Group Based Trajectory Model (GBTM) outcome report

[Maryam Bidgoli]

Hello, I’m using the Stata traj plugin for group-based trajectory modeling (logit model with a binary outcome of mammogram screening status). I'm including multiple time-varying covariates (e.g., smoking status, health insurance status, marital status, income above median, cesd (depression score), no of health conditions, ADL, and urbanity) across 7 waves using the tcov() option. My code looks like this:

traj, var(mammog_st*) indep(t*) model(logit) order(2 2) dropout(1 1) obsmar(obsmar) ///
    risk(race_2 race_3 raedyrs) weight(r3wtresp) ///
    tcov( ///
    insured3 insured5 insured7 insured9 insured11 insured13 insured15 ///
    married3 married5 married7 married9 married11 married13 married15 ///
    r3smoken r5smoken r7smoken r9smoken r11smoken r13smoken r15smoken ///
    income_abv_med3 income_abv_med5 income_abv_med7 income_abv_med9 /// income
         income_abv_med11 income_abv_med13 income_abv_med15 ///
         cesd3_high cesd5_high cesd7_high cesd9_high cesd11_high cesd13_high cesd15_high ///
          r3conde r5conde r7conde r9conde r11conde r13conde r15conde ///
          r3adl5a r5adl5a r7adl5a r9adl5a r11adl5a r13adl5a r15adl5a ///
          urban_3 urban_5 urban_7 urban_9 urban_11 urban_13 urban_15)

However, in the output, Stata only reports the coefficient for the covariate from the first wave (wave 3, eg. r3smoken). I was wondering why this happens, especially since I included the variable for all waves. Please see the attachment for the output.

Thanks,

Maryam