On the Validity of Measures of Association: The Nominal-Nominal, Two-by-Two Case

Abstract

Despite recent in the area, the problem of the selection of appropriate measures of association continues to worsen. One reason for this could be that we have thus far failed to consider some of the most important criteria which they should meet. One criterion which has never been systematically applied is validity, that is, the degree to which they measure that which they are designed to measure. Using two validation procedures frequently employed in evaluating operational difinitions of psychological and social variables, nine diferent measures of association appropiarte to nominal-nominal, 2 X 2 tables were validated on a sample of mock tables. The simple percentage difference, T , and were the most valid measures when the criterion of intuitive judgment was used, while , , and were the most valid measures when the method of factorial validation was employed. These results would suggest, then, that one ought to use or in the nominal-nominal, 2 X 2 case, as far as the validity criterion is concerned.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 20 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.