Manisha -

I am not familiar with the data you note (NSSO?), but if you are talking about official statistics from finite populations, you may be seeing survey weights, which are the inverse of the probabilities of selection.  This depends upon a random sampling design.  Here is an introduction:

 https://www.researchgate.net/publication/262971887_SURVEY_WEIGHTS

Note:  In probability proportional to size sampling, each member of a sample has its own probability of selection and weight.  Otherwise, groups of data share survey weights.

You could also be referring to regression weights, which tell us, based on heteroscedasticity, how much emphasis each collected datum should be given, based on the naturally underlying 'error' structure.  Here is an introduction to heteroscedasticity for finite populations, followed by notes on  weighted least squares regression: 

 https://www.researchgate.net/publication/262972023_HETEROSCEDASTICITY_AND_HOMOSCEDASTICITY

 https://www.researchgate.net/publication/263036348_Properties_of_Weighted_Least_Squares_Regression_for_Cutoff_Sampling_in_Establishment_Surveys

There are also calibration weights which start with survey weights, and then adjust them according to how they would apply to related "auxiliary"/(actually regressor) data which you have on the universe, or even just the sample by way of adjusting for nonresponse.  A little on calibration weights can be found in the following:

 https://www.researchgate.net/publication/261508465_Use_of_Ratios_for_Estimation_of_Official_Statistics_at_a_Statistical_Agency

My guess, especially if you have not seen anything about "auxiliary" or "regressor" data, is that you may be seeing survey weights.   There are a number of good books which cover this, including 

 Cochran, W.G(1977), Sampling Techniques, 3rd ed., John Wiley & Sons. 

and

Lohr, S.L(2010), Sampling: Design and Analysis, 2nd ed., Brooks/Cole.

When auxiliary data are available, there are other books with more emphasis on that, including

Särndal, CE, Swensson, B. and Wretman, J. (1992), Model Assisted Survey Sampling, Springer-Verlang.    

and

Brewer, KRW (2002), Combined survey sampling inference: Weighing Basu's elephants, Arnold: London and Oxford University Press.  

Almost all quantitative survey statistics texts (continuous data and proportions) in the second half of the 20th century emphasized survey weights, though I've found most, it seems, did note a little on regression modeling.  More emphasis on regression seems to appear now. 

Here is a little interesting history, where the model given at the end is for the classical ratio estimator (CRE), and many establishment surveys show greater heteriscedasticity than that, but the CRE does seem robust:

====================

Ken Brewer's Waksberg Award article: 

 Brewer, K.R.W. (2014), “Three controversies in the history of survey sampling,” Survey Methodology,

(December 2013/January 2014), Vol 39, No 2, pp. 249-262. Statistics Canada, Catalogue No. 12-001-X.

http://www.statcan.gc.ca/pub/12-001-x/2013002/article/11883-eng.htm

He believed in using probability sampling and models together, but he explains the different approaches, the pros and cons. 

====================

....

But, as I say, my guess is that you are seeing survey weights.  If you multiply each collected (i.e., sampled) datum by its survey weight, and add, then you should obtain the estimated population (or subpopulation or stratum) total estimate that they show, if this is the case.

Cheers - Jim

Article Use of Ratios for Estimation of Official Statistics at a Sta...

Article SURVEY WEIGHTS

Article HETEROSCEDASTICITY AND HOMOSCEDASTICITY

Article Properties of Weighted Least Squares Regression for Cutoff S...

Hey James,

Thanks a lot for helping me out. I am looking at the official statistics and they are survey weights.

NSSO - National Sample Survey Organisation.

I will go through the links you provided for the same, thanks a lot. I am sure they will help.

Regards,

Manisha Kapoor

Hello all,

I came across two very interesting web pages for NSSO Data.

https://nssounitlevel.wordpress.com and

https://nssodataanalysis.wordpress.com

I had queries on nsso extraction, weights and analysis, which both have helped.

To calculate the weights from multiplier, you have to go through the readme file of the purchased data.

To have a look on this, you can visit my blog:

https://nssounitlevel.wordpress.com

The use of multipliers/combined weights are for representation of the whole population parameter in a NSSO survey.