For what purpose are the American Housing Survey (AHS) "flat file" and "file flattener program" used?


Before 1997, each AHS dataset was released as a single file. Each record of this file contained all the data for a single housing unit. For those entities (such as persons) which occurred more than once in the same housing unit, this single file included one set of variables for each possible entity. For example, there were sixteen race variables (RACE1-RACE16), one for each person in the housing unit. Since most housing units did not contain this many persons (and some contained none at all), many of these variables were filled out with missing values. This arrangement of data is referred to as the flat file.

Beginning in 1997, the AHS datasets were reorganized on a relational basis. There is, for example, a separate PERSON file containing one record for each person in the survey. These records can be linked to the housing unit file using the CONTROL variable in what is known as a one-to-many relationship. There are similar files for recent mover groups, commuters, and home improvement projects. In addition, the RATIOV and MORTG/MORTGAGE files have only one record per housing unit, but they only have records for units with relevant data (for example, the MORTG/MORTGAGE file contains no records for renter-occupied households or for owner-occupied households that have no loans).

To access the SAS "Flat Files," go to the Data page and choose a survey year. Then choose either the national or metro Microdata PUF file.


Explore related topics: