Vous êtes sur la page 1sur 30

International Journal of Advances in Science and Technology,

Vol. 3, No.1, 2012


An Approach to an Unbiased Estimator for Semi Random
Sample of a Finite Population

Habib Ahmed Elsayir
(PhD Statistics)
(Habibsayiroi@Yahoo.com)
Umm Al Qura University(K.S.A)
Al Qunfudha university College
Dept.of Mathematics

Abstract

This paper provides some discussion about the problem of an unbiased variance estimator for semi
random sampling. An investigation of systematic and related sampling methods and some estimators of the
variance of the sample mean is presented. Comparison among these samples and the estimators are made
using a finite population . Results of variance estimation for semi_ random sampling are illustrated. A
method is proposed to estimate the variance of the mean of a finite population, using a systematic sample - a
part of the underlying study - and a simple random sample drawn from the remaining population .
Key words: Systematic sample, Finite population, Semi random sampling, Unbiased estimator,
Multiple-Start Systematic Sampling (MSSS)
1.0. Introduction:
Although sampling is undertaken for many purposes, interest usually focuses on certain population
characteristics, such as mean, total ,ratio of two totals or means, or proportion of units that fall into some
defined class. In this paper we try to discuss an approach to an unbiased estimator of the mean and
variance of a finite population using a specific sample can be referred to as semi random sampling ,which
in fact consists of both simple random sample and systematic sample. It is known that the precision of any
estimate made from a sample depends both on the method by which the estimate is calculated from the
sample data and on the plan of sampling .Therefore, the paper provides a general discussion on the
related sampling methods with comparison on the basis of properties and precision. A general review and
comparisons between some related sampling methods will be introduced. Theoretical methodology for
January Page 121 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
constructing estimation for underlying finite population will also be undertaken to look into unbiased and
efficient estimators.
Problem and Objectives :
Several formulae have been developed for the variance of the mean of systematic sampling, which is
related to some extent to a kind of cluster sampling in which the clusters contain n elements and the
sample consists of one cluster. Because systematic sample can be viewed as a random selection of clusters
,it is not always possible to give an unbiased or even consistent estimator of the design variance .A general
practice in survey fields is to use random sample to estimate the variance among other goals of the study
using simple random sample formula, though this may lead to unbiased estimators for the population
parameters of interest .The several attempts to get unbiased estimates of the variance using a mixture of
systematic and simple random sampling were not encouraging. Thus ,an approach to find an unbiased
estimator using a constrained mixture of two samples is needed.
Review of Literature:
Estimation in finite populations can be made under a different approach known in the literature as
model-based design in which it is supposed that the finite population is drawn from an infinite
population(super population),see (Vallian etal,2000),and (Brewer ,2002).There are some different
approaches to solve the problem of variance estimation considered in the literature. One of these
approaches is to make some assumptions about the structure of the population, where the population is
assumed randomly, which studied by (Cochran,1977) ,(Sampord ,1967) and others. These estimates
provide generalized unbiased estimation and sometimes overestimate variance. (Cochran,1977) has
discussed the models with linear trend, stratification only and smooth periodic trend. Another approach
mentioned by( Zinger ,1980),consists of taking more than one sample, a common approach is known as
multiple start sampling in which different random starts , j=1,2,,P (1 , ) and obtain
systematic samples each of size =m/ with i=0,1,, -1. (Sampford ,1962) stated that several
systematic samples, independently sited at random, will usually give a slightly less precise estimate than a
single systematic sample of the same total size .(Kirk,1984) and( Kirk,1985) has given an intensive
investigation of some estimators of variance for systematic sampling .
2.0. Comparing Systematic and other Sampling Methods:
Since our topic viewed as consisting of two methods of sampling, it persists discussing each of these
methods with their relations to other sampling methods such as stratified random sampling and cluster
sampling ,especially in the part of precision.
In practice, a simple random sample is drawn unit by unit, numbered from 1 to N.A series of random
numbers between 1 to N is then drawn by some way of selection. It is important to note that a complete
frame including all units of population is necessary and no supplementary information is required.
January Page 122 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Simple random sampling provides unbiased estimates of population means and of the variances of the
estimates, thus the sample mean is an unbiased estimator of the population mean ,(Freund,1992).
Systematic sampling is a method used when the population elements are arranged in a specific order
from which a sample can be drawn in a systematic way rather than generate a simple random sample.
Systematic sampling is especially applicable when the population to be studied is arranged in some order,
files in a cabinet, crops in a field, goods in a whorehouse, data in time etc. In such cases it may be easier to
draw random sample in a systematic way rather than generate a simple random sample. For example, a
supermarket, which wishes to investigate customer attitudes, may interview each 3rd or 10th customer
arriving to the shop. The same principle is applied to many phenomena. For example the temperatures in
a city are registered each hour in order to determine middle temperatures. It is straight forward that
systematic sampling leads to a more representative survey than simple sampling. If, for example, the
temperature is registered each day at 12 o'clock, the middle temperature would be wrongly
estimated,(Jorgen Lauridsen ,2005).
To select a systematic sample of n elements from a population of N elements, we divide the N elements
in the population to n groups of k elements and then randomly select the first element out of the first k
elements in the population, and then we select every k-th unit afterwards until we have a sample of n
elements. This type is called and every k-th systematic sample ( Cochran,1977).If the units arranged in a
circular way, and then take a random start from 1 to N and thereafter every k-th until n+1 units are
selected, this is called circular systematic sampling ,(Chakravarti etal ,1967).If the units in the aggregate
are first arranged in a linear order before selection, this is called linear systematic sampling, (Chakravarti
etal ,1967).
In some situations, systematic samples tend to be more representative than could be simple random
samples drawn from the same population..This depends upon the arrangement ,or frame from which the
sample is drawn. If a serial correlation exists (when elements in the population tend to resemble other
nearby elements more than they resemble those at greater distances),then the systematic sample will
generally tend to be more representative than the simple random sample,( Stehman S. V.,1992).
For instance, if the income of residents are investigated, and the sample is drawn by selecting every k-
th house, from a complete map of the area, we will tend to get a more representative sample than would
be obtained by a simple random sampling procedure. The systematic sample will be less subject to source
of variability ,and for that reason, will tend to be more representative than the simple random sample
and hence it is preferred in implementation because of its ease and design efficiency,( Opsomer&
Francisco ,2011). It can be stated that the degree to which the systematic sample would be more
representative than simple random sample depends upon the degree to which these tends to be a
clustering of similar elements in the list or arrangement from which the sample was drawn. It is seemed
that ,as the clustering becomes more obvious, the advantages of the systematic sample become greater
.However, it is the simple random sample to which all the formulas introduced thus for apply. The
systematic sample, on the other hand, is expected to be less representative than the simple random sample
January Page 123 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
in situations where there is a cyclical or periodic movement of the data and the length of the period of the
cycle tends to be equal to or close to k.
The method of systematic sampling has some advantages over other methods in general and over
simple random sampling in particular. The main advantage of using systematic sampling over simple
random sampling is its simplicity. Joan (2009).Some of these advantages are: (Opsomer &Francisco
,2011), (Brase,1987).
1. The actual selection of a systematic sample is easier and cheaper than that of a simple random sample
,and the population size need not be known exactly for systematic sampling.
2. Any trend, or clumping of similar units, in the population ,will tend to be more actually
represented in a systematic rather than in simple random sampling. Hence the systematic sample can
be expected to give the more precise estimate.
2.4.Systematic &Stratified Random Sampling:
Many populations are naturally divided into a number of non-overlapping subpopulations, denoted as
strata or clusters. Generally the population is divided into strata, each consisting of individuals . If
the estimator of the mean in each stratum is unbiased, then the stratified estimator of the mean for the
stratified sample is an unbiased estimator of the population mean ,. In most cases stratified sampling
leads to more efficient sampling than simple random sampling, so that higher precision(less variance) can
be obtained in the estimates. (Jorgen Lauridsen,2005).Precision can be calculated based on different
allocations of the sample, i.e. choices of , .
Viewed as cluster sample( Opsomer &Francisco ,2011), systematic sample ordering the population and
selecting each individual, can be viewed as a special case of stratified sampling:
Let stratum 1 be the first individuals, stratum 2 the next individuals etc. Choose a number ,
and select individuals numbered the

The systematic sample is spread more over the entire population than stratified sample, because in
stratified sampling the samples in the strata are drawn separately .This adds precision in some cases.
(Brase,1987). It can be noticed that cluster sampling may also be viewed as a special case of stratification,
where stratified sampling is superior to random sampling because it reduces sampling error. Stratified
random samplings generally have more statistical precision than simple random sampling.( William
M.G.,2006). Hoever,it can be stated that systematic random sampling is better than systematic sampling,
January Page 124 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
but usually is not as good as simple random sampling, and is not much easier to implement,( Philip B.
Stark,2011).
3.1. Approach of Estimation of the population parameters:
The precision of any estimate made from a sample depends both on the method by which the estimate
is calculated from the sample data and on the plan of sampling. In the following analysis, let denotes
the j-th element of the i-th systematic sample ,so that j =1,2,3,,k .The mean of the i-th sample is denoted
by .The systematic sampling estimator of the population mean, .The variance of this mean estimator
requires some assumptions as below:
1-When the population values are assumed to be in no particular order with respect to the variance of the
estimator of the mean is the same as in the case of simple random sampling. (Aczel,1989).
( )
( )
( ) 1 . 3
2 2
S
Nn
n N
Y S sy

=
2-When the mean is constant within each stratum of K elements but different from stratum to stratum,
the estimated variance of the sample mean is:
( )
( )
( )
( ) 2 . 3
1 2
1
2
2

=

=
+
n
Y Y
Nn
n N
Y S
n
i
k i i
sy

3-When the population is assumed to be either increasing or decreasing linearly in the variable of
interest, and when the sample size is large, the appropriate estimator of the variance of our estimator of
the means given by( Aczel,1989) as:
( )
( )
( )
( ) 3 . 3
2 6
2
1
2
2
2

=

=
+ +
n
Y Y Y
Nn
n N
Y S
n
i
k i k i i
sy
2 1
/
s s n I for
According to (Cochran,1977) , the variance of the mean of a systematic sample is:
( )
( )
( ) 4 . 3
1
.
1
2 2
wsy sy
S
N
n k
S
N
N
y V

=

Where :
January Page 125 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
( )
( )

= =

=
n
j
i
ij
k
i
wsy
Y y
n k
S
1
2
1
2
1
1

is the variance among units that lie within the same systematic sample. (Cochran,1977) also showed that:
( ) ( ) | ( ) 5 . 3 ] 1 1
1
2
w sy
n
N
N
n
S
y V +

=

Where :
is the correlation coefficient between pairs of units that are in the same systematic sample. It is defined
as:
( )( )
( )
2
Y y E
Y y Y y E
ij
iu ij
w


=

Which gives:
( )( )
( ) ( ) ( ) 6 . 3
1 1
2
1
2

=
<


=
k
i
u j
iu ij w
Y y Y y
S N n

And:
( )
( )
( )
2
1 1
2
1
1

= =

= =
k
i
j ij
n
j
wst sy
Y y
k n
S y V

Where
( ) ( ) u
iu
j
ij wst
Y y Y y E =
( ) ( )
( ) ( )
2
1
1 1
2
wst
u iu ij
k
i
wst
S
Y y Y y
u j
k n n

<

=

=


The above quantity is the correlation between the deviations from the stratum means of pairs of items
that are in the same systematic sample. Therefore a systematic sample has the same precision as the
corresponding stratified sample, with one unit per stratum, if 0 =
w
. The variance of the mean from
stratified sample is:
January Page 126 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
( ) ( ) 7 . 3
2
n
S
N
n N
y V
wst
st
|
.
|

\
|

=

and the variance
|
.
|

\
|

st
y V is found directly from the systematic sample totals as:
( ) ( )
( ) ( ) 8 . 3
1
1
2
1
2
1
2

=
=
=
= =
k
i
i
k
i
i sy st
Y n y n
k n
Y y
k
V y V

A great caution must be taken when considering variance estimation in systematic sampling. The
following procedure for considering variance estimation issues were given by (Kirk,1984).
1) Gather as much prior information as possible about the nature and ordering of the population.
2) If an auxiliary variable closely related to the estimation variable, is available for all units in the
population, then try several variance estimators on this variable. This investigation may provide
information about which estimator will have the best properties for estimating the variance of the
estimation variable.
3) Use the prior information in step (1) above to construct a simple model for the population.
4) Keep in mind that most surveys are multipurpose, and it may be important to use different variance
estimators for different characteristics.
Step 1-4 essentially suggest that one should know the population well before choosing a variance
estimator, which is exactly the advice most authors have suggested before using systematic sampling
(Kirk,1984).The use of systematic sampling is not generally to be recommended unless very large samples
are to be taken and a great deal of care is to be expended in the analysis of them.( Quenoienille,1966).
A danger in systematic sampling is that the characteristics being studied may have a certain pattern or
periodicity in the list .On Census record sheets, for example, the first names on the sheets tend to be
above average in income or rental value , for instance corner houses tends to have a higher rental value
than houses in the middle of the block.
Used widely in surveys of finite population, this method picks up any obvious or hidden stratification in
the population when used properly. It can be regarded as a random selection of one cluster; however, it is
not always possible to obtain an unbiased or even consistent estimator of the design variance. Hence,
biased estimators of variance must be sought if we are to estimate the precision of our investigation or
survey estimators from the sample itself.( Kirk ,1984).
January Page 127 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Finally, from all the previous discussions, it is clear that a complete frame for systematic sampling to
apply is necessary, but the construction of the frame and the drawing of the sample can proceed
simultaneously, as when every rth individual is selected in the systematic enumeration of trees in a
plantation. Furthermore, a systematic sample may provide an unbiased estimate of the population mean.
This estimate will often be more precise than that obtainable from a random sample of the same size, but
its sampling variance can not be estimated.
3.2.Estimating the Variance for Semi random Sampling(SRS):
let a finite population consist of units numbered from I to k m N . = ,where m and k are positive
integers. Let the values of the units be , ,.... 3 , 2 , 1 , N i x
i
= , with mean T. Let a random integer
0
k be
taken such that k k s s
0
1 .The sample 1 , ... , 1 , 0 ,
0
= + m i i k x
k
is called a systematic sample of
size m. There are K possible and equally probable samples. We denote by X the sample mean of a
systematic sample of size m. According to (Cochran ,1977),we have
( ) T X E
s
= and
( ) ( ) ( ) ( ) ( )
) 9 . 3 (
1 1 ( / ) 1 /
2
w s
m N N m S X V + =

Where
( )
( ) 1
1
2
2

=

=
N
T x
S
N
i
i
and
w
is the correlation coefficient between pairs of units that are in the same systematic sample, that is:
( )( )
( ) ( ) ( ) 10 . 3
1 1
2
0 0
0
1
2
T X T X
S N m
uk k
u j
jk k
k
k
w


=
+
<
+
=


it can be reached that the within-sample variance leads on the average to an underestimate of ( ) s T V if
0 >
w
or to an overestimate if 0 <
w
.
For semi random sampling, first a systematic sample of size m is drawn .Then a simple random
sample of size n is drawn from the remaining N- m units, We will denote by ( ) s X and ( ) r X the means
of these two samples respectively. This procedure will be referred to as semi random sampling (SRS).We
shall take the following weighted means of ( ) s X and ( ) r X as an estimator of T :
January Page 128 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Using the conditional expectation
r
E over random sampling with a fixed systematic sample and then
k, over systematic sampling. Following ( Zinger,1980),we have
( ) ( ) | |
( ) ( ) | | m N X m T X E
X X E E X E
s s
s
r s
r s
+ =
+ =
/ | o
| o |

Where
T N T =
To calculate
( ) | | { }
2 2
T X X E E X V r s
r s
+ = | o |
We have
( ) ( ) ( ) 11 . 3
2 2
T X V X E E s s
r s
+ =
( ) ( ) ( ) 12 . 3
2
s
r s
r s
r s
X E E
m N
m
m N
T T
X X E E

=
We can combine these results and write
( )
( )
( ) ( )
( )
( )( )
( )
( ) ( )
( )
( )( )
( )( )
( )
( ) 13 . 3
1 (
1
1
1 1
1
1 1
1 2
1
1
2
2 2
2
2
S N T N
m N m N nk
k n m N
X E E
m N k n
n m
m N k n
T T n
m N m N n
T n
X E E
s
r s
r
r s
+


+


+



=

Combining and using (3.11), (3.12) and (3.13) we can write ( ) ( ) | X V as

( ) ( ) ( ) ( ) ( ) ( ) ( ) 14 . 3 ,
3
2
2
2
1
S X V S T X V | o | o | o | + + =
Where
January Page 129 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
( )
( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( ) ( )
) 15 . 3 (
1 1 / 1 / 1 (
, 1 / 1
0
2 2 2
3
2
2
1
=
=
=
m N k n n m N k k
m N nN n m N N
| | | o
| | o
| o

To estimate the ( ) ( ) | X V let us consider the sample sum of squares.
( ) ( ) ( )
2
1

+
=
=
n m
i
i
X x | |
and calculate
( ) ( ) ( )
(

=

+
=
n m
i
r s
i r s
X X x E E E
1
2
| o |
We can then reach to :
( ) ( ) ( ) ( ) ( ) ( ) 16 . 3 , (
3
2
2
2
1
s X V C S C T C E | | | | + + =
Where

( ) 0
1
| C
( ) ( )
( )
(



+ +


+

=
2
2
1 1
2
1
| | |
m N n
n m N
n m
m N
m m N
n m
N
N
C
( )
( ) ( )
( )
( )
( ) 17 . 3
1
1
1
1
1
2
1
2 2
2
3
|
|
|
(

+
+
(



+ +

+ +
=
m N n
n m N
k
k
n m
m N
n m
nk n
k k
N kn n m
C

We can approach the problem of estimating ( ) ( ) | X V finding the values of A and
|
such
that:
( ) = | A E( ( ) ( ) , | X V
Where A is nonzero constant. We can obtain | as the solution of:
( ) ( ) ( ) ( ) ( ) 18 . 3 0
2 3 3 2
= | | o | | o C C
January Page 130 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
If m and n are fixed and k increases, we can find that:
( ) ( ) ( ) ( ) n S X V X V / 1
2
= |
and the efficiency, eff, of ( ) | X defined as
eff = ( ) ( ) | | | X V n m S ( /
2
+
tends to ( ) n m n + / . Thus this approach is unsatisfactory ,(Zinger,1980). Another way of estimating
V( X ( | ) is to consider any two values of | , say
1
| , and
2
| , such that
( ) ( ) ( ) ( )
1 3 2 2 2 3 1 2
| | | | C C C C = and use (3.16) by replacing all expected values by the corresponding
sample values. We obtain the unbiased estimators
( )
( ) ( )
( ) ( ) ( )
) 19 . 3 (
1 /
0
1
1 2
2
1 2
2

=
m N n m N n m
X X n
N
N
S
r s
o o
o o

Of
2
S and
( )
( ) ( ) ( ) ( )( )
( ) ( ) ( )
(

+
+
=
2 1
1 /
1 / 0
1
o o

n m m N n m N
X X n m n m N n m N
k V
r s

Where
N kn n m + + =
1
o
And
( ) ( ) | | 1 / 1
2
+ + = m N n m n nk o
we can obtain these results using any
1
| and
2
| as long as:
a. ( ) ( ) ( ) ( ) and C C C C
1 3 2 2 2 3 1 2
| | | | =
b. or n m =
n m if = = + 1
2 1
| |
January Page 131 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
It can be showed that 0
2
> S , which means that this procedure provides an unbiased and positive
estimator of
2
S . On the other hand, v is an unbiased estimator of ( ) s X V but is not always positive.
(Zinger ,1980).
Using (3.14) and (3.15), we obtain the unbiased estimator
( ) ( ) ( ) ( ) | | X V of X v
We now have, for all S
,
|
( ) ( ) ( ) ( ) ( ) 20 . 3
3
2
2
v S X v | o | o | + =
And
( ) ( ) ( ) ( ) | | X V X v E =
There choices for | have been considered.
1. The value of | , which immunized
( ) ( ) .
3
2
2
v S | o | o +
This produces negative estimates of ( ) ( ) | X v for some samples and is not retained. The
value ( ) n m n + = / | , corresponding to a natural weight of s X and r X fo1lowing (Zinger ,1980) , we can
say that if
( )
( ) ( )
( )
) . (
S
n m N
n m N
n m n X V
then N
w
21 3
) /
, 1 / 1
2


< +
<


2. The value which is similar to the relationship between the variances of the means for SRS and
systematic sampling. Unfortunately, some ( ) ( ) ) / n m n X v + are found negative and this choice is
therefore not recommended.
3. The value = | (), corresponding to an un-weighted average of
s
X , and
r
X produces positive
results for ( ) X v () and is recommended for this reason.
January Page 132 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
It can be shown also, by minimizing ( ) | X V ( , that if K is large, the value
2
1
= | is optimal for
( )
w
m m n 1 ( 1 / + =
which gives the best choice for n if
w
is known.
4.0.Emperical Analysis:
In order to illustrate and verify results concerning the estimation of the semi random sample, we
consider data show the main annual food crops production of the Sudan for 25 years
since1970/1971.These crops are sorghum, wheat and millet. Each of these is considered as a separate
population ,so we have population 1 see table(4-1),population 2,and population 3 for ,sorghum, wheat and
millet respectively, and the total population size N =25 for each population. For random and stratified
sampling ,analysis of variance of the population into" between rows " and "within rows " is desired. This
is presented in tables (4-2),(4-4) and (4-6).
4.1.Results for Systematic Sampling:
Now we are going to illustrate the analysis of variance of the population .The purpose for the analysis
of variance is to test for significant differences between means. The within group variability is usually
referred to as error variance. If the strata are homogeneous, the variability within-groups is expected to
be lower than the variability for the population as a whole. From the table (4-2), it is seen
that As from table (4-6), it is noticed that .
It is seen that is always less than and in population 1,2 and 3.But is less than
in two cases out of three as can be noted in population 2 and 3.This may be conducted to the
distribution of the data. Results are presented in table(4-7)
The results show that both stratified random sampling and systematic sampling are much more effective
than simple random sampling, but systematic sampling is less precise than stratified random sampling
.Systematic sampling is more precise than simple random sampling because it picks up any hidden or
obvious stratification in the population. It is clear that the variance of is less for a systematic than for a
simple random sampling since the list from which the systematic sample is drawn show a fairly consistent
trend or grouping as the data represent production during series of time. This means that there may be a
population for which systematic sampling is extremely precise and others for which it is less precise than
simple random sampling. Therefore the performance of a systematic sampling in relation to that of
stratified or simple random sampling is mainly dependent on the properties of the population. On the
other hand, a systematic sample will have the same precision as the corresponding stratified random
sample, with one unit per stratum, if which is the correlation between the deviation from the
January Page 133 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
stratum means of pairs of items that are in the same systematic sample, but the data show that the values
of for each of the three populations are negative (around - 0.2 ) .
4.2.Results for Semi Random Sampling(SRS):
In order to illustrate the variance estimation for our underlying method of sampling, let us consider
the populations 1,2, and 3 which are denoted by pop1,pop2, and pop3 respectively, with N =25,k=5.The
results of variance estimation are summarized in tables (4-8),(4-9) and (4-10).In order to obtain the
distributions of , and ( ) ( )
2
1
y V ,all possible samples were considered for n= 1,2,3
.Table (4-8) gives some results for the distribution of . Table (4-9) gives some results for the
distribution of ,as table (4-10) gives some results for the distribution of ( ) ( )
2
1
y V ,and the
distribution of using simple random sampling of size n .
It should be noted that the results in table (4-9) are the obtained unbiased estimators of of .It is
seen that > 0 or in fact 0 which means that this procedure provides an unbiased and positive
estimator, of .On the other hand , is an unbiased estimator of ,but is not always positive .For
instance, if n=1 the values of are -13.24,-15.81 and 4.6 for pop1,pop2 and pop3 respectively.
It worth remarking that the value corresponding to an unweighted average of and
producing a positive result of as can be noted in table (4-10).Therefore, is
recommended for this reason.
4.3.Comparison of Variances:
Now we compare Multiple- Start Systematic Sampling (MSSS) with Semi Random Sampling(SRS)
bases on m+n observations. In this case ,several conditions should be fulfilled .Given a finite population of
units numbered from 1 to N=mk ,We draw without replacement p systematic samples ,each of size m
so that pm = m+n (total sample size). mk =N, where m and k should be integers. We denote by
the variance of the mean, of the p systematic samples, each of size m ,Then a comparison of MSSS and
SRS will be made, Results are shown in table (4-11) using the past three underlying populations, where
the fourth population, is an artificial borrowed population of the 24 values :
13,7,6,7,19,8,8,2,5,20,18,4,0,12,9,16,10,17,15,1,1,6,20,16. Was sampled. Several remarks can be drawn
from the results of table(4-11):
1-Multiple-start systematic sampling(MSSS),in general, can lead to lower values (even zero) of the
estimated variance as can be seen in table(4-11),for population 4,and semi-random sampling(SRS) can
lead to higher values of the estimated variance. In addition, there are seen negative values of the
January Page 134 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
estimated variances for both MSSS and SRS as in both pop1 and pop2.This may not be due to the
conduct of the data but can be a disadvantage of the estimator itself.
2-There is an irregular variation of ,as a function of m , and no simple relation is evident even if m
= m ,the conclusion that SRS is better if and MSSS is better if does not hold in general.
3.The advantage of SRS is that it provides an estimator of both and .
4.4.Discussion of the Results
Semi random sampling is a method of selecting a systematic sample from a finite population and
drawing a simple random from the remaining population. The method of systematic sampling can be
alternative to simple random sampling and specially preferred when the information required to
construct a sampling is available in a list or any other organized form. Although systematic sampling is
used for its convenience of drawing and execution, but its variance of estimator requires some
assumptions about the underlying population order. Some of these assumptions are whether the
population elements follow particular order, constant mean within strata, or the population is assumed to
be either linearly increasing or decreasing in the variable of interest.
In the light of results, systematic sampling can safely be recommended in situations where the ordering
of the population is essentially random or contains at most a mild stratification, for the effects of hidden
periodicies tend to cancel out when systematic sample is drawn from each stratum.
As for a simple random sampling, it is a method which executed in a manner that every element of the
population has equal probability of being included in the sample and a complete frame including units of
population is needed. In addition, simple random sampling gives unbiased estimates of population means
and variances estimates .Thus, when we do not select our sample randomly out of the population of
interest, our sampling results may be biased. Hence, The necessity of simple random sampling arises.
When comparing simple random sampling, results show that the mean of a systematic sample can be
expected to be more precise estimate of the population mean than the mean of a random sample of the
same size. Furthermore, the sampling variance of the mean of a systematic sample from a list can be
expected to be less than that of the mean of a random sample, if there is a fairly consistent trend
throughout the list or any appreciable grouping in the values of the population ,but this sampling
variance can not be estimated without bias from a random systematic sample.
It is seen that the performance of a systematic sampling in relation to that of stratified or simple
random sampling is mainly dependent on the properties of the population. All the evidences show that
both stratified random sampling and systematic sampling are much more effective than simple random
sampling, but generally systematic sampling is less precise than stratified random sampling .The
systematic sampling is approximately as precise as stratified random sampling with one unit per stratum.
January Page 135 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
The difference between the two methods is that the systematic sample is spread more evenly over the
entire population than a stratified random sample, because in the latter, the samples in stratified
sampling the samples in the strata are drawn separately. This adds precision in some cases.
It is ,then , obvious that our SRS from which we are estimating variance is trying to get benefit from
the advantages of both systematic and simple random sampling .In this method, the stratification of a
population must be judged by the investigator's knowledge of the population and surveys carried on
similar populations.
The actual selection of a systematic sample is quicker ,easier and cheaper than that of simple random
sample, and if the population size is known exactly, this advantage is found in a part of semi random
sampling of course not the part of simple random sampling. On the other hand ,semi random sampling is
expected to lessen the danger in systematic sampling when the characteristics being studied may have a
certain pattern or periodicity or trend in the list.
Comparing multiple-start systematic sampling with semi random sampling - put into account the
required conditions - it is found that multiple-start systematic sampling can lead to lower values of the
estimated variances .However, providing more than estimator is an advantage of semi random sampling
over multiple start systematic sampling though some negative values of the estimated variance may be
encountered.
5.0.Conclusion
The problem of an unbiased variance estimator for semi random sampling is investigated for
systematic and related sampling methods and some estimators of the variance of the sample mean is
presented with comparison among these samples and the estimators for a finite population. The proposed
method to estimate the mean of a finite population and to estimate the variance of this estimate is
introduced, using a systematic sample and a simple random sample drawn from the remaining
population to get an unbiased, estimator of the population variance. Results of estimation for semi
random sampling and systematic sampling does not provide a satisfactory estimator of the variance of the
sample mean except when additional assumptions are made. Consider the problem of estimation for semi
random sampling, the followings are recommended:
1.A great caution must be taken about the nature and ordering of the population and several variance
estimators must be used to select the best and efficient estimator.
2.Semi random sampling is advised to use only if there are several samples to be taken and a great deal of
care must be taken in the analysis of data.
3.Since the semi random sampling can be regarded as a selection of one cluster, the variance estimation
method for semi random sampling can be carried in more than one cluster sampling or multistage
sampling in an attempt to get an unbiased estimator of variance.
January Page 136 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
The Tables
Table(4-1):Data for 5 Systematic Samples with N=kn=25
Population (1):Sorghum Production:
Systematic sample numbers Strata means
strata 1 2 3 4 5
1 1535 1592 1301 1691 1792 1582.2
2 2143 2606 2082 2353 1461 2129.0
3 2084 3335 1884 2006 1097 2081.2
4 3597 3277 1363 4425 1536 2839.6
5 1180 3581 4042 2386 3648 2967.4
Totals 10539 14391 10672 12861 9534 11599.4

Source: Data Analysis







January Page 137 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Table(4-2):Analysis of Variance
d.f s.s m s
Between rows(strata)
4

6634830

1658707.5
Within strata 20 16134290 806714.5 =
Totals 24 22769120 2465422 =

Source: Data Analysis














January Page 138 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Table(4-3):Data for 5 Systematic Samples with N=kn=25
Population (2):Wheat Production:
Systematic sample numbers Strata means
strata 1 2 3 4 5
1 163 124 152 235 269 188.6
2 263 289 312 165 231 252.0
3 218 142 176 157 79 154.4
4 199 157 181 247 448 238.6
5 686 895 453 475 1436 591.4
Totals 1529 1607 1274 1279 1436 1425
Source: Data Analysis
Table(4-4) :Analysis of Variance
d.f s.s m s
Between rows(strata)
4

617361.2

154340.3
Within strata 20 233282.8 11664.1 =
Totals 24 850644.5 166004.4 =
Source: Data Analysis
January Page 139 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Table(4-5):Data for 5 systematic samples with N=kn=25
Population (3):Millet Production:
Systematic sample numbers Strata means
strata 1 2 3 4 5
1 439 441 355 285 403 384.6
2 388 449 500 552 309 439.6
3 335 509 227 314 168 310.6
4 419 285 153 495 161 302.2
5 85 308 449 221 973 407.2
Totals 1664 1992 1684 1867 2014 1844.2
Source: Data Analysis
Table(4-6) :Analysis of Variance
d.f s.s m s
Between rows(strata)
4

1443246.8

36011.7
Within strata 20 683703.2 34185.2=
Totals 24 2126950 394996.9 =
Source: Data Analysis
January Page 140 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Table(4-7):Summary of Results of Mean and Variance of Specific Sampling Methods.
sorghum wheat millet
Pop 1 Pop 2 Pop 3
Population mean
N=25

2319.9

285

368.8

2465422 16664.4 394996.9

473361 26560.7 63199.5

154889.2 1866.3 5469.3

4968739.3 705.42 874.7
Source: Data Analysis











January Page 141 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Table(4-8):Distribution of
n Min Max S.e
SRS
Pop 1
m=5
1 16344.5 2057280.5 1538
2 8172.3 982172.8 1065
3 5448.2 618718.2 850
4 4086.1 436982.1 660
5 3286.9 327941.3 628
SRS
Pop 2
m=5
1 2.3204 159322.5 391
2 1.1602 75468.6 276
3 0.7735 47517.3 221
4 0.5800 33541.6 187
5 0.4640 25156.3 163









January Page 142 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Table(4-8):Distribution of (continued)
n Min Max S.e
SRS
Pop 3
m=5
1 2.8773 399145.2 616
2 1.4382 179595.2 426
3 0.9591 113078.5 340
4 0.7193 79820.1 288
5 0.5755 59865.2 251
Source: Data Analysis
Note: The skewness of :Pop1= 0.7326 Pop2 = 0.8496 Pop3 = 0.2331













January Page 143 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Table(4-9):Distribution of
n Min Max S.e
SRS
Pop 1
m=5
1 14200877 23668018 1538.4
2 16567714 982172.8 1064.9
3 9047336 618718.2 850

SRS
Pop 2
m=5
1 2.3204 159322.5 391
2 1.1602 75468.6 276
3 0.7735 47517.3 221

SRS
Pop 3
m=5
1 2.8773 379145.2 616
2 1.4382 179595.2 426
3 0.9591 113078.5 340
Source: Data Analysis







January Page 144 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Table(4-10) :Estimated Value for
Sampling Method n

Semi Random Sampling(SRS)
Pop 1
m=5
1 516375.19
2 244653.17
3 154084.04
SRS
Pop 2
m=5
1 39830.36
2 18867.15
3 11879.23
SRS
Pop 3
m=5
1 94785.99
2 44898.6o
3 28269.50
Simple random sample 5 394467.52
6 21027.22
7 40628.25
Source: Data Analysis ,Note: For simple random sample, .






January Page 145 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Table(4-11):Comparison of Estimated Variances Distributions
Pop
No.
Sampling
Method


m
p

n



Min Max S.e
1 MSSS 4 2 - 0.90 8626.28 984486.2 1064.94
SRS 4 4 - 0.30 310546.2 806952.4 720
2 MSSS 4 2 -0.97 1.101 75678.17 276.34
SRS 4 4 -0.33 705.4 83973 186.8
3 MSSS 4 2 -0.99 1.367 180094 408.86
SRS 4 4 -0.33 874.7 20077.9 288.7
1 MSSS 5 4 -0.30 - 61291.9* 436982.72 719.5
SRS 5 3 -0.45 5448.2 618718.46 850.4
2 MSSS 5 4 -0.32 0.580 33541.65 186.7
SRS 5 3 -0.49 -12.375* 47517.29 220.7
3 MSSS 5 4 -0.33 0.719 79820,14 288
SRS 5 3 -0.49 .959 113078.47 340.4





January Page 146 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012

Table(4-11):Comparison of Estimated Variances Distributions (continued)
Pop
No.
Sampling
Method


m
p

n



Min Max S.e
4 MSSS 2 2 0.10 0.00 63.80 12.42
SRS 2 2 0.10 0.05 64.84 10.66
SRS 3 1 - 0.09 0.41 48.39 10.66
4 MSSS 2 3 0.10 0.08 23.52 5.03
MSSS 3 2 - 0.09 0.08 21.33 5.20
SRS 2 4 0.10 0.02 57.87 7.89
SRS 3 3 0.09 0.13 37.76 4.89
SRS 4 2 0.21 0.37 37.33 6.59
4 MSSS 4 3 0.21 0.46 7.19 1.97
MSSS 6 2 0.5 0.01 3.13 1.19
SRS 8 4 -0.10 0.87 5.44 0.53
Source: Data Analysis




January Page 147 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
References
1.Aczel,Amir D(1989).Complete Business Statistics. Boston : Homewood.
2.Brewer,K.(2000) .Combined Survey Sampling Inference.London:Arnold.
3.Chakravarti,I.M etal(1967).Handbook of methods of Applied Statistics.vol II :Planning of surveys and
Experiments., New York: John Willey & Sons.
4.Cochran,William G.(1977).Sampling Techniques, New York: John Willey & Sons.
5.Freund,John E.(1992).Mathematical Statistics 5th edition, New Jersy :Printice -Hall int. Editions.
6.Joan ,Joseph Castillo (2009).Systematic Sampling. Experiment Resources, in htt://www.experiment-
resources.com.
7. Kirk M.Wolter (1985).Introduction to Variance Estimation. Second Edition. Series of Statistics for
Social and Behavioral Sciences. Springer.
8. Kirk M. Wolter, (1984).Variance Estimation for Systematic Sampling . Statistics for Social and
Behavioral Sciences, 298-353. Springer.com.
9.Ministry of Agriculture(1996) .Time Series of Cereals and Oil Crops 1970-71-1995-96 No.3 Sudan.
10.Opsomer,J.D &Francisco Fernandez,M.(2011).Model-based Nonparametric variance Estimation for
Systematic Sampling.in: www.stat.colostate.edu/~jopsomer/.../Nonpara_model_variance.
11. Philip B. Stark(2011). SticiGui Department of Statistics University of California, Berkeley website:
http://www.stat.berkeley.edu/~stark/SticiGui/index.htm.
12.Quenouille,M.H.(1966).Introductory Statistics. Oxford(London): Pergamon Press.
13.Richmond,Samuel B.(1957).Principles of Statistical Analysis, New York: The Ronald Press Company.
14.Sampford(1962).An Introduction to Sampling Theory with Applications to Agriculture. Edinburgh:
Oliver and Boyd.
15. Stehman S. V.(1992). Comparison of systematic and random sampling for estimating the accuracy of
maps generated from remote American Society for Photogrammetry and Remote Sensing, Bethesda,
MD.ly sensed data. Photogrammetric. 1992, vol. 58, n 9, pp. 1343-1350 .
16.Valliant,R.Dorfman,A.and Royall,R(2000).Finite population Sampling and Inference: John Wiley and
Sons, New York.
17. William M.G.(2006).Probability sampling, Research Methods Knowledge Base in Social research
methods.net/samppr.
January Page 148 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
18.Yuan Deng,Lih and Wu,C.F.J.(1987).Estimation of Variance of The Regression Estimator. Journal of
the American Statistical Association ,Vol 82,number 398 PP. 568-576.
19.Zinger,A.(1980).Variance Estimation in Partially Systematic Sampling. Journal of the American
Statistical Association,Vol.75,number 369 PP.206-211.

























January Page 149 of 150 ISSN 2229 5216
International Journal of Advances in Science and Technology,
Vol. 3, No.1, 2012
Authors Profile















Dr. habib Ahmed Elsayir received his Ph.D degree in Statistics from Omdurman
Islamic University, Sudan in 2001. He was appointed manager of Omdurman
Islamic University Branch at AlDaein 2002-2005.Now he is the Head Dept. of
Mathematics ,Alqunfudha University college, University of ummal Qura,Saudi
Arabia.
January Page 150 of 150 ISSN 2229 5216

Vous aimerez peut-être aussi