Items rare in the population can load negatively and distort a discrimination scale

Evidence

In the analysis of the Interpersonal scale, one item (“You told someone you are Muslim and afterwards they were nicer to you.”) was dropped because of inconsistent results, bringing the total number of items in this scale to nine. These inconsistent results involved the item loading being negative, despite the majority of participants responding that this has “never happened” = 5. — Naheed Ahmed, Measurement of Perceived Interpersonal…, 71

Categorization

• SEED: Angle: why measuring ‘positive’ experiences of minority identity is psychometrically treacherous — floor effects, social desirability, and valence mismatch.
• QUESTION: Would a positively framed item work with a different response scale, or does anti-Muslim discrimination simply have no positive pole to measure against?