Skip to main content
. Author manuscript; available in PMC: 2019 Aug 1.
Published in final edited form as: Int J Eat Disord. 2018 May 7;51(8):870–878. doi: 10.1002/eat.22879

Table 2.

Comparison of random coefficient models in the full sample (N = 160) to characterize the trajectory of desired weight percentage, weight difference percentage, eating disorder severity, and BMI from admission to 12-months post-discharge.

Desired Weight Percentage
Linear Quadratic
Model AIC AIC
Random Intercept 3674.07 3658.52
Random Slope 3911.02 3963.68
Random Intercept, Random Slope 3910.69 3963.68
Weight Difference Percentage
Linear Quadratic
Model AIC AIC
Random Intercept 3918.14 3895.99
Random Slope 4074.88 4076.68
Random Intercept, Random Slope 4074.79 4076.68
Eating Disorder Severity (EDE Global)
Linear Quadratic
Model AIC AIC
Random Intercept 2047.23 2046.81
Random Slope 2475.20 2540.90
Random Intercept, Random Slope 2474.64 2541.12
Body Mass Index
Linear Quadratic
Model AIC AIC
Random Intercept 2822.14 2768.58
Random Slope 2903.10 2944.02
Random Intercept, Random Slope 2902.74 2944.02

Notes. BMI = body mass index; AIC = Akaike Information Criterion (smaller AIC values indicate better fit to the data). Bold values indicate the best fitting model. The difference in AIC values for the linear and quadratic random intercept model for EDE global score was minimal (2047.23 vs. 2046.81), and the quadratic term was not statistically significant. Thus, the linear model was determined to be a better fit for EDE global score.