Comparing LPA: Real Data vs Monte Car...
Message/Author
 Jordan Tait posted on Tuesday, March 22, 2016 - 10:24 am
Hello all,

I have a data set that I have ran Latent Profile Analysis on with three (3) classes. Since the sample size is small, I would like to use a Monte Carlo simulation to create more data to run another LPA and compare results between my data and the Monte Carlo Simulated data.

My thought was to use the population estimates from my real data ( ran and saved repeatedly ) in the Monte Carlo sim. I've tried to look at manipulate examples I've found with little luck... so I have a few questions:

1. Can mplus compare two seperate LPA's ( or does this comparison make logical sense )?

2. If so, can this be done automatically with built in code ( options within monte carlo or...) or will this be a more manual process?

Any hints, tips or suggestions would be greatly appreciated!
 Jordan Tait posted on Tuesday, March 22, 2016 - 10:26 am
Here are the snippets of code I've been playing with.

==========================================
LPA with my real data, saved the estimates
==========================================
TITLE: LPA 3 Class - Perfectionism
DATA: FILE IS 150108Mplus.dat;
VARIABLE: NAMES ARE PS_mean COM_mean PPP_mean PCP_mean DAA_mean ORG_mean ID;
Usevariables = PS_mean COM_mean PPP_mean PCP_mean DAA_mean ORG_mean;
CLASSES = c (3);
IDvariable is ID;
ANALYSIS: TYPE = MIXTURE;
STARTS = 5000 500;
STITERATIONS = 100;
OUTPUT: TECH1 TECH4 TECH14;
SAVEDATA: FILE is Actual_estimates_c3.dat;
SAVE IS CPROB;
ESTIMATES= Actual_estimates_c3_for_MC_.dat

====================================================
Monte Carlo Sim using Saved estimates from real data
====================================================
TITLE: LPA 3 Class - Monte Carlo Simulation
MONTECARLO: NAMES ARE PS_mean COM_mean PPP_mean PCP_mean DAA_mean ORG_mean;
NOBSERVATIONs = 200;
NREPS = 10;
SEED = 4500;
POPULATION = Actual_estimates_c3_for_MC_.dat;
COVERAGE = Actual_estimates_c3_for_MC_.dat;
GENCLASSES = c (3);
CLASSES = c (3);
ANALYSIS: TYPE = MIXTURE;
STARTS = 1000 500;
STITERATIONS = 100;
OUTPUT: TECH1 TECH14;
 Linda K. Muthen posted on Tuesday, March 22, 2016 - 2:30 pm