Skip to main content

Table 2 Collinearity Diagnostics

From: The convergence of racial and income disparities in health insurance coverage in the United States

  Model 1 Model 2 Model 3 Model 1 (without age) Model 2 (without age) Model 3 (without age)
Condition Indices 30.4 30.2 33.8 13.4 13.6 13.0
Variance Decomposition Proportion
 Intercept 0.9716 0.9735 0.9688 0.9883 0.9897 0.9866
 group 0.0062 0.0056 0.0094 0.2239 0.2374 0.2067
 age_3c 0.8713 0.8601 0.91
 female 0.0056 0.0052 0.0039 0.0435 0.0459 0.0439
 edu 0.2839 0.2939 0.2273 0.0044 0.0041 0.0034
 employed 0.1724 0.1798 0.1427 0.0558 0.0596 0.0517
 region 0.0518 0.0531 0.0355 0.4996 0.475 0.5258
 mentalhealth53 0.0351 0.042 0.0283 0.2761 0.2975 0.2684
 adlhelp31 0.0045 0.0062 0.0003 0.0003 0 0.0001
  1. Collinearity problems indicated when: Condition Indices (CI) > 30 AND 2 or more Variance Decomposition Proportions (VDPs) > 0.5. The original models have collinearity problems that is cause by age. After dropping age, our models did not have collinearity problems (CI < 30)
\