Alternative data analysis in evaluation research: A comparison between ordinal and traditional statistical approaches
Rudd, A. and McConney, A. (2000) Alternative data analysis in evaluation research: A comparison between ordinal and traditional statistical approaches. In: EVALUATION 2000: The Annual Meeting of the American Evaluation Association, 1 - 4 November 2000, Honolulu, Hawaii.
Frequently, evaluators are charged with evaluating a program that claims its program participants will score higher over time on a given variable, or score higher than a similar group that does not receive the program. These types of hypotheses or questions can be characterized as having an ordinal nature as can data from a variety of commonly selected assessment tools such as standardized tests, questionnaires, surveys, and attitude scales. Despite the ordinal nature of many program hypotheses or program data, evaluators typically choose traditional statistical approaches to data analysis that could produce spurious conclusions. Our purpose is to examine potential differences in outcomes when using ordinal analysis versus traditional statistical analysis. We suggest that an ordinal analysis may more accurately answer and analyze ordinal situations that are common in evaluation research.
|Publication Type:||Conference Paper|
|Item Control Page|