Statistical Support - Articles

The following articles address various statistical topics of interest to SPSS users. To view, click on a title in the list below.

• My Coefficient Alpha is Negative! This article shows how negative reliability estimates can occur. It originally appeared in Keywords #68.
• Choosing an Intraclass Correlation Coefficient This article discusses the new intraclass correlation coefficient facilities in the SPSS RELIABILITY procedure, providing guidance on how to choose an ICC. It originally appeared in Keywords #67.
• My Tests Don’t Agree! This article discusses the sampling and inference issues underlying the failure of omnibus and single degree of freedom tests to yield the same logical conclusions on the same data. It originally appeared in Keywords #66.
• What Kind of Contrasts Are These? This article discusses the relationship between contrast matrices and design or basis matrices, with particular emphasis on the parameter codings section of the LOGISTIC REGRESSION and COXREG output. It originally appeared in Keywords #63.
• From MANOVA to GLM: Basics of Parameterization: The first in an ongoing series of articles designed to guide users in the transition from the older MANOVA procedure to the new GLM procedure. Begins discussion of how the procedures represent categorical factor variables and the implications for interpretation of resulting parameter estimates. It originally appeared in Keywords #64.
• From MANOVA to GLM: More Basics of Parameterization: Continues discussion of parameterization issues in MANOVA and GLM. It originally appeared in Keywords #65.
• Dichotomous Predictors in Linear Regression: Part 1 of a two part article that appeared in Keywords #56. This article explored the use of categorical predictor variables and how the estimated coefficients should be interpreted.
• Polytomous Predictors in Linear Regression: Part 2 of a two part article that appeared in Keywords #57 exploring the use of categorical predictor variables in regression.
• Interactions Among Dichotomous Predictors in Regression: Article discussing interpretation of models involving interactions among dichotomous predictors.
• Continuous by Categorical Interactions in Regression: Article that appeared in Keywords #61 exploring a regression model involving interaction between a continuous and a dichotomous predictor.
• Further Interactions with Categorical Variables in Regression: Article that appeared in Keywords #62, continuing discussion of interactions between a continuous and a dichotomous predictor.
• The Mantel-Haenszel Statistic for 2x2xK Tables: This article describes a method for obtaining the Mantel-Haenszel chi-squared statistic for 2x2xk contingency tables. The macro produces this estimator of the common odds ratio across k strata. The article originally appeared in Keywords #54.
• Interpreting MANOVA Parameter Estimates: This article, that originally appeared in Keywords #50, discusses the interpretation of the default deviation parameter estimates for a 2-way factorial between subject design.
• Testing Simple Effects in MANOVA: Part 2 of a three part series that appeared in Keywords #51. This article describes the use the the SPSS MANOVA procedure to determine what statements may be made about main effects or interactions in the presence of the two-way or higher order interaction effects.
• Nonorthogonal Contrasts on WSFACTORS in MANOVA: Part 3 of a three part series that appeared in Keywords #52. This article describes the use of the SPSS MANOVA procedure to produce nonorthogonal contrasts in repeated measures or within-subjects designs.
• My Sums of Squares Don't Add Up! This article, that originally appeared in Keywords #53, explores the situation when predictor variables are correlated and there is no unique way to compute an additive decomposition of the model sum of squares into individual parts.
• Data Setup for Comparing Means in SPSS: This article discusses the general requirements necessary for examining differences in means using SPSS.