Is it a concern or a preference? An investigation into the ability of privacy scales...
summary
400 Prolific US participants each rate statements (each statement seen by ≥40 people) drawn from common privacy scales and 39 newly written items, and results show most statements do not uniquely map to one construct; non‑negative matrix factorization reduces six theorized constructs into five overlapping meta‑constructs and reveals statements often mix concern, preference, attitude, decision, and behavior. Corpus linguistics identifies lexical cues (verbs/modals for preferences, negative expressions for concern), and newly written items achieve about an 85% match to intended constructs.
AI-written summary · about