Measurement instrument choice drastically changes forgiveness findings

Evidence

Using the Enright measure of forgiveness, the correlations with our measures were substantially weaker than in the previous survey. Importance of religion was uncorrelated with forgiveness. For Catholics, forgiveness was positively associated with more contact with the outgroup and outgroup attitudes, as in the previous study, and with the two new measures of perspective-taking and empathy. For Protestants, however, there were no significant correlates of forgiveness. We again regressed forgiveness on all the predictors. The explained variance was low for Catholics (R2 = .18) and negligible for Protestants (R2 = .05). — Nyla R. Branscombe and Bertjan Doosje (eds.), Collective Guilt, 1375

Categorization

• XREF: Connects to the measurement-validity concern about context-specific vs generic scales, echoing earlier chapters distinguishing salience from stable traits in identification measures. — connects to: Mere membership in a perpetrating society can cause shameful identity without personal guilt — Nyla R. Branscombe and Bertjan Doosje (eds.), Collective Guilt: International…, 2078; Robust causal relationships would survive differences in measurement choices; Inconsistent research findings often signal underlying measurement error rather than real variation — Bob Altemeyer, The Authoritarian Specter, 2057 (+3 more)

Broader Connections

• Likert-scale measurement artifacts may explain away the null individualism-collectivism findings #exemplifies (0.95) — Roderick M. Kramer, Geoffrey J. Leonardelli, Robert W. Livingston, Social Cognition, Social Identity, and…, 1603
• Robust causal relationships would survive differences in measurement choices #corroborates (0.85)
• Contradictory findings signal weak effects rather than mere measurement differences #contradicts (0.85) — John T. Jost and Jim Sidanius, Political Psychology, 4344
• Low internal consistency makes a scale’s findings uninterpretable #expands (0.75) — Bob Altemeyer, The Authoritarian Specter, 264
• Inconsistent research findings often signal underlying measurement error rather than real variation #qualifies (0.72) — Bob Altemeyer, The Authoritarian Specter, 2057