Psych 217 Measurements-Ch. 5 PDF

**[Quality of operational definition-ch.5 measurements]** The quality of an operational definition depends on validity and reliability. [Reliability:] Evaluation of **consistency** in an experiment's results. \*\*if an experiment is not reliable then validity is not even an issue (priority is re...

**[Quality of operational definition-ch.5 measurements]** The quality of an operational definition depends on validity and reliability. [Reliability:] Evaluation of **consistency** in an experiment's results. \*\*if an experiment is not reliable then validity is not even an issue (priority is reliability) \*\*reliabiltiy is r\~ + 0.80 How to evaluate reliability: a. [Test-retest:] Try testing and retesting the experiment to see if you receive the same results. b. [Inter-rater:] Having multiple observers complete the experiment and collect data to ensure all receive same results. This avoids any bias. c. [Internal consistency:] Similarity in results of two common/related constructs. (Ex: one who suffers depression reports low on energy and happiness) Construct = concept [True score:] [Error of measurement: ] Every measurement has an evaluated error that need to be included in calculations. Classical test theory: Phenomena in which a p's obtained score is a summation of their true score and error of measurement. Obtained score = True score + Error of Measurement [Validity:] Ensuring the accuracy of measurements. Are our measurements measuring what actually intended? Forms of validity: 1. Construct validity: concept accuracy 2. Internal validity: causality possible 3. External validity: generalization of findings beyond group and setting [Construct validity types:] A. Content 1. [Face:] does it look valid? 2. [Content:] does it capture all aspect of the measure? B. Behaviour 1. [Predictive:] does it predict future behaviour? 2. [Concurrent:] behave in a way expected according to construct. Theoretically relevant behaviours. C. Other constructs 3. [Convergent:] similar scores on relevant/other constructs 4. [Discriminant/divergent:] ensure not related to what it shouldn't be related to. (Not related to unrelated construct) In a risking questionnaire which are correct? 1. Predictive validity = predicts future, conceptually related theories Ex: over the next 2 years 2. Convergent validity = related to scores of similar constructs Ex: self-reported extroversion and self-confidence 3. Concurrent validity = able to distinguish between theoretically relevant behaviors Ex: measure of psychopathy 4. Discriminant validity = not related to what it shouldn't be related to. Ex: jump on bridge and scream at those who did it Concurrent validity= the ability to distinguish two conceptually related behaviours (not necessarily similar), not constructs. Discriminant validity= no correlation between two unrelated constructs. Psychopathy measures have no correlation with risk taking as they shouldn't!