ERIC Number: ED312298
Record Type: RIE
Publication Date: 1989-Nov
Reference Count: N/A
Dangers in Using Analysis of Covariance Procedures.
Campbell, Kathleen T.
Problems associated with the use of analysis of covariance (ANCOVA) as a statistical control technique are explained. Three problems relate to the use of "OVA" methods (analysis of variance, analysis of covariance, multivariate analysis of variance, and multivariate analysis of covariance) in general. These are: (1) the wasting of information when intervally scaled independent variables are converted to the nominal level; (2) the distortion of distribution shapes of and relationships among the non-interval predictor variables; and (3) the reduction of power against Type II error. Three other problems are associated with the use of ANCOVA as a statistical control technique: the need for very reliable measurement of the control variables; the regard many researchers have for ANCOVA as an almost magical technique for equalizing dissimilar groups; and the fact that researchers frequently disregard the critical homogeneity of regression assumption. If the regression equations of the groups are not reasonably similar, the single regression equation calculated by ignoring group membership will result in underadjustment for the experimental group. (Author/SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Note: Paper presented at the Annual Meeting of the Mid-South Educational Research Association (17th, Louisville, KY, November 9-11, 1988).