WLS Data Conventions

Structure of WLS Longitudinal Data

This longitudinal data is presented as a rectangular file rather than a hierarchical file. Each record contains data for the entire family. There is only one ID for the public - IDPUB. IDPUB is the same for graduate respondents, their sibling and their spouse. Variables that are the same for the different respondents are named uniquely. See variable naming conventions for details.

Missing Values

The following missing value codes are used throughout 1957-2007 data. 1957-1994 data only include codes -1 through -4. In the online codebook, clicking on "Inap" for value -2 expands the description to indicate which cases are inappropriate. Clicking on the expanded description reduces the label back to "Inap".

Note: From 1964-1977, if the respondent was not fielded or did not respond at
all to the survey, all variables for the respondent in that year of the survey
are coded -3 (originally labelled refused, not ascertained).
From 1993-2011, if the respondent was not fielded or did not respond to
the survey at all, all variables for the respondent in that year are coded system
missing (. for numeric variables and blank for string variables).
	

Variable Naming

If the variable name is a nmenonic, then that variable is from 1957-1977 surveys.
From 1993 onward, variable names consist of a combination of letters and numbers
that allow for convenient identification of the year and mode of the survey.

Variables that differ only by year of the survey, have identical
names for every character past the first letter.  Please note that the fielding
of the surveys spanned multiple years.  In the table below we simplify the year by
listing only the year where the greatest number of suveys were completed. 

Menu