Rc. Leaf et al., STATISTICAL-METHODS FOR ANALYSES OF INCOMPLETE CLINICAL SERVICE RECORDS - CONCURRENT USE OF LONGITUDINAL AND CROSS-SECTIONAL DATA, Journal of consulting and clinical psychology, 61(3), 1993, pp. 495-505
Statistical procedures that have become routine in other social scienc
es were used to analyze data from clinical service records. Despite th
e absence of control groups, nonrandom assignment of clients to treatm
ent conditions, and incomplete records, effective analyses of psychoth
erapeutic processes were possible. Multivariate regression models, wit
h variables that were transformed to significantly improve skewness an
d regression linearity, were controlled for heteroskedasticity and for
end-point censoring of dependent variables. They were also used to me
asure the effects of a categorical variable (gender) and a scalable va
riable (intake distress) on a reactive outcome measure (of acute distr
ess) and on an unreactive one (of long-term life satisfaction). Graphi
cal methods for summarizing large data sets helped identify intake var
iables that could control for attrition-related sampling biases. These
longitudinal covariates and corrections to adjust degrees of freedom
for cases with repeated measures were then used to construct statistic
al models that were equivalents of pure cross-sectional designs.