*Statistical Analysis; Statistical Significance IDENTIFIERS Covariance Structural Analysis; Stepwise Regression

Dissertations are the cumulative, tangible "best evidence " of interests of doctoral faculty and students in serious and incisive scholarship. Thus, dissertations are thoroughly studied by the program review teams periodically hired by boards of higher education in most states. The present paper explores seven errors in quantitative analysis in both published literature and dissertations. They are associated with beliefs concerning statistical significance testing, multivariate statistics, chi-square analysis, covariance statistical corrections, stepwise analytic methods, the psychometric integrity of instrumentation, and presentation of statistically nonsensical or impossible results. The errors are described using references to works by other authors and small hypothetical data sets to illustrate problems. Concrete examples of the errors as they occur in dissertations are cited to make clear that the errors are not hypothetical. Ten dissertations, completed since January of 1985, are cited as examples, although pseudonyms are used. These common methodological errors in dissertations may reflect poor training in research. A 76-item list of references, 14 data tables, and three figures are provided. A bibliography of the 89 dissertations from which examples are taken are appended. (TJH) Reproductions supplied by EDRS are the best that can be made from the original document. disscrit.rno 10/26/8