To some members of allstat,
Maybe that many of you have a quick answer to the question whether using
ANOVA when the dependent variable is ordinal (such was the example of Ron
Hassad) is inappropriate. It may be informative to know that there is
controversy on this topic. May I suggest the interested reader to study
Davison & Sharma, 1994, ANOVA & ANCOVA of Pre- & Post-Test, Ordinal Data,
Psychometrika 59: 593-600, Gaito & Yokubinas, 1986, An empirical basis for
the statement that measurement scale properties are irrelevant in statistical
analysis, Bul. of Psych. Soc 6: 449-450. These authors defend the use of
ANOVA when the dependent variable is ordinal.
Cheers,
Jarl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|