A Study on the Chain Ratio-Type Estimator of Finite Population Variance
Transcript of A Study on the Chain Ratio-Type Estimator of Finite Population Variance
Research ArticleA Study on the Chain Ratio-Type Estimator ofFinite Population Variance
Yunusa Olufadi1 and Cem Kadilar2
1 Department of Statistics and Mathematical Sciences Kwara State University PMB 1530 Malete Ilorin Nigeria2 Department of Statistics Hacettepe University Beytepe 06800 Ankara Turkey
Correspondence should be addressed to Yunusa Olufadi olufadi14yunusyahoocom
Received 12 August 2013 Revised 15 January 2014 Accepted 15 January 2014 Published 24 February 2014
Academic Editor Shein-chung Chow
Copyright copy 2014 Y Olufadi and C Kadilar This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited
We suggest an estimator using two auxiliary variables for the estimation of the unknown population variance The bias and themean square error of the proposed estimator are obtained to the first order of approximations In addition the problem is extendedto two-phase sampling scheme After theoretical comparisons as an illustration a numerical comparison is carried out to examinethe performance of the suggested estimator with several estimators
1 Introduction
Variations are present everywhere in our daily life It is thelaw of nature that no two things or individuals are exactlyalike For instance a physician needs a full understandingof variations in the degree of human blood pressure bodytemperature andpulse rate for adequate prescriptionAman-ufacturer needs constant knowledge of the level of variationsin peoplersquos reaction to his product to be able to knowwhetherto reduce or increase his price or improve the quality of hisproduct An agriculturist needs an adequate understandingof the variations in climatic factors especially from place toplace (or time to time) to be able to plan on when how andwhere to plant his crop
It is well known that the use of auxiliary information insample survey designs results in efficient estimators of pop-ulation parameters such as variance under some realisticconditions For example when information is available on theauxiliary variable that is positively correlated with the studyvariable the ratio estimator is a suitable estimator for theestimation of the population variance
Let 119875 be a finite population consisting of 119873 units 1198751 1198752
119875119873 The units of this finite population are identifiable in
the sense that they are uniquely labeled from 1 to 119873 and thelabel on each unit is known Let 119910 be the character understudy taking the value 119910
119894on the units 119875
119894(119894 = 1 2 119873)
and assume a sample of size 119899 is drawn by the simple randomsampling without replacement (SRSWOR)
Suppose in a survey problem that we are interested inestimating the population variance 1198782
119910 Isaki [1] presented the
ratio estimator for the population variance using the auxil-iary information The problem of estimating the populationvariance using information on single auxiliary variable hasalso been discussed by various authors including Prasad andSingh [2 3] Biradar and Singh [4] Rueda Garcia and ArcosCebrian [5] Arcos et al [6] Kadilar and Cingi [7] and Singhet al [8]
Themean square error (MSE) of the classical estimator ofthe population variance 1198782
119910 which we denote as 119905
0 is119881(119905
0) =
1198784
1199101198600 Quite often information onmany auxiliary variables is
available in the survey which can be utilized to increase theprecision of the estimate The ratio estimator of populationvariance for a single auxiliary variable denoted as 119905
1suggested
by Isaki [1] and the two-phase sampling (TPS) estimator of 1199051
denoted as 119905lowast1are as follows
1199051= 1199042
119910
1198782
1199091
1199042
1199091
MSE (1199051) = 1198784
119910(1198600+ 1198601minus 21198603)
Hindawi Publishing CorporationJournal of Probability and StatisticsVolume 2014 Article ID 723982 5 pageshttpdxdoiorg1011552014723982
2 Journal of Probability and Statistics
119905lowast
1= 1199042
119910
119904lowast2
1199091
1199042
1199091
MSE (119905lowast1) = MSE (119905
1) minus 1198784
119910(119860lowast
1minus 2119860lowast
3)
(1)
Following Olkin [9] Isaki [1] also presented the ratioestimator of variance using two auxiliary variables as follows
1199052= 1198821
1199042
119910
1199042
1199091
1198782
1199091+1198822
1199042
119910
1199042
1199091
1198782
1199092
MSE (1199052) = 1198784
119910(1198621+1198822
11198622minus 211988211198623)
119905lowast
2= 1198721
1199042
119910
1199042
1199091
119904lowast2
1199091+1198722
1199042
119910
1199042
1199092
119904lowast2
1199092(TPS approach of 119905
2)
MSE (119905lowast2) = 1198784
119910(1198631+1198722
11198632minus 211987211198633)
(2)
where119882119894and119872
119894 for 119894 = 1 2 are weights chosen tominimize
the MSE of 1199052and 119905lowast2 Furthersum119882
119894= 1 andsum119872
119894= 1 where
1199042
1199091=1
119899
119899
sum
119894=1
(1199091119894minus 1198831)2
119904lowast2
1199092=1
119899
119899
sum
119894=1
(1199092119894minus 1198832)2
119904lowast2
1199091=
1
1198991015840
1198991015840
sum
119894=1
(1199091119894minus 1199091015840
1)2
119904lowast2
1199092=
1
1198991015840
1198991015840
sum
119894=1
(1199092119894minus 1199091015840
2)2
1199091015840
1=
1
1198991015840
1198991015840
sum
119894=1
1199091119894 119909
1015840
2=
1
1198991015840
1198991015840
sum
119894=1
1199092119894
1198621= 1198600+ 1198602minus 21198604 119862
lowast
1= 119860lowast
2minus 2119860lowast
4
1198631= 1198621minus 119862lowast
1 119862
2= 1198601+ 1198602minus 21198605
119862lowast
2= 119860lowast
1+ 119860lowast
2minus 2119860lowast
5 119863
2= 1198622minus 119862lowast
2
1198623= 1198602+ 1198603minus 1198604minus 1198605 119862
lowast
3= 119860lowast
2+ 119860lowast
3minus 119860lowast
4minus 119860lowast
5
1198633= 1198623minus 119862lowast
3
1198600=1
119899(120582400
minus 1) 1198601=1
119899(120582040
minus 1)
1198602=1
119899(120582004
minus 1) 1198603=1
119899(120582220
minus 1)
1198604=1
119899(120582202
minus 1) 1198605=1
119899(120582022
minus 1)
119860lowast
1=
1
1198991015840(120582040
minus 1) 119860lowast
2=
1
1198991015840(120582004
minus 1)
119860lowast
3=
1
1198991015840(120582220
minus 1) 119860lowast
4=
1
1198991015840(120582202
minus 1)
119860lowast
5=
1
1198991015840(120582022
minus 1)
120582 =120583119886119887119888
1205831198862
2001205831198872
0201205831198882
002
120583119886119887119888
=1
119873 minus 1
119873
sum
119894=1
(119910119894minus 119884)119886
(1199091119894minus 1198831)119887
(1199092119894minus 1198832)119888
(3)
where 119886 119887 and 119888 are nonnegative integersSeveral authors (Srivastava et al [10] Upadhyaya et al
[11] and Singh et al [12]) adopted TPS procedure proposedby Chand [13] and have suggested some chain ratio-typeestimators for estimating populationmean119884of119910 In the samevein Gupta et al [14] and Singh et al [8] proposed the fol-lowing classes of estimators under the assumption that thepopulation variance of the first auxiliary variable 1198782
1199091is not
known but the population variance of another auxiliaryvariable119883
2closely related to119883
1is available TheMSEs of the
estimators suggested by Gupta et al [14] and Singh et al [8]are respectively given by
1199053= 1199042
119910(1199042
1199091
119904lowast2
1199091
)
1198681
(119904lowast2
1199092
1198782
1199092
)
1198682
MSEmin (1199053) = 1199044
119910[1198600minus 120601(
1198602
3
1198601
) minus (1198602
4
11989910158401198602
)]
1199054= 1199042
119910(1199042
1199091
119904lowast2
1199091
)
1198691
(119904lowast2
1199092
1198782
1199092
)
1198692
(1199042
1199092
1198782
1199092
)
1198693
MSEmin (1199054) = 1199044
1199101198600[1 minus 120575120574
lowast2
012minus 120579120588lowast2]
(4)
where 1198681 1198682 and 119869
119894for 119894 = 1 2 3 are constants chosen to
minimize theMSE of 1199053and 1199054 120601 = (1119899minus1119899
1015840) 120579 = 119899119899
1015840 120575 =((1198991015840minus119899)119899
1015840) 120574lowast2012
= (11986021198602
3minus2119860311986041198605+11986011198602
4)1198600(11986011198602minus
1198602
5) 120588lowast2 = 119860
2
411986001198602
In most studies several variables are considered simul-taneously either to explain or estimate (predict) the studyvariable In most cases information on several auxiliaryvariables closely related to the study variable may be easilyobtained on all units in the population For example whileconducting an educational survey the investigator may beinterested in studying characteristics such as age genderhours spent on studying per day sitting position parentrsquos edu-cational level parentrsquos income relationship with lectures andaccess to facilities (eg library internet laboratory) amongothers With the main aim of suggesting a more efficientestimator we propose in this paper under SRSWOR a chainratio-type estimator for estimating the population variancewhen information on two auxiliary variables is available Inaddition the problem is extended to the case of TPS
2 The Suggested Estimator
Following Abu-Dayyeh et al [15] we define an estimator forestimating the population variance 1198782
119910 as follows
119905 = 1199042
119910(1198782
1199091
1199042
1199091
)
1205721
(1198782
1199092
1199042
1199092
)
1205722
(5)
where 1205721and 1205722are real constants to be determined such that
the MSE of 119905 is minimum
Journal of Probability and Statistics 3
To determine the bias and MSE of 119905 we define
1199042
119910= 1198782
119910(1 + 119896
0) 119904
2
1199091= 1198782
1199091(1 + 119896
1) 119904
2
1199092= 1198782
1199092(1 + 119896
2)
(6)
such that
119864 (1198960) = 119864 (119896
1) = 119864 (119896
2) = 0
119864 (1198962
0) = 119860
0 119864 (119896
2
1) = 119860
1
119864 (1198962
2) = 119860
2 119864 (119896
01198961) = 119860
3
119864 (11989601198962) = 119860
4 119864 (119896
11198962) = 119860
5
(7)
Now expressing 119905 in terms of 119896rsquos we have
119905 = 1198782
119910(1 + 119896
0) (1 + 119896
1)minus1205721
(1 + 1198962)minus1205722
= 1198782
119910(1 + 119896
0) (1 minus 120572
11198961+1205721(1205721+ 1)
21198962
1)
times (1 minus 12057221198962+1205722(1205722+ 1)
21198962
2)
(8)
We assume that |1198961| lt 1 and |119896
2| lt 1 so that (1 + 119896
1)minus1
and (1 + 1198962)minus1 are expandable in terms of 119896rsquos By expanding
the right hand side of (8) multiplying and neglecting termsinvolving power of 119896rsquos greater than two we have
119905 minus 1198782
119910= 1198782
119910(1198960minus 12057211198961minus 12057221198962+ 1205721120572211989611198962minus 120572111989601198961
minus120572211989601198962+1205721(1205721+ 1)
21198962
1+1205722(1205722+ 1)
21198962
2)
(9)
Taking expectations on both sides of (9) we get the biasof 119905 to the first degree of approximation as
119861 (119905) = 1198782
119910(1205722
1
21198601+1205722
2
21198602+ 120572112057221198605minus 12057211198603minus 12057221198604)
(10)
Squaring both sides of (9) and neglecting terms of 119896rsquosinvolving power greater than two we have
(119905 minus 1198782
119910)2
= 1198784
119910(1198962
0+ 21205721120572211989611198962minus 2120572111989601198961
minus 2120572211989601198962+ 1205722
11198962
1+ 1205722
21198962
2)
(11)
Taking expectations on both sides of (11) we get the MSEof 119905 to the first order of approximation as
MSE (119905) = 1198784
119910(1198600+ 1205722
11198601+ 1205722
21198602minus 212057211198603
minus 212057221198604+ 2120572112057221198605)
(12)
The optimal values of 1205721and 120572
2in (12) could be obtained by
differentiating (12) with respect to 1205721and 120572
2and equalizing
to zero After a little algebraic simplification we have
120572lowast
1=11986021198603minus 11986041198605
11986011198602minus 1198602
5
120572lowast
2=11986011198604minus 11986031198605
11986011198602minus 1198602
5
(13)
We can obtain the minimumMSE of 119905 by simply substitutingthe optimal equations of 120572
1and 120572
2in (12)
3 Suggested Estimator in TPS
In certain practical situations when 1198782119909is not also known the
technique of TPS sometimes referred to as double samplingis used This scheme requires the collection of informationon 1199091and 119909
2in the first phase sample 1199041015840 of size 1198991015840 (1198991015840 lt 119873)
and on 119910 for the second phase sample 119904 of size 119899 (119899 lt 1198991015840) The
estimator 119905lowast in TPS will take the following form
119905lowast= 1199042
119910(119904lowast2
1199091
1199042
1199091
)
1205723
(119904lowast2
1199092
1199042
1199092
)
1205724
(14)
To obtain the bias and MSE of 119905lowast we write
1199042
119910= 1198782
119910(1 + 119896
0) 119904
2
1199091= 1198782
1199091(1 + 119896
1)
119904lowast2
1199091= 1198782
1199091(1 + 119896
lowast
1) 119904
2
1199092= 1198782
1199092(1 + 119896
2)
119904lowast2
1199092= 1198782
1199092(1 + 119896
lowast
2)
(15)
Note that
119864 (119896lowast
1) = 119864 (119896
lowast
2) = 0 119864 (119896
lowast2
1) = 119860
lowast
1
119864 (119896lowast2
2) = 119860
lowast
2 119864 (119896
1119896lowast
1) = 119860lowast
1
119864 (1198962119896lowast
2) = 119860lowast
2 119864 (119896
0119896lowast
1) = 119860lowast
3
119864 (1198960119896lowast
2) =
1
1198991015840(120582202
minus 1) = 119860lowast
4
119864 (1198961119896lowast
1) = 119864 (119896
2119896lowast
1) = 119864 (119896
lowast
1119896lowast
2) = 119860lowast
5
(16)
Expressing 119905lowast in terms of 119896rsquos and following the procedure
explained in Section 2 we get the bias and MSE of the esti-mator 119905lowast respectively as
119861 (119905lowast) = 1198782
119910(1205723(1205723+ 1)
21198643+1205724(1205724+ 1)
21198644
+120572312057241198645minus 12057231198641minus 12057241198642)
MSE (119905lowast) = MSE (119905) minus 1198784119910(1205722
3119860lowast
1+ 1205722
4119860lowast
2+ 212057231205724119860lowast
5
+ 21205723119860lowast
3+ 21205724119860lowast
4)
(17)
where1198641= 1198603minus 119860lowast
3 119864
2= 1198604minus 119860lowast
4
1198643= 1198601minus 119860lowast
1 119864
4= 1198602minus 119860lowast
2
1198645= 1198605minus 119860lowast
5
(18)
4 Journal of Probability and Statistics
Minimization of (17) with respect to 1205723and 120572
4 yields their
optimum values as
120572lowast
3=11986411198644minus 11986421198645
11986431198644minus 1198642
5
120572lowast
4=11986421198643minus 11986411198645
11986431198644minus 1198642
5
(19)
Substitution of 120572lowast3and 120572lowast
4in (17) gives the minimum value of
the MSE of 119905lowast
4 Efficiency Comparisons
In this section we considered the theoretical comparisonsof the performances of the suggested estimators (119905 and 119905
lowast)with respect to the traditional estimator (119905
0) Isaki [1] ratio
estimators 1199051 119905lowast1 1199052 and 119905
lowast
2(for single and double auxiliary
variables) Gupta et al [14] estimator (1199053) and Singh et al [8]
estimator (1199054) which are investigated We have the following
conditions(i) MSE (119905) minusMSE (119905
0) lt 0 997904rArr 119867
1lt 0
(ii) MSE (119905) minusMSE (1199051) lt 0 997904rArr 119867
2lt 0
(iii) MSE (119905) minusMSE (1199052) lt 0 997904rArr 119867
3+ 1198674lt 0
(iv) MSE (119905) minusMSE (1199053) lt 0 997904rArr 119867
5lt 1198676
(v) MSE (119905) minusMSE (1199054) lt 0 997904rArr 119867
5lt 1198677
(vi) MSE (119905lowast) minusMSE (119905lowast1) lt 0
997904rArr MSE (119905) minusMSE (1199051) lt 1198784
1199101198678
(vii) MSE (119905lowast) minusMSE (119905lowast2) lt 0
997904rArr MSE (119905) minusMSE (1199052) lt 1198784
1199101198679
(20)
where1198671= 1205722
11198601+ 1205722
21198602+ 2120572112057221198605minus 212057211198603minus 212057221198604
1198672= (1205722
1minus 1)119860
1+ 1205722
21198602minus 2 (120572
1minus 1)119860
3
minus 212057221198604+ 2120572112057221198605
1198673= (1205721minus1198821) [(1205721+1198821) 1198601minus 21198603]
+ (1205722+1198821) [(1205722minus1198821) 1198602minus 21198604] + 2119860
4
1198674= 2 (120572
11205722+1198822
1minus1198821)1198605minus (1 minus 2119882
1) 1198602
1198675= 120572lowast2
11198601+ 120572lowast2
21198602minus 2120572lowast
11198603minus 2120572lowast
21198604+ 2120572lowast
1120572lowast
21198605
1198676= 120601(
1198602
3
1198601
) minus (1198602
4
11989910158401198602
)
1198677= 120575120574lowast2
012minus 120579120588lowast2
1198678= (1205722
3minus 1)119860
lowast
1+ 1205722
4119860lowast
2
minus 2 (1205723minus 1)119860
lowast
3+ 21205724119860lowast
4+ 212057231205724119860lowast
5
1198679= 119867 + 119862
lowast
1+1198722
1119862lowast
2+ 21198721119862lowast
3
(21)
Table 1 The MSE and PRE of the different estimators with respectto 1199050
Estimators MSE PRE1199050
01267 1001199051
01241 103119905lowast
101251 101
1199052(opt) 00586 217119905lowast
2(opt) 00859 1471199053
00543 2331199054
00479 265119905(opt) 00451 281119905lowast
(opt) 00774 164
5 Numerical Illustration
In this section we illustrate the performance of variousestimators of the population variance 1198782
119910 by considering the
data about 119884 output 119883 number of workers and 119885 fixedcapital given in Murthy [16] The data summary is brieflypresented as follows
119873 = 80 119899 = 10 120582400
= 22667
120582040
= 36500 120582004
= 28664
120582220
= 23377 120582202
= 22208
120582022
= 31400
(22)
The MSE and percent relative efficiency (PRE) of variousestimators of 1198782
119910 with respect to the conventional estimator
1199050 have been computed and presented in Table 1 Note that
for the calculation of the MSE of 119905lowast we take 1198991015840 = 25 and alsonote that the minimum MSE of 119905
2and 119905lowast2is obtained using
MSEmin(1199052) = 1198784
119910(1198621minus 1198622
31198622) and MSEmin(119905
lowast
2) = 119878
4
119910(1198631minus
1198632
31198632)
Table 1 reveals that the suggested estimator 119905 has thesmallest MSE and thus the highest PRE among other esti-mators considered in this study The suggested estimator inTPS 119905
lowast also provides a sufficient improvement in varianceestimation compared to the existing ones (119905lowast
1and 119905lowast2) It is also
observed fromTable 1 that the TPS estimators are less efficientthan their corresponding
6 Conclusion
We have developed a new estimator for estimating the finitepopulation variance under SRSWOR which is found to bemore efficient than the traditional estimator Isaki [1] ratioestimators (using single and double auxiliary variables)Gupta et al [14] estimator and Singh et al [8] estimator whencertain conditions as outlined in Section 4 are satisfiedThistheoretical inference is also supported by the result of anapplication with original data In future we hope to extendthe estimators suggested here for the development of a newestimator in the stratified random sampling
Journal of Probability and Statistics 5
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] C T Isaki ldquoVariance estimation using auxiliary informationrdquoJournal of the American Statistical Association vol 78 no 381pp 117ndash123 1983
[2] B Prasad and H P Singh ldquoSome improved ratio-type estima-tors of finite population variance in sample surveysrdquo Commu-nications in Statistics vol 19 no 3 pp 1127ndash1139 1990
[3] B Prasad and H P Singh ldquoUnbiased estimators of finite popu-lation variance using auxiliary information in sample surveysrdquoCommunications in Statistics vol 21 no 5 pp 1367ndash1376 1992
[4] R S Biradar and H P Singh ldquoAn alternative to ratio estimatorof population Variancerdquo Assam Statistical Review vol 8 no 2pp 18ndash33 1994
[5] M Rueda Garcia and A Arcos Cebrian ldquoRepeated substitu-tion method the ratio estimator for the population variancerdquoMetrika vol 43 no 2 pp 101ndash105 1996
[6] A Arcos M Rueda M D Martınez S Gonzalez and YRoman ldquoIncorporating the auxiliary information available invariance estimationrdquo Applied Mathematics and Computationvol 160 no 2 pp 387ndash399 2005
[7] C Kadilar andH Cingi ldquoImprovement in estimating the popu-lation mean in simple random samplingrdquo Applied MathematicsLetters vol 19 no 1 pp 75ndash79 2006
[8] H P Singh S Singh and J M Kim ldquoEfficient use of auxiliaryvariables in estimating finite population variance in two-phasesamplingrdquoCommunications of the Korean Statistical Society vol17 no 2 pp 165ndash181 2010
[9] I Olkin ldquoMultivariate ratio estimation for finite populationsrdquoBiometrika vol 45 pp 154ndash165 1958
[10] S R Srivastava S R Srivastava and B B Khare ldquoChain ratiotype estimator for ratio of two populationmeans using auxiliarycharactersrdquo Communications in Statistics vol 18 no 10 pp3917ndash3926 1989
[11] L N Upadhyaya K S Kushwaha and H P Singh ldquoA modifiedchain ratio-type estimator in two-phase sampling using multiauxiliary informationrdquo Metron vol 48 no 1ndash4 pp 381ndash3931990
[12] V K Singh H P Singh H P Singh and D Shukla ldquoA generalclass of chain estimators for ratio and product of two means ofa finite populationrdquo Communications in Statistics vol 23 no 5pp 1341ndash1355 1994
[13] L Chand Some ratio-type estimators based on two or more aux-iliary variables [PhD thesis] Iowa State University Ames IowaUSA 1975
[14] R K Gupta S Singh and N S Mangat ldquoSome chain ratio typeestimators for estimating finite population variancerdquo AligarhJournal of Statistics vol 12-13 pp 65ndash69 1992-1993
[15] W A Abu-Dayyeh M S Ahmed R A Ahmed and H AMuttlak ldquoSome estimators of a finite population mean usingauxiliary informationrdquo Applied Mathematics and Computationvol 139 no 2-3 pp 287ndash298 2003
[16] M N Murthy Sampling Theory and Methods Statistical Pub-lishing Society Calcutta India 1967
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Mathematical Problems in Engineering
Hindawi Publishing Corporationhttpwwwhindawicom
Differential EquationsInternational Journal of
Volume 2014
Journal ofApplied Mathematics
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
ProbabilityandStatistics
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Advances in
Mathematical Physics
Complex AnalysisJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
OptimizationJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
International Journal of
Combinatorics
OperationsResearch
Advances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Journal of Function Spaces
Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of Mathematics and Mathematical Sciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Algebra
Discrete Dynamics in Nature and Society
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
DecisionSciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Discrete MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Stochastic AnalysisInternational Journal of
2 Journal of Probability and Statistics
119905lowast
1= 1199042
119910
119904lowast2
1199091
1199042
1199091
MSE (119905lowast1) = MSE (119905
1) minus 1198784
119910(119860lowast
1minus 2119860lowast
3)
(1)
Following Olkin [9] Isaki [1] also presented the ratioestimator of variance using two auxiliary variables as follows
1199052= 1198821
1199042
119910
1199042
1199091
1198782
1199091+1198822
1199042
119910
1199042
1199091
1198782
1199092
MSE (1199052) = 1198784
119910(1198621+1198822
11198622minus 211988211198623)
119905lowast
2= 1198721
1199042
119910
1199042
1199091
119904lowast2
1199091+1198722
1199042
119910
1199042
1199092
119904lowast2
1199092(TPS approach of 119905
2)
MSE (119905lowast2) = 1198784
119910(1198631+1198722
11198632minus 211987211198633)
(2)
where119882119894and119872
119894 for 119894 = 1 2 are weights chosen tominimize
the MSE of 1199052and 119905lowast2 Furthersum119882
119894= 1 andsum119872
119894= 1 where
1199042
1199091=1
119899
119899
sum
119894=1
(1199091119894minus 1198831)2
119904lowast2
1199092=1
119899
119899
sum
119894=1
(1199092119894minus 1198832)2
119904lowast2
1199091=
1
1198991015840
1198991015840
sum
119894=1
(1199091119894minus 1199091015840
1)2
119904lowast2
1199092=
1
1198991015840
1198991015840
sum
119894=1
(1199092119894minus 1199091015840
2)2
1199091015840
1=
1
1198991015840
1198991015840
sum
119894=1
1199091119894 119909
1015840
2=
1
1198991015840
1198991015840
sum
119894=1
1199092119894
1198621= 1198600+ 1198602minus 21198604 119862
lowast
1= 119860lowast
2minus 2119860lowast
4
1198631= 1198621minus 119862lowast
1 119862
2= 1198601+ 1198602minus 21198605
119862lowast
2= 119860lowast
1+ 119860lowast
2minus 2119860lowast
5 119863
2= 1198622minus 119862lowast
2
1198623= 1198602+ 1198603minus 1198604minus 1198605 119862
lowast
3= 119860lowast
2+ 119860lowast
3minus 119860lowast
4minus 119860lowast
5
1198633= 1198623minus 119862lowast
3
1198600=1
119899(120582400
minus 1) 1198601=1
119899(120582040
minus 1)
1198602=1
119899(120582004
minus 1) 1198603=1
119899(120582220
minus 1)
1198604=1
119899(120582202
minus 1) 1198605=1
119899(120582022
minus 1)
119860lowast
1=
1
1198991015840(120582040
minus 1) 119860lowast
2=
1
1198991015840(120582004
minus 1)
119860lowast
3=
1
1198991015840(120582220
minus 1) 119860lowast
4=
1
1198991015840(120582202
minus 1)
119860lowast
5=
1
1198991015840(120582022
minus 1)
120582 =120583119886119887119888
1205831198862
2001205831198872
0201205831198882
002
120583119886119887119888
=1
119873 minus 1
119873
sum
119894=1
(119910119894minus 119884)119886
(1199091119894minus 1198831)119887
(1199092119894minus 1198832)119888
(3)
where 119886 119887 and 119888 are nonnegative integersSeveral authors (Srivastava et al [10] Upadhyaya et al
[11] and Singh et al [12]) adopted TPS procedure proposedby Chand [13] and have suggested some chain ratio-typeestimators for estimating populationmean119884of119910 In the samevein Gupta et al [14] and Singh et al [8] proposed the fol-lowing classes of estimators under the assumption that thepopulation variance of the first auxiliary variable 1198782
1199091is not
known but the population variance of another auxiliaryvariable119883
2closely related to119883
1is available TheMSEs of the
estimators suggested by Gupta et al [14] and Singh et al [8]are respectively given by
1199053= 1199042
119910(1199042
1199091
119904lowast2
1199091
)
1198681
(119904lowast2
1199092
1198782
1199092
)
1198682
MSEmin (1199053) = 1199044
119910[1198600minus 120601(
1198602
3
1198601
) minus (1198602
4
11989910158401198602
)]
1199054= 1199042
119910(1199042
1199091
119904lowast2
1199091
)
1198691
(119904lowast2
1199092
1198782
1199092
)
1198692
(1199042
1199092
1198782
1199092
)
1198693
MSEmin (1199054) = 1199044
1199101198600[1 minus 120575120574
lowast2
012minus 120579120588lowast2]
(4)
where 1198681 1198682 and 119869
119894for 119894 = 1 2 3 are constants chosen to
minimize theMSE of 1199053and 1199054 120601 = (1119899minus1119899
1015840) 120579 = 119899119899
1015840 120575 =((1198991015840minus119899)119899
1015840) 120574lowast2012
= (11986021198602
3minus2119860311986041198605+11986011198602
4)1198600(11986011198602minus
1198602
5) 120588lowast2 = 119860
2
411986001198602
In most studies several variables are considered simul-taneously either to explain or estimate (predict) the studyvariable In most cases information on several auxiliaryvariables closely related to the study variable may be easilyobtained on all units in the population For example whileconducting an educational survey the investigator may beinterested in studying characteristics such as age genderhours spent on studying per day sitting position parentrsquos edu-cational level parentrsquos income relationship with lectures andaccess to facilities (eg library internet laboratory) amongothers With the main aim of suggesting a more efficientestimator we propose in this paper under SRSWOR a chainratio-type estimator for estimating the population variancewhen information on two auxiliary variables is available Inaddition the problem is extended to the case of TPS
2 The Suggested Estimator
Following Abu-Dayyeh et al [15] we define an estimator forestimating the population variance 1198782
119910 as follows
119905 = 1199042
119910(1198782
1199091
1199042
1199091
)
1205721
(1198782
1199092
1199042
1199092
)
1205722
(5)
where 1205721and 1205722are real constants to be determined such that
the MSE of 119905 is minimum
Journal of Probability and Statistics 3
To determine the bias and MSE of 119905 we define
1199042
119910= 1198782
119910(1 + 119896
0) 119904
2
1199091= 1198782
1199091(1 + 119896
1) 119904
2
1199092= 1198782
1199092(1 + 119896
2)
(6)
such that
119864 (1198960) = 119864 (119896
1) = 119864 (119896
2) = 0
119864 (1198962
0) = 119860
0 119864 (119896
2
1) = 119860
1
119864 (1198962
2) = 119860
2 119864 (119896
01198961) = 119860
3
119864 (11989601198962) = 119860
4 119864 (119896
11198962) = 119860
5
(7)
Now expressing 119905 in terms of 119896rsquos we have
119905 = 1198782
119910(1 + 119896
0) (1 + 119896
1)minus1205721
(1 + 1198962)minus1205722
= 1198782
119910(1 + 119896
0) (1 minus 120572
11198961+1205721(1205721+ 1)
21198962
1)
times (1 minus 12057221198962+1205722(1205722+ 1)
21198962
2)
(8)
We assume that |1198961| lt 1 and |119896
2| lt 1 so that (1 + 119896
1)minus1
and (1 + 1198962)minus1 are expandable in terms of 119896rsquos By expanding
the right hand side of (8) multiplying and neglecting termsinvolving power of 119896rsquos greater than two we have
119905 minus 1198782
119910= 1198782
119910(1198960minus 12057211198961minus 12057221198962+ 1205721120572211989611198962minus 120572111989601198961
minus120572211989601198962+1205721(1205721+ 1)
21198962
1+1205722(1205722+ 1)
21198962
2)
(9)
Taking expectations on both sides of (9) we get the biasof 119905 to the first degree of approximation as
119861 (119905) = 1198782
119910(1205722
1
21198601+1205722
2
21198602+ 120572112057221198605minus 12057211198603minus 12057221198604)
(10)
Squaring both sides of (9) and neglecting terms of 119896rsquosinvolving power greater than two we have
(119905 minus 1198782
119910)2
= 1198784
119910(1198962
0+ 21205721120572211989611198962minus 2120572111989601198961
minus 2120572211989601198962+ 1205722
11198962
1+ 1205722
21198962
2)
(11)
Taking expectations on both sides of (11) we get the MSEof 119905 to the first order of approximation as
MSE (119905) = 1198784
119910(1198600+ 1205722
11198601+ 1205722
21198602minus 212057211198603
minus 212057221198604+ 2120572112057221198605)
(12)
The optimal values of 1205721and 120572
2in (12) could be obtained by
differentiating (12) with respect to 1205721and 120572
2and equalizing
to zero After a little algebraic simplification we have
120572lowast
1=11986021198603minus 11986041198605
11986011198602minus 1198602
5
120572lowast
2=11986011198604minus 11986031198605
11986011198602minus 1198602
5
(13)
We can obtain the minimumMSE of 119905 by simply substitutingthe optimal equations of 120572
1and 120572
2in (12)
3 Suggested Estimator in TPS
In certain practical situations when 1198782119909is not also known the
technique of TPS sometimes referred to as double samplingis used This scheme requires the collection of informationon 1199091and 119909
2in the first phase sample 1199041015840 of size 1198991015840 (1198991015840 lt 119873)
and on 119910 for the second phase sample 119904 of size 119899 (119899 lt 1198991015840) The
estimator 119905lowast in TPS will take the following form
119905lowast= 1199042
119910(119904lowast2
1199091
1199042
1199091
)
1205723
(119904lowast2
1199092
1199042
1199092
)
1205724
(14)
To obtain the bias and MSE of 119905lowast we write
1199042
119910= 1198782
119910(1 + 119896
0) 119904
2
1199091= 1198782
1199091(1 + 119896
1)
119904lowast2
1199091= 1198782
1199091(1 + 119896
lowast
1) 119904
2
1199092= 1198782
1199092(1 + 119896
2)
119904lowast2
1199092= 1198782
1199092(1 + 119896
lowast
2)
(15)
Note that
119864 (119896lowast
1) = 119864 (119896
lowast
2) = 0 119864 (119896
lowast2
1) = 119860
lowast
1
119864 (119896lowast2
2) = 119860
lowast
2 119864 (119896
1119896lowast
1) = 119860lowast
1
119864 (1198962119896lowast
2) = 119860lowast
2 119864 (119896
0119896lowast
1) = 119860lowast
3
119864 (1198960119896lowast
2) =
1
1198991015840(120582202
minus 1) = 119860lowast
4
119864 (1198961119896lowast
1) = 119864 (119896
2119896lowast
1) = 119864 (119896
lowast
1119896lowast
2) = 119860lowast
5
(16)
Expressing 119905lowast in terms of 119896rsquos and following the procedure
explained in Section 2 we get the bias and MSE of the esti-mator 119905lowast respectively as
119861 (119905lowast) = 1198782
119910(1205723(1205723+ 1)
21198643+1205724(1205724+ 1)
21198644
+120572312057241198645minus 12057231198641minus 12057241198642)
MSE (119905lowast) = MSE (119905) minus 1198784119910(1205722
3119860lowast
1+ 1205722
4119860lowast
2+ 212057231205724119860lowast
5
+ 21205723119860lowast
3+ 21205724119860lowast
4)
(17)
where1198641= 1198603minus 119860lowast
3 119864
2= 1198604minus 119860lowast
4
1198643= 1198601minus 119860lowast
1 119864
4= 1198602minus 119860lowast
2
1198645= 1198605minus 119860lowast
5
(18)
4 Journal of Probability and Statistics
Minimization of (17) with respect to 1205723and 120572
4 yields their
optimum values as
120572lowast
3=11986411198644minus 11986421198645
11986431198644minus 1198642
5
120572lowast
4=11986421198643minus 11986411198645
11986431198644minus 1198642
5
(19)
Substitution of 120572lowast3and 120572lowast
4in (17) gives the minimum value of
the MSE of 119905lowast
4 Efficiency Comparisons
In this section we considered the theoretical comparisonsof the performances of the suggested estimators (119905 and 119905
lowast)with respect to the traditional estimator (119905
0) Isaki [1] ratio
estimators 1199051 119905lowast1 1199052 and 119905
lowast
2(for single and double auxiliary
variables) Gupta et al [14] estimator (1199053) and Singh et al [8]
estimator (1199054) which are investigated We have the following
conditions(i) MSE (119905) minusMSE (119905
0) lt 0 997904rArr 119867
1lt 0
(ii) MSE (119905) minusMSE (1199051) lt 0 997904rArr 119867
2lt 0
(iii) MSE (119905) minusMSE (1199052) lt 0 997904rArr 119867
3+ 1198674lt 0
(iv) MSE (119905) minusMSE (1199053) lt 0 997904rArr 119867
5lt 1198676
(v) MSE (119905) minusMSE (1199054) lt 0 997904rArr 119867
5lt 1198677
(vi) MSE (119905lowast) minusMSE (119905lowast1) lt 0
997904rArr MSE (119905) minusMSE (1199051) lt 1198784
1199101198678
(vii) MSE (119905lowast) minusMSE (119905lowast2) lt 0
997904rArr MSE (119905) minusMSE (1199052) lt 1198784
1199101198679
(20)
where1198671= 1205722
11198601+ 1205722
21198602+ 2120572112057221198605minus 212057211198603minus 212057221198604
1198672= (1205722
1minus 1)119860
1+ 1205722
21198602minus 2 (120572
1minus 1)119860
3
minus 212057221198604+ 2120572112057221198605
1198673= (1205721minus1198821) [(1205721+1198821) 1198601minus 21198603]
+ (1205722+1198821) [(1205722minus1198821) 1198602minus 21198604] + 2119860
4
1198674= 2 (120572
11205722+1198822
1minus1198821)1198605minus (1 minus 2119882
1) 1198602
1198675= 120572lowast2
11198601+ 120572lowast2
21198602minus 2120572lowast
11198603minus 2120572lowast
21198604+ 2120572lowast
1120572lowast
21198605
1198676= 120601(
1198602
3
1198601
) minus (1198602
4
11989910158401198602
)
1198677= 120575120574lowast2
012minus 120579120588lowast2
1198678= (1205722
3minus 1)119860
lowast
1+ 1205722
4119860lowast
2
minus 2 (1205723minus 1)119860
lowast
3+ 21205724119860lowast
4+ 212057231205724119860lowast
5
1198679= 119867 + 119862
lowast
1+1198722
1119862lowast
2+ 21198721119862lowast
3
(21)
Table 1 The MSE and PRE of the different estimators with respectto 1199050
Estimators MSE PRE1199050
01267 1001199051
01241 103119905lowast
101251 101
1199052(opt) 00586 217119905lowast
2(opt) 00859 1471199053
00543 2331199054
00479 265119905(opt) 00451 281119905lowast
(opt) 00774 164
5 Numerical Illustration
In this section we illustrate the performance of variousestimators of the population variance 1198782
119910 by considering the
data about 119884 output 119883 number of workers and 119885 fixedcapital given in Murthy [16] The data summary is brieflypresented as follows
119873 = 80 119899 = 10 120582400
= 22667
120582040
= 36500 120582004
= 28664
120582220
= 23377 120582202
= 22208
120582022
= 31400
(22)
The MSE and percent relative efficiency (PRE) of variousestimators of 1198782
119910 with respect to the conventional estimator
1199050 have been computed and presented in Table 1 Note that
for the calculation of the MSE of 119905lowast we take 1198991015840 = 25 and alsonote that the minimum MSE of 119905
2and 119905lowast2is obtained using
MSEmin(1199052) = 1198784
119910(1198621minus 1198622
31198622) and MSEmin(119905
lowast
2) = 119878
4
119910(1198631minus
1198632
31198632)
Table 1 reveals that the suggested estimator 119905 has thesmallest MSE and thus the highest PRE among other esti-mators considered in this study The suggested estimator inTPS 119905
lowast also provides a sufficient improvement in varianceestimation compared to the existing ones (119905lowast
1and 119905lowast2) It is also
observed fromTable 1 that the TPS estimators are less efficientthan their corresponding
6 Conclusion
We have developed a new estimator for estimating the finitepopulation variance under SRSWOR which is found to bemore efficient than the traditional estimator Isaki [1] ratioestimators (using single and double auxiliary variables)Gupta et al [14] estimator and Singh et al [8] estimator whencertain conditions as outlined in Section 4 are satisfiedThistheoretical inference is also supported by the result of anapplication with original data In future we hope to extendthe estimators suggested here for the development of a newestimator in the stratified random sampling
Journal of Probability and Statistics 5
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] C T Isaki ldquoVariance estimation using auxiliary informationrdquoJournal of the American Statistical Association vol 78 no 381pp 117ndash123 1983
[2] B Prasad and H P Singh ldquoSome improved ratio-type estima-tors of finite population variance in sample surveysrdquo Commu-nications in Statistics vol 19 no 3 pp 1127ndash1139 1990
[3] B Prasad and H P Singh ldquoUnbiased estimators of finite popu-lation variance using auxiliary information in sample surveysrdquoCommunications in Statistics vol 21 no 5 pp 1367ndash1376 1992
[4] R S Biradar and H P Singh ldquoAn alternative to ratio estimatorof population Variancerdquo Assam Statistical Review vol 8 no 2pp 18ndash33 1994
[5] M Rueda Garcia and A Arcos Cebrian ldquoRepeated substitu-tion method the ratio estimator for the population variancerdquoMetrika vol 43 no 2 pp 101ndash105 1996
[6] A Arcos M Rueda M D Martınez S Gonzalez and YRoman ldquoIncorporating the auxiliary information available invariance estimationrdquo Applied Mathematics and Computationvol 160 no 2 pp 387ndash399 2005
[7] C Kadilar andH Cingi ldquoImprovement in estimating the popu-lation mean in simple random samplingrdquo Applied MathematicsLetters vol 19 no 1 pp 75ndash79 2006
[8] H P Singh S Singh and J M Kim ldquoEfficient use of auxiliaryvariables in estimating finite population variance in two-phasesamplingrdquoCommunications of the Korean Statistical Society vol17 no 2 pp 165ndash181 2010
[9] I Olkin ldquoMultivariate ratio estimation for finite populationsrdquoBiometrika vol 45 pp 154ndash165 1958
[10] S R Srivastava S R Srivastava and B B Khare ldquoChain ratiotype estimator for ratio of two populationmeans using auxiliarycharactersrdquo Communications in Statistics vol 18 no 10 pp3917ndash3926 1989
[11] L N Upadhyaya K S Kushwaha and H P Singh ldquoA modifiedchain ratio-type estimator in two-phase sampling using multiauxiliary informationrdquo Metron vol 48 no 1ndash4 pp 381ndash3931990
[12] V K Singh H P Singh H P Singh and D Shukla ldquoA generalclass of chain estimators for ratio and product of two means ofa finite populationrdquo Communications in Statistics vol 23 no 5pp 1341ndash1355 1994
[13] L Chand Some ratio-type estimators based on two or more aux-iliary variables [PhD thesis] Iowa State University Ames IowaUSA 1975
[14] R K Gupta S Singh and N S Mangat ldquoSome chain ratio typeestimators for estimating finite population variancerdquo AligarhJournal of Statistics vol 12-13 pp 65ndash69 1992-1993
[15] W A Abu-Dayyeh M S Ahmed R A Ahmed and H AMuttlak ldquoSome estimators of a finite population mean usingauxiliary informationrdquo Applied Mathematics and Computationvol 139 no 2-3 pp 287ndash298 2003
[16] M N Murthy Sampling Theory and Methods Statistical Pub-lishing Society Calcutta India 1967
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Mathematical Problems in Engineering
Hindawi Publishing Corporationhttpwwwhindawicom
Differential EquationsInternational Journal of
Volume 2014
Journal ofApplied Mathematics
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
ProbabilityandStatistics
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Advances in
Mathematical Physics
Complex AnalysisJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
OptimizationJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
International Journal of
Combinatorics
OperationsResearch
Advances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Journal of Function Spaces
Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of Mathematics and Mathematical Sciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Algebra
Discrete Dynamics in Nature and Society
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
DecisionSciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Discrete MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Stochastic AnalysisInternational Journal of
Journal of Probability and Statistics 3
To determine the bias and MSE of 119905 we define
1199042
119910= 1198782
119910(1 + 119896
0) 119904
2
1199091= 1198782
1199091(1 + 119896
1) 119904
2
1199092= 1198782
1199092(1 + 119896
2)
(6)
such that
119864 (1198960) = 119864 (119896
1) = 119864 (119896
2) = 0
119864 (1198962
0) = 119860
0 119864 (119896
2
1) = 119860
1
119864 (1198962
2) = 119860
2 119864 (119896
01198961) = 119860
3
119864 (11989601198962) = 119860
4 119864 (119896
11198962) = 119860
5
(7)
Now expressing 119905 in terms of 119896rsquos we have
119905 = 1198782
119910(1 + 119896
0) (1 + 119896
1)minus1205721
(1 + 1198962)minus1205722
= 1198782
119910(1 + 119896
0) (1 minus 120572
11198961+1205721(1205721+ 1)
21198962
1)
times (1 minus 12057221198962+1205722(1205722+ 1)
21198962
2)
(8)
We assume that |1198961| lt 1 and |119896
2| lt 1 so that (1 + 119896
1)minus1
and (1 + 1198962)minus1 are expandable in terms of 119896rsquos By expanding
the right hand side of (8) multiplying and neglecting termsinvolving power of 119896rsquos greater than two we have
119905 minus 1198782
119910= 1198782
119910(1198960minus 12057211198961minus 12057221198962+ 1205721120572211989611198962minus 120572111989601198961
minus120572211989601198962+1205721(1205721+ 1)
21198962
1+1205722(1205722+ 1)
21198962
2)
(9)
Taking expectations on both sides of (9) we get the biasof 119905 to the first degree of approximation as
119861 (119905) = 1198782
119910(1205722
1
21198601+1205722
2
21198602+ 120572112057221198605minus 12057211198603minus 12057221198604)
(10)
Squaring both sides of (9) and neglecting terms of 119896rsquosinvolving power greater than two we have
(119905 minus 1198782
119910)2
= 1198784
119910(1198962
0+ 21205721120572211989611198962minus 2120572111989601198961
minus 2120572211989601198962+ 1205722
11198962
1+ 1205722
21198962
2)
(11)
Taking expectations on both sides of (11) we get the MSEof 119905 to the first order of approximation as
MSE (119905) = 1198784
119910(1198600+ 1205722
11198601+ 1205722
21198602minus 212057211198603
minus 212057221198604+ 2120572112057221198605)
(12)
The optimal values of 1205721and 120572
2in (12) could be obtained by
differentiating (12) with respect to 1205721and 120572
2and equalizing
to zero After a little algebraic simplification we have
120572lowast
1=11986021198603minus 11986041198605
11986011198602minus 1198602
5
120572lowast
2=11986011198604minus 11986031198605
11986011198602minus 1198602
5
(13)
We can obtain the minimumMSE of 119905 by simply substitutingthe optimal equations of 120572
1and 120572
2in (12)
3 Suggested Estimator in TPS
In certain practical situations when 1198782119909is not also known the
technique of TPS sometimes referred to as double samplingis used This scheme requires the collection of informationon 1199091and 119909
2in the first phase sample 1199041015840 of size 1198991015840 (1198991015840 lt 119873)
and on 119910 for the second phase sample 119904 of size 119899 (119899 lt 1198991015840) The
estimator 119905lowast in TPS will take the following form
119905lowast= 1199042
119910(119904lowast2
1199091
1199042
1199091
)
1205723
(119904lowast2
1199092
1199042
1199092
)
1205724
(14)
To obtain the bias and MSE of 119905lowast we write
1199042
119910= 1198782
119910(1 + 119896
0) 119904
2
1199091= 1198782
1199091(1 + 119896
1)
119904lowast2
1199091= 1198782
1199091(1 + 119896
lowast
1) 119904
2
1199092= 1198782
1199092(1 + 119896
2)
119904lowast2
1199092= 1198782
1199092(1 + 119896
lowast
2)
(15)
Note that
119864 (119896lowast
1) = 119864 (119896
lowast
2) = 0 119864 (119896
lowast2
1) = 119860
lowast
1
119864 (119896lowast2
2) = 119860
lowast
2 119864 (119896
1119896lowast
1) = 119860lowast
1
119864 (1198962119896lowast
2) = 119860lowast
2 119864 (119896
0119896lowast
1) = 119860lowast
3
119864 (1198960119896lowast
2) =
1
1198991015840(120582202
minus 1) = 119860lowast
4
119864 (1198961119896lowast
1) = 119864 (119896
2119896lowast
1) = 119864 (119896
lowast
1119896lowast
2) = 119860lowast
5
(16)
Expressing 119905lowast in terms of 119896rsquos and following the procedure
explained in Section 2 we get the bias and MSE of the esti-mator 119905lowast respectively as
119861 (119905lowast) = 1198782
119910(1205723(1205723+ 1)
21198643+1205724(1205724+ 1)
21198644
+120572312057241198645minus 12057231198641minus 12057241198642)
MSE (119905lowast) = MSE (119905) minus 1198784119910(1205722
3119860lowast
1+ 1205722
4119860lowast
2+ 212057231205724119860lowast
5
+ 21205723119860lowast
3+ 21205724119860lowast
4)
(17)
where1198641= 1198603minus 119860lowast
3 119864
2= 1198604minus 119860lowast
4
1198643= 1198601minus 119860lowast
1 119864
4= 1198602minus 119860lowast
2
1198645= 1198605minus 119860lowast
5
(18)
4 Journal of Probability and Statistics
Minimization of (17) with respect to 1205723and 120572
4 yields their
optimum values as
120572lowast
3=11986411198644minus 11986421198645
11986431198644minus 1198642
5
120572lowast
4=11986421198643minus 11986411198645
11986431198644minus 1198642
5
(19)
Substitution of 120572lowast3and 120572lowast
4in (17) gives the minimum value of
the MSE of 119905lowast
4 Efficiency Comparisons
In this section we considered the theoretical comparisonsof the performances of the suggested estimators (119905 and 119905
lowast)with respect to the traditional estimator (119905
0) Isaki [1] ratio
estimators 1199051 119905lowast1 1199052 and 119905
lowast
2(for single and double auxiliary
variables) Gupta et al [14] estimator (1199053) and Singh et al [8]
estimator (1199054) which are investigated We have the following
conditions(i) MSE (119905) minusMSE (119905
0) lt 0 997904rArr 119867
1lt 0
(ii) MSE (119905) minusMSE (1199051) lt 0 997904rArr 119867
2lt 0
(iii) MSE (119905) minusMSE (1199052) lt 0 997904rArr 119867
3+ 1198674lt 0
(iv) MSE (119905) minusMSE (1199053) lt 0 997904rArr 119867
5lt 1198676
(v) MSE (119905) minusMSE (1199054) lt 0 997904rArr 119867
5lt 1198677
(vi) MSE (119905lowast) minusMSE (119905lowast1) lt 0
997904rArr MSE (119905) minusMSE (1199051) lt 1198784
1199101198678
(vii) MSE (119905lowast) minusMSE (119905lowast2) lt 0
997904rArr MSE (119905) minusMSE (1199052) lt 1198784
1199101198679
(20)
where1198671= 1205722
11198601+ 1205722
21198602+ 2120572112057221198605minus 212057211198603minus 212057221198604
1198672= (1205722
1minus 1)119860
1+ 1205722
21198602minus 2 (120572
1minus 1)119860
3
minus 212057221198604+ 2120572112057221198605
1198673= (1205721minus1198821) [(1205721+1198821) 1198601minus 21198603]
+ (1205722+1198821) [(1205722minus1198821) 1198602minus 21198604] + 2119860
4
1198674= 2 (120572
11205722+1198822
1minus1198821)1198605minus (1 minus 2119882
1) 1198602
1198675= 120572lowast2
11198601+ 120572lowast2
21198602minus 2120572lowast
11198603minus 2120572lowast
21198604+ 2120572lowast
1120572lowast
21198605
1198676= 120601(
1198602
3
1198601
) minus (1198602
4
11989910158401198602
)
1198677= 120575120574lowast2
012minus 120579120588lowast2
1198678= (1205722
3minus 1)119860
lowast
1+ 1205722
4119860lowast
2
minus 2 (1205723minus 1)119860
lowast
3+ 21205724119860lowast
4+ 212057231205724119860lowast
5
1198679= 119867 + 119862
lowast
1+1198722
1119862lowast
2+ 21198721119862lowast
3
(21)
Table 1 The MSE and PRE of the different estimators with respectto 1199050
Estimators MSE PRE1199050
01267 1001199051
01241 103119905lowast
101251 101
1199052(opt) 00586 217119905lowast
2(opt) 00859 1471199053
00543 2331199054
00479 265119905(opt) 00451 281119905lowast
(opt) 00774 164
5 Numerical Illustration
In this section we illustrate the performance of variousestimators of the population variance 1198782
119910 by considering the
data about 119884 output 119883 number of workers and 119885 fixedcapital given in Murthy [16] The data summary is brieflypresented as follows
119873 = 80 119899 = 10 120582400
= 22667
120582040
= 36500 120582004
= 28664
120582220
= 23377 120582202
= 22208
120582022
= 31400
(22)
The MSE and percent relative efficiency (PRE) of variousestimators of 1198782
119910 with respect to the conventional estimator
1199050 have been computed and presented in Table 1 Note that
for the calculation of the MSE of 119905lowast we take 1198991015840 = 25 and alsonote that the minimum MSE of 119905
2and 119905lowast2is obtained using
MSEmin(1199052) = 1198784
119910(1198621minus 1198622
31198622) and MSEmin(119905
lowast
2) = 119878
4
119910(1198631minus
1198632
31198632)
Table 1 reveals that the suggested estimator 119905 has thesmallest MSE and thus the highest PRE among other esti-mators considered in this study The suggested estimator inTPS 119905
lowast also provides a sufficient improvement in varianceestimation compared to the existing ones (119905lowast
1and 119905lowast2) It is also
observed fromTable 1 that the TPS estimators are less efficientthan their corresponding
6 Conclusion
We have developed a new estimator for estimating the finitepopulation variance under SRSWOR which is found to bemore efficient than the traditional estimator Isaki [1] ratioestimators (using single and double auxiliary variables)Gupta et al [14] estimator and Singh et al [8] estimator whencertain conditions as outlined in Section 4 are satisfiedThistheoretical inference is also supported by the result of anapplication with original data In future we hope to extendthe estimators suggested here for the development of a newestimator in the stratified random sampling
Journal of Probability and Statistics 5
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] C T Isaki ldquoVariance estimation using auxiliary informationrdquoJournal of the American Statistical Association vol 78 no 381pp 117ndash123 1983
[2] B Prasad and H P Singh ldquoSome improved ratio-type estima-tors of finite population variance in sample surveysrdquo Commu-nications in Statistics vol 19 no 3 pp 1127ndash1139 1990
[3] B Prasad and H P Singh ldquoUnbiased estimators of finite popu-lation variance using auxiliary information in sample surveysrdquoCommunications in Statistics vol 21 no 5 pp 1367ndash1376 1992
[4] R S Biradar and H P Singh ldquoAn alternative to ratio estimatorof population Variancerdquo Assam Statistical Review vol 8 no 2pp 18ndash33 1994
[5] M Rueda Garcia and A Arcos Cebrian ldquoRepeated substitu-tion method the ratio estimator for the population variancerdquoMetrika vol 43 no 2 pp 101ndash105 1996
[6] A Arcos M Rueda M D Martınez S Gonzalez and YRoman ldquoIncorporating the auxiliary information available invariance estimationrdquo Applied Mathematics and Computationvol 160 no 2 pp 387ndash399 2005
[7] C Kadilar andH Cingi ldquoImprovement in estimating the popu-lation mean in simple random samplingrdquo Applied MathematicsLetters vol 19 no 1 pp 75ndash79 2006
[8] H P Singh S Singh and J M Kim ldquoEfficient use of auxiliaryvariables in estimating finite population variance in two-phasesamplingrdquoCommunications of the Korean Statistical Society vol17 no 2 pp 165ndash181 2010
[9] I Olkin ldquoMultivariate ratio estimation for finite populationsrdquoBiometrika vol 45 pp 154ndash165 1958
[10] S R Srivastava S R Srivastava and B B Khare ldquoChain ratiotype estimator for ratio of two populationmeans using auxiliarycharactersrdquo Communications in Statistics vol 18 no 10 pp3917ndash3926 1989
[11] L N Upadhyaya K S Kushwaha and H P Singh ldquoA modifiedchain ratio-type estimator in two-phase sampling using multiauxiliary informationrdquo Metron vol 48 no 1ndash4 pp 381ndash3931990
[12] V K Singh H P Singh H P Singh and D Shukla ldquoA generalclass of chain estimators for ratio and product of two means ofa finite populationrdquo Communications in Statistics vol 23 no 5pp 1341ndash1355 1994
[13] L Chand Some ratio-type estimators based on two or more aux-iliary variables [PhD thesis] Iowa State University Ames IowaUSA 1975
[14] R K Gupta S Singh and N S Mangat ldquoSome chain ratio typeestimators for estimating finite population variancerdquo AligarhJournal of Statistics vol 12-13 pp 65ndash69 1992-1993
[15] W A Abu-Dayyeh M S Ahmed R A Ahmed and H AMuttlak ldquoSome estimators of a finite population mean usingauxiliary informationrdquo Applied Mathematics and Computationvol 139 no 2-3 pp 287ndash298 2003
[16] M N Murthy Sampling Theory and Methods Statistical Pub-lishing Society Calcutta India 1967
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Mathematical Problems in Engineering
Hindawi Publishing Corporationhttpwwwhindawicom
Differential EquationsInternational Journal of
Volume 2014
Journal ofApplied Mathematics
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
ProbabilityandStatistics
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Advances in
Mathematical Physics
Complex AnalysisJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
OptimizationJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
International Journal of
Combinatorics
OperationsResearch
Advances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Journal of Function Spaces
Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of Mathematics and Mathematical Sciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Algebra
Discrete Dynamics in Nature and Society
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
DecisionSciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Discrete MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Stochastic AnalysisInternational Journal of
4 Journal of Probability and Statistics
Minimization of (17) with respect to 1205723and 120572
4 yields their
optimum values as
120572lowast
3=11986411198644minus 11986421198645
11986431198644minus 1198642
5
120572lowast
4=11986421198643minus 11986411198645
11986431198644minus 1198642
5
(19)
Substitution of 120572lowast3and 120572lowast
4in (17) gives the minimum value of
the MSE of 119905lowast
4 Efficiency Comparisons
In this section we considered the theoretical comparisonsof the performances of the suggested estimators (119905 and 119905
lowast)with respect to the traditional estimator (119905
0) Isaki [1] ratio
estimators 1199051 119905lowast1 1199052 and 119905
lowast
2(for single and double auxiliary
variables) Gupta et al [14] estimator (1199053) and Singh et al [8]
estimator (1199054) which are investigated We have the following
conditions(i) MSE (119905) minusMSE (119905
0) lt 0 997904rArr 119867
1lt 0
(ii) MSE (119905) minusMSE (1199051) lt 0 997904rArr 119867
2lt 0
(iii) MSE (119905) minusMSE (1199052) lt 0 997904rArr 119867
3+ 1198674lt 0
(iv) MSE (119905) minusMSE (1199053) lt 0 997904rArr 119867
5lt 1198676
(v) MSE (119905) minusMSE (1199054) lt 0 997904rArr 119867
5lt 1198677
(vi) MSE (119905lowast) minusMSE (119905lowast1) lt 0
997904rArr MSE (119905) minusMSE (1199051) lt 1198784
1199101198678
(vii) MSE (119905lowast) minusMSE (119905lowast2) lt 0
997904rArr MSE (119905) minusMSE (1199052) lt 1198784
1199101198679
(20)
where1198671= 1205722
11198601+ 1205722
21198602+ 2120572112057221198605minus 212057211198603minus 212057221198604
1198672= (1205722
1minus 1)119860
1+ 1205722
21198602minus 2 (120572
1minus 1)119860
3
minus 212057221198604+ 2120572112057221198605
1198673= (1205721minus1198821) [(1205721+1198821) 1198601minus 21198603]
+ (1205722+1198821) [(1205722minus1198821) 1198602minus 21198604] + 2119860
4
1198674= 2 (120572
11205722+1198822
1minus1198821)1198605minus (1 minus 2119882
1) 1198602
1198675= 120572lowast2
11198601+ 120572lowast2
21198602minus 2120572lowast
11198603minus 2120572lowast
21198604+ 2120572lowast
1120572lowast
21198605
1198676= 120601(
1198602
3
1198601
) minus (1198602
4
11989910158401198602
)
1198677= 120575120574lowast2
012minus 120579120588lowast2
1198678= (1205722
3minus 1)119860
lowast
1+ 1205722
4119860lowast
2
minus 2 (1205723minus 1)119860
lowast
3+ 21205724119860lowast
4+ 212057231205724119860lowast
5
1198679= 119867 + 119862
lowast
1+1198722
1119862lowast
2+ 21198721119862lowast
3
(21)
Table 1 The MSE and PRE of the different estimators with respectto 1199050
Estimators MSE PRE1199050
01267 1001199051
01241 103119905lowast
101251 101
1199052(opt) 00586 217119905lowast
2(opt) 00859 1471199053
00543 2331199054
00479 265119905(opt) 00451 281119905lowast
(opt) 00774 164
5 Numerical Illustration
In this section we illustrate the performance of variousestimators of the population variance 1198782
119910 by considering the
data about 119884 output 119883 number of workers and 119885 fixedcapital given in Murthy [16] The data summary is brieflypresented as follows
119873 = 80 119899 = 10 120582400
= 22667
120582040
= 36500 120582004
= 28664
120582220
= 23377 120582202
= 22208
120582022
= 31400
(22)
The MSE and percent relative efficiency (PRE) of variousestimators of 1198782
119910 with respect to the conventional estimator
1199050 have been computed and presented in Table 1 Note that
for the calculation of the MSE of 119905lowast we take 1198991015840 = 25 and alsonote that the minimum MSE of 119905
2and 119905lowast2is obtained using
MSEmin(1199052) = 1198784
119910(1198621minus 1198622
31198622) and MSEmin(119905
lowast
2) = 119878
4
119910(1198631minus
1198632
31198632)
Table 1 reveals that the suggested estimator 119905 has thesmallest MSE and thus the highest PRE among other esti-mators considered in this study The suggested estimator inTPS 119905
lowast also provides a sufficient improvement in varianceestimation compared to the existing ones (119905lowast
1and 119905lowast2) It is also
observed fromTable 1 that the TPS estimators are less efficientthan their corresponding
6 Conclusion
We have developed a new estimator for estimating the finitepopulation variance under SRSWOR which is found to bemore efficient than the traditional estimator Isaki [1] ratioestimators (using single and double auxiliary variables)Gupta et al [14] estimator and Singh et al [8] estimator whencertain conditions as outlined in Section 4 are satisfiedThistheoretical inference is also supported by the result of anapplication with original data In future we hope to extendthe estimators suggested here for the development of a newestimator in the stratified random sampling
Journal of Probability and Statistics 5
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] C T Isaki ldquoVariance estimation using auxiliary informationrdquoJournal of the American Statistical Association vol 78 no 381pp 117ndash123 1983
[2] B Prasad and H P Singh ldquoSome improved ratio-type estima-tors of finite population variance in sample surveysrdquo Commu-nications in Statistics vol 19 no 3 pp 1127ndash1139 1990
[3] B Prasad and H P Singh ldquoUnbiased estimators of finite popu-lation variance using auxiliary information in sample surveysrdquoCommunications in Statistics vol 21 no 5 pp 1367ndash1376 1992
[4] R S Biradar and H P Singh ldquoAn alternative to ratio estimatorof population Variancerdquo Assam Statistical Review vol 8 no 2pp 18ndash33 1994
[5] M Rueda Garcia and A Arcos Cebrian ldquoRepeated substitu-tion method the ratio estimator for the population variancerdquoMetrika vol 43 no 2 pp 101ndash105 1996
[6] A Arcos M Rueda M D Martınez S Gonzalez and YRoman ldquoIncorporating the auxiliary information available invariance estimationrdquo Applied Mathematics and Computationvol 160 no 2 pp 387ndash399 2005
[7] C Kadilar andH Cingi ldquoImprovement in estimating the popu-lation mean in simple random samplingrdquo Applied MathematicsLetters vol 19 no 1 pp 75ndash79 2006
[8] H P Singh S Singh and J M Kim ldquoEfficient use of auxiliaryvariables in estimating finite population variance in two-phasesamplingrdquoCommunications of the Korean Statistical Society vol17 no 2 pp 165ndash181 2010
[9] I Olkin ldquoMultivariate ratio estimation for finite populationsrdquoBiometrika vol 45 pp 154ndash165 1958
[10] S R Srivastava S R Srivastava and B B Khare ldquoChain ratiotype estimator for ratio of two populationmeans using auxiliarycharactersrdquo Communications in Statistics vol 18 no 10 pp3917ndash3926 1989
[11] L N Upadhyaya K S Kushwaha and H P Singh ldquoA modifiedchain ratio-type estimator in two-phase sampling using multiauxiliary informationrdquo Metron vol 48 no 1ndash4 pp 381ndash3931990
[12] V K Singh H P Singh H P Singh and D Shukla ldquoA generalclass of chain estimators for ratio and product of two means ofa finite populationrdquo Communications in Statistics vol 23 no 5pp 1341ndash1355 1994
[13] L Chand Some ratio-type estimators based on two or more aux-iliary variables [PhD thesis] Iowa State University Ames IowaUSA 1975
[14] R K Gupta S Singh and N S Mangat ldquoSome chain ratio typeestimators for estimating finite population variancerdquo AligarhJournal of Statistics vol 12-13 pp 65ndash69 1992-1993
[15] W A Abu-Dayyeh M S Ahmed R A Ahmed and H AMuttlak ldquoSome estimators of a finite population mean usingauxiliary informationrdquo Applied Mathematics and Computationvol 139 no 2-3 pp 287ndash298 2003
[16] M N Murthy Sampling Theory and Methods Statistical Pub-lishing Society Calcutta India 1967
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Mathematical Problems in Engineering
Hindawi Publishing Corporationhttpwwwhindawicom
Differential EquationsInternational Journal of
Volume 2014
Journal ofApplied Mathematics
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
ProbabilityandStatistics
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Advances in
Mathematical Physics
Complex AnalysisJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
OptimizationJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
International Journal of
Combinatorics
OperationsResearch
Advances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Journal of Function Spaces
Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of Mathematics and Mathematical Sciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Algebra
Discrete Dynamics in Nature and Society
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
DecisionSciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Discrete MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Stochastic AnalysisInternational Journal of
Journal of Probability and Statistics 5
Conflict of Interests
The authors declare that there is no conflict of interestsregarding the publication of this paper
References
[1] C T Isaki ldquoVariance estimation using auxiliary informationrdquoJournal of the American Statistical Association vol 78 no 381pp 117ndash123 1983
[2] B Prasad and H P Singh ldquoSome improved ratio-type estima-tors of finite population variance in sample surveysrdquo Commu-nications in Statistics vol 19 no 3 pp 1127ndash1139 1990
[3] B Prasad and H P Singh ldquoUnbiased estimators of finite popu-lation variance using auxiliary information in sample surveysrdquoCommunications in Statistics vol 21 no 5 pp 1367ndash1376 1992
[4] R S Biradar and H P Singh ldquoAn alternative to ratio estimatorof population Variancerdquo Assam Statistical Review vol 8 no 2pp 18ndash33 1994
[5] M Rueda Garcia and A Arcos Cebrian ldquoRepeated substitu-tion method the ratio estimator for the population variancerdquoMetrika vol 43 no 2 pp 101ndash105 1996
[6] A Arcos M Rueda M D Martınez S Gonzalez and YRoman ldquoIncorporating the auxiliary information available invariance estimationrdquo Applied Mathematics and Computationvol 160 no 2 pp 387ndash399 2005
[7] C Kadilar andH Cingi ldquoImprovement in estimating the popu-lation mean in simple random samplingrdquo Applied MathematicsLetters vol 19 no 1 pp 75ndash79 2006
[8] H P Singh S Singh and J M Kim ldquoEfficient use of auxiliaryvariables in estimating finite population variance in two-phasesamplingrdquoCommunications of the Korean Statistical Society vol17 no 2 pp 165ndash181 2010
[9] I Olkin ldquoMultivariate ratio estimation for finite populationsrdquoBiometrika vol 45 pp 154ndash165 1958
[10] S R Srivastava S R Srivastava and B B Khare ldquoChain ratiotype estimator for ratio of two populationmeans using auxiliarycharactersrdquo Communications in Statistics vol 18 no 10 pp3917ndash3926 1989
[11] L N Upadhyaya K S Kushwaha and H P Singh ldquoA modifiedchain ratio-type estimator in two-phase sampling using multiauxiliary informationrdquo Metron vol 48 no 1ndash4 pp 381ndash3931990
[12] V K Singh H P Singh H P Singh and D Shukla ldquoA generalclass of chain estimators for ratio and product of two means ofa finite populationrdquo Communications in Statistics vol 23 no 5pp 1341ndash1355 1994
[13] L Chand Some ratio-type estimators based on two or more aux-iliary variables [PhD thesis] Iowa State University Ames IowaUSA 1975
[14] R K Gupta S Singh and N S Mangat ldquoSome chain ratio typeestimators for estimating finite population variancerdquo AligarhJournal of Statistics vol 12-13 pp 65ndash69 1992-1993
[15] W A Abu-Dayyeh M S Ahmed R A Ahmed and H AMuttlak ldquoSome estimators of a finite population mean usingauxiliary informationrdquo Applied Mathematics and Computationvol 139 no 2-3 pp 287ndash298 2003
[16] M N Murthy Sampling Theory and Methods Statistical Pub-lishing Society Calcutta India 1967
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Mathematical Problems in Engineering
Hindawi Publishing Corporationhttpwwwhindawicom
Differential EquationsInternational Journal of
Volume 2014
Journal ofApplied Mathematics
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
ProbabilityandStatistics
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Advances in
Mathematical Physics
Complex AnalysisJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
OptimizationJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
International Journal of
Combinatorics
OperationsResearch
Advances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Journal of Function Spaces
Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of Mathematics and Mathematical Sciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Algebra
Discrete Dynamics in Nature and Society
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
DecisionSciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Discrete MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Stochastic AnalysisInternational Journal of
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Mathematical Problems in Engineering
Hindawi Publishing Corporationhttpwwwhindawicom
Differential EquationsInternational Journal of
Volume 2014
Journal ofApplied Mathematics
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
ProbabilityandStatistics
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Advances in
Mathematical Physics
Complex AnalysisJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
OptimizationJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
International Journal of
Combinatorics
OperationsResearch
Advances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Journal of Function Spaces
Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of Mathematics and Mathematical Sciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Algebra
Discrete Dynamics in Nature and Society
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
DecisionSciences
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Discrete MathematicsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom
Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Stochastic AnalysisInternational Journal of