Trend tests on the same sets of data having unequally spaced intervals of the independent variable were conducted under conditions of adjustment and non-adjustment of the orthogonal coefficients for linear and quadratic regression. The major result was an inflated F value for the linear component and potentially spurious values for the quadratic component. It was recommended that authors explicitly cite procedures when adjustments of orthogonal coefficients have been made in trend analyses for unevenly spaced levels of the independent variable.
Get full access to this article
View all access options for this article.
References
1.
EdwardsA. L.Experimental design in psychological research. New York: Holt, Rinehart, & Winston, 1960.
2.
GaitoJ.Unequal intervals and unequal n in trend analyses. Psychological Bulletin, 1965, 63, 125–127.
3.
GrandageA.Orthogonal coefficients for unequal intervals. Biometrics, 1958, 14, 287–289.
4.
GrantD. A.Analysis-of-variance tests in the analysis and comparison of curves. Psychological Bulletin, 1956, 53, 141–154.
5.
HaysW. L.Statistics for psychologists. New York: Holt, Rinehart, & Winston, 1963.
6.
MintzJ.A correlational method for the investigation of systematic trends in serial data. Educational and Psychological Measurement, 1970, 30, 575–578.
7.
MyersJ. L.Fundamentals of experimental design. Boston: Allyn & Bacon, 1966.
8.
RobsonD. S.A simple method for constructing orthogonal polynomials when the independent variable is unequally spaced. Biometrics, 1959, 15, 187–191.
9.
RoeckeleinJ. E.Magnitude estimations of filled time as a function of the statistical properties of tone sequences. Unpublished doctoral dissertation, Arizona State Univer., 1971.
10.
VaughanG.CorballisM.Beyond tests of significance: Estimating strength of effects in selected ANOVA designs. Psychological Bulletin, 1969, 72, 204–213.
