Exploring Inter-Variable Relationships

Banking institutions can employ the Distance Correlation procedure available in IBM® SPSS® Statistics to assess the dependencies among the selected variables.

Go to Analyze > Correlate > Distance Correlation.

In the main dialog, select the following four variables for analysis:
  • Age in years
  • Years with current employer
  • Household income (in thousands)
  • Debt-to-income ratio (×100)
These variables are chosen due to their relevance in evaluating customer financial stability and loan repayment potential.
Figure 1. Choosing variables

Criteria

Within the Criteria subdialog, the following analytical options are specified:
Normalization Method: Min-Max Scaling
This normalization technic rescales each variable to the [0, 1] interval by subtracting the minimum value and dividing by the range. This approach can ensure that all variables contribute proportionally to the distance calculations, regardless of their original scales or units.
Confidence Interval Percentage: 95% (Default)
Select a 95% confidence level for the estimation of confidence intervals around the distance correlation coefficients. This setting allows for interval-based inference, indicating the range within which the true population distance correlation is expected to lie with 95% confidence.

Print

In the Print subdialog, select the following options.
Print Details
Shows all configuration settings in the output, including variables selected, normalization method, and test parameters.
Distance Correlation Coefficients
Displays pairwise distance correlation values, indicating the strength of dependence (linear or nonlinear) between variables.
Distance Covariance Estimates
Reports the pairwise distance covariance values, which quantify the magnitude of joint variability.
Distance Metrics
Details the computed pairwise distances for each variable, useful for diagnostics and interpretation.
Significance Estimates
Provides p-values associated with each distance correlation coefficient, helping assess statistical significance.
These outputs collectively facilitate both numerical and inferential interpretation of inter-variable dependencies.

Plot

In the Plot subdialog, set a bivariate scatter plot with the following axes.
  • X-axis: Age in years
  • Y-axis: Debt-to-income ratio (×100)
This plot visually illustrates the nature of the association between a respondent's age and their debt burden relative to income. It allows for immediate inspection of whether the relationship exhibits linear, curvilinear, or more complex nonobvious patterns thereby complementing the formal numerical results.