Survey data from coerced group settings is virtually worthless as individual measurement.

Evidence

We shall leave the story at this point, but in case you care, the party leaders decided to let me give out my surveys, which I did to twenty-four members following an interesting demonstration on the Fourth of July across from the White House. I also administered a second questionnaire to forty-four Nazis on September 5, following another demonstration. At the time I was thrilled by the data. But over the years I have seen their virtual uselessness since the participants, sitting in their Stormtrooper uniforms in front of the party headquarters, filling out a psychological survey only because they had practically been ordered to do so by the party leaders, were hardly answering as individuals. — Bob Altemeyer, The Authoritarian Specter, 4301

Categorization

• XREF: Reinforces the book’s core methodological point that RWA scores require free individual responding, not in-group conformity pressure — echoes the chapter’s earlier discussion of measurement integrity.
• SEED: Article angle: when does field research on extremists produce data that is worse than no data at all? The difference between studying a group’s members and studying its coerced audience.