Nonorthogonal Analysis of Variance: Putting the Question before the Answer

Abstract

There has been considerable debate in the literature concerning the "appropriate" method for calculating least squares analysis of variance with unequal sample sizes. It is argued here that the choice of the appropriate method depends upon the question the experimenter wants to answer about the data. The different questions reflect different null hypotheses. An example is presented to demonstrate the practical effects of using two alternative methods of analysis (weighted and unweighted means) and the null hypotheses tested under each method. It is shown that the different methods address substantially different experimental questions. Procedures advocated by Appelbaum and Cramer (1974) are challenged on several grounds. An argument is also presented against use of Overall and Spiegel's Method II, based on the general uninterpretability of the hypotheses tested.

References

Appelbaum, M. I. and Cramer, E. M. Some problems in the nonorthogonal analysis of variance. Psychological Bulletin, 1974, 81, 335-343.

Appelbaum, M. I. and Cramer, E. M. Balancing-analysis of variance by another name. Journal of Educational Statistics, 1976, 1, 233-252.

Blair, R. C. and Higgins, J. J.

Comments on "contrast coding in least squares regression analysis."

American Educational Research Journal, 1978, 15, 149-151. (a)

Blair, R. C. and Higgins, J. J. Tests of hypotheses for unbalanced factorial designs under various regression/coding method combinations. Educational and Psychological Measurement, 1978, 38, 621-631. (b)

Carlson, J. E. and Timm, N. H. Analysis of nonorthogonal fixed-effects designs. Psychological Bulletin, 1974, 81, 563-570.

Cohen, J. The statistical power of abnormal-social psychological research . Journal of Abnormal and Social Psychology, 1962, 65, 145-153.

Cohen, J. Multiple regression as a general data-analytic system . Psychological Bulletin, 1968, 70, 426-443.

Herr, D. G. and Gaebelein, J. Nonorthogonal two way analysis of variance. Psychological Bulletin, 1978, 85, 1, 207-216.

Hocking, R. R. and Speed, F. M. Letter to the editor. The American Statistician, 1976, 30, 206.

10.

Lewis, C. and Keren, G. You can't have your cake and eat it too: Some considerations of the error term. Psychological Bulletin, 1977, 84, 1150-1154.

11.

Lord, F. M. On the statistical treatment of football numbers . American Psychologist, 1953, 8, 750-751.

12.

Nie, N. H. , Hull, C. H. , Jenkins, J. G. , Steinbrenner, K. , and Bent, D. H. SPSS: Statistical Package for the Social Sciences (2nd ed.). New York: McGraw-Hill, Inc., 1975.

13.

O'Brien, R. G.

Comment on "Some problems in the nonorthogonal analysis of variance."

Psychological Bulletin, 1976, 83, 72-74.

14.

Overall, J. E. and Spiegel, D. K. Concerning least squares analysis of experimental data. Psychological Bulletin, 1969, 72, 311-322.

15.

Overall, J. E. , Spiegel, D. K. , and Cohen, J. Equivalence of orthogonal and nonorthogonal analysis of variance. Psychological Bulletin, 1975, 82, 182-186.

16.

Scheffe, H. The analysis of variance. New York: John Wiley & Sons, Inc., 1959.

17.

Speed, F. M. and Hocking, R. R. The use of the R()-notation with unbalanced data. The American Statistician, 1976, 30, 30-33.

18.

Winer, B. J. Statistical principles in experimental design (2nd ed.). New York: McGraw-Hill, Inc., 1971.