Table 1.
Model 1 (n = 4,027) |
Model 2 (n = 2,562) |
Model 3 (n = 2,609) |
Model 4 (n = 1,989) |
||||
Item | Neighborhood and school strength | Alcohol use | Home alcohol access | Protective family management | Deviant peer affiliations | Beliefs favorable to use | Alcohol use |
Neighborhood context | |||||||
Perceived neighborhood strength | .750 | ||||||
Neighborhood and police preventive action | .761 | ||||||
Perceived neighborhood problems | .309 | ||||||
School strength | .254 | ||||||
Home and family management | |||||||
Last time drank, received alcohol from parent | .187 | ||||||
Easy to get alcohol from parent | .641 | ||||||
Easy to get alcohol from home | .750 | ||||||
Parent ask about school | .591 | ||||||
Parent praise when do a good job | .549 | ||||||
Eat dinner with parent | .455 | ||||||
Parent ask who with | .453 | ||||||
Parent/child conversations | .577 | ||||||
Parent talk about problems alcohol can cause | .805 | ||||||
Parent talk about family rules against drinking | .728 | ||||||
Parent talk about consequences of drinking | .739 | ||||||
Parent talk about influence of ads/commercials | .603 | ||||||
Deviant peer affiliations | |||||||
Past month, how often friends asked to drink | .902 | ||||||
Past month, how often friends asked to get drunk | .839 | ||||||
Beliefs favorable to use | |||||||
Normative estimates and expectations | .807 | ||||||
Outcome expectancies | .593 | ||||||
Self efficacy | .535 | ||||||
Alcohol use | |||||||
Past-year alcohol use | .765 | .858 | |||||
Past-month alcohol use | .833 | .987 | |||||
Past-week alcohol use | .681 | .824 | |||||
Heavy episodic use | .697 | .709 | |||||
Ever drunk | .527 | .536 | |||||
Fit indices | |||||||
CFI | .945 | .925 | .995 | .954 | |||
TLI | .924 | .906 | .986 | .908 | |||
RMSEA | .019 | .064 | .037 | .101 | |||
SRMR | .021 | .039 | .015 | .034 |
Notes: Other neighborhood contextual items (exposure to alcohol advertisements, off-sale alcohol outlet density, commercial alcohol accessibility, and area deprivation) did not load sufficiently with the two identified factors or with each other in Model 1; therefore, each was included separately in structural models. CFI = comparative fit index; TLI = Tucker–Lewis index; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual.