Measurement in the Testing of Ideas
- A construct is simply an idea, concept, or theoretical abstraction, which has multiple referents that in psychology are known as measurable variables.
- To “play” with theoretical constructs, such as intelligence, happiness, and personality, requires objective instruments of measurement. No one measurement can fully capture a construct.
- A tool of measurement is judged by the extent to which it provides reliable (consistent) and valid (meaningful) scores.
- The correlation coefficient statistic is used to assess reliability and validity.
- A reliable instrument is internally consistent and stable over time, as reflected by high interitem and test–retest reliability coefficients.
- An instrument has convergent validity when it correlates with other measures of the same construct, and it has discriminant validity when it does not correlate with measures of a different construct.