Reviews methods for analyzing outcome data when some data are missing. Presents approach for comparing outcome results in two-group, randomized design with missing data. Proposed analysis is based on concept of "modeling our ignorance" by examining all possible outcomes, given known number of missing results with binary outcome, and then…

Develops random-effects probit model for case in which outcome of interest is series of correlated binary responses, obtained as product of longitudinal response process where individual is repeatedly classified on binary outcome variable or in multilevel or clustered problems in which individuals within groups are considered to share…

Addresses methodological and ethical issues in context of three related school-based studies of primary and targeted prevention of depressive symptoms and disorder in high school adolescents. Issues discussed include obtaining subject consent and protection of confidentiality, minimizing attrition, unit of assignment issue, and ensuring therapist…

Describes two research projects linking family processes with elementary school children's competence in school to provide examples of methodological issues related to investigations that cross systems of family and school. Issues discussed include gaining cooperation from school, recruitment and consent procedures, data collection and measurement…