I am a novice working with Structural Equation Modeling and would like an interpretation of my Multiple Indicators Multiple Causes (MIMIC Model) results. I have recoded my independent variables as dichotomous (0, 1). Most of my dependent variables have also been coded as dichotomous with (Bernoulli, profit), with one of the indicators being ordinal / logit. I'm using Stata as my software.

I would like to know how to read these results and if there are particular statistical tests that I should be considering to interpret the model, I would be grateful for a journal article(s) lead, as appropriate.

Thanks a million for the steer. // John

= = = = =

Model:

gsem ( socunr2 -> Onset, ) (regrep2 -> Onset, ) (massv12 -> Onset, ) (gvinst2 -> Onset, ) (allycap2 -> Onset, ) (crismg2 -> Onset, ) (regime2 -> Onset, ) (durreg2 -> Onset, ) (pow sta2 -> Onset, ) (econdt2 -> Onset, ) (Onset -> triggr2, family(bernoulli) link(probit)) (Onset -> gravty, family(ordinal) link(logit) ) (Onset -> majres2, family(bernoulli) link( probit) ) (Onset -> issue2, family(bernoulli) link(probit) ) (Onset -> viol2, family(bernou lli) link(probit) ), latent (Onset ) nocapslatent Fitting fixed-effects model: Iteration 0: log likelihood = -4078.6247 Iteration 1: log likelihood = -4076. 1871 Iteration 2: log likelihood = -4076. 187 Refining starting values: Grid node 0: log likelihood = -3851.2078 Fitting full model: Iteration 0: log likelihood = -3851. 2078 (not concave) Iteration 1: log likelihood = -3521.8276 (not concave) Iteration 2: log likelihood = -3465. 1151 Iteration 3: log likelihood = -3421.3373 Iteration 4: log likelihood = -3389. 035 (not concave) Iteration 5: log likelihood = -3384. 1975 Iteration 6: log likelihood = -3380.4375 Iteration 7: log likelihood = -3379. 8093 Iteration 8: log likelihood = -3379.6104 Iteration 9: log likelihood = -3379.5903 Iteration 10: log likelihood = -3379.5902 Generalized structural equation model Number of obs = 912 Response: triggr2 Family : Bernoulli Link: Probit Response: gravty Family : Ordinal Link: Logit Response: majres2 Family : Bernoulli Link: Probit Response: issue2 Family : Bernoulli Link: Probit Response: viol2 Family: Bernoulli Link: Probit Log likelihood = -3379.5902( 1) [triggr2]Onset = 1 Coefficient Std. err. Z P> |Z| [95% conf. interval] triggr2 Onset 1 ( constrained) cons -. 8547608 . 1025825 -8.33 0.000 -1. 055819 -.6537027 gravty Onset . 1092931 . 0972495 1. 12 0. 261 -. 0813124 . 2998986 Onset socunr2 0421248 . 0483554 0.87 0. 384 -. 05265 . 1368995 regrep2 -. 0039367 0506385 -0. 08 0. 938 -. 1031864 . 0953129 massv12 -. 0310275 058931 -0. 53 0. 599 . 14653 . 084475 gvinst2 . 0940059 0478043 1.97 0 . 049 . 0003112 1877006 allycap2 -. 0773902 0396268 -1. 95 0 . 051 . 1550574 . 000277 crismg2 1. 15766 . 0967554 11. 96 0 . 000 . 9680226 1. 347297 regime2 . 0641422 0387452 1. 66 0 . 098 -. 0117971 . 1400814 durreg2 012374 . 0361616 0.34 0. 732 -. 0585014 . 0832494 powsta2 0574698 . 0412221 1.39 0. 163 -. 023324 . 1382636 econdt2 . 0204254 . 0400312 0. 51 0. 610 -. 0580343 . 098885 maj res2 Onset 3. 730029 6105417 6. 11 0. 000 2. 533389 4. 926668 _cons -3. 789195 . 6136217 -6. 18 0 . 000 -4. 991872 -2. 586519 issue2 Onset -. 1775436 . 0725495 -2. 45 0 . 014 -. 319738 -. 0353493 cons -. 3869011 . 0654419 -5. 91 0 . 000 -. 5151649 -. 2586373 viol2 Onset 3. 997689 669914 5.97 0 . 000 2 . 684682 5. 310697 cons -1. 297775 . 3272486 -3.97 0 . 000 -1. 939171 -. 6563799 /gravty cut1 -3. 238754 . 1919053 -3. 614881 -2. 862626 cut2 -1. 763788 1175525 -1. 994187 -1. 533389 cut 3 -. 7942502 0999485 -. 9901456 -. 5983548 cut 4 . 1651359 0959439 -. 0229107 . 3531825 cut 5 1. 168955 . 1028236 . 9674248 1. 370486 cut6 2. 504187 . 1401033 2. 22959 2. 778785 cut7 4. 315098 . 2883872 3. 74987 4. 880327 var (e. Onset) . 1005831 . 0242422 . 0627149 . 1613164crismg2 regrep2 massv/2 gvinst2 allycap2 regime2 socunr2 durreg2 powsta2 econdt2 1.2 -.0039 -.031 .094 -.077 1.064 .042 .012 .057 02 Onset 3.7 .11 -.18 4 Bernoulli Bernoulli ordinal Bernoulli Bernoulli triggr2 majres2 gravty issue2 ..85 -3.8 viol2 -.39 -1.3 probit probit logit probit probit