The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Topics » Service Provision Assessment (SPA) » Understanding Tanzania 2014/15 SPA weights
Understanding Tanzania 2014/15 SPA weights Wed, 19 July 2017 12:08
 asheffel Messages: 1Registered: July 2017 Member
I am working with the 2014/15 Tanzania SPA survey and looking to calculate indicators using the facility, provider, and client observation/exit data files. In order to correctly specify the survey design, I am trying to understand how the survey weights in each dataset are calculated. Can anyone provide any detailed information on how each survey weight (facility, provider, and client) is calculated? In particular, I haven't been able to find any information on how the provider weights are calculated. I am also unsure if the facility weights have been incorporated into the provider weight and if the facility and provider weights have been incorporated into the client weight. Any insights into each weight variable would be much appreciated!

Many thanks!
Ashley
Re: Understanding Tanzania 2014/15 SPA weights [message #13205 is a reply to message #12827] Tue, 03 October 2017 11:46
 Liz-DHS Messages: 1392Registered: February 2013 Senior Member
Dear User,
A response from Senior Sampler, Dr. Ruilin Ren:
Quote:

For all SPA surveys (including the Tanzania SPA 2014/15), the health facilities are selected by a stratified equal probability systematic sampling, stratification is achieved by crossing region and facility type. The health facility weight is calculated based on the selection probability, and then adjusted for non-response at the sampling stratum level. The provider weight is calculated based on the facility weight multiplied by the inverse of the selection probability of providers; the providers are selected with equal probability by provider type within each of the sampling stratum, and then adjusted for provider non-response. The client weights (ANC visit weight, Family Planning visit weight and Sick child visit weight) are calculated in the same way as for the provider weight, based on the facility weight multiplied by the inverse of the selection probability of clients; the clients are selected with equal probability by client type within each of the sampling stratum, and then adjusted for client non-response. All the sampling weights presented in the data files are normalized weights which are relative weights and are valid for estimating means, proportions, ratios and rates, but not valid for estimating population totals and not valid for pooled data.
 Previous Topic: Senegal SPA 2012-2015, Drug Availability in Health Facilities Next Topic: Replicating Indicators Tanzania 2015 SPA
Goto Forum:

Current Time: Mon Mar 18 10:23:34 Eastern Daylight Time 2019