Table 3.
Summary of results from MLM growth curve analyses
| Percent Days Abstinent | Drinks per Drinking Day | |||||
|---|---|---|---|---|---|---|
|
| ||||||
| b | SE | p-value | b | SE | p-value | |
| Intercept | 25.968 | 5.710 | <.001 | 2.180 | .094 | <.001 |
| Intercept X Change | 4.182 | 8.382 | .620 | .063 | .152 | .683 |
| Time | -3.945 | 3.283 | .230 | -.067 | .025 | .008 |
| Time X Change | 29.256 | 4.446 | <.001 | -.551 | .093 | <.001 |
| Time2 | 1.397 | .686 | .042 | .000 | .003 | .972 |
| Time2 X Change | -4.820 | .853 | <.001 | .051 | .011 | <.001 |
| Time3 | -.095 | .046 | .040 | .000 | .000 | .429 |
| Time3 X Change | .282 | .056 | <.001 | -.0013 | .0003 | <.001 |
| Time4 | .002 | .001 | .055 | -- | -- | -- |
| Time4 X Change | -.005 | .001 | <.001 | -- | -- | -- |
Note: Change is a level 2 moderator with 0 = gradual changers and 1 = rapid changers. Runtime level 1 n=990. Percent Days Abstinent was estimated with restricted maximum likelihood with robust standard errors and Drinks per Drinking Day with Poisson estimation with robust standard errors. To reduce the number of intervals and to fit the most parsimonious models, past history intervals were entered as a summary variable due to minimal changes in drinking over the 166 days; therefore, 23 time intervals were included in all final analyses. All time effects were estimated as fixed effects.