Choose the Analysis > Explore relation of variable pair menu or the Analysis > Explore relation of variables menu, and choose the variables you want to analyze, then hit OK.

Differences between the relation of variable pair and the relation of variables menus:

  • In the relation of variable pair menu, if you choose more than two variables, all possible pairs will be explored. In a variable pair, the first one will be the predictor and the second one is the predicted variable. If you want to have more than one predictor variables, use the relation of variables menu.
    • If the two variables in a pair have different measurement level, then lower measurement level will be used for both variables. For example, for an interval-ordinal pair, CogStat handles them if both variables were ordinal.
  • In the relation of variables menu, you can choose one or more predictor variables. If you want to see the relation of the variable pairs of the chosen set of variables, use the relation of variable pair menu.

In the Display options... dialog, the minimum and maximum of the x and y axes can be set.

The analyses can be used for test-retest reliability analysis. See all the other reliability analyses that are supported by CogStat.

The following values and graphs will be displayed (see also the common elements of the results for more details):

Raw data

Results For interval data - one predictor For interval data - several predictors (New in v2.4) For ordinal data For nominal data
Sample size Number of observed and missing cases Number of observed and missing cases Number of observed and missing cases Number of observed and missing cases
Raw data Scatter plot of the raw data Scatter plot matrix of the predicted and predictor variables Scatter plot of the raw order data Mosaic plot

On scatter plots, the size of a dot is proportional with the number of cases with that value pair.

Sample properties

Results For interval data - one predictor For interval data - several predictors (New in v2.4) For ordinal data For nominal data
Descriptive data Equation of the linear regression line Equation of the linear regression line   Contingency table (case count)
Contingency table (percentage)
Graphs of the data Scatter plot with regression line Scatter plot with regression line
Partial regression plots with regression line		    
Standardized effect sizes Pearson and Spearman correlation coefficient R2, Log-likelihood, AIC, BIC, and Pearson’s partial correlations Spearman correlation coefficient Cramér’s V
Residual analysis (New in v2.3) Residual plot and histogram of residuals Residual plot and histogram of residuals    

On scatter plots the size of a dot is proportional with the number of cases with that value pair.

Population properties

Results For interval data - one predictor For interval data - several predictors (New in v2.4) For ordinal data For nominal data
Assumptions of inferential methods (New in v2.3) Henze-Zirkler test of multivariate normality

(New in v2.3) Koenker’s test and White’s test of homoscedasticity		 Henze-Zirkler test of multivariate normality

Koenker’s test and White’s test of homoscedasticity

Variance inflation factors (VIF) to test multicollinearity

If VIF is larger than 10 for a variable, then (a) betas of the other regressors predicting that variable, and (b) correlations of the predictors are also displayed.		    
Population parameters (New in v2.3) Slope and intercept point estimations with 95% confidence intervals Slopes and intercept point estimations with 95% confidence intervals   Contingency table (confidence interval, multinomial proportions with Goodman method)
Graphs of the population parameters (New in v2.3) Linear regression line with 95% confidence interval      
Standardized effect sizes Pearson and Spearman correlation coefficient, and their confidence intervals Adjusted R2, Pearson’s partial correlations, and their confidence intervals Spearman correlation coefficient and its confidence interval  
Hypothesis test and sensitivity power analyses p values for correlation coefficients

(New in v2.3) Bayesian hypothesis test for Pearson correlation		 Model F-test and test for the regressor slopes p value for correlation coefficient Chi-squared test and sensitivity power analyses