1998
DOI: 10.2307/146433
|View full text |Cite
|
Sign up to set email alerts
|

An Analysis of Sample Attrition in Panel Data: The Michigan Panel Study of Income Dynamics

Abstract: coefficients. We provide a statistical framework for conducting tests for attrition bias that draws a sharp distinction between selection on unobservables and on observables and that shows that weighted least squares can generate consistent parameter estimates when selection is based on observables, even when they are endogenous. Our empirical analysis shows that attrition is highly selective and is concentrated among lower socioeconomic status individuals. We also show that attrition is concentrated among tho… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

15
298
0
3

Year Published

2010
2010
2021
2021

Publication Types

Select...
5
3

Relationship

0
8

Authors

Journals

citations
Cited by 732 publications
(323 citation statements)
references
References 27 publications
15
298
0
3
Order By: Relevance
“…Nevertheless, in line with the evidence reported in Cheng and Trivedi (2015), a comparison of the adjusted and unadjusted estimates reveals that, despite the missing-at-random assumption is strongly rejected, withdrawal attrition does not significantly affect the estimates of the TC/HDL ratio equation for the selected incident patients. This evidence is also corroborated by a comparison between the total and nonattriting sample á la Fitzgerald, Gottschalk, and Moffitt (1998) and by the Hausman-like test proposed by Verbeek and Nijman (1992).…”
Section: Robustness Check: Working With Subsamples and Dealing With Asupporting
confidence: 60%
See 1 more Smart Citation
“…Nevertheless, in line with the evidence reported in Cheng and Trivedi (2015), a comparison of the adjusted and unadjusted estimates reveals that, despite the missing-at-random assumption is strongly rejected, withdrawal attrition does not significantly affect the estimates of the TC/HDL ratio equation for the selected incident patients. This evidence is also corroborated by a comparison between the total and nonattriting sample á la Fitzgerald, Gottschalk, and Moffitt (1998) and by the Hausman-like test proposed by Verbeek and Nijman (1992).…”
Section: Robustness Check: Working With Subsamples and Dealing With Asupporting
confidence: 60%
“…Even though the extent of attrition due to both death and patients' withdrawal appears to be small, in what follows, we further investigate if our main results are in some way affected by the presence of attrition bias. Following the approach in Fitzgerald et al (1998), we compare the estimates reported in column 3 of Table 4, obtained using the total sample, with those obtained using only the nonattriting sample, the latter excluding 1,281 (death or withdrawal) patients for 21,011 observations. By looking at Table AI, we notice that only the coefficient of the Age:59-64 dummy is significantly different between the total and nonattriting samples.…”
Section: Discussionmentioning
confidence: 99%
“…Fitzgerald et al 1998). One particular reason for attrition is that some groups of respondents cannot be reached easily or may lose interest in participating in later waves of the survey.…”
Section: Panel Attritionmentioning
confidence: 99%
“…Sampling partners was an innovation of both HRS and ELSA and stands in sharp contrast to typical panels such as the PSID and BHPS which rely on a single respondent who answers questions for both partners (Fitzgerald et al, 1998). Existing research has shown that the quality of health information reported for the partner is much lower (Weir and Smith, 2007;Smith, 2007).…”
Section: Explaining the Higher Attrition Rate In Elsamentioning
confidence: 99%
“…Most of the existing literature has focused on non-ageing panels in the United States, especially during earlier time periods when attrition rates typically were considerably lower (Becketti et al, 1988;Fitzgerald et al, 1998;Lillard and Panis, 1998; Zapel, 1998).…”
mentioning
confidence: 99%