Survey of Income and Program Participation

SIP 1996: Longitudinal Weight Adjustment Factor (WGT-19)

August 09, 2005

Purpose:

In the SIPP 1996 Panel, the 1997, 1998 and 1999 Calendar year weights sum to totals that exceeded the estimated March population controls of 1997, 1998 and 1999 respectively.

Statement of Problem 1:

The longitudinal population controls for March 1997, 1998 and 1999 are much higher than the corresponding cross sectional controls. Table 1 presents the March population controls and total weights for the cross sectional and longitudinal files. The longitudinal weights had been raked to controls that have a few million more than there should be. For more information, please see the attached document at the bottom of the page.

Resolution of Problem 1:

Scalar adjustment factors for each calendar year have been calculated and should be applied to every person’s weight included in Calendar years 1997-1999 estimates in order to correct the problem. These factors are listed below. Given the effect of the problem is random and not particular to any certain group, the scalar adjustment was determined to be sufficient in resolving the problem. For much information, please see the attached document at the bottom of the page.

Statement of Problem 2:

In the longitudinal 1997 Calendar year weighting, there are some cases in which people received non zero weights when they should have received zero weights. For most of these cases, the sample person is not present in any month of the year 1997. The sum of weights for these cases is approximately 1.3 million. For much information, please see the attached document at the bottom of the page.

Resolution of Problem 2:

Assign a value of zero to the 1997 Calendar year weight for all persons who are not present in March of 1997. These cases should be considered not in sample for the 1997 Calendar year. For much information, please see the attached document at the bottom of the page.

Click below to see more information about this User Note.