| Abstract: | In practice, analysts must deal with the limitations of Likert-scale scores, one of which is nonreversibility. These practical limitations should not obscure the fact that the parameters of the underlying population distributions are the essential objects of interest. These were the foci of the investigation by Gregoire and Driver (1987), because the pertinent issue concerns the extent to which transformed data are informative of actual, albeit unknowable and unmeasurable, population structures. (PsycINFO Database Record (c) 2010 APA, all rights reserved) |
