Survey Methodology #2006-02 - Use Of Dependent Interviewing Procedures To Improve Data Quality In The Measurement Of Change - U.s. Census Bureau Page 2
ADVERTISEMENT
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25Use of Dependent Interviewing Procedures
to Improve Data Quality in the Measurement of Change
Jeffrey Moore, Nancy Bates, Joanne Pascale, Julia Klein Griffiths, Aniekan Okon
U.S. Census Bureau
1. Introduction
This chapter examines the impact of dependent interviewing procedures on “seam bias,” a form of
measurement error peculiar to longitudinal surveys. “Seam bias” refers to the tendency for
estimates of change measured across the “seam” between two successive survey administrations
to far exceed change estimates measured within a single interview. Except in very special
circumstances, the presence of seam bias is a clear signal of measurement error. Much research
over the past two decades has documented the existence of seam bias in longitudinal surveys, and
has also shed light on the essential nature of the phenomenon – too little reporting of change
within the reference period of a single interview, and too much reporting of change at the seam.
Attempts to control seam bias have met with some success, but have been limited primarily to
employment-related characteristics.
Following an extensive research and development program, the U.S. Census Bureau recently
implemented new procedures in the Survey of Income and Program Participation (SIPP) 2004
panel questionnaire in an attempt to significantly reduce seam bias. The primary questionnaire
revision which was intended to accomplish this was a more extensive and more focused use of
dependent interviewing (DI) procedures than had been employed in prior SIPP panels. This
chapter describes those procedures, and examines their impact on estimates of month-to-month
change across the initial waves of the 2004 panel for a number of characteristics (e.g.,
participation in government transfer programs, school enrollment, employment, earnings, and
health insurance coverage) through a comparison with similar estimates derived from 2001 SIPP
panel data. We find evidence of significant positive change with the new procedures – estimates
of month-to-month change from the initial interview waves of the 2004 panel are in general much
less afflicted with seam bias than their 2001 counterparts. Even with the improvement, however,
much seam bias still remains.
The remainder of this chapter is organized as follows: In Section 2 we briefly describe the seam
bias phenomenon, and summarize work which has attempted to understand and ameliorate it.
Section 3 provides background on SIPP, the U.S. Census Bureau survey that served as the vehicle
for the research presented here. We describe the DI procedures used in SIPP in the most recent
panel (2001) before the current one, and the development and refinement of new DI procedures
via a research and development effort called the SIPP Methods Panel project. Section 4 presents
the primary research results, which consist of comparisons of 2004 SIPP panel seam bias results,
across a variety of characteristics, with results for the same characteristics derived from the old
questionnaire used in the 2001 SIPP panel. Finally, in Section 5 we present our conclusions,
including implications of the current findings for future research.
-1-
ADVERTISEMENT
0 votes
Related Articles
Related forms
Survey Methodology #2012-02 - New Relationship And Marital Questions: A Reflection Of Changes To The Social And Legal Recognition Of Same-sex Couples In The U.s. - U.s. Census Bureau
Legal
Early Childhood Program Participation, From The National Household Education Surveys Program Of 2012 - U.s. Department Of Education
Legal
Form Hud-9887 And 9887-a - Document Pachagefor Applicant's/tenant's Concent To The Release Of Information - U.s. Department Of Housing And Urban Development
Legal
Related Categories
Parent category: Legal