SOCIOL 5 Lecture Notes - Lecture 7: Convergent Validity, Discriminant Validity, Predictive Validity
Document Summary
Vertically: classification, rank order, equal intervals, nonarbitrary zero. If we measured the same thing over and over do we get the same answer. Validity and reliability do not always hang together. Face validity: does your measure resemble your idea of the concept. Predictive validity: does our measure correlate with behavior. Construct validity: do your measures capture your theoretical constructs. Degree to which inferences we have ade from our study match the concepts underlying our research in the first place. Two main components: convergent validity and discriminant validity. If a measure is reliable it does not mean it is valid. Degree to which a measurement instrument (a survey question or a series of survey questions, a direct behavior observation. Reliable measure would be narrowed down and unreliable measures more broad and general. Inter-rater: 2 observers watching the same behavior, their scores should agree with each other. Internal consistency: within a test people should respond consistently to all questions.