Approaches for Addressing Missing Data in Statistical Analyses of Female and Male Adolescent Fertility
Dudley L. Poston, Jr., Texas A&M University
Eugenia Conde, Texas A&M University
Missing data is a pervasive problem in social science research. In this paper we use data from the National Longitudinal Survey of Adolescent Health (Add Health) and undertake two separate analyses, one for females and the other for males, of the likelihood of having had a teen birth. We handle the problem of missing data using several approaches, namely, 1: listwise deletion, 2: mean substitution, 3: mean substitution for subgroups, 4: the proxy method, 5: dropping the variables with excessive amounts of missing data, and 6-8: three variants of multiple imputation. We show in our analyses that depending on the method used, many of the independent variables in the sex-specific models vary in whether they are, or are not, statistically significant in predicting the log odds of a person having had a teen birth, and in the ranking of the magnitude of their relative effects on the outcome.