Unidimensional attitude scales allow reliable cross-scale inference from single survey items

Evidence

In fact, with three exceptions all of the items on all four of these scales correlated significantly with the summed scores of the other three scales, in both samples. That’s a 489 out of 492, or .994, batting average. You can look at any item on any of those scales, and almost always correctly infer that people who answered it in a certain way also tended to be more authoritarian, fundamentalist, dogmatic, or zealous. You might make the same kind of inferences when interpreting other instruments’ relationships with things, but it will often be wrong even with very large samples (RWA, pp. 205-208). Relatively unidimensional scales bring you this kind of consistent performance and ease of interpretation on a silver platter. — Bob Altemeyer, The Authoritarian Specter, 4276

Categorization

• SEED: Angle: the extraordinary interpretability of narrow, unidimensional scales versus the misleading inferences drawn from broad heterogeneous instruments.