Research Design in Clinical Psychology

This chapter presented information to supplement statistical significance testing. Among this is the inclusion of a measure of the strength or magnitude of the relation whenever significant tests are presented. Completer analyses, intent-to-treat, and models of imputation were discussed as ways of handling missing data in research where there are repeated measures and participants drop out before all measures are completed. Outliers in the data and deleting data were also discussed. Multiple comparison tests and controlling for error rates were presented as well. Many decision points in the data presentation and analyses can influence the findings. I discussed robustness of findings and encouraged evaluating the results in different ways to see if critical decisions influenced the results and conclusions. Also, the use exploratory data analyses was encouraged to examine a variety of relations and special findings that might generate hypotheses for future research. Meta-analysis was also discussed. This now is very commonly used as a way of combining studies to summarize a literature and to ask novel questions that usually go beyond what any single study has examined. Finally, secondary data analyses were discussed. This is conducting studies based on data that other people have collected. There are rich databases that provide special opportunities for further evaluation beyond the original goals of the study.