A shared psychological trait connects RWA Scale items despite diverse content

Evidence

Responses to the RWA Scale reflect many different things, because each item talks about something in particular and because there are many sources of measurement error. But there is a psychological connection among the test’s items that accounts for about 30% of the variance the whole bunch produces. When we try to pull that 30% apart, sometimes we cannot because of the “‘strong force” (the cohesion among the items on this test is too strong). But when we can pull the common variance apart, the best we can do is get the protrait and contrait items into largely separate piles. These piles show that the direction of the items’ wording also constitutes a force to be reckoned with. But the correlation between these two factors indicates that the RWA Scale is essentially unidimensional. — Bob Altemeyer, The Authoritarian Specter, 691

Categorization

• XREF: Factor analysis and wording-direction method effects echo psychometrics discussions from the broader research-methods literature the reader curates.
• DEFINE: Explains what item-wording direction (protrait vs contrait) does to factor structure and why the scale still measures one trait.