Custom Programming SAS Code for Merging SPD Files

/*
PROGRAM: SPDMERGE.SAS

DESCRIPTION: The first SPD longitudinal file contains edited information on
employment, earnings, income sources, income amounts, health insurance, and
poverty. It also contains some information regarding means-tested government
transfer programs, such as TANF and Food Stamps. Due to time constraints
and in the interest of providing users with a consistent and current file in
the most timely manner, a number of variables from the SPD core and topical
modules were not included on the first version of the longitudinal file.

In addition to the core data on the SPD longitudinal file, the SPD 1998
instrument collected a wealth of information that allows for an in-depth
analysis of household and family characteristics pertaining directly to
welfare reform. These data include questions on new welfare programs such
as transportation assistance, childcare vouchers, and newly instituted work
training and child well-being modules. Although these data are not currently
included on the SPD longitudinal file, they are available on the SPD 1998
Experimental file.

Users wishing to use the full complement of these data may do so by merging
SPD Longitudinal file with the 1998 file. Users should be warned, however,
that these data underwent less rigorous editing procedures than did data on
other SPD files. In addition, the data available are for 1997 only, so
comparisons using the same variable over time, in most cases cannot be made.

The merging procedure is actually quite simple and the program below can be
directly downloaded and executed with a minimum amount of work. The program
merges the SPD longitudinal file with the SPD 1998 Experimental file at the
person level. The variables required for this type of merge are:

SIPP_PNL SIPP Panel Number
PP_ID Sample Unit Identification Number
PP_PNUM Person Identification Number
PP_ENTRY Entry ID.

For merges at other units of analysis, please see
http://www.sipp.census.gov/spd/long/fileinfo-1stl.htm for more information.

To execute the file merge, simply select a directory path for the library
statement, enter the appropriate name for each file where specified in the
program, and select a name for the merged file. It is also recommended that
users give some thought to the size of the files they are working with.
There are 2047 variables on the longitudinal file and 3122 on the experimental
file. Keeping only the variables from the experimental file that one wishes
to work with will help reduce file size and increase the processing speed
of programs.

TYPE: SAS PROC/DATA STEP

CREATED: 1/23/01

BY: AFJ

libname spd '' ;

data temp_lng ;
set spd. ;

MERGEKEY = (sipp_pnl||pp_id||pp_pnum||pp_entry) ;

run ;

data temp_exp ;
set spd.
(keep = sipp_pnl pp_id pp_pnum pp_entry plus specify which variables to
keep from the experimental file) ;

MERGEKEY = (sipp_pnl||pp_id||pp_pnum||pp_entry) ;

run ;

proc sort data = temp_lng ;
by MERGEKEY;
run ;

proc sort data = temp_est ;
by MERGEKEY;
run ;

data spd. ;
merge temp_lng (in=a) temp_est (in=b) ;
by mergekey ;
if a ;
run ;

Linking Files

Contact: (dsd.survey.program.dynamics@census.gov)
URL: http://www.sipp.census.gov/spd/