Answers are grounded in this paper and cite their sources.
Analysts of social science data often need the flexibility and power of intervallevel statistics, even though strictly interval measurement has not been achieved. In such cases, analysts should attempt to assign category values that minimize distortion of the underlying variables. This paper describes a number of methods for category value estimation. The methods include estimation from the observed frequencies based on an assumption about the underlying distribution, estimation from single or multiple criterion variables, and estimation from item text. These techniques can be easily used, and their sensible application can avoid needless measurement error and resulting statistical bias.
No takes yet. Share an insight, caveat, or question.
Hensler et al. (1979) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: