Microdata Access
The Cross-Section Archive has a customized subset download feature. This will enable you to download selected variables from the microdata.
For researchers affiliated with an ICPSR member institution, additional data is available through the ICPSR Data Access archive. Note that not all years are available for direct download; please reach out directly to ICPSR if you are interested in those missing years.
The standard cross-section archive and the ICPSR archive both include survey weights. When using the survey weights, users should consider two key points:
Household vs. Adult Level Survey Weights
The Surveys of Consumers weight adjustment includes base weight adjustments and calibration to a target population. Household weights are designed to yield a representative sample of all U.S. households, whereas adult weights are intended to produce a representative sample of all adults living in private households.
The decision on the appropriate weights to use should be driven by the data user's research objectives. It is not merely a matter of the measurement of household characteristics (e.g., size, location, income, wealth) versus individual characteristics (e.g., attitudes, expectations, education, employment, wages). Instead, this decision should hinge on the preferred unit of analysis or the "population" that the analysis aims to represent. For instance, studies focusing on population attitudes often use adult weights to examine the prevalence and dynamics of attitude changes among individuals. Conversely, studies on economic behavior might use household weights, as they often consider the household the primary decision-making unit for analysis.
Variance Estimation Method
Data users should use the linearized estimator for variance computations. This method is included in the specialized survey procedures found in commonly used software packages such as SAS® and R. Additionally, STATA® has recently incorporated linearized estimator computations across all its analysis procedures.
