Research Design in Clinical Psychology

Construct validity pertains to interpreting the basis for the causal relation between the independent variable and the dependent variable. The investigator may conclude that the experimental manipulation was responsible for group differences (internal validity was well handled), but the study may not permit a conclusion about why the effect occurred. Other factors embedded in the manipulation alone or in combination with the manipulation might account for the findings (construct validity is in question). Data-evaluation validity refers to those aspects of the study that affect the quantitative evaluation and can lead to misleading or false conclusions. Several concepts basic to statistical evaluation were mentioned and include the probability of accepting and rejecting the null hypothesis, the probability of making such decisions when they are false, and effect size. Major factors that commonly serve as threats to data-evaluation validity include low statistical power, subject heterogeneity, variability in the procedures of an investigation, unreliability of the measures, restricted range of the measure, and multiple statistical comparisons and their error rates. All four types of validity including internal, external, construct, and data-evaluation validity need to be considered at the design stage of an investigation.