Skip to main content
. Author manuscript; available in PMC: 2020 Jul 1.
Published in final edited form as: Prev Sci. 2019 Jul;20(5):753–764. doi: 10.1007/s11121-018-0966-6

Table 2.

Model Fit Indices

k=3 k=4 k=5

AIC BIC AIC BIC AIC BIC

Intercept-only for Will 5938.54 5983.09 5913.15* 5963.26* 5911.32 5967.00
Intercept-only for Int 3930.70* 3975.24* 3931.40 3981.51 3932.05 3987.73
Anx on Will 5831.64 5914.90 5807.00* 5901.36* 5804.60 5910.07
Anx on Int 3847.91* 3931.16* 3850.68 3945.03 3851.90 3957.35
Dep on Will 5944.46 6027.97 5919.63* 6014.29* 5913.74 6019.53
Dep on Int 3917.07* 4000.58* 3919.08 4013.73 3919.79 4025.57
PA on Will 5924.08 6007.59 5896.04* 5990.68* 5895.86 6001.64
PA on Int 3933.70 4017.21 3935.88* 4030.52* 3938.13 4043.90
PA* low Anx on Int 1566.05* 1637.28* 1564.16 1644.89 1558.82 1649.05
PA* high Anx on Int 2336.61* 2411.38* 2337.34 2421.50 2338.02 2432.73

Notes.: Will = willingness to drink; Int = intentions to drink; Anx = trait anxiety; Dep = depressive symptoms; PA = positive affect. Model fit indices are based on examination of Akaike information criterion (AIC) and Bayesian information criterion (BIC) for k = 3, 4, and 5 knots. Smaller values indicate better fit. Values with an asterisk (*) indicate best model fit (i.e., final chosen model).