UCLA Academic Technology Services HomeServicesClassesContactJobs
Search

LEM Textbook Examples
Latent Class Scaling Analysis by C. Mitchell Dayton

Table 3.2 on page 28 using pleural thickening data.

Model I: unconstrained

 lat 1
 man 3 
 dim 2 2 2 2 
 lab X A B C 
 mod X   
     A|X 
     B|X 
     C|X   

 dat [1513 23
        59 12
        21 19
        11 34]
*** STATISTICS ***
  Number of iterations = 122
  Converge criterion   = 0.0000006663
  Seed random values   = 2610
  X-squared            = 0.0000 (0.0000)
  L-squared            = 0.0000 (0.0000)
  Cressie-Read         = 0.0000 (0.0000)
  Dissimilarity index  = 0.0000
  Degrees of freedom   = 0
  Log-likelihood       = -891.14280
  Number of parameters = 7 (+1)
  Sample size          = 1692.0
  BIC(L-squared)       = 0.0000
  AIC(L-squared)       = 0.0000
  BIC(log-likelihood)  = 1834.3213
  AIC(log-likelihood)  = 1796.2856
*** LATENT CLASS OUTPUT ***
          X  1    X  2
         0.9455  0.0545
  A  1   0.9900  0.2510
  A  2   0.0100  0.7490
  B  1   0.9646  0.3566
  B  2   0.0354  0.6434
  C  1   0.9891  0.2350
  C  2   0.0109  0.7650

Model II: Homogeneous

 lat 1
 man 3 
 dim 2 2 2 2 
 lab X A B C 
 mod X   
     A|X 
     B|X eq1 A|X
     C|X eq1 B|X
 dat [1513 23
        59 12
        21 19
        11 34]
*** STATISTICS ***
  Number of iterations = 95
  Converge criterion   = 0.0000006431
  Seed random values   = 257
  X-squared            = 29.3556 (0.0000)
  L-squared            = 27.4114 (0.0000)
  Cressie-Read         = 28.5695 (0.0000)
  Dissimilarity index  = 0.0175
  Degrees of freedom   = 4
  Log-likelihood       = -904.84848
  Number of parameters = 3 (+1)
  Sample size          = 1692.0
  BIC(L-squared)       = -2.3233
  AIC(L-squared)       = 19.4114
  BIC(log-likelihood)  = 1831.9980
  AIC(log-likelihood)  = 1815.6970
*** LATENT CLASS OUTPUT ***
          X  1    X  2
         0.0546  0.9454
  A  1   0.2835  0.9812
  A  2   0.7165  0.0188
  B  1   0.2835  0.9812
  B  2   0.7165  0.0188
  C  1   0.2835  0.9812
  C  2   0.7165  0.0188

Model III: Reader B, heterogeneous

   lat 1
   man 3
   dim 2 2 2 2 
   lab X A B C 
   mod X   
       A|X 
       B|X 
       C|X eq1 A|X   
   dat [1513 23
          59 12
          21 19
          11 34]
*** STATISTICS ***
  Number of iterations = 168
  Converge criterion   = 0.0000006857
  Seed random values   = 1327
  X-squared            = 0.1344 (0.9350)
  L-squared            = 0.1344 (0.9350)
  Cressie-Read         = 0.1344 (0.9350)
  Dissimilarity index  = 0.0009
  Degrees of freedom   = 2
  Log-likelihood       = -891.21001
  Number of parameters = 5 (+1)
  Sample size          = 1692.0
  BIC(L-squared)       = -14.7329
  AIC(L-squared)       = -3.8656
  BIC(log-likelihood)  = 1819.5884
  AIC(log-likelihood)  = 1792.4200
*** LATENT CLASS OUTPUT ***
          X  1    X  2
         0.9455  0.0545
  A  1   0.9896  0.2431
  A  2   0.0104  0.7569
  B  1   0.9646  0.3566
  B  2   0.0354  0.6434
  C  1   0.9896  0.2431
  C  2   0.0104  0.7569

Model IV: Reader B, false negative

  lat 1
  man 3
  dim 2 2 2 2 
  lab X A B C 
  mod X   
      A|X eq2
      B|X eq2
      C|X eq2 
  des [1 0 2 0
       1 0 3 0
       1 0 2 0]
  dat [1513 23
         59 12
         21 19
         11 34]

*** STATISTICS ***

  Number of iterations = 142
  Converge criterion   = 0.0000009692
  Seed random values   = 4135

  X-squared            = 29.1815 (0.0000)
  L-squared            = 27.2505 (0.0000)
  Cressie-Read         = 28.3916 (0.0000)
  Dissimilarity index  = 0.0171
  Degrees of freedom   = 3
  Log-likelihood       = -904.76807
  Number of parameters = 4 (+1)
  Sample size          = 1692.0
  BIC(L-squared)       = 4.9495
  AIC(L-squared)       = 21.2505
  BIC(log-likelihood)  = 1839.2708
  AIC(log-likelihood)  = 1817.5361

*** LATENT CLASS OUTPUT ***

          X  1    X  2
         0.9467  0.0533
  A  1   0.9808  0.2626
  A  2   0.0192  0.7374
  B  1   0.9808  0.2952
  B  2   0.0192  0.7048
  C  1   0.9808  0.2626
  C  2   0.0192  0.7374

Model V: Reader B, false positive

    lat 1
    man 3
    dim 2 2 2 2 
    lab X A B C 
    mod X   
        A|X eq2
        B|X eq2
        C|X eq2 
    
    des [1 0 2 0
         3 0 2 0
         1 0 2 0]
    
    dat [1513 23
           59 12
           21 19
           11 34]

*** STATISTICS ***

  Number of iterations = 116
  Converge criterion   = 0.0000008031
  Seed random values   = 584

  X-squared            = 3.1099 (0.3750)
  L-squared            = 2.9870 (0.3936)
  Cressie-Read         = 3.0641 (0.3818)
  Dissimilarity index  = 0.0039
  Degrees of freedom   = 3
  Log-likelihood       = -892.63628
  Number of parameters = 4 (+1)
  Sample size          = 1692.0
  BIC(L-squared)       = -19.3140
  AIC(L-squared)       = -3.0130
  BIC(log-likelihood)  = 1815.0072
  AIC(log-likelihood)  = 1793.2726

*** LATENT CLASS OUTPUT ***

          X  1    X  2
         0.9447  0.0553
  A  1   0.9892  0.2868
  A  2   0.0108  0.7132
  B  1   0.9659  0.2868
  B  2   0.0341  0.7132
  C  1   0.9892  0.2868
  C  2   0.0108  0.7132

Cheating Data Example on page 30.

We have skipped column SE(J) and SE(B) for the estimation of standard error using Jackknife and bootstrapping methods in the table. Table 3.3 on page 32 corresponds to the part of the output labeled as "(CONDITIONAL) PROBABILITIES". Table 3.4 on page 33 corresponds to the part of the output labeled as "FREQUENCIES". The last two columns from Table 3.4 can be produced using any statistical software such as Stata.

  lat 1
  man 4
  dim 2 2 2 2 2
  lab X A B C D
  mod X   
      A|X 
      B|X 
      C|X 
      D|X
  
  dat [207 46
         7  5
        13  4
         1  2
        10  3
         1  2
        11  4
         1  2]


*** STATISTICS ***

  Number of iterations = 367
  Converge criterion   = 0.0000009825
  Seed random values   = 1326

  X-squared            = 8.3209 (0.2155)
  L-squared            = 7.7643 (0.2559)
  Cressie-Read         = 8.0766 (0.2325)
  Dissimilarity index  = 0.0315
  Degrees of freedom   = 6
  Log-likelihood       = -440.02713
  Number of parameters = 9 (+1)
  Sample size          = 319.0
  BIC(L-squared)       = -26.8269
  AIC(L-squared)       = -4.2357
  BIC(log-likelihood)  = 931.9410
  AIC(log-likelihood)  = 898.0543

*** FREQUENCIES ***

  A B C D     observed  estimated  std. res.
  1 1 1 1     207.000    205.718      0.089
  1 1 1 2      46.000     47.412     -0.205
  1 1 2 1       7.000      8.955     -0.653
  1 1 2 2       5.000      2.450      1.629
  1 2 1 1      13.000     12.299      0.200
  1 2 1 2       4.000      5.118     -0.494
  1 2 2 1       1.000      1.956     -0.684
  1 2 2 2       2.000      1.091      0.871
  2 1 1 1      10.000      9.334      0.218
  2 1 1 2       3.000      4.343     -0.644
  2 1 2 1       1.000      1.770     -0.579
  2 1 2 2       2.000      1.017      0.974
  2 2 1 1      11.000      8.620      0.811
  2 2 1 2       4.000      5.156     -0.509
  2 2 2 1       1.000      2.347     -0.879
  2 2 2 2       2.000      1.413      0.494

*** (CONDITIONAL) PROBABILITIES ***

* P(X) *

  1              0.8390  (0.0781)
  2              0.1610  (0.0781)

* P(A|X) *

  1 | 1          0.9835  (0.0289)
  2 | 1          0.0165  (0.0289)
  1 | 2          0.4239  (0.1806)
  2 | 2          0.5761  (0.1806)

* P(B|X) *

  1 | 1          0.9709  (0.0306)
  2 | 1          0.0291  (0.0306)
  1 | 2          0.4118  (0.1753)
  2 | 2          0.5882  (0.1753)

* P(C|X) *

  1 | 1          0.9629  (0.0152)
  2 | 1          0.0371  (0.0152)
  1 | 2          0.7843  (0.0846)
  2 | 2          0.2157  (0.0846)

* P(D|X) *

  1 | 1          0.8181  (0.0263)
  2 | 1          0.1819  (0.0263)
  1 | 2          0.6240  (0.1004)
  2 | 2          0.3760  (0.1004)

The last two columns of Table 3.4 on page 33 using Stata:

clear
input   A B C D     observed  estimated  std_res
  1 1 1 1     207.000    205.718      0.089
  1 1 1 2      46.000     47.412     -0.205
  1 1 2 1       7.000      8.955     -0.653
  1 1 2 2       5.000      2.450      1.629
  1 2 1 1      13.000     12.299      0.200
  1 2 1 2       4.000      5.118     -0.494
  1 2 2 1       1.000      1.956     -0.684
  1 2 2 2       2.000      1.091      0.871
  2 1 1 1      10.000      9.334      0.218
  2 1 1 2       3.000      4.343     -0.644
  2 1 2 1       1.000      1.770     -0.579
  2 1 2 2       2.000      1.017      0.974
  2 2 1 1      11.000      8.620      0.811
  2 2 1 2       4.000      5.156     -0.509
  2 2 2 1       1.000      2.347     -0.879
  2 2 2 2       2.000      1.413      0.494
end
gen x2 = (observed-estimated)^2/estimated
sum x2
gen per_x2 = x2/(16*r(mean))
sort D C B A
list A B C D observed estimated x2 per_x2, clean
       A   B   C   D   observed   estima~d         x2     per_x2  
  1.   1   1   1   1        207    205.718   .0079892   .0009601  
  2.   2   1   1   1         10      9.334   .0475205    .005711  
  3.   1   2   1   1         13     12.299   .0399546   .0048017  
  4.   2   2   1   1         11       8.62    .657123   .0789728  
  5.   1   1   2   1          7      8.955   .4268034   .0512931  
  6.   2   1   2   1          1       1.77   .3349717   .0402568  
  7.   1   2   2   1          1      1.956   .4672474   .0561536  
  8.   2   2   2   1          1      2.347   .7730758    .092908  
  9.   1   1   1   2         46     47.412   .0420514   .0050537  
 10.   2   1   1   2          3      4.343   .4153002   .0499106  
 11.   1   2   1   2          4      5.118   .2442212   .0293504  
 12.   2   2   1   2          4      5.156   .2591808   .0311483  
 13.   1   1   2   2          5       2.45   2.654082   .3189666  
 14.   2   1   2   2          2      1.017   .9501368   .1141871  
 15.   1   2   2   2          2      1.091   .7573612   .0910194  
 16.   2   2   2   2          2      1.413   .2438563   .0293066  

Table 3.5, 3.6 and 3.7 are omitted for now since LEM does not have those options for calculating these quantities.


Table 3.8 on page 39 using the academic cheating data with a single latent variable of two classes.

We first create a data file containing the latent classification probabilities and the modal class using the output option wla of LEM.

  lat 1
  man 4
  dim 2 2 2 2 2
  lab X A B C D
  mod X   
      A|X 
      B|X 
      C|X 
      D|X
  
  dat [207 46
         7  5
        13  4
         1  2
        10  3
         1  2
        11  4
         1  2]

 wla cheat_lca.dat

Next, we simply copy and paste the content of data file cheat_lca.dat to Stata do file editor and input it as a data set. Based on the output from the LEM run, we create a sequence of variables representing the conditional probabilities. Based on the conditional probabilities and the latent class probabilities, we then create column 3 and 4 of Table 3.8.

clear
input a b c d observed p1 p2 class error
  1  1  1  1  207.0000 0.9787 0.0213   1 0.0213
  1  1  1  2   46.0000 0.9442 0.0558   1 0.0558
  1  1  2  1    7.0000 0.8652 0.1348   1 0.1348
  1  1  2  2    5.0000 0.7032 0.2968   1 0.2968
  1  2  1  1   13.0000 0.4904 0.5096   2 0.4904
  1  2  1  2    4.0000 0.2621 0.7379   2 0.2621
  1  2  2  1    1.0000 0.1187 0.8813   2 0.1187
  1  2  2  2    2.0000 0.0473 0.9527   2 0.0473
  2  1  1  1   10.0000 0.3612 0.6388   2 0.3612
  2  1  1  2    3.0000 0.1726 0.8274   2 0.1726
  2  1  2  1    1.0000 0.0733 0.9267   2 0.0733
  2  1  2  2    2.0000 0.0284 0.9716   2 0.0284
  2  2  1  1   11.0000 0.0117 0.9883   2 0.0117
  2  2  1  2    4.0000 0.0044 0.9956   2 0.0044
  2  2  2  1    1.0000 0.0017 0.9983   2 0.0017
  2  2  2  2    2.0000 0.0006 0.9994   2 0.0006
end

gen a11 = .9835 
gen a12 = .4239
gen b11 = .9709
gen b12 = .4118
gen c11 = .9629
gen c12 = .7843
gen d11 = .8181
gen d12 = .6240

gen px1 = .8390

gen py1 = 1
gen py2 = 1

foreach var of varlist a b c d {
   replace py1 = py1*`var'11 if `var'==1
   replace py2 = py2*`var'12 if `var'==1

   replace py1 = py1*(1-`var'11) if `var'==2
   replace py2 = py2*(1-`var'12) if `var'==2
}

replace py1 = py1*px1
replace py2 = py2*(1-px1)
gen odds = p2/p1

sort d c b a

list a b c d observed py1 py2 class p2 odds, clean
       a   b   c   d   observed        py1        py2   class      p2       odds  
  1.   1   1   1   1        207   .6311003   .0137544       1   .0213   .0217636  
  2.   2   1   1   1         10   .0105879   .0186929       2   .6388   1.768549  
  3.   1   2   1   1         13   .0189155   .0196463       2   .5096   1.039152  
  4.   2   2   1   1         11   .0003173   .0267003       2   .9883   84.47009  
  5.   1   1   2   1          7    .024316   .0037828       1   .1348   .1558021  
  6.   2   1   2   1          1   .0004079    .005141       2   .9267   12.64257  
  7.   1   2   2   1          1   .0007288   .0054032       2   .8813     7.4246  
  8.   2   2   2   1          1   .0000122   .0073432       2   .9983   587.2353  
  9.   1   1   1   2         46   .1403217   .0082879       1   .0558   .0590976  
 10.   2   1   1   2          3   .0023542   .0112637       2   .8274   4.793743  
 11.   1   2   1   2          4   .0042057   .0118382       2   .7379   2.815338  
 12.   2   2   1   2          4   .0000706   .0160886       2   .9956   226.2727  
 13.   1   1   2   2          5   .0054065   .0022794       1   .2968   .4220705  
 14.   2   1   2   2          2   .0000907   .0030978       2   .9716   34.21127  
 15.   1   2   2   2          2    .000162   .0032558       2   .9527   20.14165  
 16.   2   2   2   2          2   2.72e-06   .0044247       2   .9994   1665.667 

Table 3. 9 is omitted since we don't have the individual data file.


Table 3.10 is omitted since LEM requires that the data be inputted as response vectors and this model requires that the data be inputted as frequencies of scores.


Table 4.3 on page 50 using left-right clinical scale data.

Model I: Proctor

  lat 1
  man 3
  dim 4 2 2 2
  lab X A B C 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2

  dat [170  0
         6  0
        73  1
       254 69]
  
  des [1 0 0 1 0 1 0 1    *A|X
       1 0 1 0 0 1 0 1    *B|X
       1 0 1 0 1 0 0 1 ]  *C|X

*** FREQUENCIES ***
  A B C     observed  estimated  std. res.
  1 1 1     170.000    169.437      0.043
  1 1 2       0.000      1.478     -1.216
  1 2 1       6.000      3.689      1.203
  1 2 2       0.000      0.607     -0.779
  2 1 1      73.000     72.887      0.013
  2 1 2       1.000      1.208     -0.190
  2 2 1     254.000    255.395     -0.087
  2 2 2      69.000     68.298      0.085

Model II: Intrusion-Omission Error 

  lat 1
  man 3
  dim 4 2 2 2
  lab X A B C 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2

  dat [170  0
         6  0
        73  1
       254 69]
  
  des [1 0 0 2 0 2 0 2    *A|X
       1 0 1 0 0 2 0 2    *B|X
       1 0 1 0 1 0 0 2 ]  *C|X
*** FREQUENCIES ***

  A B C     observed  estimated  std. res.
  1 1 1     170.000    170.000     -0.000
  1 1 2       0.000      0.021     -0.146
  1 2 1       6.000      4.550      0.680
  1 2 2       0.000      1.204     -1.097
  2 1 1      73.000     73.000      0.000
  2 1 2       1.000      1.204     -0.186
  2 2 1     254.000    255.450     -0.091
  2 2 2      69.000     67.571      0.174

Model III: Variable-Specific Error

Notice that we didn't impose constraint on the third variable since the program didn't converge in LEM. 

  lat 1
  man 3
  dim 4 2 2 2
  lab X A B C 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
  
 dat [170  0
        6  0
       73  1
      254 69]
  
  des [1 0 0 1 0 1 0 1    *A|X
       2 0 2 0 0 2 0 2    *B|X
       3 0 3 0 3 0 0 3 ]  *C|X
*** FREQUENCIES ***

  A B C     observed  estimated  std. res.
  1 1 1     170.000    170.992     -0.076
  1 1 2       0.000      0.000     -0.002
  1 2 1       6.000      5.008      0.443
  1 2 2       0.000      0.000     -0.010
  2 1 1      73.000     73.000     -0.000
  2 1 2       1.000      1.992     -0.703
  2 2 1     254.000    254.000     -0.000
  2 2 2      69.000     68.008      0.120

Model IV: Latent-Class Specific Error

As mentioned in the text, we have imposed constraint on the model.

  lat 1
  man 3
  dim 4 2 2 2
  lab X A B C 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
  
dat [170  0
       6  0
      73  1
     254 69]
  
   des [-1 0  0 2  0 3  0 -1   *A|X
        -1 0  2 0  0 3  0 -1   *B|X
        -1 0  2 0  3 0  0 -1]  *C|X
  
  sta A|X [0 1 .5 .5 .5 .5 1 0]
  sta B|X [0 1 .5 .5 .5 .5 1 0]
  sta C|X [0 1 .5 .5 .5 .5 1 0]

wla table4_1_out.dat
  A B C     observed  estimated  std. res.
  1 1 1     170.000    170.000      0.000
  1 1 2       0.000      0.014     -0.117
  1 2 1       6.000      5.736      0.110
  1 2 2       0.000      0.130     -0.360
  2 1 1      73.000     73.012     -0.001
  2 1 2       1.000      0.976      0.024
  2 2 1     254.000    254.132     -0.008
  2 2 2      69.000     69.000     -0.000

Last two columns of Table 4.1 from the output data file table4_1_out.dat created from the LEM code above. Notice that the order of the latent classes depends on the software and is reversed here as to the result in the book.

  1  1  1  170.0000 0.0000 0.0050 0.0008 0.9943   4 0.0057
  1  2  1    6.0000 0.0000 0.0019 0.9981 0.0000   3 0.0019
  2  1  1   73.0000 0.0000 0.9216 0.0784 0.0000   2 0.0784
  2  1  2    1.0000 0.0000 0.8674 0.1326 0.0000   2 0.1326
  2  2  1  254.0000 0.0000 0.0033 0.9967 0.0000   3 0.0033
  2  2  2   69.0000 0.9169 0.0002 0.0830 0.0000   1 0.0831

Table 4.2 on page 51 using the clinical scale data.

Model I: Proctor model

    lat 1
    man 3
    dim 4 2 2 2
    lab X A B C 
    mod X 
        A|X eq2
        B|X eq2
        C|X eq2
  
    dat [170  0
           6  0
          73  1
         254 69]
    
    des [1 0 0 1 0 1 0 1    *A|X
         1 0 1 0 0 1 0 1    *B|X
         1 0 1 0 1 0 0 1 ]  *C|X


*** STATISTICS ***

  Number of iterations = 125
  Converge criterion   = 0.0000008523
  Seed random values   = 5398

  X-squared            = 3.5860 (0.3098)
  L-squared            = 5.4403 (0.1423)
  Cressie-Read         = 3.9194 (0.2703)
  Dissimilarity index  = 0.0064
  Degrees of freedom   = 3
  Log-likelihood       = -746.10161
  Number of parameters = 4 (+1)
  Sample size          = 573.0
  BIC(L-squared)       = -13.6123
  AIC(L-squared)       = -0.5597
  BIC(log-likelihood)  = 1517.6068
  AIC(log-likelihood)  = 1500.2032

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4
         0.3024  0.1240  0.4553  0.1184
  A  1   0.9914  0.0086  0.0086  0.0086
  A  2   0.0086  0.9914  0.9914  0.9914
  B  1   0.9914  0.9914  0.0086  0.0086
  B  2   0.0086  0.0086  0.9914  0.9914
  C  1   0.9914  0.9914  0.9914  0.0086
  C  2   0.0086  0.0086  0.0086  0.9914

Model II: Intrusion-omission error model

  lat 1
    man 3
    dim 4 2 2 2
    lab X A B C 
    mod X 
        A|X eq2
        B|X eq2
        C|X eq2
  
    dat [170  0
           6  0
          73  1
         254 69]
    
    des [1 0 0 2 0 2 0 2    *A|X
         1 0 1 0 0 2 0 2    *B|X
         1 0 1 0 1 0 0 2 ]  *C|X


*** STATISTICS ***

  Number of iterations = 88
  Converge criterion   = 0.0000009873
  Seed random values   = 3689

  X-squared            = 1.7600 (0.4148)
  L-squared            = 2.9444 (0.2294)
  Cressie-Read         = 1.9907 (0.3696)
  Dissimilarity index  = 0.0050
  Degrees of freedom   = 2
  Log-likelihood       = -744.85363
  Number of parameters = 5 (+1)
  Sample size          = 573.0
  BIC(L-squared)       = -9.7574
  AIC(L-squared)       = -1.0556
  BIC(log-likelihood)  = 1521.4617
  AIC(log-likelihood)  = 1499.7073

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4
         0.2944  0.1216  0.4597  0.1243
  A  1   1.0000  0.0175  0.0175  0.0175
  A  2   0.0000  0.9825  0.9825  0.9825
  B  1   1.0000  1.0000  0.0175  0.0175
  B  2   0.0000  0.0000  0.9825  0.9825
  C  1   1.0000  1.0000  1.0000  0.0175
  C  2   0.0000  0.0000  0.0000  0.9825

Model III: Variable-specific error model

    lat 1
    man 3
    dim 4 2 2 2
    lab X A B C 
    mod X 
        A|X eq2
        B|X eq2
        C|X eq2
    
   dat [170  0
          6  0
         73  1
        254 69]
    
    des [1  0  0 1  0 1 0  1    *A|X
         2  0  2 0  0 2 0  2    *B|X
         -1 0 -1 0 -1 0 0 -1]   *C|X
  sta C|X [.9999 .0001 .9999 .0001 .9999 .0001 .0001 .9999]


*** STATISTICS ***

  Number of iterations = 241
  Converge criterion   = 0.0000009855
  Seed random values   = 6016

  X-squared            = 0.7334 (0.6930)
  L-squared            = 0.8522 (0.6531)
  Cressie-Read         = 0.7666 (0.6816)
  Dissimilarity index  = 0.0035
  Degrees of freedom   = 2
  Log-likelihood       = -743.80752
  Number of parameters = 5 (+1)
  Sample size          = 573.0
  BIC(L-squared)       = -11.8496
  AIC(L-squared)       = -3.1478
  BIC(log-likelihood)  = 1519.3695
  AIC(log-likelihood)  = 1497.6150

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4
         0.3072  0.1179  0.4528  0.1221
  A  1   1.0000  0.0000  0.0000  0.0000
  A  2   0.0000  1.0000  1.0000  1.0000
  B  1   0.9716  0.9716  0.0284  0.0284
  B  2   0.0284  0.0284  0.9716  0.9716
  C  1   0.9999  0.9999  0.9999  0.0001
  C  2   0.0001  0.0001  0.0001  0.9999

Model IV: Latent-class-specific model

  lat 1
  man 3
  dim 4 2 2 2
  lab X A B C 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
  
  
  rec 573
  dat table4_1_raw.dat
  
   des [-1 0  0 2  0 3  0 -1   *A|X
        -1 0  2 0  0 3  0 -1   *B|X
        -1 0  2 0  3 0  0 -1]  *C|X
  
  sta A|X [0 1 .5 .5 .5 .5 1 0]
  sta B|X [0 1 .5 .5 .5 .5 1 0]
  sta C|X [0 1 .5 .5 .5 .5 1 0]


*** STATISTICS ***

  Number of iterations = 30
  Converge criterion   = 0.0000006655
  Seed random values   = 5339

  X-squared            = 0.1559 (0.9250)
  L-squared            = 0.2989 (0.8612)
  Cressie-Read         = 0.1845 (0.9119)
  Dissimilarity index  = 0.0005
  Degrees of freedom   = 2
  Log-likelihood       = -743.53087
  Number of parameters = 5 (+1)
  Sample size          = 573.0
  BIC(L-squared)       = -12.4029
  AIC(L-squared)       = -3.7011
  BIC(log-likelihood)  = 1518.8162
  AIC(log-likelihood)  = 1497.0617

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4
         0.1104  0.1219  0.4727  0.2950
  A  1   0.0000  0.0124  0.0221  1.0000
  A  2   1.0000  0.9876  0.9779  0.0000
  B  1   0.0000  0.9876  0.0221  1.0000
  B  2   1.0000  0.0124  0.9779  0.0000
  C  1   0.0000  0.9876  0.9779  1.0000
  C  2   1.0000  0.0124  0.0221  0.0000

Table 4.3 on page 56 using Lazarsfeld-Stouffer Attitude data

Model I: One intrinsically unscalable class model

  lat 1
  man 4
  dim 6 2 2 2 2
  lab X A B C D 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
    
  dat [75  3
       42 10
       55  8
       45 16
       69 16
       60 25
       96 52
      199 229]  
  
    des [-1 0  0 -1  0 -1  0 -1 0 -1 0 0
         -1 0 -1  0  0 -1  0 -1 0 -1 0 0
         -1 0 -1  0 -1  0  0 -1 0 -1 0 0
         -1 0 -1  0 -1  0 -1  0 0 -1 0 0]
    
    sta A|X [1 0 0 1 0 1 0 1 0 1 .3 .7]
    sta B|X [1 0 1 0 0 1 0 1 0 1 .3 .7]
    sta C|X [1 0 1 0 1 0 0 1 0 1 .3 .7]
    sta D|X [1 0 1 0 1 0 1 0 0 1 .3 .7]


*** STATISTICS ***

  Number of iterations = 148
  Converge criterion   = 0.0000009812
  Seed random values   = 5371

  X-squared            = 26.0852 (0.0002)
  L-squared            = 26.5005 (0.0002)
  Cressie-Read         = 26.0954 (0.0002)
  Dissimilarity index  = 0.0412
  Degrees of freedom   = 6
  Log-likelihood       = -2357.21059
  Number of parameters = 9 (+1)
  Sample size          = 1000.0
  BIC(L-squared)       = -14.9460
  AIC(L-squared)       = 14.5005
  BIC(log-likelihood)  = 4776.5910
  AIC(log-likelihood)  = 4732.4212

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4    X  5    X  6
         0.0498  0.0112  0.0000  0.0789  0.1880  0.6721
  A  1   1.0000  0.0000  0.0000  0.0000  0.0000  0.3039
  A  2   0.0000  1.0000  1.0000  1.0000  1.0000  0.6961
  B  1   1.0000  1.0000  0.0000  0.0000  0.0000  0.3556
  B  2   0.0000  0.0000  1.0000  1.0000  1.0000  0.6444
  C  1   1.0000  1.0000  1.0000  0.0000  0.0000  0.4657
  C  2   0.0000  0.0000  0.0000  1.0000  1.0000  0.5343
  D  1   1.0000  1.0000  1.0000  1.0000  0.0000  0.7456
  D  2   0.0000  0.0000  0.0000  0.0000  1.0000  0.2544

Model II: Two intrinsically unscalable classes model

  lat 1
  man 4
  dim 7 2 2 2 2
  lab X A B C D 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  
  dat [75  3
       42 10
       55  8
       45 16
       69 16
       60 25
       96 52
      199 229]  
  
  des [-1 0  0 -1  0 -1  0 -1 0 -1 0 0 0 0
       -1 0 -1  0  0 -1  0 -1 0 -1 0 0 0 0
       -1 0 -1  0 -1  0  0 -1 0 -1 0 0 0 0
       -1 0 -1  0 -1  0 -1  0 0 -1 0 0 0 0]
  
  sta A|X [1 0 0 1 0 1 0 1 0 1 .3 .7 .4 .6]
  sta B|X [1 0 1 0 0 1 0 1 0 1 .3 .7 .4 .6]
  sta C|X [1 0 1 0 1 0 0 1 0 1 .3 .7 .4 .6]
  sta D|X [1 0 1 0 1 0 1 0 0 1 .3 .7 .4 .6]

*** STATISTICS ***

  Number of iterations = 3400
  Converge criterion   = 0.0000010000
  Seed random values   = 4829

  X-squared            = 4.0302 (0.0447)
  L-squared            = 3.5974 (0.0579)
  Cressie-Read         = 3.8694 (0.0492)
  Dissimilarity index  = 0.0079
  Degrees of freedom   = 1
  Log-likelihood       = -2345.75903
  Number of parameters = 14 (+1)
  Sample size          = 1000.0
  BIC(L-squared)       = -3.3104
  AIC(L-squared)       = 1.5974
  BIC(log-likelihood)  = 4788.2266
  AIC(log-likelihood)  = 4719.5181

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4    X  5    X  6    X  7
         0.0213  0.0000  0.0146  0.1145  0.1419  0.3008  0.4069
  A  1   1.0000  0.0000  0.0000  0.0000  0.0000  0.1565  0.4562
  A  2   0.0000  1.0000  1.0000  1.0000  1.0000  0.8435  0.5438
  B  1   1.0000  1.0000  0.0000  0.0000  0.0000  0.2355  0.5109
  B  2   0.0000  0.0000  1.0000  1.0000  1.0000  0.7645  0.4891
  C  1   1.0000  1.0000  1.0000  0.0000  0.0000  0.3571  0.5670
  C  2   0.0000  0.0000  0.0000  1.0000  1.0000  0.6429  0.4330
  D  1   1.0000  1.0000  1.0000  1.0000  0.0000  0.3108  0.9760
  D  2   0.0000  0.0000  0.0000  0.0000  1.0000  0.6892  0.0240

Model III: intrusion-omission error model

  lat 1
  man 4
  dim 5 2 2 2 2
  lab X A B C D 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  
  dat [75  3
       42 10
       55  8
       45 16
       69 16
       60 25
       96 52
      199 229]  
  
  des [1 0 0 2 0 2 0 2 0 2
       1 0 1 0 0 2 0 2 0 2
       1 0 1 0 1 0 0 2 0 2
       1 0 1 0 1 0 1 0 0 2]

*** STATISTICS ***

  Number of iterations = 109
  Converge criterion   = 0.0000009896
  Seed random values   = 2808

  X-squared            = 63.3048 (0.0000)
  L-squared            = 71.5053 (0.0000)
  Cressie-Read         = 65.0494 (0.0000)
  Dissimilarity index  = 0.0868
  Degrees of freedom   = 9
  Log-likelihood       = -2379.71298
  Number of parameters = 6 (+1)
  Sample size          = 1000.0
  BIC(L-squared)       = 9.3355
  AIC(L-squared)       = 53.5053
  BIC(log-likelihood)  = 4800.8725
  AIC(log-likelihood)  = 4771.4260

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4    X  5
         0.1927  0.0804  0.1274  0.3376  0.2619
  A  1   0.7870  0.1279  0.1279  0.1279  0.1279
  A  2   0.2130  0.8721  0.8721  0.8721  0.8721
  B  1   0.7870  0.7870  0.1279  0.1279  0.1279
  B  2   0.2130  0.2130  0.8721  0.8721  0.8721
  C  1   0.7870  0.7870  0.7870  0.1279  0.1279
  C  2   0.2130  0.2130  0.2130  0.8721  0.8721
  D  1   0.7870  0.7870  0.7870  0.7870  0.1279
  D  2   0.2130  0.2130  0.2130  0.2130  0.8721

Model IV: variable-specific error model

  lat 1
  man 4
  dim 5 2 2 2 2
  lab X A B C D 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  
  dat [75  3
       42 10
       55  8
       45 16
       69 16
       60 25
       96 52
      199 229]  
  
  des [1 0 0 1 0 1 0 1 0 1
       2 0 2 0 0 2 0 2 0 2
       3 0 3 0 3 0 0 3 0 3
       4 0 4 0 4 0 4 0 0 4]

*** STATISTICS ***

  Number of iterations = 192
  Converge criterion   = 0.0000009553
  Seed random values   = 778

  X-squared            = 42.6358 (0.0000)
  L-squared            = 43.6218 (0.0000)
  Cressie-Read         = 42.8195 (0.0000)
  Dissimilarity index  = 0.0767
  Degrees of freedom   = 7
  Log-likelihood       = -2365.77124
  Number of parameters = 8 (+1)
  Sample size          = 1000.0
  BIC(L-squared)       = -4.7325
  AIC(L-squared)       = 29.6218
  BIC(log-likelihood)  = 4786.8045
  AIC(log-likelihood)  = 4747.5425

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4    X  5
         0.1585  0.0620  0.0673  0.3562  0.3560
  A  1   0.8602  0.1398  0.1398  0.1398  0.1398
  A  2   0.1398  0.8602  0.8602  0.8602  0.8602
  B  1   0.8187  0.8187  0.1813  0.1813  0.1813
  B  2   0.1813  0.1813  0.8187  0.8187  0.8187
  C  1   0.7656  0.7656  0.7656  0.2344  0.2344
  C  2   0.2344  0.2344  0.2344  0.7656  0.7656
  D  1   0.9896  0.9896  0.9896  0.9896  0.0104
  D  2   0.0104  0.0104  0.0104  0.0104  0.9896

Model V: Intrusion-omission error and one intrinsically unscalable class model

  lat 1
  man 4
  dim  6 2 2 2 2
  lab X A B C D 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  
  dat [75  3
       42 10
       55  8
       45 16
       69 16
       60 25
       96 52
      199 229]  
  
    des [1 0 0 2 0 2 0 2 0 2 0 0
         1 0 1 0 0 2 0 2 0 2 0 0
         1 0 1 0 1 0 0 2 0 2 0 0
         1 0 1 0 1 0 1 0 0 2 0 0]


*** STATISTICS ***

  Number of iterations = 556
  Converge criterion   = 0.0000009916
  Seed random values   = 597

  X-squared            = 5.3657 (0.2518)
  L-squared            = 5.6425 (0.2275)
  Cressie-Read         = 5.4356 (0.2455)
  Dissimilarity index  = 0.0145
  Degrees of freedom   = 4
  Log-likelihood       = -2346.78156
  Number of parameters = 11 (+1)
  Sample size          = 1000.0
  BIC(L-squared)       = -21.9886
  AIC(L-squared)       = -2.3575
  BIC(log-likelihood)  = 4769.5484
  AIC(log-likelihood)  = 4715.5631

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4    X  5    X  6
         0.0197  0.0216  0.0660  0.2837  0.1252  0.4838
  A  1   1.0000  0.4537  0.4537  0.4537  0.4537  0.0256
  A  2   0.0000  0.5463  0.5463  0.5463  0.5463  0.9744
  B  1   1.0000  1.0000  0.4537  0.4537  0.4537  0.0791
  B  2   0.0000  0.0000  0.5463  0.5463  0.5463  0.9209
  C  1   1.0000  1.0000  1.0000  0.4537  0.4537  0.1708
  C  2   0.0000  0.0000  0.0000  0.5463  0.5463  0.8292
  D  1   1.0000  1.0000  1.0000  1.0000  0.4537  0.3993
  D  2   0.0000  0.0000  0.0000  0.0000  0.5463  0.6007

Model VI: Variable-specific error and one intrinsically unscalable class model

  lat 1
  man 4
  dim  6 2 2 2 2
  lab X A B C D 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  
  dat [75  3
       42 10
       55  8
       45 16
       69 16
       60 25
       96 52
      199 229]  
  
    des [1 0 0 1 0 1 0 1 0 1 0 0
         2 0 2 0 0 2 0 2 0 2 0 0
         3 0 3 0 3 0 0 3 0 3 0 0
         4 0 4 0 4 0 4 0 0 4 0 0]

*** STATISTICS ***

  Number of iterations = 536
  Converge criterion   = 0.0000009061
  Seed random values   = 2114

  X-squared            = 1.5933 (0.4508)
  L-squared            = 1.6237 (0.4440)
  Cressie-Read         = 1.6020 (0.4489)
  Dissimilarity index  = 0.0084
  Degrees of freedom   = 2
  Log-likelihood       = -2344.77216
  Number of parameters = 13 (+1)
  Sample size          = 1000.0
  BIC(L-squared)       = -12.1919
  AIC(L-squared)       = -2.3763
  BIC(log-likelihood)  = 4779.3451
  AIC(log-likelihood)  = 4715.5443

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4    X  5    X  6
         0.1800  0.0513  0.0902  0.1839  0.1449  0.3497
  A  1   0.7576  0.2424  0.2424  0.2424  0.2424  0.0103
  A  2   0.2424  0.7576  0.7576  0.7576  0.7576  0.9897
  B  1   0.6288  0.6288  0.3712  0.3712  0.3712  0.0000
  B  2   0.3712  0.3712  0.6288  0.6288  0.6288  1.0000
  C  1   0.6650  0.6650  0.6650  0.3350  0.3350  0.1356
  C  2   0.3350  0.3350  0.3350  0.6650  0.6650  0.8644
  D  1   1.0000  1.0000  1.0000  1.0000  0.0000  0.3879
  D  2   0.0000  0.0000  0.0000  0.0000  1.0000  0.6121

Table 4.4 on page 58 using Model VI in the example above. Notice that it should be model VI as explained in previous page instead of model V. The estimated frequency column is shown as a part of standard output of LEM. We used option wla to write the classification table to a text file.

    lat 1
    man 4
    dim  6 2 2 2 2
    lab X A B C D 
    mod X 
        A|X eq2
        B|X eq2
        C|X eq2
        D|X eq2
    
    dat [75  3
         42 10
         55  8
         45 16
         69 16
         60 25
         96 52
        199 229]  
    
      des [1 0 0 1 0 1 0 1 0 1 0 0
           2 0 2 0 0 2 0 2 0 2 0 0
           3 0 3 0 3 0 0 3 0 3 0 0
           4 0 4 0 4 0 4 0 0 4 0 0]
  
  wla table4_4.out

*** FREQUENCIES ***

  A B C D     observed  estimated  std. res.
  1 1 1 1      75.000     73.185      0.212
  1 1 1 2       3.000      4.370     -0.655
  1 1 2 1      42.000     45.070     -0.457
  1 1 2 2      10.000      8.679      0.448
  1 2 1 1      55.000     55.431     -0.058
  1 2 1 2       8.000      7.708      0.105
  1 2 2 1      45.000     42.958      0.312
  1 2 2 2      16.000     16.594     -0.146
  2 1 1 1      69.000     68.652      0.042
  2 1 1 2      16.000     13.649      0.636
  2 1 2 1      60.000     60.266     -0.034
  2 1 2 2      25.000     27.113     -0.406
  2 2 1 1      96.000     96.501     -0.051
  2 2 1 2      52.000     51.849      0.021
  2 2 2 1     199.000    198.937      0.004
  2 2 2 2     229.000    229.037     -0.002

We now copy and paste the content of data file table4_4.out to Stata do file editor and input them as data set. The model posterior probability is the largest of six possible class probabilities.

clear
input A B C D observed p1 p2 p3 p4 p5 p6 class error
  1  1  1  1   75.0000 0.7796 0.0708 0.0737 0.0758 0.0000 0.0000   1 0.2204
  1  1  1  2    3.0000 0.0002 0.0000 0.0000 0.0001 0.9997 0.0000   5 0.0003
  1  1  2  1   42.0000 0.6373 0.0579 0.0603 0.2445 0.0000 0.0000   1 0.3627
  1  1  2  2   10.0000 0.0000 0.0000 0.0000 0.0001 0.9999 0.0000   5 0.0001
  1  2  1  1   55.0000 0.6064 0.0551 0.1653 0.1699 0.0000 0.0033   1 0.3936
  1  2  1  2    8.0000 0.0001 0.0000 0.0000 0.0001 0.9620 0.0379   5 0.0380
  1  2  2  1   45.0000 0.3939 0.0358 0.1074 0.4355 0.0000 0.0274   4 0.5645
  1  2  2  2   16.0000 0.0000 0.0000 0.0000 0.0000 0.8877 0.1122   5 0.1123
  2  1  1  1   69.0000 0.2660 0.2359 0.2456 0.2525 0.0000 0.0000   1 0.7340
  2  1  1  2   16.0000 0.0000 0.0000 0.0000 0.0001 0.9999 0.0000   5 0.0001
  2  1  2  1   60.0000 0.1526 0.1353 0.1408 0.5713 0.0000 0.0000   4 0.4287
  2  1  2  2   25.0000 0.0000 0.0000 0.0000 0.0001 0.9999 0.0000   5 0.0001
  2  2  1  1   96.0000 0.1115 0.0989 0.2966 0.3049 0.0000 0.1882   4 0.6951
  2  2  1  2   52.0000 0.0000 0.0000 0.0000 0.0000 0.4468 0.5532   6 0.4468
  2  2  2  1  199.0000 0.0272 0.0241 0.0724 0.2938 0.0000 0.5824   6 0.4176
  2  2  2  2  229.0000 0.0000 0.0000 0.0000 0.0000 0.2009 0.7991   6 0.2009
end
egen postp=rmax(p1-p6)
sort D C B A
list A B C D observed class postp, clean
       A   B   C   D   observed   class   postp  
  1.   1   1   1   1         75       1   .7796  
  2.   2   1   1   1         69       1    .266  
  3.   1   2   1   1         55       1   .6064  
  4.   2   2   1   1         96       4   .3049  
  5.   1   1   2   1         42       1   .6373  
  6.   2   1   2   1         60       4   .5713  
  7.   1   2   2   1         45       4   .4355  
  8.   2   2   2   1        199       6   .5824  
  9.   1   1   1   2          3       5   .9997  
 10.   2   1   1   2         16       5   .9999  
 11.   1   2   1   2          8       5    .962  
 12.   2   2   1   2         52       6   .5532  
 13.   1   1   2   2         10       5   .9999  
 14.   2   1   2   2         25       5   .9999  
 15.   1   2   2   2         16       5   .8877  
 16.   2   2   2   2        229       6   .7991 

Result on page 60 using Stouffer-Toby role conflict data of Table 4.6 on page 61. In this example, we show how to specify using a data set with individual records. You can download the data file following the link here.

  lat 1
  man 4
  dim 5 2 2 2 2
  lab X A B C D 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  
  rec 216
  dat table4_6_raw.dat
  
  des [1 0 0 1 0 1 0 1 0 1
       3 0 3 0 0 4 0 4 0 4
       5 0 5 0 5 0 6 0 6 0
       7 0 7 0 7 0 7 0 0 7] 

*** STATISTICS ***

  Number of iterations = 158
  Converge criterion   = 0.0000009505
  Seed random values   = 3566

  X-squared            = 0.8954 (0.9706)
  L-squared            = 0.9210 (0.9687)
  Cressie-Read         = 0.9032 (0.9700)
  Dissimilarity index  = 0.0194
  Degrees of freedom   = 5
  Log-likelihood       = -503.56819
  Number of parameters = 10 (+1)
  Sample size          = 216.0
  BIC(L-squared)       = -25.9554
  AIC(L-squared)       = -9.0790
  BIC(log-likelihood)  = 1060.8892
  AIC(log-likelihood)  = 1027.1364

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4    X  5
         0.2391  0.0179  0.1026  0.4390  0.2014
  A  1   0.8639  0.1361  0.1361  0.1361  0.1361
  A  2   0.1361  0.8639  0.8639  0.8639  0.8639
  B  1   0.9477  0.9477  0.3638  0.3638  0.3638
  B  2   0.0523  0.0523  0.6362  0.6362  0.6362
  C  1   0.9398  0.9398  0.9398  0.2531  0.2531
  C  2   0.0602  0.0602  0.0602  0.7469  0.7469
  D  1   0.9884  0.9884  0.9884  0.9884  0.0116
  D  2   0.0116  0.0116  0.0116  0.0116  0.9884

  Model of one intrinsically unscalable class is fitted to Stouffer-Toby data (page 61).

  lat 1
  man 4
  dim 6 2 2 2 2
  lab X A B C D 
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  rec 216
  dat table4_6_raw.dat
  des [-1 0  0 -1  0 -1  0 -1 0 -1 0 0
       -1 0 -1  0  0 -1  0 -1 0 -1 0 0
       -1 0 -1  0 -1  0  0 -1 0 -1 0 0
       -1 0 -1  0 -1  0 -1  0 0 -1 0 0]      
  sta A|X [1 0 0 1 0 1 0 1 0 1 .3 .7]
  sta B|X [1 0 1 0 0 1 0 1 0 1 .3 .7]
  sta C|X [1 0 1 0 1 0 0 1 0 1 .3 .7]
  sta D|X [1 0 1 0 1 0 1 0 0 1 .3 .7]
*** STATISTICS ***
  Number of iterations = 168
  Converge criterion   = 0.0000009558
  Seed random values   = 455
  X-squared            = 1.0053 (0.9854)
  L-squared            = 0.9885 (0.9860)
  Cressie-Read         = 0.9988 (0.9857)
  Dissimilarity index  = 0.0135
  Degrees of freedom   = 6
  Log-likelihood       = -503.60197
  Number of parameters = 9 (+1)
  Sample size          = 216.0
  BIC(L-squared)       = -31.2631
  AIC(L-squared)       = -11.0115
  BIC(log-likelihood)  = 1055.5815
  AIC(log-likelihood)  = 1025.2039
*** LATENT CLASS OUTPUT ***
          X  1    X  2    X  3    X  4    X  5    X  6
         0.1771  0.0350  0.0257  0.0318  0.0484  0.6820
  A  1   1.0000  0.0000  0.0000  0.0000  0.0000  0.1951
  A  2   0.0000  1.0000  1.0000  1.0000  1.0000  0.8049
  B  1   1.0000  1.0000  0.0000  0.0000  0.0000  0.4425
  B  2   0.0000  0.0000  1.0000  1.0000  1.0000  0.5575
  C  1   1.0000  1.0000  1.0000  0.0000  0.0000  0.3845
  C  2   0.0000  0.0000  0.0000  1.0000  1.0000  0.6155
  D  1   1.0000  1.0000  1.0000  1.0000  0.0000  0.7655
  D  2   0.0000  0.0000  0.0000  0.0000  1.0000  0.2345

Latent markov model example --  to be done


Located latent class model example -- to be done


T-class mixture model example -- to be done


Table 5.1 on page 69 using IEA bus data

Model I: linear scale

  lat 1
  man 4
  dim 5 2 2 2 2
  lab X A B C D
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  
  dat [1138  13
         75  15
        502   9
        198  23
       1532  43
        200  59
       1354  37
        852 309]
  
   des [1 0  0 1  0 1  0 1  0  1   *A|X
        2 0  2 0  0 2  0 2  0  2   *B|X
        3 0  3 0  3 0  0 3  0  3   *C|X
        4 0  4 0  4 0  4 0  0  4]  *D|X


*** STATISTICS ***

  Number of iterations = 105
  Converge criterion   = 0.0000009275
  Seed random values   = 2318

  X-squared            = 40.3671 (0.0000)
  L-squared            = 46.8495 (0.0000)
  Cressie-Read         = 42.2199 (0.0000)
  Dissimilarity index  = 0.0163
  Degrees of freedom   = 7
  Log-likelihood       = -12930.79076
  Number of parameters = 8 (+1)
  Sample size          = 6359.0
  BIC(L-squared)       = -14.4539
  AIC(L-squared)       = 32.8495
  BIC(log-likelihood)  = 25931.6425
  AIC(log-likelihood)  = 25877.5815


*** FREQUENCIES ***

  A B C D     observed  estimated  std. res.
  1 1 1 1    1138.000   1148.738     -0.317
  1 1 1 2      13.000     23.308     -2.135
  1 1 2 1      75.000     69.801      0.622
  1 1 2 2      15.000     13.094      0.527
  1 2 1 1     502.000    467.044      1.617
  1 2 1 2       9.000     10.886     -0.572
  1 2 2 1     198.000    182.668      1.134
  1 2 2 2      23.000     57.460     -4.546
  2 1 1 1    1532.000   1532.935     -0.024
  2 1 1 2      43.000     32.269      1.889
  2 1 2 1     200.000    220.747     -1.396
  2 1 2 2      59.000     60.592     -0.205
  2 2 1 1    1354.000   1376.752     -0.613
  2 2 1 2      37.000     34.955      0.346
  2 2 2 1     852.000    852.315     -0.011
  2 2 2 2     309.000    275.435      2.022

To create the column next to the estimated frequencies, we again use Stata.

clear
input   A B C D     observed  estimated  std_res
  1 1 1 1    1138.000   1148.738     -0.317
  1 1 1 2      13.000     23.308     -2.135
  1 1 2 1      75.000     69.801      0.622
  1 1 2 2      15.000     13.094      0.527
  1 2 1 1     502.000    467.044      1.617
  1 2 1 2       9.000     10.886     -0.572
  1 2 2 1     198.000    182.668      1.134
  1 2 2 2      23.000     57.460     -4.546
  2 1 1 1    1532.000   1532.935     -0.024
  2 1 1 2      43.000     32.269      1.889
  2 1 2 1     200.000    220.747     -1.396
  2 1 2 2      59.000     60.592     -0.205
  2 2 1 1    1354.000   1376.752     -0.613
  2 2 1 2      37.000     34.955      0.346
  2 2 2 1     852.000    852.315     -0.011
  2 2 2 2     309.000    275.435      2.022
end
gen pearson = (observed - estimated)^2/estimated
egen ttl = sum(pearson)
sort D C B A
list, clean
       A   B   C   D   observed   estima~d   std_res    pearson        ttl  
  1.   1   1   1   1       1138   1148.738     -.317   .1003757   40.36703  
  2.   2   1   1   1       1532   1532.935     -.024   .0005704   40.36703  
  3.   1   2   1   1        502    467.044     1.617   2.616288   40.36703  
  4.   2   2   1   1       1354   1376.752     -.613   .3759947   40.36703  
  5.   1   1   2   1         75     69.801      .622   .3872376   40.36703  
  6.   2   1   2   1        200    220.747    -1.396   1.949914   40.36703  
  7.   1   2   2   1        198    182.668     1.134   1.286872   40.36703  
  8.   2   2   2   1        852    852.315     -.011   .0001164   40.36703  
  9.   1   1   1   2         13     23.308    -2.135    4.55873   40.36703  
 10.   2   1   1   2         43     32.269     1.889   3.568575   40.36703  
 11.   1   2   1   2          9     10.886     -.572   .3267495   40.36703  
 12.   2   2   1   2         37     34.955      .346     .11964   40.36703  
 13.   1   1   2   2         15     13.094      .527   .2774428   40.36703  
 14.   2   1   2   2         59     60.592     -.205   .0418283   40.36703  
 15.   1   2   2   2         23      57.46    -4.546    20.6664   40.36703  
 16.   2   2   2   2        309    275.435     2.022   4.090292   40.36703  

Model II: biform scale

  lat 1
  man 4
  dim 6 2 2 2 2
  lab X A B C D
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  
  dat [1138  13
         75  15
        502   9
        198  23
       1532  43
        200  59
       1354  37
        852 309]
  
   des [1 0  0 1  1 0  0 1 0 1 0 1     *A|X
        2 0  2 0  0 2  0 2 0 2 0 2     *B|X
        3 0  3 0  3 0  3 0 0 3 0 3     *C|X
        4 0  4 0  4 0  4 0 4 0 0 4]    *D|X
  
*** STATISTICS ***

  Number of iterations = 178
  Converge criterion   = 0.0000009704
  Seed random values   = 2653

  X-squared            = 35.2821 (0.0000)
  L-squared            = 39.6006 (0.0000)
  Cressie-Read         = 36.5025 (0.0000)
  Dissimilarity index  = 0.0127
  Degrees of freedom   = 6
  Log-likelihood       = -12927.16632
  Number of parameters = 9 (+1)
  Sample size          = 6359.0
  BIC(L-squared)       = -12.9452
  AIC(L-squared)       = 27.6006
  BIC(log-likelihood)  = 25933.1513
  AIC(log-likelihood)  = 25872.3326

*** FREQUENCIES ***

  A B C D     observed  estimated  std. res.
  1 1 1 1    1138.000   1130.527      0.222
  1 1 1 2      13.000     22.054     -1.928
  1 1 2 1      75.000     73.252      0.204
  1 1 2 2      15.000     10.609      1.348
  1 2 1 1     502.000    503.105     -0.049
  1 2 1 2       9.000     11.647     -0.775
  1 2 2 1     198.000    169.048      2.227
  1 2 2 2      23.000     52.757     -4.097
  2 1 1 1    1532.000   1538.874     -0.175 
  2 1 1 2      43.000     31.549      2.039
  2 1 2 1     200.000    213.575     -0.929
  2 1 2 2      59.000     54.525      0.606
  2 2 1 1    1354.000   1353.022      0.027
  2 2 1 2      37.000     36.893      0.018
  2 2 2 1     852.000    869.598     -0.597
  2 2 2 2     309.000    287.966      1.240

We skip the part of creating the column labeled "Discr II" next to the estimated frequencies as it is the same way as we have shown for model I.

Model III: augmented biform

  lat 1
  man 4
  dim 7 2 2 2 2
  lab X A B C D
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  
  dat [1138  13
         75  15
        502   9
        198  23
       1532  43
        200  59
       1354  37
        852 309]
  
   des [1 0  0 1  0 1  0 1 0 1 1 0 1 0     *A|X
        2 0  2 0  0 2  0 2 0 2 0 2 0 2     *B|X
        3 0  3 0  3 0  0 3 0 3 3 0 0 3     *C|X
        4 0  4 0  4 0  4 0 0 4 4 0 4 0]    *D|X

*** STATISTICS ***

  Number of iterations = 174
  Converge criterion   = 0.0000009458
  Seed random values   = 5390

  X-squared            = 20.7938 (0.0009)
  L-squared            = 18.5343 (0.0023)
  Cressie-Read         = 19.8692 (0.0013)
  Dissimilarity index  = 0.0060
  Degrees of freedom   = 5
  Log-likelihood       = -12916.63316
  Number of parameters = 10 (+1)
  Sample size          = 6359.0
  BIC(L-squared)       = -25.2539
  AIC(L-squared)       = 8.5343
  BIC(log-likelihood)  = 25920.8426
  AIC(log-likelihood)  = 25853.2663

*** FREQUENCIES ***

  A B C D     observed  estimated  std. res.
  1 1 1 1    1138.000   1130.049      0.237
  1 1 1 2      13.000     22.502     -2.003
  1 1 2 1      75.000     74.036      0.112
  1 1 2 2      15.000      6.614      3.261
  1 2 1 1     502.000    500.219      0.080
  1 2 1 2       9.000     10.783     -0.543
  1 2 2 1     198.000    198.526     -0.037
  1 2 2 2      23.000     30.282     -1.323
  2 1 1 1    1532.000   1539.779     -0.198
  2 1 1 2      43.000     32.363      1.870
  2 1 2 1     200.000    208.673     -0.600
  2 1 2 2      59.000     61.000     -0.256
  2 2 1 1    1354.000   1354.613     -0.017
  2 2 1 2      37.000     36.699      0.050
  2 2 2 1     852.000    845.105      0.237
  2 2 2 2     309.000    307.757      0.071

Table 5.2 on page 70 based on model III in the previous example.

  lat 1
  man 4
  dim 7 2 2 2 2
  lab X A B C D
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  dat [1138  13
         75  15
        502   9
        198  23
       1532  43
        200  59
       1354  37
        852 309]
   des [1 0  0 1  0 1  0 1 0 1 1 0 1 0     *A|X
        2 0  2 0  0 2  0 2 0 2 0 2 0 2     *B|X
        3 0  3 0  3 0  0 3 0 3 3 0 0 3     *C|X
        4 0  4 0  4 0  4 0 0 4 4 0 4 0]    *D|X
*** LATENT CLASS OUTPUT ***
          X  1    X  2    X  3    X  4    X  5    X  6    X  7
         0.2078  0.2678  0.2273  0.1695  0.0630  0.0409  0.0238
  A  1   0.9170  0.0830  0.0830  0.0830  0.0830  0.9170  0.9170
  A  2   0.0830  0.9170  0.9170  0.9170  0.9170  0.0830  0.0830
  B  1   0.8365  0.8365  0.1635  0.1635  0.1635  0.1635  0.1635
  B  2   0.1635  0.1635  0.8365  0.8365  0.8365  0.8365  0.8365
  C  1   0.9670  0.9670  0.9670  0.0330  0.0330  0.9670  0.0330
  C  2   0.0330  0.0330  0.0330  0.9670  0.9670  0.0330  0.9670
  D  1   0.9806  0.9806  0.9806  0.9806  0.0194  0.9806  0.9806
  D  2   0.0194  0.0194  0.0194  0.0194  0.9806  0.0194  0.0194

Table 5.3 on page 71 based on model III in the previous example. We use option wla to write out the latent classification and error probabilities to a file and we show the content of the file after the LEM code. The first five columns are manifest variables and their frequencies. The next seven columns are posterior probabilities for each class. Based on the posterior probabilities, the classification is determined. The modal posterior probabilities is the maximum of the seven probabilities. The last column is the error probabilities, which is 1-modal posterior probabilities.

  lat 1
  man 4
  dim 7 2 2 2 2
  lab X A B C D
  mod X 
      A|X eq2
      B|X eq2
      C|X eq2
      D|X eq2
  
  dat [1138  13
         75  15
        502   9
        198  23
       1532  43
        200  59
       1354  37
        852 309]
  
   des [1 0  0 1  0 1  0 1 0 1 1 0 1 0     *A|X
        2 0  2 0  0 2  0 2 0 2 0 2 0 2     *B|X
        3 0  3 0  3 0  0 3 0 3 3 0 0 3     *C|X
        4 0  4 0  4 0  4 0 0 4 4 0 4 0]    *D|X

wla table5_3.out
  1  1  1  1 1138.0000 0.8505 0.0993 0.0165 0.0004 0.0000 0.0327 0.0007   1 0.1495
  1  1  1  2   13.0000 0.8439 0.0985 0.0163 0.0004 0.0078 0.0324 0.0006   1 0.1561
  1  1  2  1   75.0000 0.4432 0.0517 0.0086 0.1874 0.0014 0.0170 0.2907   1 0.5568
  1  1  2  2   15.0000 0.0980 0.0114 0.0019 0.0415 0.7791 0.0038 0.0643   5 0.2209
  1  2  1  1  502.0000 0.3755 0.0438 0.1903 0.0048 0.0000 0.3780 0.0075   6 0.6220
  1  2  1  2    9.0000 0.3441 0.0402 0.1744 0.0044 0.0835 0.3465 0.0069   6 0.6535
  1  2  2  1  198.0000 0.0323 0.0038 0.0164 0.3577 0.0026 0.0325 0.5547   7 0.4453
  1  2  2  2   23.0000 0.0042 0.0005 0.0021 0.0463 0.8708 0.0042 0.0719   5 0.1292
  2  1  1  1 1532.0000 0.0565 0.8044 0.1334 0.0034 0.0000 0.0022 0.0000   2 0.1956
  2  1  1  2   43.0000 0.0531 0.7562 0.1254 0.0032 0.0600 0.0020 0.0000   2 0.2438
  2  1  2  1  200.0000 0.0142 0.2026 0.0336 0.7342 0.0054 0.0005 0.0093   4 0.2658
  2  1  2  2   59.0000 0.0010 0.0137 0.0023 0.0496 0.9328 0.0000 0.0006   5 0.0672
  2  2  1  1 1354.0000 0.0126 0.1787 0.7760 0.0198 0.0001 0.0126 0.0003   3 0.2240
  2  2  1  2   37.0000 0.0092 0.1303 0.5659 0.0144 0.2708 0.0092 0.0002   3 0.4341
  2  2  2  1  852.0000 0.0007 0.0098 0.0425 0.9278 0.0068 0.0007 0.0118   4 0.0722
  2  2  2  2  309.0000 0.0000 0.0005 0.0023 0.0503 0.9461 0.0000 0.0006   5 0.0539

Table 6. 2 on page 76 using table6_1_g.dat, table6_1_f.dat and table6_1_m.dat.

Model I: Combined group analysis

  lat 1
  man 5
  dim 2 2 2 2 2 2
  lab X A B C D G
  mod   
      G
      A|X 
      B|X 
      C|X 
      D|X
  
  rec 317
  dat table6_1_g.dat
  
*** STATISTICS ***

  Number of iterations = 441
  Converge criterion   = 0.0000009597
  Seed random values   = 1059

  X-squared            = 24.8712 (0.2528)
  L-squared            = 28.8872 (0.1167)
  Cressie-Read         = 25.4289 (0.2291)
  Dissimilarity index  = 0.0856
  Degrees of freedom   = 21
  Log-likelihood       = -653.10120
  Number of parameters = 10 (+1)
  Sample size          = 317.0
  BIC(L-squared)       = -92.0497
  AIC(L-squared)       = -13.1128
  BIC(log-likelihood)  = 1363.7914
  AIC(log-likelihood)  = 1326.2024
*** (CONDITIONAL) PROBABILITIES ***
* P(X) *
  1              0.8353  (0.0796)
  2              0.1647  (0.0796)
* P(G) *
  1              0.4322  (0.0278)
  2              0.5678  (0.0278)
* P(A|X) *
  1 | 1          0.9837  (0.0291)
  2 | 1          0.0163  (0.0291)
  1 | 2          0.4314  (0.1779)
  2 | 2          0.5686  (0.1779)
* P(B|X) *
  1 | 1          0.9761  (0.0312)
  2 | 1          0.0239  (0.0312)
  1 | 2          0.4128  (0.1754)
  2 | 2          0.5872  (0.1754)
* P(C|X) *
  1 | 1          0.9629  (0.0153)
  2 | 1          0.0371  (0.0153)
  1 | 2          0.7858  (0.0838)
  2 | 2          0.2142  (0.0838)
* P(D|X) *
  1 | 1          0.8174  (0.0264)
  2 | 1          0.1826  (0.0264)
  1 | 2          0.6236  (0.0996)
  2 | 2          0.3764  (0.0996)
Model II: Analysis on female group
  lat 1
  man 4
  dim 2 2 2 2 2 
  lab X A B C D 
  mod   
      X
      A|X 
      B|X 
      C|X 
      D|X
  
  rec 180
  dat table6_1_f.dat

*** STATISTICS ***

  Number of iterations = 360
  Converge criterion   = 0.0000009611
  Seed random values   = 5364

  X-squared            = 7.3025 (0.2938)
  L-squared            = 8.6605 (0.1936)
  Cressie-Read         = 7.4207 (0.2837)
  Dissimilarity index  = 0.0362
  Degrees of freedom   = 6
  Log-likelihood       = -273.20739
  Number of parameters = 9 (+1)
  Sample size          = 180.0
  BIC(L-squared)       = -22.4972
  AIC(L-squared)       = -3.3395
  BIC(log-likelihood)  = 593.1514
  AIC(log-likelihood)  = 564.4148

*** (CONDITIONAL) PROBABILITIES ***

* P(X) *

1 0.8534 (0.1022)
2 0.1466 (0.1022)

* P(A|X) *

1 | 1 0.9802 (0.0421)
2 | 1 0.0198 (0.0421)
1 | 2 0.3570 (0.2975)
2 | 2 0.6430 (0.2975)

* P(B|X) *

1 | 1 0.9365 (0.0514)
2 | 1 0.0635 (0.0514)
1 | 2 0.3085 (0.2616)
2 | 2 0.6915 (0.2616)

* P(C|X) *

1 | 1 0.9410 (0.0214)
2 | 1 0.0590 (0.0214)
1 | 2 0.8126 (0.1053)
2 | 2 0.1874 (0.1053)

* P(D|X) *

1 | 1 0.7909 (0.0355)
2 | 1 0.2091 (0.0355)
1 | 2 0.6255 (0.1313)
2 | 2 0.3745 (0.1313)

Model III: Analysis on male group

  lat 1
  man 4
  dim 2 2 2 2 2 
  lab X A B C D 
  mod   
      X
      A|X 
      B|X 
      C|X 
      D|X
  
  
  rec 137
  dat table6_1_m.dat
  
*** STATISTICS ***

  Number of iterations = 157
  Converge criterion   = 0.0000008803
  Seed random values   = 4014

  X-squared            = 5.5237 (0.4786)
  L-squared            = 6.3978 (0.3801)
  Cressie-Read         = 5.5771 (0.4722)
  Dissimilarity index  = 0.0284
  Degrees of freedom   = 6
  Log-likelihood       = -156.17714
  Number of parameters = 9 (+1)
  Sample size          = 137.0
  BIC(L-squared)       = -23.1220
  AIC(L-squared)       = -5.6022
  BIC(log-likelihood)  = 356.6341
  AIC(log-likelihood)  = 330.3543

*** (CONDITIONAL) PROBABILITIES ***

* P(X) *

  1              0.8545  (0.0595)
  2              0.1455  (0.0595)

* P(A|X) *

  1 | 1          0.9775  (0.0281)
  2 | 1          0.0225  (0.0281)
  1 | 2          0.4299  (0.1702)
  2 | 2          0.5701  (0.1702)

* P(B|X) *

  1 | 1          1.0000  (0.0000) *
  2 | 1          0.0000  (0.0000) *
  1 | 2          0.5486  (0.1943)
  2 | 2          0.4514  (0.1943)

* P(C|X) *

  1 | 1          0.9944  (0.0184)
  2 | 1          0.0056  (0.0184)
  1 | 2          0.6817  (0.1341)
  2 | 2          0.3183  (0.1341)

* P(D|X) *

  1 | 1          0.8553  (0.0364)
  2 | 1          0.1447  (0.0364)
  1 | 2          0.5457  (0.1425)
  2 | 2          0.4543  (0.1425)

Table 6.4 on page 80 using spatial data, table6_3.dat, table6_3_f.dat and table6_3_m.dat.

Model I: analysis on male group

    lat 1
    man 3
    dim 4 2 2 2 
    lab X A B C 
    mod 
        A|X eq2
        B|X eq2
        C|X eq2
          
    rec 266
    dat table6_3_m.dat
    
    des [ 1 0  0 2  0 2  0 2    
          1 0  1 0  0 2  0 2 
          1 0  1 0  1 0  0 2]

*** STATISTICS ***

  Number of iterations = 71
  Converge criterion   = 0.0000008951
  Seed random values   = 5672

  X-squared            = 2.8375 (0.2420)
  L-squared            = 4.4355 (0.1089)
  Cressie-Read         = 3.1347 (0.2086)
  Dissimilarity index  = 0.0132
  Degrees of freedom   = 2
  Log-likelihood       = -360.82348
  Number of parameters = 5 (+1)
  Sample size          = 266.0
  BIC(L-squared)       = -6.7315
  AIC(L-squared)       = 0.4355
  BIC(log-likelihood)  = 749.5644
  AIC(log-likelihood)  = 731.6470

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4
         0.3033  0.1579  0.4226  0.1162
  A  1   1.0000  0.0294  0.0294  0.0294
  A  2   0.0000  0.9706  0.9706  0.9706
  B  1   1.0000  1.0000  0.0294  0.0294
  B  2   0.0000  0.0000  0.9706  0.9706
  C  1   1.0000  1.0000  1.0000  0.0294
  C  2   0.0000  0.0000  0.0000  0.9706

Model II: analysis on female group

    lat 1
    man 3
    dim 4 2 2 2 
    lab X A B C 
    mod 
        A|X eq2
        B|X eq2
        C|X eq2
          
    rec 307
    dat table6_3_f.dat
    
    des [ 1 0  0 2  0 2  0 2    
          1 0  1 0  0 2  0 2 
          1 0  1 0  1 0  0 2]

*** STATISTICS ***

  Number of iterations = 127
  Converge criterion   = -0.0000000092
  Seed random values   = 642

  X-squared            = 1.6669 (0.4345)
  L-squared            = 1.5951 (0.4504)
  Cressie-Read         = 1.5679 (0.4566)
  Dissimilarity index  = 0.0031
  Degrees of freedom   = 2
  Log-likelihood       = -378.81571
  Number of parameters = 5 (+1)
  Sample size          = 307.0
  BIC(L-squared)       = -9.8585
  AIC(L-squared)       = -2.4049
  BIC(log-likelihood)  = 786.2657
  AIC(log-likelihood)  = 767.6314

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4
         0.2858  0.0910  0.4917  0.1314
  A  1   1.0000  0.0087  0.0087  0.0087
  A  2   0.0000  0.9913  0.9913  0.9913
  B  1   1.0000  1.0000  0.0087  0.0087
  B  2   0.0000  0.0000  0.9913  0.9913
  C  1   1.0000  1.0000  1.0000  0.0087
  C  2   0.0000  0.0000  0.0000  0.9913

Model III: Combined analysis. The results here are a little off here, especially the p-value.

    lat 1
    man 4
    dim 4 2 2 2 2
    lab X A B C G
    mod G   eq2
        A|X eq2
        B|X eq2
        C|X eq2
          
    rec 573
    dat table6_3.dat
    
    des [-1 0
          1 0  0 2  0 2  0 2    
          1 0  1 0  0 2  0 2 
          1 0  1 0  1 0  0 2]
  
  sta G [.5 .5]
  
  *** STATISTICS ***

  Number of iterations = 87
  Converge criterion   = 0.0000009668
  Seed random values   = 3147

  X-squared            = 18.1788 (0.0520)
  L-squared            = 19.3957 (0.0355)
  Cressie-Read         = 18.2839 (0.0504)
  Dissimilarity index  = 0.0700
  Degrees of freedom   = 10
  Log-likelihood       = -1142.02696
  Number of parameters = 5 (+1)
  Sample size          = 573.0
  BIC(L-squared)       = -44.1132
  AIC(L-squared)       = -0.6043
  BIC(log-likelihood)  = 2315.8084
  AIC(log-likelihood)  = 2294.0539

*** LATENT CLASS OUTPUT ***

          X  1    X  2    X  3    X  4
         0.2944  0.1216  0.4597  0.1243
  A  1   1.0000  0.0175  0.0175  0.0175
  A  2   0.0000  0.9825  0.9825  0.9825
  B  1   1.0000  1.0000  0.0175  0.0175
  B  2   0.0000  0.0000  0.9825  0.9825
  C  1   1.0000  1.0000  1.0000  0.0175
  C  2   0.0000  0.0000  0.0000  0.9825
  G  1   0.5000  0.5000  0.5000  0.5000
  G  2   0.5000  0.5000  0.5000  0.5000

How to cite this page

Report an error on this page

UCLA Researchers are invited to our Statistical Consulting Services
We recommend others to our list of Other Resources for Statistical Computing Help
These pages are Copyrighted (c) by UCLA Academic Technology Services


The content of this web site should not be construed as an endorsement of any particular web site, book, or software product by the University of California