Family studies have seen a dramatic increase in the use of statistical tools for the analysis of nominal-level variables. Such models are categorized as log-linear models often known as logit models or logistic-regression models. Despite logistic regressions growing popularity there is still confusion about the nature and proper use in family studies. The authors present a nontechnical discussion of logistic regression with illustrations and comparisons to better-known procedures such as percentaging tables and ordinary least squares regression. They contend that logistic regression can be a powerful statistical procedure when used appropriately. Nominal-level dependent variables are common in family research and logistic-regression models appropriately model the impact of predictor variables on these outcomes. With the proliferation of computer software for estimating logistic-regression models use of logistic regression is likely to increase. Though some time and attention is required to master it the advantages of logistic regression make the effort worthwhile.
[1]
David A. Wise,et al.
College Choice In America
,
1983
.
[2]
P. Schmidt,et al.
Limited-Dependent and Qualitative Variables in Econometrics.
,
1984
.
[3]
E. Hanushek.
Statistical methods for social scientists
,
1977
.
[4]
Stephen E. Fienberg,et al.
The analysis of cross-classified categorical data
,
1980
.
[5]
G. Judge,et al.
The Theory and Practice of Econometrics
,
1981
.
[6]
Peter Burke,et al.
Log-linear models
,
1980
.
[7]
Alan L. Ginsburg,et al.
The Role of Responsibility and Knowledge in Reducing Teenage Out-of-Wedlock Childbearing
,
1987
.
[8]
John H. Aldrich,et al.
Linear probability, logit and probit models
,
1984
.