What are the four standard scale types and what are each used for (most often)?
What are the four standard scale types and what are each used for (most often)?
1. Likert scales have sets of agree/disagree items (usually an odd number of options with a neutral center); a summary (e.g., the mean) of all items is the subject's score. Very reliable, so good for small effects or small samples.
2. Guttman scales have sets of ascending items for (yes/no) agreement; the score for the subject is the highest item that he or she agreed to. Adaptive in a way to reduce demand for "average" behavior, so good for situations with strong "norms" for behavior or "touchy" subjects where continuing to ask questions after the subject says "no" would be bad.
3. Thurstone scales are sets of items to (yes/no) check off; the items are pre-rated for value and the subject's score is the sum of values. These often have the best construct validity, so good for situations where a fine distinction (and, therefore, discriminant validity) is needed.
4. Semantic differentials are sets of pairs of opposites where the subject indicates where they are between extremes on each pair. These are much harder to analyze, but work best for overlapping constructs.
Comments
Post a Comment