Table 2.
Regression coefficients and significance for alcohol/drug group, twin contact, and group × contact interaction effects on child problem behavior, with regression equation: Y = b0 + b1(AD/DDa) + b2(AD/DDb) + b3(Twin Contact) + b4(Contact × AD/DDa) + b5(Contact × AD/DDb)
| Effect | Child outcome variable |
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Child AD symptoms |
Child ND symptoms |
Child CD symptoms |
Problem index |
|||||||||
| b (SE) | Beta | p | b(SE) | Beta | p | b(SE) | Beta | p | b(SE) | Beta | p | |
| MODEL 1 | ||||||||||||
| b1: AD/DD group 3 vs. 2a | −.21 (.14) | −.06 | .15 | −.21 (.15) | −.06 | .16 | −.34 (.14) | −.10 | .01 | −.76 (.33) | −.09 | .02 |
| b2: AD/DD group 4 vs. 2a | −.37 (.12) | −.12 | .002 | −.43 (.13) | −.13 | .001 | −.19 (.12) | −.06 | .11 | −1.0 (.28) | −.14 | < .001 |
| MODEL 2 | ||||||||||||
| b1: AD/DD group 3 vs. 2a | −.22 (.14) | −.06 | .12 | −.23 (.15) | −.06 | .13 | −.35 (.14) | −.10 | .01 | −.81 (.33) | −.09 | .01 |
| b2: AD/DD group 4 vs. 2a | −.37 (.12) | −.12 | .003 | −.43 (.13) | −.13 | .001 | −.18 (.12) | −.06 | .12 | −.99 (.28) | −.13 | < .001 |
| b3: Twin contact | −.05 (.05) | −.04 | .24 | −.07 (.05) | −.05 | .17 | −.06 (.04) | −.04 | .18 | −.18 (.10) | −.06 | .09 |
| MODEL 3 | ||||||||||||
| b1: AD/DD group 3 vs. 2a | −.20 (.14) | −.05 | .17 | −.22 (.15) | −.06 | .15 | −.34 (.14) | −.10 | .01 | −.76 (.33) | −.08 | .02 |
| b2: AD/DD group 4 vs. 2a | −.36 (.12) | −.11 | .004 | −.42 (.13) | −.13 | .001 | −.18 (.12) | −.06 | .14 | −.97 (.28) | −.13 | .001 |
| b3: Twin contactb | −.06 (.05) | −.04 | .21 | −.07 (.05) | −.05 | .16 | −.06 (.04) | −.05 | .17 | −.18 (.10) | −.06 | .08 |
| b4: Interaction 1c | .04 (.12) | .02 | .72 | −.02 (.13) | −.01 | .88 | .02 (.12) | .01 | .85 | .03 (.28) | .01 | .90 |
| b5: Interaction 2c | −.15 (.11) | −.07 | .18 | −.14 (.11) | −.06 | .24 | −.11 (.10) | −.06 | .29 | −.39 (.25) | −.08 | .11 |
AD alcohol dependence, DD drug dependence, ND nicotine dependence, CD conduct disorder. Child age, child gender, father age, father income, father education, mother’s AD symptoms, and missing mother data variables are also in the model
AD/DD group effects reported are for dummy coded variables, instead of weighted effect codes with Group 4 as the reference group, so that the Group 2 versus Group 3 effect size could be shown
The twin contact effect is a test of the weighted average of the contact slopes, from a regression equation with weighted effect codes for the alcohol/drug groups, instead of the dummy coded variables given in the table
Interaction 1 refers to the interaction between AD/DD group 3 versus 2 and Twin Contact, while Interaction 2 refers to the interaction between AD/DD group 4 versus 2 and Twin contact