Skip to main content
. 2017 May 19;3(2):e28. doi: 10.2196/publichealth.7149

Table 3.

Parameter estimates from mixed-effects models for change in sexual risk from T1 to T2.

Variables Unconditional model
Beta (SE)
Model 1: level 1
Beta (SE)
Model 2a: main effects model
Adjusted beta (SE)
Model 3a: interaction model
Adjusted beta (SE)
Fixed effect




Intercept (γ00) 0.584 (0.040) (P<.001) 0.432 (0.045) (P<.001) 0.801 (0.263) (P=.004) 0.758 (0.256) (P=.003)

Time (γ10)
0.303 (0.046) (P<.001) 0.256 (0.061) (P<.001) 0.252 (0.058) (P<.001)
More than 100 SMS per day

0.384 (0.080) (P<.001) 1.088 (0.300) (P<.001)
SMS to whom




Parents

−0.084 (0.074)

Boyfriend or girlfriend

0.160 (0.080) (P=.01) 0.148 (0.079)
Number of SNS accounts

0.018 (0.024)
Number of SNS activities performed

0.028 (0.024)
Parental monitoring

−0.140 (0.053) (P=.006) −0.077 (0.064)
100 SMS × parental monitoring


−0.237 (0.098) (P=.009)
Female

−0.219 (0.090) (P=.01) −0.218 (0.089) (P=.008)
Age

0.177 (0.046) (P=.001) 0.187 (0.046) (P<.001)
Random effect




Participant 0.771 (0.035) 0.785 (0.034) 0.729 (0.044) 0.746 (0.041)

Residual σ2 0.737 (0.022) 0.705 (0.021) 0.616 (0.034) 0.612 (0.032)
LL ratio −1547.279 −1523.1006 −813.11182 −853.34291

a Models 2 and 3 control for age, gender, survey language, and intervention type. Age and gender were statistically significant so these variables are presented in the tables. The other control variables were not statistically significant. SMS use and having an SNS account were omitted from the final model due to collinearity with high SMS and number of SNS accounts.