Dear colleagues
I have tried to conduct a Monte Carlo simulation in Mplus for a latent three-factor model with correlations between the factors. I want to test whether n = 500 would be enough to construct the factors properly and to detect correlations as low as r = .3. (I know it propably is, I just wanted to finally explore Monte Carlo simulations).
However, I keep receiving the following error message, no matter what I do:
*** ERROR in MONTECARLO command
Probabilities of observations for patterns must sum up to 1.
Unfortunately, the Mplus website is no help. I have enclosed my syntax. It is a lot to ask, but those of you who are familiar with Monte Carlo simulations for categorical indicators, would you consider pointing my in the right direction? What do I need to change?
Thank you in advance for your insight.
Best
Marcel
P.S. I have attached a more reader friendly PDF of the syntax as well.
-------------------------------------------------------------------------------------------
TITLE:
Monte Carlo study
three factors: subjective content knowledge (geography, history, civics)
categorical indicators (4-point Likert scale)
MONTECARLO:
NAMES =
f18_a f18_b f18_c f18_d f18_e f18_f f18_g
f19_a f19_b f19_c f19_d f19_e
f20_a f20_b f20_c f20_d f20_e f20_f f20_h f20_i;
CATEGORICAL =
f18_a f18_b f18_c f18_d f18_e f18_f f18_g
f19_a f19_b f19_c f19_d f19_e
f20_a f20_b f20_c f20_d f20_e f20_f f20_h f20_i;
NOBSERVATIONS = 500;
NREPS = 1000;
!Items measured on a 4-point Likert scale
GENERATE = f18_a-f20_i (3 p);
!5 % missings on each item assumed
PATMISS = f18_a-f20_i (.05);
!50 % of cases assumed to have missings
PATPROBS = .50;
!The next part is probably where I get it wrong
MODEL POPULATION:
SCK_Geo BY f18_a-f18_g*.1
SCK_Geo@1;
SCK_His BY f19_a-f19_e*.1
SCK_His@1;
SCK_Civ BY f20_a-f20_i*.1
SCK_Civ@1;
SCK_Civ WITH SCK_Geo*.3;
SCK_Civ WITH SCK_His* .3;
SCK_His WITH SCK_Geo* .3;
ANALYSIS:
TYPE = general;
ESTIMATOR = WLSMV;
PARAMETERIZATION=THETA;
MODEL:
SCK_Geo BY f18_a-f18_g*.1
SCK_Geo@1;
SCK_His BY f19_a-f19_e*.1
SCK_His@1;
SCK_Civ BY f20_a-f20_i*.1
SCK_Civ@1;
SCK_Civ WITH SCK_Geo*.3;
SCK_Civ WITH SCK_His* .3;
SCK_His WITH SCK_Geo * .3;
OUTPUT: Tech9;