Link (http://rpubs.com/winterwang/ndnsd1) to the preliminary results.

1 Mixed effect LCA with carbohydrates intake groups redefined: (cutoff = 25%, 50%, 75%)

  1. Not eating;
  2. Eating low carbohydrate food (energy contribution less than 25%);
  3. Eating low-to-medium carbohydrate food (energy contribution between 25% and 50%);
  4. Eating medium-to-high carbohydrate food (energy contribution between 50% and 75%);
  5. Eating high carbohydrate food (energy contribution higher or equal to 75%).

1.1 Model comparison and selection

Model Comparison. (All data, n = 6155)
N of classes log-likelihood G^2 AIC BIC cAIC aBIC Entropy
1 -452486.6 412947.4 413139.4 413917.5 414013.5 413612.4 —
2 -449820.0 407614.2 408000.2 409564.6 409757.6 408951.3 0.46189054
3 -446750.6 401475.4 402055.4 404406.1 404696.1 403484.4 0.667683942
4 -446604.8 401183.7 401957.7 405094.6 405481.6 403864.8 0.621539464
5 -444384.7 396743.6 397711.6 401634.8 402118.8 400096.6 0.685273251
6 -443459.9 394893.9 396055.9 400765.4 401346.4 398919.0 0.660860479
7 -442430.7 392835.4 394191.4 399687.1 400365.1 397532.4 0.746801851
8 -441588.3 391150.6 392700.6 398982.6 399757.6 396519.7 0.706370885
9 -440937.6 389849.4 391593.4 398661.6 399533.6 395890.4 0.700263271
10 -440387.2 388748.6 390686.6 398541.0 399510.0 395461.6 0.704468767
Note:
Abbreviation: N, number; AIC, Akaike information criterion; BIC, Bayesian information criterion; aBIC, adjusted BIC; cAIC, consistent AIC; Entropy, a pseudo-r-squared index.

1.2 Visualisation of the 3-class model

1.3 Visualisation of the classes by Carbohydrate eating types (3 classes)

1.4 Visualisation of the 7-class model

1.5 Visualisation of the classes by Carbohydrate eating types (7 classes)

2 Mixed effect LCA with carbohydrates intake groups redefined: (cutoff = 50%)

  1. Not eating;
  2. Eating low carbohydrate food (energy contribution less than or equal to 25%);
  3. Eating low-to-medium carbohydrate food (energy contribution between < 50%);
  4. Eating medium-to-high carbohydrate food (energy contribution between > = 50%);
  5. Eating high carbohydrate food (energy contribution higher or equal to 75%).

2.1 Model comparison and selection

Model Comparison. (All data, n = 6155)
N of classes log-likelihood G^2 AIC BIC cAIC aBIC Entropy
1 -372017.3 264558.5 264654.5 265043.5 265091.5 264891.0 —
2 -369643.0 259809.9 260003.9 260790.1 260887.1 260481.9 0.445955616
3 -368727.6 257979.1 258271.1 259454.5 259600.5 258990.5 0.585828111
4 -366000.3 252524.5 252914.5 254495.1 254690.1 253875.4 0.67693015
5 -364998.3 250520.5 251008.5 252986.3 253230.3 252210.9 0.720780262
6 -363641.8 247807.4 248393.4 250768.4 251061.4 249837.2 0.701241468
7 -362787.7 246099.3 246783.3 249555.5 249897.5 248468.6 0.732843141
8 -362217.6 244959.1 245741.1 248910.4 249301.4 247667.8 0.739100667
9 -362177.3 244878.4 245758.4 249324.9 249764.9 247926.6 0.740787409
10 -361580.8 243685.5 244663.5 248627.2 249116.2 247073.2 0.696270112
Note:
Abbreviation: N, number; AIC, Akaike information criterion; BIC, Bayesian information criterion; aBIC, adjusted BIC; cAIC, consistent AIC; Entropy, a pseudo-r-squared index.

2.2 Visualisation of the 5-class model

2.3 Visualisation of the classes by Carbohydrate eating types (5 classes)

3 Some comments on changing the cut-off values

In the above analyses, cut-off values were 25%, 50%, 75% or simply 50% only. However, both analyses produced worse models compared with cut-offs (25%, 75% or simply 25%) in the preliminary results. Entropy dropped largely, BIC, AIC etc. also increased.