Skip to main content

Table 3 Results of linear regression predicting alcohol consumption in high school a (n = 1,100).

From: High school drinking mediates the relationship between parental monitoring and college drinking: A longitudinal analysis

  Bivariate Models   Multivariate Model c
  b SE t ( df ) sr 2 p   b SE t ( df ) sr 2 p
Parental Monitoring Score -.13 .01 -10.49 (1,194) .08 <.0001   -.12 .01 -9.26 (1,092) .07 <.0001
Sex [Reference = Female] .97 .16 6.14 (1,249) .03 <.0001   .69 .16 4.32 (1,092) .01 <.0001
Race [Reference = Non-White] 1.37 .17 7.91 (1,246) .05 <.0001   1.26 .18 7.02 (1,092) .04 <.0001
Religiosityb [Reference = Slightly/Not Important] -.56 .16 -3.46 (1,243) .01 .0006   -.10 .16 -.59 (1,092) <.01 .56
R 2        .14
F (df, df) p        24.82 (7, 1,092) p < .0001
  1. Effects were evaluated using the null hypothesis test of b = 0 (tested as: b/SE) which evaluates the unique contribution of a variable in a regression equation.
  2. a High school alcohol consumption was defined as the typical number of drinks per drinking day during the past year at the screener.
  3. b Religiosity was dichotomized into a binary variable (i.e., extremely/moderately vs. slightly/not).
  4. c As a proxy for socioeconomic status, the effect of mother's education was held constant in the multivariate model. Effect size (sr2) for each explanatory variable was as follows: parental monitoring score (.07), sex (.01), race (.04), religiosity (<.01), mother's education (<.01).