NotesFAQContact Us
Collection
Advanced
Search Tips
ERIC Number: ED456162
Record Type: Non-Journal
Publication Date: 1998-Feb
Pages: 40
Abstractor: N/A
Reference Count: N/A
ISBN: N/A
ISSN: N/A
Methodological and Theoretical Issues in Studying Nonresident Fathers: A Selective Review.
Schaeffer, Nora Cate; Seltzer, Judith A.; Dykema, Jennifer
This paper reviews some of the principal methodological issues that arise in studying nonresident fathers using surveys, including sample frame, survey participation, and response error. The review focuses on issues involving divorced families. Results are highlighted from analyses of important national data sets and of local data sets that include a record-check component. The findings reviewed point to some of the challenges that arise when attempting to describe fathers' points of view using survey data. The most obvious is the challenge of improving participation rates among both fathers and mothers, but particularly among fathers and particularly among some subgroups of fathers, such as those who do not pay support. Nonparticipation bias is clearly an important issue for some estimates, such as the mean amount of child support paid. Although the evidence is less conclusive, it appears that nonparticipation bias may be less of a problem in estimating relationships among variables, such as the factors predicting child support payments. In addition to the problems raised by nonparticipation, the studies reviewed here suggest that for some variables, such as the amount of child support paid, response error may be more serious for fathers than for mothers. Because social desirability pressures appear to contribute to this response error, it is plausible that such errors affect constructs such as "contact" for which there is no criterion in the same way that they affect constructs for which a criterion is available. Research using social-information-processing models to study response errors may suggest ways to improve the accuracy of answers provided by both fathers and mothers. None of the studies reviewed here estimate the impact of response errors on model estimates, but such effects can be expected to be substantial. (Contains 15 tables and 28 references.) (Author/SLD)
Publication Type: Information Analyses
Education Level: N/A
Audience: N/A
Language: English
Sponsor: Annie E. Casey Foundation, Baltimore, MD.; Ford Foundation, New York, NY.; Mott (C.S.) Foundation, Flint, MI.
Authoring Institution: Pennsylvania Univ., Philadelphia. National Center on Fathers and Families.