Skip to main content
. Author manuscript; available in PMC: 2018 May 1.
Published in final edited form as: Oncol Nurs Forum. 2017 Nov 1;44(6):739–750. doi: 10.1188/17.ONF.739-750

TABLE 3.

Model Fit for Latent Growth Mixture Models

Modela Log Likelihood BIC Entropy Posterior Probability Class Proportions
Fatigue
Cycle 2
 1 class −4,309.41 8,651.7 NA NA NA
 2 class −3,954.18 7,958.98 0.944 0.988, 0.989 69%, 31%

 3 class −3,847.23 7,742.81 0.944 0.977, 0.95, 0.982 59%, 30%, 11%

 4 class −3,800.81 7,657.7 0.943 0.95, 0.96, 0.984, 0.935 57%, 12%, 22%, 10%
Cycle 3
 1 class −4,062.06 8,156.06 NA NA NA
 2 class −3,652.62 7,344.68 0.963 0.998, 0.968 72%, 28%

 3 class −3,570.33 7.187.62 0.941 0.986, 0.95, 0.96 64%, 25%, 11%

 4 class −3,538.36 7,131.18 0.881 0.935, 0.869, 1, 0.956 53%, 22%, 22%, 3%

Disturbed sleep
Cycle 2
 1 class −4,067.91 8,168.7 NA NA NA

 2 class −3,906.86 7,854.34 0.984 0.997, 0.995 89%, 11%

 3 class −3,855.66 7,759.66 0.958 0.945, 0.989, 0.968 84%, 10%, 5%
Cycle 3
 1 class −3,582.73 7,197.38 NA NA NA

 2 class −3,441.79 6,923.02 0.925 0.981, 0.98 81%, 19%

 3 class −3,385.73 6,818.42 0.962 0.987, 0.995, 0.969 82%, 14%, 5%
a

Classes were added stepwise to the model to determine the solution that best fit the data, evaluated on the BIC and entropy, and examined for the number of participants in each class and graphed visually.

BIC—Bayesian information criterion; NA—not applicable

Note. Shading indicates the selected model solution.