Survey data analysis
Stata is the most complete software available for the statistical analysis
of survey data.
Survey data requires special statistical techniques, and Stata is at the
forefront in bringing those techniques to those who analyze survey data.
Stata let’s you make proper inferences with data from complex surveys, including
surveys with multistage designs, stratification, clustering, and sampling
weights. In addition to means, totals, and ratios estimators, Stata
provides over 40 regression-based estimators for survey data, including
estimators for continuous outcomes, binary outcomes, ordered outcomes,
multinomial outcomes, count outcomes, censored outcomes, models with sample
selection, models with endogenous covariates, and models for survival data.
Stata can perform linear and nonlinear hypothesis tests of the estimated
parameters and can compute predictive margins to further analyze your fitted
models.
With Stata, you describe your survey design once. After that, you simply
prefix your Stata command with svy: (or select “Survey data
estimation” in the dialog boxes), and Stata makes the correct
statistical adjustments for your complex survey design.
Stata provides the five variance estimators for survey
data—Taylor-series linearization (Huber/White/sandwich), balanced and
repeated replications (BRR), survey bootstrap, survey jackknife, and
successive difference replication (SDR). Stata also has facilities to
support surveys that provide their own replicate weight variables. The
survey bootstrap and SDR variance estimators require replicate weight
variables.
Explore Stata 12’s resources on survey data analysis
Stata is a complete integrated statistical package that provides
everything you need for data analysis, data management, and graphics.
To learn more, click here.
|
;)