peopleanalyst

how-to · evidence-based answer

How do I write survey items for a new scale?

The short answer

Write items the way psychometricians do: one construct per scale, one idea per item, plain language at the respondents' reading level, no double-barreled or leading wording, and a response format chosen deliberately. Then treat the draft as a hypothesis — pilot it, check reliability and factor structure, and revise. A scale drafted in a meeting and never validated measures the meeting, not the construct.

The problem underneath

Item writing is a psychometric craft with known failure modes — double-barreled items, acquiescence bias, reversed-item artifacts — yet most organizational surveys are drafted in a meeting and never validated.

The evidence

Every claim on this site traces to a graded source — see the proof graph.

Go deeper

Related questions

scale developmentitem writingLikert scalesurvey designpsychometrics

← All answers