Testing between-family associations in within-family comparisons

Using behaviourally discordant siblings to test for gene-behavior associati ons is a common tool in molecular genetics, because the within-family contr ast offers a research design that avoids confounds inevitable in all betwee n-family comparisons of unrelated individuals. We propose a similar strateg y to assess the behaviour-behaviour associations on which much of psycholog ical science is built. Between-family correlations of personality test scor es (e.g., sensation seeking) and behavioural outcomes (e.g., substance use) may be mediated by variables that differ between families (e.g., social cl ass or religiosity) and correlate with both personality and outcome. Contra sting twin and nontwin siblings who were highly discordant for behavioral c orrelates of substance use, we tested whether between-family behavioral cor relations replicated within families. Some, but not all, did. Within-family analysis of behaviorally discordant siblings may find wide application in efforts to clarify the meaning of correlational research data.