Table 3.
Comparison of R2 Values for Models Testing Four Interaction Terms for Each Domaina
| Model | CM domain | Base R2 | Interaction R2 | p Value |
|---|---|---|---|---|
| Male | Infectious disease | .3379 | .3392 | .3637 |
| Time 1 | Chronic disease | .3379 | .3385 | .6061 |
| Impairment | .3379 | .3385 | .4308 | |
| SES | .3379 | .3386 | .3579 | |
| Parental behavior | .3379 | .3390 | .1641 | |
| Male | Infectious disease | .6518 | .6522 | .1642 |
| Time 2 | Chronic disease | .6518 | .6523 | .3275 |
| Impairment | .6518 | .6519 | .2593 | |
| SES | .6518 | .6521 | .4846 | |
| Parental behavior | .6518 | .6520 | .6617 | |
| Female | Infectious disease | .3148 | .3155 | .1940 |
| Time 1 | Chronic disease | .3148 | .3156 | .2384 |
| Impairment | .3148 | .3153 | .6356 | |
| SES | .3148 | .3154 | .6205 | |
| Parental behavior | .3148 | .3156 | .0573 | |
| Female | Infectious disease | .6453 | .6455 | .7395 |
| Time 2 | Chronic disease | .6453 | .6457 | .1701 |
| Impairment | .6453 | .6456 | .0204 | |
| SES | .6453 | .6454 | .9617 | |
| Parental behavior | .6453 | .6456 | .6461 |
Note: CM = childhood misfortune; SES = socioeconomic status.
aFour product terms were tested for each model (Black × CM1, Black × CM2+, Hispanic × CM1, Hispanic × CM2+). Base R2 is the corrected R2 from the fully adjusted M3. Each group of four product term interactions was entered into the adjusted model and tested for joint significance by assessing whether the set of interactions significantly increased the R2 compared with the base model R2. Due to the high number of related hypothesis tests being conducted, an adjusted significance cutoff (0.00256) was obtained using a step-down procedure to control for the family-wise error rate (Benjamini & Liu, 1999; multproc command in Stata 14 using method=liu1). None of the product term interaction groups were significant.