Latent Transition Analysis1
Appendix A
Summary of PROC LTA Statements and Options
Syntax / Required / DescriptionPROC LTA / √ / Invokes the procedure
options:
DATA / √ / Specifies SAS data file to be analyzed
VERBOSE_OUTPUT / Shows starting values, parameter restrictions and maximum absolute deviation and loglikelihood at each iteration
OUTEST / Saves parameter estimates to SAS data file in one record
OUTPARAM / Saves parameter estimates to SAS data file
OUTPOST / Saves posterior probabilities to SAS data file
NOPRINT / Suppresses printing of output
START / Allows user to provide starting values
RESTRICT / Allows user to specify parameter restrictions for item-response probabilities
NOBETATEST / Suppresses tests of significance for covariates
statements:
NSTATUS / √ / Specifies number of latent statuses at each time
NTIMES / √ / Specifies number of times of measurement
ITEMS / √ / Declares variables that indicate latent status variable at each time
CATEGORIES / √ / Specifies number of response categories in items
ID / Declares identifier and other variables to retain in posterior probabilities SAS file
GROUPS / Declares categorical grouping variable
GROUPNAMES / Specifies a label for each group
COVARIATES1 / Declares variables to include as covariates for Time 1 status
COVARIATES2 / Declares variables to include as covariates for transitions
REFERENCE1 / Specifies latent status to be used as reference group in prediction from COVARIATES1
REFERENCE2 / Specifies latent statuses to be used as reference groups in prediction from COVARIATES2
BINARY1 / Specifies latent status to be used as comparison group in prediction from COVARIATES1, and that binary logistic regression is to be used
BINARY2 / Specifies latent statuses to be used as comparison groups in prediction from COVARIATES2, and that binary logistic regression is to be used
STABILIZE / Invokes a stabilizing prior for the logistic regressions
FREQ / Declares frequency count variable, to use when data are aggregated
ESTIMATION / Specifies estimation procedure
SEED / √ a / Specifies seed for random number generator
MEASUREMENT / Invokes measurement invariance across groups and/or times
MAXITER / Specifies maximum number of iterations
CRITERION / Specifies convergence criterion for maximum absolute deviation
a SEED statement is required only if the START option is not included in the PROC LTA statement.
Appendix B
SAS Syntax Used to Fit LTA Models
*Model 1. Five-status Model for Dating and Sexual Behavior;
proclta data=pgm.sex_recoded;
title1 'Model 1';
nstatus 5;
ntimes 3;
items datepar_alt_98 sex_yr_98 part_98 expos_98
datepar_alt_99 sex_yr_99 part_99 expos_99
datepar_alt_00 sex_yr_00 part_00 expos_00;
categories 3232;
measurement times;
seed 592667;
run;
*Model 2. Five-status Model for Dating and Sexual Behavior, by Sex;
proclta data=pgm.sex_recoded;
title1 'Model 2';
nstatus 5;
ntimes 3;
items datepar_alt_98 sex_yr_98 part_98 expos_98
datepar_alt_99 sex_yr_99 part_99 expos_99
datepar_alt_00 sex_yr_00 part_00 expos_00;
categories 3232;
groups gender;
groupnames male female;
measurement times groups;
seed 592667;
run;
*Model 3. Time 1 Substance Use Predicting Time 1 Deltas;
procltadata=pgm.sex_recoded start=lta5_start;
title1 'Model 3';
nstatus 5;
ntimes 3;
items datepar_alt_98 sex_yr_98 part_98 expos_98
datepar_alt_99 sex_yr_99 part_99 expos_99
datepar_alt_00 sex_yr_00 part_00 expos_00;
categories 3232;
covariates1 cig_yr_98 drunk_98 mar_yr_98;
reference1 1;
measurement times;
run;
*Model 4. Time 1 Substance Use Predicting Time 1 Deltas, by Sex;
proclta data=pgm.sex_recoded start=lta5gend_start;
title1 'Model 4';
nstatus 5;
ntimes 3;
items datepar_alt_98 sex_yr_98 part_98 expos_98
datepar_alt_99 sex_yr_99 part_99 expos_99
datepar_alt_00 sex_yr_00 part_00 expos_00;
categories 3232;
covariates1 cig_yr_98 drunk_98 mar_yr_98;
reference11;
groups gender;
groupnames male female;
measurement times groups;
run;
*Model 5. Drunkenness Predicting Dating and Sexual Behavior Over Time;
proclta data=pgm.sex_recoded start=time2_start_tau rest=time2_rest_tau;
title1 'Model 5';
nstatus 5;
ntimes 2;
items datepar_alt_98 sex_yr_98 part_98 expos_98
datepar_alt_99 sex_yr_99 part_99 expos_99;
categories 3232;
covariates1 drunk_98;
covariates2 drunk_98;
reference11;
reference21 2 3 0 5;
measurement times;
run;
Appendix C
Starting Values Provided for Model 3
*Starting Values for Five-class Models;
datalta5_start;
inputparam $ group variable $ time statusrespcatestls1 estls2 estls3 estls4 estls5;
datalines;
DELTA 1 . 1 . . 0.2 0.2 0.2 0.2 0.2
TAU 1 . 1 1 . 0.2 0.2 0.2 0.2 0.2
TAU 1 . 1 2 . 0.2 0.2 0.2 0.2 0.2
TAU 1 . 1 3 . 0.2 0.2 0.2 0.2 0.2
TAU 1 . 1 4 . 0.2 0.2 0.2 0.2 0.2
TAU 1 . 1 5 . 0.2 0.2 0.2 0.2 0.2
TAU 1 . 2 1 . 0.2 0.2 0.2 0.2 0.2
TAU 1 . 2 2 . 0.2 0.2 0.2 0.2 0.2
TAU 1 . 2 3 . 0.2 0.2 0.2 0.2 0.2
TAU 1 . 2 4 . 0.2 0.2 0.2 0.2 0.2
TAU 1 . 2 5 . 0.2 0.2 0.2 0.2 0.2
RHO 1 datepar_alt_98 1 . 1 0.6 0.2 0.2 0.2 0.2
RHO 1 sex_yr_98 1 . 1 0.6 0.6 0.4 0.4 0.4
RHO 1 part_98 1 . 1 0.6 0.6 0.2 0.2 0.2
RHO 1 expos_98 1 . 1 0.6 0.6 0.4 0.6 0.4
RHO 1 datepar_alt_98 1 . 2 0.2 0.2 0.6 0.2 0.2
RHO 1 sex_yr_98 1 . 2 0.4 0.4 0.6 0.6 0.6
RHO 1 part_98 1 . 2 0.2 0.2 0.6 0.2 0.2
RHO 1 expos_98 1 . 2 0.4 0.4 0.6 0.4 0.6
RHO 1 datepar_alt_98 1 . 3 0.2 0.6 0.2 0.6 0.6
RHO 1 sex_yr_98 1 . 3 0 0 0 0 0
RHO 1 part_98 1 . 3 0.2 0.2 0.2 0.6 0.6
RHO 1 expos_99 1 . 3 0 0 0 0 0
RHO 1 datepar_alt_99 2 . 1 0.6 0.2 0.2 0.2 0.2
RHO 1 sex_yr_99 2 . 1 0.6 0.6 0.4 0.4 0.4
RHO 1 part_99 2 . 1 0.6 0.6 0.2 0.2 0.2
RHO 1 expos_99 2 . 1 0.6 0.6 0.4 0.6 0.4
RHO 1 datepar_alt_99 2 . 2 0.2 0.2 0.6 0.2 0.2
RHO 1 sex_yr_99 2 . 2 0.4 0.4 0.6 0.6 0.6
RHO 1 part_99 2 . 2 0.2 0.2 0.6 0.2 0.2
RHO 1 expos_99 2 . 2 0.4 0.4 0.6 0.4 0.6
RHO 1 datepar_alt_99 2 . 3 0.2 0.6 0.2 0.6 0.6
RHO 1 sex_yr_99 2 . 3 0 0 0 0 0
RHO 1 part_99 2 . 3 0.2 0.2 0.2 0.6 0.6
RHO 1 expos_99 2 . 3 0 0 0 0 0
RHO 1 datepar_alt_00 3 . 1 0.6 0.2 0.2 0.2 0.2
RHO 1 sex_yr_00 3 . 1 0.6 0.6 0.4 0.4 0.4
RHO 1 part_00 3 . 1 0.6 0.6 0.2 0.2 0.2
RHO 1 expos_00 3 . 1 0.6 0.6 0.4 0.6 0.4
RHO 1 datepar_alt_00 3 . 2 0.2 0.2 0.6 0.2 0.2
RHO 1 sex_yr_00 3 . 2 0.4 0.4 0.6 0.6 0.6
RHO 1 part_00 3 . 2 0.2 0.2 0.6 0.2 0.2
RHO 1 expos_00 3 . 2 0.4 0.4 0.6 0.4 0.6
RHO 1 datepar_alt_00 3 . 3 0.2 0.6 0.2 0.6 0.6
RHO 1 sex_yr_00 3 . 3 0 0 0 0 0
RHO 1 part_00 3 . 3 0.2 0.2 0.2 0.6 0.6
RHO 1 expos_00 3 . 3 0 0 0 0 0
;