Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: weight in a field survey

From   Steve Samuels <>
Subject   Re: st: weight in a field survey
Date   Tue, 23 Mar 2010 11:14:10 -0400

Estelle, I think that by "stratum weight", you mean the first-stage,
selection of villages within strata, and that by "cluster weight", you
mean the second-stage selection of households within village.  Please
use standard terminology to avoid confusion. The general rule for
constructing weights is as you say: multiply together the first and
second-stage probabilities, then invert. (Or invert and then

The 5 HH per village design might not be PPS.  Suppose the measure of
size at the first stage is the estimated population size of each
village. Two villages in the same stratum, both with estimated
population N = 2000, would have the same first-stage probability of
selection. If there are 500 HH in the first village, 600 in the
second, and a simple random sample of HH is taken in each, the second
stage probabilities will differ.  See Kish, Survey Sampling, 1965,
Wiley, page 223.

Michael is not correct about having to add a weight for household size
to analyze individual data.  Each person in a selected household
shares the selection probability of the household and gets the same
sampling weight.

If there is external information about the stratum characteristics,
including the distribution of age, village sizes, gender, sizes of
households, for example, the sampling weights can be "raked" so that
totals of these characteristics estimated from the sample match the
known population totals. See Nick Winter's -survwgt- package,
available at SSC.

The description of your selection algorithms is vague.  Please
describe the algorithm that you used or intend to use.

On Mon, Mar 22, 2010 at 5:43 PM, Michael I. Lichter
<> wrote:
> Estelle,
> If this is a true PPS design, you shouldn't have to weight for anything
> except for unequal probabilities of selection across strata. (Unequal
> probabilities for village selection are balanced out by unequal
> probabilities for household selection.) That weight would be, as I think
> you're saying, weight for stratum i = 1/(proportion of total population in
> stratum i). Of course, you will also want to add a weight for household size
> when doing individual-level tabulations.
> Michael
> Estelle PASQUIER wrote:
>> Dear statlist,
>>  We are conducting a field survey on
>> malaria. Our population is sampled by stratifying it in rural and urban
>> settings (first step) then selecting villages with a probability
>> proportional
>> to their population sizes. The last step consists in selecting randomly a
>> fixed
>> number of households in the village. I am a little bit concern with the
>> sample
>> weights I have to choose with svy: my suggestion would be to calculate the
>> stratum weight as the ratio of total population on population in each
>> stratum;
>> to calculate the cluster weight as the inverse of the sampling
>> probability, and
>> then to use the product of these two weights as final weight. Am I right?
>> My second concern is to know if I there is
>> anything to do for the last step of this sampling design? Thanks for any
>> help,
>>  Estelle
>> *
>> *   For searches and help try:
>> *
>> *
>> *
> --
> Michael I. Lichter, Ph.D. <>
> Research Assistant Professor & NRSA Fellow
> UB Department of Family Medicine / Primary Care Research Institute
> UB Clinical Center, 462 Grider Street, Buffalo, NY 14215
> Office: CC 126 / Phone: 716-898-4751 / FAX: 716-898-3536
> *
> *   For searches and help try:
> *
> *
> *

Steven Samuels
18 Cantine's Island
Saugerties NY 12477
Voice: 845-246-0774
Fax:    206-202-4783

*   For searches and help try:

© Copyright 1996–2016 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index