A questionnaire’s items agree: that does not settle what its score measures

Find the scale the number describes

A report can contain several questionnaires or several subscales within one questionnaire. Locate which items produced the stated statistic and whether the authors intended them to form one score. A coefficient for one subscale cannot automatically validate all the other questions in the same survey or a shortened app version with different wording.

COSMIN distinguishes item interrelatedness from other measurement properties. That limited definition is used here, without supplying a scoring threshold or a formal instrument appraisal. The reading task concerns what the reported statistic addresses. It does not recommend a questionnaire, authorize a diagnosis or establish that No Vape uses a validated research scale.

COSMIN Initiative: COSMIN Manual Version2.0: measurement-property definitions

Use an original wrong-target thought experiment

Imagine an invented questionnaire whose items all ask closely related questions about whether an interface is easy to navigate. They might produce related answers. Calling that score quitting success would still change the target: agreement among usability items does not make their content an abstinence measure. This imaginary exercise contains no actual questionnaire or numerical coefficient.

Write the intended concept in one box, the item content in another and the property tested in a third. Then ask whether the claim moves between those boxes. If the content or scoring structure is unavailable, keep the gap visible rather than assuming an impressive coefficient settles it. You need not obtain participants’ private responses to identify a mismatch in what a public claim says.

Do not exchange one measurement property for another

Consistency among items at one administration is different from similar scores on repeated occasions. Both differ from evidence about what the score actually represents. A paper may investigate several properties, so attach each finding to its actual question. Do not infer all the others from whichever result appears first in the abstract.

If the report discusses the scale’s dimensional structure, retain that explanation alongside the item statistic. A broad overall coefficient does not by itself explain whether a collection of different concepts should be treated as one score. This guide does not ask you to fit a model or calculate a replacement coefficient; unresolved technical questions can remain questions for an appropriately qualified reader.

Repeat a narrow measurement statement

Your note can identify the questionnaire version, subscale, tested property and relevant population, leaving clinical meaning and intervention benefit on separate lines. Avoid calling the instrument useless merely because one statistic is insufficient for a broader claim. The appropriate response is to find the additional evidence needed for the claim, or refrain from repeating it.

For personal health decisions, consult a qualified professional about the actual measure and context. A research questionnaire is also different from your own informal trigger labels. You can organize everyday observations without awarding your personal note a validity certificate. The useful output is a correctly scoped measurement statement, with the intended concept still attached to the items that were actually examined.