Strongly worded items measure no better than mild ones.

Evidence

The modifications produced no change in the mean interitem correlations of the two versions, which was . in each case. Nor did the correlation with RWA scores differ (. and .64). Thus I was wrong. Extreme items did not interconnect better than milder statements. But neither does strong wording seem to hurt much. — Bob Altemeyer, The Authoritarian Specter, 828 (#corroborates, via linked claim)

Since the items “naturally selected” themselves, from the lot I supplied for “choosing up,” I do not know. I suspect that strongly phrased statements are more definite (less ambiguous), which reduces their measurement error and gives them a better chance to intercorrelate and make the team. Contraits in particular might prosper if sharply phrased, because it appears you have to whack High RWAs some (cognitively speaking) to get them to notice they disagree. — Bob Altemeyer, The Authoritarian Specter, 822 (#complicates, via linked claim)

Categorization

• XREF: connects to: Strongly worded survey items correlate no better than mild ones — Bob Altemeyer, The Authoritarian Specter, 828; Strongly worded statements measure more reliably than mild ones.; Extreme wording makes scale items more definite and less ambiguous — Bob Altemeyer, The Authoritarian Specter, 822 (+1 more)

Broader Connections

• Strongly worded survey items correlate no better than mild ones #corroborates (1.00) — Bob Altemeyer, The Authoritarian Specter, 828
• Strongly worded statements measure more reliably than mild ones. #contradicts (1.00)
• Extreme wording makes scale items more definite and less ambiguous #complicates (0.78) — Bob Altemeyer, The Authoritarian Specter, 822
• Vague evaluative wording makes survey items unreliable across respondents #complicates (0.75) — Naheed Ahmed, Measurement of Perceived Interpersonal…, 50