Chapter 17 Making Sense of Advanced Statistical Procedures in Research Articles
Brief Review of Multiple Regression Predicting scores on a criterion variable from two or more predictor variables Proportion of variance accounted for ( R 2 )
Hierarchical and Stepwise Multiple Regression Hierarchical multiple regression –Examine contribution to the prediction of each variable added in a sequential fashion Stepwise Multiple regression –Controversial exploratory procedure –Predictor variable with best prediction located –Find next predictor variable that gives highest R 2 with first predictor variable –Repeat until best predictor variable does not give significant improvement
Hierarchical and Stepwise Multiple Regression Both involve adding variables a stage at a time and checking for significant improvement of prediction Theory/plan determines order of variables in hierarchical regression No initial plan in stepwise regression –Useful in exploratory and applied research
Partial Correlation Association between two variables, over and above influence of one or more other variables Holding constant, partialing out, controlling for, adjusting for Partial correlation coefficient
Reliability Reliability –Test-retest reliability –Split-half reliability –Cronbach’s alpha (α) –Interrater reliability
Factor Analysis Measured large number of variables Identifies variables that clump together Factor Factor loading Several approaches to factor analysis Naming the factors
Causal Modeling Measured large number of variables Does the pattern of correlations match theory of which variables cause which? Path analysis –Path –Path coefficient
Causal Modeling Path analysis
Causal Modeling Structural equation modeling –Elaboration of path analysis –Fit index e.g., RMSEA –Latent variable –Measured variable
Causal Modeling Structural equation modeling
Causal Modeling Structural equation modeling
Causal Modeling Limitations –Other patterns of causality possible –Alternative theories –Correlation and causality –Linear relationships –Restriction in range
Independent and Dependent Variables Independent variable –Predictor variable Dependent variable –Criterion variable
Analysis of Covariance (ANCOVA) ANOVA adjusting the dependent variable for effect of additional variables Analogous to partial correlation Covariate Adjusted means
Multivariate Analysis of Variance (MANOVA) and Covariance (MANCOVA) Multivariate statistics –More than one dependent variable Multivariate analysis of variance (MANOVA) –ANOVA with more than one dependent variable –Univariate ANOVA
Multivariate Analysis of Variance (MANOVA) and Covariance (MANCOVA) Multivariate analysis of covariance (MANCOVA) –ANCOVA with more than one dependent variable –MANOVA with covariates
Overview of Statistical Techniques
Controversy: Should Statistics be Controversial? Fisher Neyman Pearson
Reading Results Using Unfamiliar Techniques Don’t panic! Look for a p level Look for pattern of results that is considered significant Look for degree of association or size of the difference Look up in statistics book Take more statistics courses!