recurring lack of BLRT results for model 6 - why?

[4da...@gmail.com]

Hello,

after several checks it turns out every now and then I generate a subsample of my data that gives no results in BLRT using tidyLPA for model 6

is there any reason for that related to e.g. data format etc.? I import data from excel and spss files using RStudio, but they look allright

e.g.

Compare tidyLPA solutions:

 Model Classes LogLik    BIC      SABIC    AIC         CAIC         ICL           Entropy      BLRT_val   BLRT_p     Warnings

 2     1       -20642.54 41359.55 41327.79 41305.08 41369.55 -41359.55 1.0000000                               

 2     2       -19900.91 39958.21 39891.49 39843.81 39979.21 -40279.63 0.7198330 1483.26804 0.00990099         

 2     3       -19660.32 39558.95 39457.29 39384.64 39590.95 -40061.40 0.7207942  481.17255 0.00990099         

 2     4       -19584.58 39489.39 39352.78 39255.16 39532.39 -40070.32 0.7463098  151.48006 0.00990099         

 2     5       -19505.69 39413.53 39241.98 39119.39 39467.53 -40231.25 0.7020809  157.77470 0.00990099         

 2     6       -19472.67 39429.40 39222.90 39075.33 39494.40 -40378.66 0.6878820   66.05598 0.00990099         

 6     1       -19706.56 39562.07 39498.53 39453.12 39582.07 -39562.07 1.0000000                               

 6     2       -19389.39 39084.10 38953.85 38860.77 39125.10 -39595.46 0.5534877                               

 6     3       -19303.83 39069.38 38872.42 38731.66 39131.38 -39903.59 0.5637605                               

 6     4       -19228.10 39074.32 38810.64 38622.20 39157.32 -40211.54 0.5497164                               

 6     5       -19192.03 39158.56 38828.17 38592.06 39262.56 -40306.10 0.5998436                               

 6     6                                                                                               Warning 

An analytic hierarchy process, based on the fit indices AIC, AWE, BIC, CLC, and KIC (Akogul & Erisoglu, 2017), suggests the best solution is Model 6 with 4 classes.

[Caspar van Lissa]

Dear 4daren,

Model 6 is a very complex model, and you're estimating it with a lot of classes. My guess is that the model is not converging. The fit is hardly better than for the simpler models you're estimating, so I'd recommend settling for one of those.

Best,
CJ

[4da...@gmail.com]

Hello Caspar,

do you mean to ignore BIC and AHP for model 6 and settle best solution among model 2 results? To get AHP I could run the analysis only for model 2... Do I get the suggestion right?

Thanks for prompt answer Caspar!