Effects of Measurement Error on Occupational Event History Analysis

Written by:
Working Paper Number: SEHSD-WP1991-10 or SIPP-WP-151

Event history analysis is one of the more promising new tools available to academic and policy analysts interested in the dynamic processes underlying employment and program participation behavior.  The monthly dating of occupation and industry employment, along with its large size and national representativeness, potentially make the SIPP one of the most attractive data bases for such analyses.  The extent to which the well documented measurement errors in the SIPP (and all other panel studies with reference periods longer than the basic unit of measurement) detract from this potential is not well understood.  The natural experiment resulting from the overlap in the 1985 and 1986 SIPP Panels, which differ only in the method of collecting occupation and industry data, provides a unique opportunity to gain insight into the effect of measurement errors on event history analysis.

Page Last Revised - October 8, 2021