Analysis of Ordinal Data

by David K. Hildebrand, James D. Laing, Howard Rosenthal

Ordinal data can be rank ordered but not assumed to have equal distances between categories. Using support by judges for civil rights measures and bussing as the primary example, this paper indicates how such data can best be analyzed.

80 pages, Paperback

First published May 1, 1977

Reviews

[Steven Peterson · Rating 3 out of 5]

A useful addition to the Sage series, "Quantitative Applications in the Social Sciences." This slender volume explores statistical techniques for analyzing ordinal data. What are ordinal data? Measures whose categories can be ranked as greater or lesser--but the differences are not necessarily the same across categories. An answer to a question such as "low, medium, or high" would be an example. We can't say that the differences between each category are equal.

Now, I learned my statistics in an era that was "rough and ready." Ah, just assume that ordinal data can be handled by higher level interval data statistical techniques. Purists would blanch at that! Volumes such as this note the technically appropriate statistics to use with ordinal-level data.

Examples? Somer's d, Kendall's tau, Spearman's rho, gamma. . . .

The book tends to be a bit technical, and would turn off those not familiar with statistics. Still, a good resource for those who want a brief technical introduction to ordinal statistics.