Stata Capabilities - Survey Methods

New in Stata 10

Survey regression models
  • Linear regression
  • Logistic regression
  • Multinomial logistic regression
  • Negative binomial regression
  • Ordered logistic regression
  • Probit regression
  • Ordered probit regression
  • Poisson regression
  • Censored and interval regression
  • Instrumental variables regression
  • Heckman selection model
  • Probit estimation with selection

Variance and standard error estimates

  • Taylor-series linearization (Huber/White/sandwich)
  • Balanced and repeated replications (BRR)
  • Survey jackknife

Features

  • Design effects
  • Misspecification effects
  • Effects for linear combinations
  • Estimate linear/nonlinear combinations of parameters
  • Hypotheses tests for survey data
  • Poststratification
  • Estimation with linear constraints






Summary statistics
  • Population and subpopulation means
  • Population and subpopulation proportions
  • Population and subpopulation ratios
  • Population and subpopulation totals
  • Provide full covariance estimates across subpopulations

Summary tables

  • Two-way contingency tables with tests of independence
  • One-way tables
  • Table describing the sampling design of survey data

Estimators that support linear constraints

Sampling designs

  • Sampling (probability) weights
  • Stratification
  • Clustering
  • Multistage designs
  • Finite population correction in all stages

Marginal effects

  • Marginal effects and elasticities
  • Standard errors and confidence intervals
  • Computed at means or specified covariate values
  • Computed for any predicted statistic

Maximum pseudolikelihood estimation

  • User-defined likelihoods
  • Survey characteristics automatically handled
Back to Capabilities Home

Back to Stata homepage
Back to Timberlake Consultants

©Timberlake Consultants Limited
Last revised:18/06/2007