguish the Burying from the Blaring factor. C 2001 Academic Press

Impression management has been widely studied during the past 20 years for its effects on career outcomes (Giacolone & Rosenfeld, 1989). Previous research has addressed the inuence of impression management on selection interviews (Baron, 1986; Fletcher, 1989; Kacmar, Delery, & Ferris, 1992; Kinicki & Lockwood, 1985; Stevens & Kristof, 1995; Wade & Kinicki, 1997), performance ratings (Wayne & Ferris, 1990; Wayne & Kacmar, 1991; Wayne & Liden, 1995), career success (Judge & Bretz, 1994), assessments of promotability (Thacker & Wayne, 1995), and feedback seeking (Ashford & Northcraft, 1992; Morrison & Bies, 1991). These studies have examined several categories of direct impression management tactics including assertive and defensive (Stevens & Kristoff, 1995; Wayne & Kacmar, 1991). Assertive tactics are proactive behaviors undertaken by
We thank William P. Anthony, Pamela L. Perrew , Jane Reimers, and Nagaraj Sivasubramaniam for e helpful suggestions on a previous version of the manuscript. Address correspondence and reprint requests to Martha Andrews, West Virginia University, College of Business and Economics, Department, of Management, Morgantown, WV 26506-6025. E-mail: andrews@be.wvu.edu. 142 0001-8791/01 $35.00
Copyright C 2001 by Academic Press All rights of reproduction in any form reserved.

IMAS

143

individuals to create a specic identity to further their careers. For example, the use of assertive tactics has been related to increased performance ratings (Wayne & Ferris 1990; Wayne & Liden, 1995), promotions (Thacker & Wayne, 1995; Wayne, Liden, Graf, & Ferris, 1997), and increased ratings in selection interviews (Gilmore & Ferris, 1989; Stevens & Kristof, 1995). Defensive tactics are reactive behaviors used by individuals following actions that may portray them negatively (Tedeschi & Melburg, 1984). Defensive tactics seek to avoid negative career implications. One group of tactics that largely has been ignored is indirect impression management tactics (Cialdini, 1989) which assist in positive career advancements and outcomes. Unlike direct tactics, indirect tactics create desirable impressions by including a third party. Individuals may impression manage an absent target through a third person in hopes of having the information passed on to the intended target at a later time (Schlenker, 1980). Alternatively, the third party may be absent yet associated with the impression manager. For example, people may boast about their relationships with favorable others to present themselves in a positive light. Likewise, people may attempt to distance themselves or blur their relationships with unfavorable others. Cialdini and Richardson (1980) were among the rst to acknowledge that people draw attention to themselves through their connections with popular or successful others and divert attention from or deny their connections with negative or unsuccessful others. They dubbed these activities impression management by association. Impression management by association refers to the behaviors people use at work in managing their associations with other people and things to create favorable impressions of themselves (Cialdini & Richardson, 1980). In their early work, Cialdini and Richardson discussed two forms: basking and blasting. They suggested people bask in the reected glory (BIRG) of others and blast the opposition in order to enhance their prestige. Studying the pronouns college students used to describe the wins or losses of their school football teams, Cialdini et al. (1976) provided evidence of the BIRG phenomenon. Students used the word we in describing a win signicantly more often than in describing a loss. In later work, Cialdini (1989) expanded basking into four forms of connection-focused tactics (Boasting, Burying, Blaring, and Blurring). Each of these tactics is described below. Boasting An actors tendency to boast about his or her positive connections with favorable others was called Basking earlier by Cialdini and Richardson (1980). They later replaced Basking with Boasting. For example, someone may boast about having attended the same high school as a professional football star. In organizations boasting can be seen as an individual strives to receive credit for anothers success or capitalize on his or her association with a high performing group in an effort to secure a strong performance appraisal or a promotion (Cialdini, 1989). Cialdini (1989) warns that these tendencies may diminish ones motivation for self-attained goals.

144 Burying

ANDREWS AND KACMAR

Just as people have connections with favorable others, they often have connections with unfavorable others. It is these connections people wish to bury. For example, someone may Bury a relationship by failing to mention his or her past association with a convicted criminal. At work, members of departments which have performed poorly may try to distance themselves from the group, hence, burying the relationship. When a manager recommends a subordinate for promotion who subsequently turns out to be ineffective, the manager may protect his or her own image by explaining they never worked that closely together. Blaring There also are connections with unfavorable others that people wish to publicly minimize. Actors make a point of mentioning that they are in no way associated with a negative person (i.e., blaring these connections) to make sure others are aware of how weak the connection between them is. In the organization, this is easily illustrated as members often stray far from those who challenge the opinions and authority of leaders (Cialdini, 1989). Blurring Workers blur their links with favorable others by way of strategic omissions. Specically, workers make a point of not mentioning differences between themselves and successful others as a means of increasing others perceptions of them. For example, consider two groups, Group A and Group B, within a single division of an organization. Group A is clearly responsible for the successful launch of a new product. While interviewing for a high level position, a member from Group B discusses the successful new product but neglects to mention his or her group had very little involvement in the launch. Hence, the actor clouds his or her weak link to Group A. While theory and common sense predict that organizational members actively use impression management by association to secure valuable career-oriented rewards, empirical conrmation of these predictions is lacking. Our goal was to develop a scale to measure the frequency of organizational members use of association tactics following the scale development procedure outlined by Hinkin (1998). Because there were no existing scales to measure this phenomenon, we thought it appropriate to approach item writing by holding focus groups. STAGE 1: ITEM GENERATION We began with three focus groups. The rst focus group, conducted orally, consisted of 17 participants from an executive management program at a university located in the southeast. The 13 men and 4 women were employed by various public and private organizations throughout the state. Two of the men were AfricanAmerican and the remaining participants were Caucasian. The second focus group, also conducted orally, consisted of 5 male and 2 female part-time MBA students, all employed full-time with various organizations. All of the participants were

IMAS

145

Caucasian. The third focus group, in which participants wrote their answers, was held at an insurance organization in the southeast. This focus group consisted of 6 men and 5 women. The rationale for holding a written focus group was to capture anonymously the thoughts of those who are easily intimidated or do not like to speak in front of others. Procedure Each focus group began with an introduction and discussion of the four connection-focused tactics of impression management (i.e., Boasting, Burying, Blaring, and Blurring). Participants were then asked to describe a time at work when they or someone they knew used any of the four tactics. Extensive notes were taken during the two oral focus groups and both were audio-taped to ensure that all information was collected. Content Analyses Following the focus groups, the audio tapes were transcribed and comments from the written focus group were typed. Both authors independently analyzed the transcripts and categorized the behaviors as Boasting, Burying, Blaring, Blurring, or other. The items generated through content analysis were then classied by six doctoral students. This classication was conducted to ensure content adequacy of the impression management by association scale (IMAS) using the quantitative approach suggested by Schriesheim, Powers, Scandura, Gardiner, and Lankau (1993). The students were provided with the theoretical denitions of each of the four tactics and an other category. They then placed each item next to the most appropriate denition. The ratings were collected and combined into a matrix. The matrix consisted of how many times each item was assigned to each of the four tactics. Following the completion of the matrix from the rst round of coding, the denitions were slightly modied and a second round of coding was conducted using six new coders. Results Responses from each of the six content adequacy coders were combined into a matrix. For the Boasting tactic, there was at least 83% (i.e., ve of six coders) agreement for 9 of the 10 items, 6 of which had perfect agreement. For the Blurring tactic, there was 83% agreement for 2 of the 9 items and less than majority agreement for the remaining Blurring items. For the Blaring tactic, there was majority agreement for 1 of the 8 items. For the Burying tactic, there also was majority agreement for 1 of the 8 items. Items that were sorted primarily into two categories or more were discarded or modied. Items primarily placed into a category other than the one for which they were intended also were modied or discarded. This procedure resulted in a total of 32 items being retained. This set of items was then given to a new group of six raters with the same instructions but with modied denitions of each of the four tactics. These ratings were collected and a new matrix was constructed. There was at least 83% agreement

146

ANDREWS AND KACMAR

for 6 of the Boasting items, 4 of which had perfect agreement. For the Blurring tactic, there was majority agreement (four of six coders) for 5 of the 8 items. For the Blaring tactic, there was majority agreement for 2 of the 8 items. For the Burying tactic, 3 of the 8 items had majority agreement. After discussions with the raters we determined that the low agreement among the raters for Burying and Blaring could be attributed to the denitions used to distinguish Burying from Blaring and not the items. Hence, the Burying and Blaring items were retained. Overall, there was majority agreement for 19 of the 32 items. STAGE 2: ITEM SELECTION Participants To conrm expectations about the structure of IMAS, the items developed in Stage 1 were administered to 189 employees working at multiple levels of various organizations throughout the United States. The organizations included a large retail furniture company (n = 24), a banking organization (n = 12), a computer software company (n = 76), an insurance company (n = 13), a data processing company (n = 4), members of the Society for Human Resource Management (n = 54), and a school board (n = 6). The participants consisted of 89 males and 97 females. Three of the respondents did not indicate their gender. Organizational tenure ranged from 1 year to 41 years with a mean tenure of 5.8 years. The average age of the participants was 40.5 years and consisted of 4 African-Americans, 4 Asians, 152 Caucasians, 22 Hispanics, 5 Other, and 2 of the respondents did not indicate their race. The level of education of the respondents was high school (n = 12), some college (n = 52), Bachelor of Science (n = 83), Masters degree (n = 31), and higher than a Masters degree (n = 8). Three of the respondents did not indicate their level of education. Finally, the respondents were asked to indicate the number of levels between them and the highest level of management. This gure ranged from 0, indicating that the respondent was at the highest level of management, to 8, indicating that there were 8 levels of management between the respondent and the highest ranking manager. The mean number of levels was 3.2. Procedure Surveys were sent by U.S. mail to participants at all of the organizations except those employed by the computer software company. These employees received the survey by electronic mail. Participants who received the survey by U.S. mail were assured complete anonymity. Those who received the survey electronically were assured of condentiality and were asked to e-mail their responses directly back to the researchers while those who received surveys by the U.S. mail were asked to mail their responses back to the researchers in the enclosed stamped envelope. Measures IMAS. To measure the use of association tactics, the 32 IMAS items developed in Stage 1 were included in a survey administered to the participants. Respondents

IMAS

147

were asked to indicate how frequently they used the behaviors indicated by the items in IMAS on a 5-point Likert-type scale with the following anchors: (1) never do it, (2) rarely do it, (3) occasionally do it, (4) often do it, and (5) nearly always do it. Social desirability. To assess whether participants responded to the IMAS items in a socially desirable manner, Shultz and Chavez (1994) 11-item social desirability scale was included in the survey. A sample item is I always return money when I nd it. These items were responded to on a 5-point Likert-type scale with anchors of (1) strongly disagree to (5) strongly agree. The internal reliability estimate for this scale was .73. Results Correlational analyses revealed that 3 of the IMAS items correlated signicantly with the social desirability scale: Bury 3 (r = .17, p < .02) , Bury 5 (r = .19, p < .01), and Bury 8 (r = .23, p < .002). These items were deleted from the scale. The remaining 29 items were examined using principal components analysis with an oblimin rotation. Using an oblimin rotation was considered appropriate at this stage since the primary purpose was to obtain theoretically relevant dimensions (Hair, Anderson, & Tatham, 1987). This analysis resulted in eight factors. Various criteria were used to determine which items and factors to retain. First, eigenvalues greater than 1 and scree plots were examined to determine which factors to interpret. Next, factors composed of items representing several dimensions were deleted. With respect to items, any items with loadings of .40 or above were retained (Ford, MacCallum, & Tait, 1986). However, items designed to represent one dimension but that had their highest loading on a factor representing another dimension were eliminated to increase the future factor stability and nonredundancy of the items. Finally, items that loaded on multiple dimensions (i.e., loaded greater than .30 on more than one factor) were eliminated. Following these decisions rules, 17 items were retained. These items were reanalyzed following the same procedure resulting in a six-factor solution. Two of the six factors were eliminated, as those factors included items from multiple dimensions. One additional Blur item was removed because it loaded on a factor containing two Bury items. After these deletions, 12 items remained and another principal components analysis was conducted. Results indicated four factors each representing one of the four types of IMAS. However, one item cross-loaded on two factors and was deleted prior to conducting the nal factor analysis. Results from this nal analysis, shown in Table 1, indicate the 11 items that were retained loaded on their respective factors with none of the items cross-loading on other factors. Reliability estimates were computed for each of the IMAS dimensions and then for the full scale. Cronbachs alpha for the subscales were .64 for Boasting, .58 for Blurring, .74 for Blaring, and .48 for Burying. The internal reliability estimate for the full IMAS scale was .75. To ensure that the respondents were not responding in a socially desirable manner, a correlation between each of the IMAS subscales

148

ANDREWS AND KACMAR TABLE 1 Principal Components Analysis Results Using 11 Retained IMAS Itemsa Factor Item Blare .8446 .8032 .7790 Blur .0837 .0187 .0949 Boast .0222 .0727 .0744 Bury .0492 .0159 .0201 .0702 .1300

I make sure my supervisor knows I am not like poor performers in the ofce. When someone else does a poor job, I let others know I maintain a higher level of performance. When my boss discusses a problem with me brought on by a troublemaker, I make sure he or she knows Im nothing like the troublemaker. I just smile and nod if ever complimented on work for which another group is responsible. When a superior compliments me on good work for which someone else is responsible, I dont bother to explain otherwise. When others ask me about my relationships with a successful person in the organization, I dont let on that we barely know each other. I let others know about my friendships with superiors in my organization. I let others know that I am friends with people in informative or powerful departments. I bring up my past experience with well-known previous employers to make others aware of my competence. When my peer has a major problem with his or her work, I try to disassociate from him or her so that others wont think I am involved. I try to disconnect myself from unproductive employees in the ofce, even though some of them are my friends. Eigenvalue Variance explained
aN

.0063 .0166

.8770 .7578

.0929 .1603

.0360

.7527

.0883

.0346 .0923 .0155 .1236

.0335 .1008 .0725 .0012

.0171 .0677 .0043 .0080

.8393 .7714 .6823

.0538

.8711

.0629

.0309

.1262

.7290

3.29844 30.0

1.67925 15.3

1.24276 11.3

1.01738 9.2

= 189.

and the social desirability scale was calculated. The correlations ranged from .01 to .11 and were not signicant. STAGE 3: SCALE DEVELOPMENT The scale that emerged at the end of Stage 2 was composed of 11 items, 3 for each factor except for Bury which only had 2 items. In order to have an equal number of items in each subscale, 3 additional Bury items were created: When my peer receives a poor performance evaluation, I try to distance myself from him or her so others wont think my performance is similar to his or hers, I try to distance myself from peers who are incompetent, and When a peer develops a negative reputation, I try to disassociate from him or her. These items were added to the 11 retained items and included in the second survey.

IMAS

149

Participants A second round of data was collected from employees of hotels (n = 153) and a human resource management list serve (n = 89). The employees of the hotels consisted of supervisors and managers from 12 different hotels. Surveys were administered at their respective weekly management meetings, which yielded a response rate of 100%. Ten of these hotels were associated with 10 different national hotel chains while 2 were independently owned and operated. The full group consisted of 115 males and 124 females. The age of the participants ranged from 20 to 61 years with a mean of 35.58 years. Organizational tenure ranged from 1 to 45 years with an average tenure of 6.92 years. With respect to race, 151 of the respondents were Caucasian, 33 were African-American, 11 were Asian, 27 were Hispanic, and 12 selected the other category. The level of education of the respondents was high school (n = 12), some college (n = 61), Bachelors degree (n = 97), Masters degree (n = 50), and higher than a Masters degree (n = 18). Four of the respondents did not indicate their level of education. The number of levels between respondents and the highest level of management ranged from 0, indicating that the respondent was at the highest level of management, to 8, indicating that there were 8 levels of management between the respondent and the highest ranking manager. The mean number of levels was 2.8. Procedure Members of the human resource management list serve returned their surveys by e-mail. Surveys were collected on site from the hospitality employees. E-mail respondents were assured condentiality while the hotel employees were assured their responses would be anonymous. Results The 11 items retained in Stage 2 were subjected to principal components analyses with an oblimin rotation. This was done to cross-validate the results from Stage 1. This analysis produced the same four-factor solution found in Stage 2. Next, three principal components analyses in which each of the new Bury items were added 1 at a time were run. Results revealed that 1 of the new Bury items (When a peer develops a negative reputation, I try to disassociate from him or her) loaded on the factor containing the previous 2 Bury items. Hence, it was retained in the revised scale. The alpha reliabilities for each of the factors were .69 for Boasting, .53 for Burying, .74 for Blaring, and .51 for Blurring. The reliability for the full 12-item scale was .76. Results from the factor analysis appear in Table 2. With the addition of the new Bury item, two of the Bury items cross-loaded on the Blare factor above .30. They were I try to disconnect myself from unproductive employees in the ofce, even though some of them are my friends, and When a peer develops a negative reputation, I try to disassociate from him or her. For an item to be considered a Bury tactic, an established relationship between the actors must be present. It appears that the relationship component of these two Bury items (i.e., friend and

150

ANDREWS AND KACMAR TABLE 2 Principal Components Analysis Results Using Sample 2 Factor Item Blare Blur .13487 .84922 Boast .06327 .00191 Bury .15370 .00015 .02269 .78809

1. I make sure my supervisor knows I am not like .79974 poor performers in the ofce. 2. When a superior compliments me on good work .22850 for which someone else is responsible, I dont bother to explain otherwise. 3. I let others know that I am friends with people .13980 in informative or powerful departments. 4. When my peer has a major problem with his or her .01830 work, I try to disassociate from him or her so that others wont think I am involved. 5. When my boss discusses a problem with me .76330 brought on by a troublemaker, I make sure he or she knows Im nothing like the troublemaker. 6. I just smile and nod if ever complimented on work .14636 for which another group is responsible. 7. I let others know about my friendships with .13337 superiors in my organization. 8. I try to disconnect myself from unproductive .38465 employees in the ofce, even though some of them are my friends. 9. When someone else does a poor job, I let others .74251 know I maintain a higher level of performance. 10. When others ask me about my relationships with .11074 a successful person in the organization, I dont let on that we barely know each other. 11. I bring up my past experience with well-known .05629 previous employers to make others aware of my competence. 12. When a peer develops a negative reputation, .39212 I try to disassociate from him or her. Eigenvalue 3.3818 Variance explained 28.2 Note. N = 221.

.01633 .00633

.81151 .19121

.18414

.15299

.14994

.82299 .00892 .03786

.10333 .79310 .24636

.14493 .06788 .67837

.13780 .56679

.22913 .00337

.01276 .31930

.32458

.44649

.29878

.21000 1.4999 12.5

.21488 1.4795 12.3

.47540 1.0746 9.0

peer) was not strong enough to force them to fall on the appropriate factor. The inuence of the unfavorable component in these items (i.e., negative reputation and unproductive employee) had almost as strong an impact as the relationship component. Conversely, one of the Blur items (When others ask me about my relationship with a successful person in the organization, I dont let on that we barely know each other) cross-loaded on the Boast factor, indicating that the relationship described in the item was nearly as important as the actual blurring activity. To conrm the factor structure of the scale, LISREL 8.12a utilizing a maximum likelihood estimation was used (Joreskog & Sorbom, 1993). The chi-square was

IMAS TABLE 3 Results of Model Comparisons Model: 4-Factor: Boast Bury Blur Blare 3-Factor: (Bury/Blur) Blare Boast 3-Factor: (Boast/Bury) Blur Blare 3-Factor: (Boast/Blur) Bury Blare 3-Factor: (Boast/Blare) Bury Blur 3-Factor: (Bury/Blare) Boast Blur 3-Factor: (Blur/Blare) Boast Bury 2-Factor: (Boast/Bury) (Blur/Blare) 2-Factor: (Boast/Blur) (Bury/Blare) 2-Factor: (Boast/Blare) (Bury/Blur) 1-Factor: (Boast/Blare/Bury/Blur) GFI AGFI NFI CFI PNFI .94 .89 .88 .87 .88 .92 .88 .83 .86 .84 .83 .90 .83 .81 .80 .82 .88 .81 .75 .79 .77 .75 .85 .77 .73 .70 .71 .81 .73 .61 .67 .63 .57 .91 .82 .78 .75 .77 .88 .78 .66 .73 .68 .62 .62 .59 .56 .54 .55 .63 .56 .49 .54 .51 .47 2 105.97 160.75 187.11 204.23 195.97 127.21 186.02 267.10 223.45 250.68 292.83 df 48 51 51 51 51 51 51 53 53 53 54
2 diff

151

d f diff

54.78 81.14 98.26 90.00 21.25 80.05 161.13 117.48 144.71 186.86

3 3 3 3 3 3 5 5 5 6

2 Note. All diff scores are signicant at the p < .01 level.

105.97 (d f = 48, p = .000). A chi-square value that is two to three times the degrees of freedom is considered an acceptable t (McIver & Carmines, 1981). The root mean square error of approximation was .071, which meets the minimum acceptable level of .08 (Price & Mueller, 1986). The t indices for the model were: goodness of t index = .94, adjusted goodness of t index = .90, normed t index = .85, comparative t index = .87. In addition, the parsimony normed t index was .62, which exceeds the acceptable level of .60 (Mulaik et al., 1989). All of the path loadings were signicant ( p < .05) and ranged from .41 to .86. To verify that the four-factor structure is the best representation of the items in IMAS, all possible alternative models were estimated (Mathieu & Farr, 1991). The models included a one-factor, 3 two-factor, and 6 three-factor models. The two-factor models represented all possible parings of the four dimensions. The three-factor models represented all possible three-way combinations of the four dimensions. The results for these models are shown in Table 3. In addition, chisquare difference tests were conducted to compare all of the alternative models to the four-factor solution. These results also are shown in Table 3. Results from the alternative model estimations support the four-factor structure as originally designed, as do the chi-square difference tests. However, results for one of the three-factor models (Boast, Blur, and Bury/Blare) were also fairly strong. While the Boast and Blur factors were independent, the Bury and Blare items may not be completely distinct. This may be a result of individuals attempting to ensure that they are not associated with undesirable others regardless of the connection. The strong or weak connection, which distinguishes Bury from Blare, is not as important to respondents as disassociating from negative others. STAGE 4: SCALE EVALUATION To provide further evidence of the construct validity of IMAS, convergent and discriminant validity of IMAS was examined. A scale displays convergent validity

152

ANDREWS AND KACMAR

when it correlates with other theoretically relevant variables. Hence, a survey was administered that included IMAS and three other variables expected to be positively related to the use of impression management: Self-Monitoring (Snyder, 1974), Locus of Control (Spector, 1988), and Need for Power (Steers & Braunstein, 1976). A construct displays discriminant validity when it is related positively to other similar variables, but not too highly. To explore the discriminant validity of IMAS, Wayne and Ferriss (1990) measure of impression management and the coalitions and assertiveness subscales from Schriesheim and Hinkin (1990) also were measured in the nal data collection. The rationale behind the selection of each of these variables is provided below. Self-Monitoring Individuals who are high in self-monitoring are able to adapt their behavior to the situation (Snyder, 1974). They search for more information and can easily assess social situations. For example, high self-monitors should perceive more opportunities to Boast about their favorable relationships to others. A positive relationship between self-monitoring and impression management has been suggested by several researchers (e.g., Gardner & Martinko, 1988). Further, empirical research has shown high self-monitors to be more effective at impression managing than low self-monitors (Schlenker & Leary, 1982). Given this discussion, a positive, moderate correlation is expected between self-monitoring and IMAS. Locus of Control Individuals with an internal locus of control believe that their successes and failures result from their own efforts and ability, whereas individuals with an external locus of control believe that these events are beyond their control (Rotter, 1966). Internals believe that they have control over what happens to them while externals feel that most outcomes are due to luck. Internals may actively manage their associations with others in order to be viewed more favorably. Externals, believing their reputation is somewhat beyond their control, may be less likely to take an active role in managing their associations. For example, an internal may Blare about his or her lack of involvement in a performance mishap to ensure that he or she is not blamed. However, an external may take the blame because he or she does not believe an explanation will make a difference. In discussing factors that contribute to ones propensity to engage in impression management behaviors, Ralston (1985) suggested that internals would make more active attempts to impress superiors to obtain desired career-related outcomes. Hence, a positive moderate correlation is expected between internal locus of control and IMAS. Need for Power People with a high need for power try to gain control and be in command when working in a group. They attempt to inuence those around them and try to direct others in their work activities (Steers & Braunstein, 1976). People construct

IMAS

153

impressions consistent with their desired identity images (Leary & Kowalski, 1990). Thus, those with a high need for power would likely try to associate and be seen with high-powered others. This way they may be viewed as more powerful. We expect a positive correlation between need for power and IMAS. Inuence Scales In order for IMAS to be considered a unique construct, it should demonstrate discriminant validity. Since IMAS is a unique form of impression management, it should not correlate too strongly with scales designed to measure inuence or other forms of impression management. Three scales were included in the present study to test these predictions: Wayne and Ferriss (1990) and Schriesheim and Hinkins coalition and assertiveness subscales (1990). The coalitions subscale was selected because both IMAS and coalitions involve relationships with co-workers. IMAS involves managing relationships with co-workers while coalitions involve gaining support of co-workers. Assertiveness was selected because assertive individuals would be more inclined to manage their associations. Wayne and Ferriss (1990) impression management scale was included to ensure that IMAS tap a separate and unique form of impression management. While IMAS is expected to be positively related to these scales, we anticipate only moderate correlations between them. Method Participants A third round of data collection was conducted at a large university located in the southeastern United States. The participants consisted of 221 nonfaculty employees working at multiple levels of the university. These respondents were employed in various administrative jobs throughout the university ranging from copy center employees to associate deans. The group contained 81 males and 139 females. One person did not indicate his or her sex. The age of the sample ranged from 20 to 64 years with a mean age of 41.46 years. Organizational tenure ranged from 1 to 34 years with a mean tenure of 10 years. Of the total sample, 176 were Caucasian, 22 were African-American, 2 were Asian, 4 were Hispanic, and 4 were other ethnic minorities. Educational levels included high school diploma (n = 11), some schooling following high school (n = 33), bachelors degree (n = 70), masters degree (n = 67), and doctoral degree (n = 39). Procedure Nonfaculty mailing labels were purchased for administrative professionals. Approximately 950 surveys were sent to a random sample of these employees by campus mail with an enclosed introductory letter and a return envelope. Participants were informed that the survey was voluntary and complete anonymity was assured. Respondents were asked to complete and return the survey by campus mail within 3 weeks. A total of 221 surveys were returned for a response rate of 23%.

154 Measures

ANDREWS AND KACMAR

IMAS. The 12-item scale resulting from Stage 3 was included to measure the use of association tactics. These items were responded to on a 5-point Likert-type scale with the following anchors: (1) never do it, (2) rarely do it, (3) occasionally do it, (4) often do it, and (5) nearly always do it. The alpha reliability was .86. Self-monitoring. Snyders (1974) 25-item scale was used to measure the respondents self-monitoring ability. Sample items from this scale are When I am uncertain of how to act in a social situation, I look to the behavior of others for cues and I can make impromptu speeches even on topics about which I have almost no information. Items were responded to on a 5-point Likert-type scale with anchors of (1) strongly disagree to (5) strongly agree. The internal consistency reliability for this scale was .75. Locus of control. Spectors (1988) seven-item scale was used as a measure of locus of control. Items were coded such that higher scores indicated an internal locus of control and lower scores indicated an external locus of control. Sample items from this scale are The main difference between people who make a lot of money and people who make a little money is luck (reverse-coded) and It takes a lot of luck to be an outstanding employee (reverse-coded). Items were responded to on a 5-point Likert-type scale with anchors of (1) strongly disagree to (5) strongly agree. The alpha reliability for this scale was .87. Need for power. Steers and Braunsteins (1976) Manifest Needs Questionnaire provided the items designed to measure need for power. This scale consists of ve items (e.g., I strive to gain more control over the events around me at work and I nd myself organizing and directing the activities of others). Items were responded to on a 5-point Likert-type scale with anchors of (1) never to (5) always. Cronbachs alpha for this scale was .73. Wayne and Ferriss (1990) impression management scale. A 10-item version of Wayne and Ferriss (1990) scale used by Wayne and Liden (1995) was used to measure impression management. Sample items included Try to act as a model employee in front of your supervisor by, for example, never taking longer than the established time for lunch and Work hard when you know the results will be seen by your supervisor. This scale was responded to on a 5-point Likert-type scale with anchors of (1) never do it to (5) nearly always do it. The alpha reliability for this scale was .74. Coalitions. Schriesheim and Hinkins (1990) revision of the coalition subscale from the Kipnis, Schmidt, and Wilkinson (1980) inuence scale was used to measure the use of coalitions as a means of inuence. The two items contained in this subscale were Obtained the support of co-workers to back up my request and Obtained the support of my subordinates to back up my request. This scale was responded to on a 5-point Likert-type scale with anchors of (1) never do it to (5) nearly always do it. The alpha reliability estimate for this scale was .83. Assertiveness. Assertiveness was measured using Schriesheim and Hinkins (1990) revision of Kipnis et al.s (1980) assertiveness subscale. This subscale consists of six items. Examples are Set a time deadline for him or her to do what I

IMAS

155

asked and Had a showdown in which I confronted him or her face-to-face. This scale was responded to on a 5-point Likert-type scale with anchors of (1) never do it to (5) nearly always do it. The internal consistency reliability estimate for this scale was .78. Results Correlational Analysis The correlations between each of the variables included in the nal data collection appear in Table 4 along with the means and standard deviations. The alpha reliabilities for each of the factors were .87 for Boasting, .77 for Blurring, and .79 for Burying and .77 for Blaring. The reliability for IMAS ( = .86) is also included in Table 4. Conrmatory Factor Analysis To conrm the four factor structure of IMAS, LISREL 8.12a (Joreskog & Sorbom, 1993), utilizing a maximum likelihood estimation, was used. The model estimated and the completely standardized path loadings are shown in Fig. 1. The chi-square was 93.07 (d f = 48, p = .000), indicating acceptable t (McIver & Carmines, 1981). The goodness-of-t (GFI) and the adjusted goodness-of-t indices (AGFI) were .94 and .90, respectively, indicating strong model to data t (Medsker, Williams, & Holohan, 1994). The NFI for the present sample was .92, exceeding the acceptable level of .90 (Medsker et al., 1994). The comparative t index (CFI) (Bentler, 1990) produced by the current sample was .96, also exceeding the acceptable level of .90 (Mulaik et al., 1989). The parsimony normed t index (PNFI) was .67, exceeding the acceptable level of .60 (Mulaik et al., 1989). Discriminant Validity Chi-square difference tests were conducted to check the discriminant validity of IMAS (Anderson & Gerbing, 1988). Three separate models were estimated
TABLE 4 Means, Standard Deviations, and Correlations for Scales of Interest for Sample 3
Variable 1. Self-Monitoring 2. Locus of Control 3. Need for Power 4. Wayne and Ferris 5. Coalitions 6. Assertiveness 7. Boast 8. Blur 9. Blare 10. Bury 11. IMAS Mean 2.82 2.35 3.30 3.34 2.94 2.49 1.91 1.40 2.01 1.75 1.77 SD 1 2 3 4 5 6 7 8 9 10 11

.44 (.75) .80 .02 .68 .37 .58 .15 1.07 .23 .69 .19 .86 .26 .61 .15 .90 .19 .79 .19 .61 .26

(.87) .04 (.73) .02 .15 (.74) .12 .35 .10 (.83) .07 .40 .06 .43 (.78) .12 .24 .29 .35 .29 (.87) .19 .08 .05 .19 .18 .33 (.77) .06 .22 .18 .32 .25 .48 .36 (.77) .12 .24 .11 .31 .37 .38 .38 .53 (.79) .17 .25 .23 .40 .38 .75 .64 .82 .77 (.86)

Note. N = 221. Reliabilities on the diagonal. Correlations >.17 are signicant at p < .01. Correlations >.13 are signicant at p < .05.

156

ANDREWS AND KACMAR

FIG. 1. Completely standardized path loadings for a four-factor IMAS model using sample 3 (one-factor model coefcients appear in parentheses).

utilizing a maximum likelihood estimation in LISREL 8.12a (Joreskog & Sorbom, 1993). Each model included the 12 IMAS items along with the items from one other impression management scale. All items were assigned to their respective underlying factors. A chi-square was computed for each model in which there were no constraints on the estimated correlation parameter. An additional chi-square was computed for each model in which the estimated correlation parameter was constrained to 1. A chi-square difference test was then computed for each of the three models. A signicantly lower chi-square for the initial model compared to the constrained model indicates discriminant validity (Bagozzi & Phillips, 1982). In each comparison, the chi-square difference test was signicant, indicating discriminant 2 validity for IMAS (Wayne & Ferris (1990) scale (diff (1) = 351, p < .001), As2 2 sertiveness (diff (1) = 217, p < .001), and Coalitions (diff (1) = 208, p < .001). Condence intervals around the correlations between IMAS and each of the impression management scales also were computed to ensure the discriminant

IMAS

157

validity of IMAS (Anderson & Gerbing, 1988). The condence intervals were computed by adding and subtracting twice the standard error to each of the correlations between IMAS and the other scales. Condence intervals that do not include 1 provide further evidence of discriminant validity (Anderson & Gerbing, 1988). The condence intervals for the correlations between IMAS and the three scales were Wayne & Ferris (1990) r = .23 (.098, .362), Assertiveness r = .37 (.220, .520), and Coalition r = .40 (.252, .548). Convergent Validity The convergent validity of IMAS was assessed rst by examining the correlations between IMAS and other variables thought to be positively related to the use of association tactics. Signicant correlations between theoretically relevant variables and the construct of interest have been suggested by many researchers to be an indication of convergent validity of new constructs (e.g., Anastasi, 1982; Hinkin, 1998). In the present study, these variables included Self-Monitoring, Locus of Control, and Need for Power. As indicated in Table 4, the correlations between IMAS and these variables were .26 ( p < .01), .17 ( p < .05), and .25 ( p < .01), respectively. These correlations indicate convergent validity for the full scale because of the signicant, positive relationships between IMAS and internal Locus of Control, Need for Power, and Self-Monitoring. Correlations also were computed for each of the subscales with these variables. All were signicant ( p < .05) between the IMAS subscales and Self-Monitoring and Need for Power except Blur and Need for Power (r = .08). However, only Blur was signicantly related to an internal locus of control ( p < .01). A second method for assessing the convergent validity of a construct is to estimate the path coefcients between each of the indicators and the underlying construct factor (Anderson & Gerbing, 1988). A signicant path coefcient between each item and the underlying factor provides evidence of convergent validity (Anderson & Gerbing, 1988). The estimated model including completely standardized path coefcients appears in Fig. 1. All of the path loadings were signicant ( p < .05), providing additional evidence of the convergent validity of IMAS. Finally, we conducted a second-order factor analysis to determine if the four subscales were better represented by a higher-order factor structure. This model specied that 2 a single higher-order factor underlied the data. Results (diff (2) = 10.96, p < .01) indicated that the second-order model t the data better than the four-factor solution. GENERAL DISCUSSION This study developed a scale that measures connection-focused tactics of impression management by association. Scale development followed the process identied by Hinkin (1998): item generation, scale development, and scale evaluation. The nal scale consisted of 12 items, 3 for each IMAS factor, and showed acceptable reliability and convergent and discriminant validity. Results generally

158

ANDREWS AND KACMAR

supported the four dimensions of the connection-focused tactics conceptualized by Cialdini (1989). While our results clearly show that the items in IMAS differentiate between favorable and unfavorable others, cross-loadings found in Stage 2 raised concern about whether a few of the items sufciently represented the strength of the relationships the actor had with others. It is the strength of the relationship that distinguishes Boasting from Blurring and Blaring from Burying. Unfortunately, some of the items in IMAS may fail to adequately distinguish between strong and weak relationships with others. In regard to the Blur item that loaded on the Boast factor, it may be that individuals Boast about their relationships with favorable others regardless of whether the relationships are legitimate. The strength of a personal relationship is largely perceptual by both parties. For example, one person may feel close to another but the other does not share the sentiment. This may be especially true in a work environment where some employees feel close to peers while others consider these relationships to be merely work relationships. With respect to the Bury items that loaded on the Blare factor, it is possible that individuals may claim they are in no way connected to unfavorable others regardless of whether a tight connection exists. These individuals may constantly redene their identities by establishing and/or breaking relationships with others based on their current reputation. Alternatively, they may simply fail to acknowledge relationships with others in order to improve their own images. Another possibility is that the negative reputation of the other party outweighs the action of either distancing oneself (i.e., Burying) or stating that no connection ever existed (i.e., Blaring). Moreover, individuals may be very sensitive about protecting their own reputations. They may be more inclined to deny associations with unfavorable others regardless of the type of relationship in order to protect their image. While individuals may not go out of their way to improve their image by boasting about their connections with favorable others, they certainly will not risk their reputation by being associated with unfavorable others. It is this notion that may be responsible for the relatively strong t of the three-factor model in Stage 3 (i.e., Boast, Blur, and Bury/Blare). Results of the tests for discriminant and convergent validity indicated that IMAS appears to tap a new and separate construct that is different from other types of impression management. The correlations between IMAS and Locus of Control, Need for Power, and Self-Monitoring were signicant, providing evidence of convergent validity. Finally, the strong results for the second-order factor analysis indicate that the items that compose IMAS may be best used as a full scale rather than four individual subscales. This scale development has both strengths and weaknesses. Beginning with limitations, the lack of distinction between Blaring and Burying was evidenced early in the development process. In Stage 1, item generation, the coders appeared to be mixed with regard to the separation of these two groups of items. Perhaps a third round of rening and coding prior to the rst administration of the scale might have produced a consistent four-factor solution representative of

IMAS

159

Cialdini and Richardsons (1980) original conception of connection-focused tactics of impression management by association. Second, the nal group consisted of nonfaculty employees at a large state university. Normally, state government working environments are bureaucratic in nature. The abundance of rules and highly structured procedures for promotions and the limited resources available for merit raises may squelch the need for individuals to impression manage. Thus, results from this group may be limited in their generalizability to other types of organizations. Since we found signicant results in an environment with potential range restriction, some condence can be placed in the fact that indirect impression management does exist in organizations. This study also has strengths. First, the development of IMAS allowed some conrmation of the connection-focused tactics of impression management by association. Boasting and Blurring were distinguished from one another and both were distinguished from the Blaring/Burying tactic. Second, IMAS displayed discriminant and convergent validity, lending support to the construct validity of this type of indirect impression management. Third, IMAS is short and easy to use, thereby providing researchers with an efcient tool for measuring impression management by association. Finally, an accepted scale development procedure was followed to develop IMAS that incorporated a variety of samples and data analysis techniques. This study has provided researchers a way to measure a theoretically and intuitively appealing concept. However, it is only the rst step in a long stream of investigations that can address this phenomenon. Further renements of IMAS also should be undertaken. This could involve altering the Bury and Blare items to better distinguish the two types of tactics in an effort to determine whether Burying and Blaring truly are conceptually different. An additional extension to IMAS would be to add items to the scale to ensure that the four domains of impression management by association dened by Cialdini (1989) are fully represented. While existing scales (e.g., Bolino & Turnley, 1999; Wayne & Ferris, 1990) measure direct tactics of impression management, the development of IMAS provides the rst measure of association tactics. When used in full, IMAS should provide an accurate and consistent measure of Cialdinis (1989) connection-focused tactics. Including IMAS in studies that also measure direct impression management tactics will produce a more complete picture of the impression management process. REFERENCES
Anastasi, A. (1982). Psychological testing (5th ed.). New York: Macmillan. Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103, 411423. Ashford, S. J., & Northcraft, G. B. (1992). Conveying more (or less) than we realize: The role of impression management in feedback seeking. Organizational Behavior and Human Decision Processes, 53, 310334. Bagozzi, R. P., & Phillips, L. W. (1982). Representing and testing organizational theories: A holistic construal. Administrative Science Quarterly, 27, 459489.

160

ANDREWS AND KACMAR

Baron, R. (1986). Self presentation in job interviews: When there can be too much of a good thing. Journal of Applied Psychology, 16, 1628. Bentler, P. M. (1990). Comparative t indices in structural models. Psychological Bulletin, 107, 238 246. Bolino, M. C., & Turnley, W. H. (1999). Measuring impression management in organizations: A scale development based on the Jones and Pittman taxonomy. Organizational Research Methods, 2, 187206. Cialdini, R. B. (1989). Indirect tactics of image management: Beyond basking. In R. A. Giacolone & P. Rosenfeld (Eds.), Impression management in the organization (pp. 4556). Hillsdale, NJ: Erlbaum. Cialdini, R. B., Borden, R. J., Thorne, A., Walker, M. R., Freeman, S. I., & Sloan, L. R. (1976). Basking in reected glory: Three (football) eld studies. Journal of Personality and Social Psychology, 34, 366375. Cialdini, R. B., & Richardson, K. D. (1980). Two indirect tactics of image management: Basking and blasting. Journal of Personality and Social Psychology, 39, 406415. Fletcher, C. (1989). Impression management in the employment interview. In R. A. Giacolone & P. Rosenfeld (Eds.), Impression management in the organization (pp. 269281). Hillsdale, NJ: Erlbaum. Ford, J. K., MacCallum, R. C., & Tait, M. (1986). The application of exploratory factor analysis in applied psychology: A critical review and analysis. Personnel Psychology, 39, 291314. Gardner, W. L., & Martinko, M. J. (1988). Impression management in organizations. Journal of Management, 14, 321338. Giacolone, R. A., & Rosenfeld, P. (1989). Impression management in the organization. Hillsdale, NJ:Erlbaum. Gilmore, D. C., & Ferris, G. R. (1989). The politics of the employment interview. In R. W. Eder & G. R. Ferris (Eds.), The employment interview: Theory, research, and practice. Beverly Hills, CA: Sage. Hair, J., Anderson, R. O., & Tatham, R. (1987). Multidimensional data analysis. New York: Macmillan. Hinkin, T. R. (1998). A brief tutorial on the development of measures for use in survey questionnaires. Organizational Research Methods, 1, 104121. Joreskog, K. G., & Sorbom, D. (1993). LISREL 8: Structural equation modeling with the SIMPLIS command language. Hillsdale, NJ: Erlbaum. Judge, T. A., & Bretz, R. D., Jr. (1994). Political inuence behavior and career success. Journal of Management, 20, 4365. Kacmar, K. M., Delery, J. E., & Ferris, G. R. (1992). Differential effectiveness of applicant IM tactics on employment interview decisions. Journal of Applied Psychology, 22, 12501272. Kinicki, A. J., & Lockwood, C. A. (1985). The interview process: An examination of factors recruiters use in evaluating job applicants. Journal of Vocational Behavior, 26, 117125. Kipnis D., Schmidt S. M., & Wilkinson, I. (1980). Intraorganizational inuence tactics: Explorations in getting ones way. Journal of Applied Psychology, 65, 440452. Leary, M. R., & Kowalski, R. M. (1990). Impression management: A literature review and two component model. Psychological Bulletin, 107, 3447. Mathieu, J. E., & Farr, J. L. (1991). Further evidence for the discriminant validity of measures of organizational commitment, job involvement, and job satisfaction. Journal of Applied Psychology, 76, 127133. McIver, J. P., & Carmines, E. G. (1981). Unidimensional Scaling. Beverly Hills, CA: Sage. Medsker, G. J., Williams, L. J., & Holohan, P. J. (1994). A review of current practices for evaluating causal models in organizational behavior and human resources management research. Journal of Management, 20, 439464. Morrison, E. W., & Bies, R. J. (1991). Impression management in the feedbackseeking process: A literature review and research agenda. Academy of Management Review, 16, 522541. Mulaik, S. A., James, L. R., Van Alstine, J., Bennett, N., Lind, S., & Stilwell, C. D. (1989). Evaluation of goodness-of-t indices for structural equation models. Psychological Bulletin, 105, 430445.

IMAS

161

Price, J. L., & Mueller, C. W. (1986). Handbook of organizational measurement. Marsheld, MA: Pitman. Ralston, D. A. (1985). Employee ingratiation: The role of management. Academy of Management Review, 10, 477487. Richardson, K. D., & Cialdini, R. B. (1981). Basking and blasting: Tactics of indirect self-presentation. In J. T. Tedeschi (Ed.), Impression management theory and social psychological research (pp. 4153). New York: Academic Press. Rotter, J. B. (1966). Generalized expectancies for internal versus external locus of control. Psychological Bulletin, 92, 641669. Schlenker, B. R. (1980). Impression management: The self concept, social identity, and interpersonal relations. Melbourne, FL: Krieger. Schlenker, B. R., & Leary, M. R. (1982). Social anxiety and self presentation: A conceptualization and model. Psychological Bulletin, 92, 641669. Schriesheim, C. A., & Hinkin, T. R. (1990). Inuence tactics used by subordinates: A theoretical and empirical analysis and renement of the Kipnis, Schmidt, and Wilkinson subscales. Journal of Applied Psychology, 75, 246257. Schriesheim, C. A., Powers, K. J., Scandura, T. A., Gardiner, C. C., & Lankau, M. J. (1993). Improving construct measurement in management research: Comments and a quantitative approach for assessing the theoretical content adequacy of paper-and-pencil survey-type instruments. Journal of Management, 19, 385418. Shultz, K. S., & Chavez, D. V. (1994). The reliability and factor structure of a social desirability scale in English and in Spanish. Educational and Psychological Measurement, 54, 934940. Snyder, M. (1974). Self-monitoring of expressive behavior. Journal of Personality and Social Psychology, 4, 526537. Spector, P. E. (1988). Development of the work locus of control scale. Journal of Occupational Psychology, 61, 335340. Steers, R. M., & Braunstein, D. N. (1976). A behaviorally based measure of manifest needs in a work setting. Journal of Vocational Behavior, 9, 251266. Stevens, C. K., & Kristof, A. L. (1995). Making the right impression: A eld study of applicant impression management during job interviews. Journal of Applied Psychology, 80, 587606. Tedeschi, J. T., & Melburg, V. (1984). Impression management and inuence in the organization. In S. Bacharach & E. J. Lawler (Eds.), Research in the sociology of organizations (Vol. 3, pp. 3158). Greenwich, CT: JAI Press. Thacker, R., & Wayne, S. (1995). An examination of the relationship between upward inuence tactics and assessments of promotability. Journal of Management, 21, 739756. Wade, K. J., & Kinicki, A. J. (1997). Subjective applicant qualications and interpersonal attraction as mediators within a process model of interview selection decisions. Journal of Vocational Behavior, 50, 2340. Wayne, S. J., & Ferris, G. R. (1990). Inuence tactics, affect, and exchange quality in supervisorsubordinate interactions: A laboratory experiment and eld study. Journal of Applied Psychology, 75, 487499. Wayne, S. J., & Kacmar, K. M. (1991). The effects of impression management on the performance appraisal process. Organizational Behavior and Human Decision Processes, 48, 7088. Wayne, S. J., & Liden, R. C. (1995). Effects of impression management on performance ratings: A longitudinal study. Academy of Management Journal, 38, 232260. Wayne, S. J, Liden, R. C., Graf, I. K., & Ferris, G. R. (1997). The role of upward inuence tactics in human resource decisions. Personnel Psychology, 50, 979997. Received: August 31, 1999

Vous aimerez peut-être aussi