Scribd
Upload
7 views5 pages

Lecture 13

The document discusses ratio and regression estimation methods for improving population mean estimates using auxiliary information. It highlights that the ratio method is most effective when the relationship between variables is linear, while regression estimation is used when the regression line does not pass through the origin. Additionally, it provides formulas for calculating estimators and their variances, along with an example of estimating maize cultivation area.

Uploaded by

Jorams Barasa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
7 views5 pages

Lecture 13

The document discusses ratio and regression estimation methods for improving population mean estimates using auxiliary information. It highlights that the ratio method is most effective when the relationship between variables is linear, while regression estimation is used when the regression line does not pass through the origin. Additionally, it provides formulas for calculating estimators and their variances, along with an example of estimating maize cultivation area.

Uploaded by

Jorams Barasa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

( S y2 + R 2 S x2 − 2 R ρ S x S y )

V1 (Yˆ ) = N 2 (1 − f ) ......................(6.2.2.10)
n

Judging by this approximation, the ratio method will give a more precise result whenever

RS x 
2ρ > 
Sy


or  ....................................(6.2.2.11)
C.V ( x) 
ρ> 
2C.V ( y ) 

Thus, the issue depends by and large on the strength of correlation between y and x . If x is
the same character as y but has been measured on a previous occasion, the coefficient of
variation may be taken to be equal. In that case it pays to use the ratio method of estimation if
ρ exceeds 0.5. But one should not be dogmatic about the inequality (6.2.2.11) since it is
based on an approximation. Ratio estimator is at its best when the relation between y and x
is a straight line through the origin i.e. y = kx .

6.2 Regression Estimation


This is another method of using auxiliary information to improve the estimate of the
population mean or total of the character under study. As has been noted, the ratio method is
at its best when y = kx .

Conversely, the regression estimation is used to estimate the population mean when the
regression line of y on x does not pass through the origin but makes an intercept along the
y − axis( y ≠ kx) .

6.2.1 Regression Estimation of Population Mean


Let yi , xi (i = 1, 2,L, n) be the sample values of the main character y and auxiliary character
x respectively obtained on a simple random sample of size n selected without replacement
from a population of size N. An estimator of the population mean of y is

yd = y − k ( x − X )............(6.2.2.12)

k is a suitably chosen constant.

82
The estimator y d is called difference estimator, which is unbiased for Y .

Its variance is

V ( yd ) = V ( y ) + k 2V ( x ) − 2kCov( x , y )
1− f 2
= ( S y + k 2 S x2 − 2kS xy )..................................(6.2.2.13)
n

To find the best value of k to use, we differentiate V ( yd ) with respect to k and equate it to

zero. This gives k optimum.

S xy
kop = = β , the population regression coefficient.
S x2

The sample estimator of the V ( yd ) is

1− f n
∑ {( yi − y ) − k ( xi − x )}
2
Vˆ ( yd ) =
n(n − 1) i =1
1− f 2
=
n
( s y + k 2 sx2 − 2ksxy ) .....................................(6.2.2.14)

When we substitute β , the regression coefficient, in place of k in (6.2.2.12) we have the


estimator

yd = y − β ( x − X )............(6.2.2.15)

In practice β is not known but can be estimated from the sample at hand. The sample
sxy
estimator of β is βˆ = .
sx2

Using this estimated value of β we have the linear regression estimator of the mean

ylr = y − βˆ ( x − X ).......................(6.2.2.16)

In large sample, the sample estimator of the variance of ylr is

1− f 2 ˆ2 2
Vˆ ( ylr ) =
n
( )
s y + β s x − 2 βˆ sxy ...............(6.2.2.17)

Or

83

1− f 2
Vˆ ( ylr ) = s y (1 − ρˆ 2 ) 
n 

where  .......................(6.2.2.18)
sxy2 
ρ = 2 2
ˆ 2 
s y sx 

Example 6.2

For the data in example 6.1, estimate the total area under maize cultivation using the method
of regression estimation.

Solution

Since the population regression coefficient is not known, it is estimated from the sample data.
From example 6.1

sxy 88302.982
βˆ = 2
= = 0.46387
s x 190361.93

Using (6.2.2.16) the regression mean is

ylr = y − βˆ ( x − X )
= 530.0755 − 0.46387(829.1885 − 952.4995) = 587.2758

The estimate of the total area under maize cultivation is

Yˆlr = Nylr = 88 × 587.2758 = 51, 680.268ha

The variance of this total is given by

Vˆ (Yˆlr ) = N 2Vˆ ( ylr )

Using (6.2.2.18) we need to obtain ρ̂ 2

sxy2 (88302.982) 2
ρˆ 2 = = = 0.66256
sx2 s y2 (190361.93)(61822.741)

84
1− f 2
Vˆ (Yˆlr ) = N 2 s y (1 − ρˆ 2 )
n
= (88) (0.03864)(61822.742)(0.33744)
2

= 6, 242,338.104

In-Text Question
The ratio method is at its best when y = kx . True or False

In-Text Answer
True

Summary for study session 6


In study session 6, you have learnt that:

1. Ration estimation is a technique that uses available auxiliary information which is


connected with the variable of interest. In a sample survey, in addition to estimating
the means, totals and proportions, you may wish to estimate the ratio of two
characters.
2. Ratio estimates are prejudiced and modifications must be made when they are used in
experimental or survey.
3. Ratio estimators, although biased, are consistent, and with simple random sampling
for moderately large samples, the bias is negligible.
4. Regression estimation is another method of using auxiliary information to improve
the estimate of the population mean or total of the character under study

Self-Assessment Questions (SAQs) for Study Session 6


Now that you have completed this study session, you can assess how well you have achieved
its Learning outcomes by answering the following questions. You can check your answers
with the Notes on the Self-Assessment questions at the end of this Module.

SAQ 6.1
Define ratio estimation

SAQ 6.2
Explain the use of regression estimation

85
Notes on SAQ
SAQ 6.1
Ration estimation is a technique that uses available auxiliary information which is connected
with the variable of interest

SAQ 6.2
This is another method of using auxiliary information to improve the estimate of the
population mean or total of the character under study

References
Cochran, W.G, (1977); Sampling Techniques third edition, New York: John Wiley & Sons

Daroga Singh and Chaudhary F.S, (1986); Theory and Analysis of Sample Survey Design,
New Delhi: Wiley Eastern Limited

Des Raj and Promod Chandhok (1998); Sample Survey Theory, New Delhi: Narosa
Publishing House

Kish L. (1965); Survey Sampling, New York: John Wiley & Sons

Okafor F.C (2002); Sampling Survey Theory with Applications, Nsukka: Afro-Orbis
Publishers

Mukhopadhyay P. (2005): Theory and Methods of Survey Sampling, New Delhi:


Prentice-Hall of India Private Limited

86

You might also like