The study of psychological measurement, including questionnaire construction, measurement reliability and construct validity. A score can be repeatable without supporting the intended interpretation; consistency and validity answer different questions.
Clustering at the upper limit of a measurement instrument can prevent it from distinguishing higher underlying levels. A bounded score need not exhibit a ceiling effect, and a sample maximum alone does not establish one. The issue concerns what variation the instrument can measure, rather than a proof that a treatment has no effect.
Consistency or precision of a psychological measure under specified conditions. Unreliable measures can weaken observed associations; high reliability alone does not establish construct validity.
A systematic tendency in responding that can alter a score independently of the target ability. Its interpretation depends on the task; a willingness to endorse questionnaire items is not automatically an accuracy difference.
A change score scaled by its expected measurement error. With equal marginal standard deviation , reliability and independent measurement errors at the two assessments, the difference has error standard deviation , givingA conventional normal-error criterion compares its magnitude with 1.96. The assumptions and instrument-specific reliability matter; reliable change is not itself diagnostic remission or proof that treatment caused the change.
Improvement exceeding a specified measurement-error criterion, with any additional no-deterioration rule made explicit. It differs from crossing a symptom cutoff and from improvement attributable to treatment.
A service outcome combining a reliable improvement criterion with movement from baseline caseness below the relevant clinical cutoffs. The instruments and denominator must be specified; this is not automatically remission established by diagnostic interview.
Covariation attributable to shared measurement procedures rather than solely to the intended traits. Item format, response style or a common informant can make measures correlate even when their substantive contents differ.
A measurement design applying multiple methods to multiple proposed traits. Cross-method agreement can support convergent validity, while comparison among traits supports discriminant validity. Trait and method must not be completely confounded in the design.
The extent to which a measurement model has the same relevant interpretation across groups or occasions. Equal factor patterns, loadings and intercepts or thresholds are distinct restrictions. Comparing latent means generally needs stronger evidence than merely fitting similar factor patterns.
Consistency or precision of measurements under stated sources of variation. Reliability depends on the population, conditions and intended score; it is not a permanent property of an instrument alone. Inter-rater reliability and test-retest reliability address different sources of variation. Reliability does not establish construct validity.
The degree of association among items or components of a measure under a specified model. High internal consistency can arise from overlapping content or common method variance; it does not by itself establish one substantive dimension or construct validity.
Consistency of scores obtained from the same units on different occasions, under a justified interval and conditions. A low value may reflect measurement error, genuine change or both; the intended stability of the construct matters.
Consistency or agreement among raters assessing the same units. The statistic should match the scale and whether consistency or absolute agreement matters. Repeated ratings improve information about a unit without creating additional independently sampled units.
A structured set of items used to obtain reports or judgments. Item wording, response options, respondent perspective and scoring determine what the resulting measurements mean. Content validity, measurement reliability and construct validity require evidence beyond a convenient scoring rule.
A seven-item anxiety symptom questionnaire used in routine outcome monitoring. It should not be assumed to capture every disorder-specific difficulty or replace a clinical assessment.[ref]
Feedback-informed treatment versus usual psychological treatment for depression and anxiety: a multisite, open-label, cluster randomised controlled trial by Jaime Delgadillo and others. Lancet Psychiatry 5, 564–572; 77 participating therapists, 2233 patients, cluster allocation and not-on-track outcomes
A nine-item depression symptom questionnaire. A score, a change score, a caseness threshold and a clinical diagnosis are distinct uses or interpretations.[ref]
Feedback-informed treatment versus usual psychological treatment for depression and anxiety: a multisite, open-label, cluster randomised controlled trial by Jaime Delgadillo and others. Lancet Psychiatry 5, 564–572; 77 participating therapists, 2233 patients, cluster allocation and not-on-track outcomes
A method for investigating how intended respondents understand an item and construct their answer, using prompts about their reasoning. It helps diagnose ambiguous terms, mismatched assumptions and unsuitable response options before large-scale data collection.
Articles by others on the same topic
Psychometrics is a field of study concerned with the theory and technique of psychological measurement. This includes the assessment of knowledge, abilities, attitudes, personality traits, and other psychological constructs. The goals of psychometrics include developing reliable and valid instruments for measuring these constructs, analyzing the data obtained from these instruments, and interpreting the results.