Lavaan Warning during CFA Analysis with Ordinal Data (vcov does not appear to be positive definite)

681 views
Skip to first unread message

Erdem Uygun

unread,
Jan 2, 2021, 4:28:26 PM1/2/21
to lavaan
Hi to all,

In AMOS, I performed a CFA with poor fit on 5-point-Likert Scale data(From Strongly Disagree to Strongly Agree). In fact, results are mostly accumulated on Agree (4) and Strongly Agree (5) options giving skewness and kurtosis to data. I performed CFA with lavaan by putting "ordered" command to treat answers as ordinal and performed calculations with automatically selected WLSMV estimator.

Results enhanced significantly except the warning below:

Warning message:
In lav_model_vcov(lavmodel = lavmodel, lavsamplestats = lavsamplestats,  :
  lavaan WARNING:
    The variance-covariance matrix of the estimated parameters (vcov)
    does not appear to be positive definite! The smallest eigenvalue
    (= -4.193201e-18) is smaller than zero. This may be a symptom that
    the model is not identified.

What does that mean? Should I ignore the warning and report what I have found or should I do something else to eliminate it? I put the whole lavaan results below. I spent hours on that issue but could not find any practical steps to be taken so far. Thank you in advance I really appreciate your help.

Lavaan Result
model<-'
+ IGA=~Q1+Q2+Q3+Q4+Q5+Q6+Q7+Q8
+ KATW=~Q9+Q10+Q11+Q12
+ KATC=~Q13+Q14+Q15+Q16+Q17+Q18+Q19+Q20+Q21+Q22+Q23
+ IM=~Q24+Q25+Q26+Q27+Q28+Q29+Q30+Q31
+ IS=~Q32+Q33+Q34+Q35+Q36'
>
> fit<-cfa(model,data=CFA_full,ordered=c("Q1","Q2","Q3","Q4","Q5","Q6","Q7","Q8","Q9","Q10","Q11","Q12","Q13","Q14","Q15","Q16","Q17","Q18","Q19","Q20","Q21","Q22","Q23","Q24","Q25","Q26","Q27","Q28","Q29","Q30","Q31","Q32","Q33","Q34","Q35","Q36"))
Warning message:
In lav_model_vcov(lavmodel = lavmodel, lavsamplestats = lavsamplestats,  :
  lavaan WARNING:
    The variance-covariance matrix of the estimated parameters (vcov)
    does not appear to be positive definite! The smallest eigenvalue
    (= -4.193201e-18) is smaller than zero. This may be a symptom that
    the model is not identified.

>
> summary(fit,fit.measures=TRUE,standardized = TRUE)
lavaan 0.6-7 ended normally after 125 iterations

  Estimator                                       DWLS
  Optimization method                           NLMINB
  Number of free parameters                        189
                                                      
  Number of observations                          1165
                                                      
Model Test User Model:
                                              Standard      Robust
  Test Statistic                              3339.255    4326.233
  Degrees of freedom                               584         584
  P-value (Chi-square)                           0.000       0.000
  Scaling correction factor                                  0.836
  Shift parameter                                          332.453
       simple second-order correction                             

Model Test Baseline Model:

  Test statistic                            971449.445  110645.100
  Degrees of freedom                               630         630
  P-value                                        0.000       0.000
  Scaling correction factor                                  8.824

User Model versus Baseline Model:

  Comparative Fit Index (CFI)                    0.997       0.966
  Tucker-Lewis Index (TLI)                       0.997       0.963
                                                                  
  Robust Comparative Fit Index (CFI)                            NA
  Robust Tucker-Lewis Index (TLI)                               NA

Root Mean Square Error of Approximation:

  RMSEA                                          0.064       0.074
  90 Percent confidence interval - lower         0.062       0.072
  90 Percent confidence interval - upper         0.066       0.076
  P-value RMSEA <= 0.05                          0.000       0.000
                                                                  
  Robust RMSEA                                                  NA
  90 Percent confidence interval - lower                        NA
  90 Percent confidence interval - upper                        NA

Standardized Root Mean Square Residual:

  SRMR                                           0.044       0.044

Parameter Estimates:

  Standard errors                           Robust.sem
  Information                                 Expected
  Information saturated (h1) model        Unstructured

Latent Variables:
                   Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
  IGA =~                                                                
    Q1                1.000                               0.677    0.677
    Q2                1.082    0.035   31.114    0.000    0.733    0.733
    Q3                1.194    0.036   33.603    0.000    0.809    0.809
    Q4                0.980    0.036   27.119    0.000    0.664    0.664
    Q5                1.180    0.035   33.485    0.000    0.799    0.799
    Q6                1.340    0.040   33.683    0.000    0.908    0.908
    Q7                1.277    0.041   31.423    0.000    0.865    0.865
    Q8                1.189    0.037   32.327    0.000    0.805    0.805
  KATW =~                                                               
    Q9                1.000                               0.875    0.875
    Q10               0.797    0.027   29.347    0.000    0.698    0.698
    Q11               1.030    0.026   40.159    0.000    0.901    0.901
    Q12               0.772    0.030   25.445    0.000    0.676    0.676
  KATC =~                                                               
    Q13               1.000                               0.791    0.791
    Q14               1.184    0.018   64.011    0.000    0.937    0.937
    Q15               1.168    0.018   65.770    0.000    0.924    0.924
    Q16               1.212    0.020   60.515    0.000    0.959    0.959
    Q17               1.220    0.020   60.296    0.000    0.965    0.965
    Q18               1.128    0.017   65.376    0.000    0.892    0.892
    Q19               1.172    0.018   64.040    0.000    0.928    0.928
    Q20               1.215    0.021   58.712    0.000    0.962    0.962
    Q21               1.226    0.020   59.882    0.000    0.970    0.970
    Q22               1.102    0.019   59.138    0.000    0.872    0.872
    Q23               1.160    0.019   62.383    0.000    0.918    0.918
  IM =~                                                                 
    Q24               1.000                               0.880    0.880
    Q25               0.933    0.011   81.340    0.000    0.821    0.821
    Q26               1.042    0.010  100.520    0.000    0.916    0.916
    Q27               1.044    0.010  103.791    0.000    0.919    0.919
    Q28               1.067    0.011  100.736    0.000    0.938    0.938
    Q29               0.993    0.013   78.005    0.000    0.873    0.873
    Q30               1.070    0.011   96.475    0.000    0.942    0.942
    Q31               0.931    0.014   65.183    0.000    0.819    0.819
  IS =~                                                                 
    Q32               1.000                               0.871    0.871
    Q33               1.032    0.013   79.794    0.000    0.899    0.899
    Q34               1.072    0.014   76.660    0.000    0.933    0.933
    Q35               1.077    0.014   74.990    0.000    0.938    0.938
    Q36               1.082    0.017   63.681    0.000    0.942    0.942

Covariances:
                   Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
  IGA ~~                                                                
    KATW              0.405    0.020   19.854    0.000    0.683    0.683
    KATC              0.452    0.020   22.629    0.000    0.844    0.844
    IM                0.524    0.021   25.216    0.000    0.879    0.879
    IS                0.443    0.020   22.412    0.000    0.751    0.751
  KATW ~~                                                               
    KATC              0.498    0.020   25.141    0.000    0.720    0.720
    IM                0.499    0.020   24.522    0.000    0.647    0.647
    IS                0.541    0.019   28.412    0.000    0.711    0.711
  KATC ~~                                                               
    IM                0.606    0.018   34.258    0.000    0.871    0.871
    IS                0.509    0.019   26.659    0.000    0.738    0.738
  IM ~~                                                                 
    IS                0.613    0.018   34.553    0.000    0.801    0.801

Intercepts:
                   Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
   .Q1                0.000                               0.000    0.000
   .Q2                0.000                               0.000    0.000
   .Q3                0.000                               0.000    0.000
   .Q4                0.000                               0.000    0.000
   .Q5                0.000                               0.000    0.000
   .Q6                0.000                               0.000    0.000
   .Q7                0.000                               0.000    0.000
   .Q8                0.000                               0.000    0.000
   .Q9                0.000                               0.000    0.000
   .Q10               0.000                               0.000    0.000
   .Q11               0.000                               0.000    0.000
   .Q12               0.000                               0.000    0.000
   .Q13               0.000                               0.000    0.000
   .Q14               0.000                               0.000    0.000
   .Q15               0.000                               0.000    0.000
   .Q16               0.000                               0.000    0.000
   .Q17               0.000                               0.000    0.000
   .Q18               0.000                               0.000    0.000
   .Q19               0.000                               0.000    0.000
   .Q20               0.000                               0.000    0.000
   .Q21               0.000                               0.000    0.000
   .Q22               0.000                               0.000    0.000
   .Q23               0.000                               0.000    0.000
   .Q24               0.000                               0.000    0.000
   .Q25               0.000                               0.000    0.000
   .Q26               0.000                               0.000    0.000
   .Q27               0.000                               0.000    0.000
   .Q28               0.000                               0.000    0.000
   .Q29               0.000                               0.000    0.000
   .Q30               0.000                               0.000    0.000
   .Q31               0.000                               0.000    0.000
   .Q32               0.000                               0.000    0.000
   .Q33               0.000                               0.000    0.000
   .Q34               0.000                               0.000    0.000
   .Q35               0.000                               0.000    0.000
   .Q36               0.000                               0.000    0.000
    IGA               0.000                               0.000    0.000
    KATW              0.000                               0.000    0.000
    KATC              0.000                               0.000    0.000
    IM                0.000                               0.000    0.000
    IS                0.000                               0.000    0.000

Thresholds:
                   Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
    Q1|t1            -2.096    0.088  -23.827    0.000   -2.096   -2.096
    Q1|t2            -1.906    0.075  -25.428    0.000   -1.906   -1.906
    Q1|t3            -1.526    0.057  -26.584    0.000   -1.526   -1.526
    Q1|t4            -0.122    0.037   -3.308    0.001   -0.122   -0.122
    Q2|t1            -2.285    0.105  -21.762    0.000   -2.285   -2.285
    Q2|t2            -2.059    0.085  -24.183    0.000   -2.059   -2.059
    Q2|t3            -1.561    0.059  -26.610    0.000   -1.561   -1.561
    Q2|t4            -0.303    0.037   -8.100    0.000   -0.303   -0.303
    Q3|t1            -2.285    0.105  -21.762    0.000   -2.285   -2.285
    Q3|t2            -2.116    0.090  -23.630    0.000   -2.116   -2.116
    Q3|t3            -1.777    0.068  -26.156    0.000   -1.777   -1.777
    Q3|t4            -0.168    0.037   -4.537    0.000   -0.168   -0.168
    Q4|t1            -2.116    0.090  -23.630    0.000   -2.116   -2.116
    Q4|t2            -1.699    0.064  -26.437    0.000   -1.699   -1.699
    Q4|t3            -1.145    0.047  -24.369    0.000   -1.145   -1.145
    Q4|t4             0.170    0.037    4.596    0.000    0.170    0.170
    Q5|t1            -2.422    0.121  -20.058    0.000   -2.422   -2.422
    Q5|t2            -1.880    0.073  -25.599    0.000   -1.880   -1.880
    Q5|t3            -1.467    0.055  -26.474    0.000   -1.467   -1.467
    Q5|t4             0.168    0.037    4.537    0.000    0.168    0.168
    Q6|t1            -2.383    0.116  -20.552    0.000   -2.383   -2.383
    Q6|t2            -2.077    0.087  -24.011    0.000   -2.077   -2.077
    Q6|t3            -1.718    0.065  -26.382    0.000   -1.718   -1.718
    Q6|t4            -0.154    0.037   -4.186    0.000   -0.154   -0.154
    Q7|t1            -2.230    0.100  -22.403    0.000   -2.230   -2.230
    Q7|t2            -2.077    0.087  -24.011    0.000   -2.077   -2.077
    Q7|t3            -1.673    0.063  -26.504    0.000   -1.673   -1.673
    Q7|t4            -0.355    0.038   -9.439    0.000   -0.355   -0.355
    Q8|t1            -2.348    0.112  -20.995    0.000   -2.348   -2.348
    Q8|t2            -2.008    0.081  -24.639    0.000   -2.008   -2.008
    Q8|t3            -1.351    0.052  -26.011    0.000   -1.351   -1.351
    Q8|t4             0.044    0.037    1.201    0.230    0.044    0.044
    Q9|t1            -2.383    0.116  -20.552    0.000   -2.383   -2.383
    Q9|t2            -1.673    0.063  -26.504    0.000   -1.673   -1.673
    Q9|t3            -1.003    0.044  -22.637    0.000   -1.003   -1.003
    Q9|t4             0.467    0.038   12.221    0.000    0.467    0.467
    Q10|t1           -2.137    0.091  -23.419    0.000   -2.137   -2.137
    Q10|t2           -1.065    0.045  -23.460    0.000   -1.065   -1.065
    Q10|t3           -0.079    0.037   -2.138    0.033   -0.079   -0.079
    Q10|t4            1.108    0.046   23.969    0.000    1.108    1.108
    Q11|t1           -2.383    0.116  -20.552    0.000   -2.383   -2.383
    Q11|t2           -1.809    0.070  -26.004    0.000   -1.809   -1.809
    Q11|t3           -1.187    0.048  -24.794    0.000   -1.187   -1.187
    Q11|t4            0.513    0.039   13.316    0.000    0.513    0.513
    Q12|t1           -1.788    0.068  -26.109    0.000   -1.788   -1.788
    Q12|t2           -0.725    0.040  -17.908    0.000   -0.725   -0.725
    Q12|t3           -0.027    0.037   -0.732    0.464   -0.027   -0.027
    Q12|t4            0.941    0.043   21.728    0.000    0.941    0.941
    Q13|t1           -2.383    0.116  -20.552    0.000   -2.383   -2.383
    Q13|t2           -1.821    0.070  -25.946    0.000   -1.821   -1.821
    Q13|t3           -1.218    0.049  -25.076    0.000   -1.218   -1.218
    Q13|t4            0.504    0.038   13.086    0.000    0.504    0.504
    Q14|t1           -2.315    0.108  -21.396    0.000   -2.315   -2.315
    Q14|t2           -2.116    0.090  -23.630    0.000   -2.116   -2.116
    Q14|t3           -1.880    0.073  -25.599    0.000   -1.880   -1.880
    Q14|t4           -0.122    0.037   -3.308    0.001   -0.122   -0.122
    Q15|t1           -2.315    0.108  -21.396    0.000   -2.315   -2.315
    Q15|t2           -1.992    0.080  -24.773    0.000   -1.992   -1.992
    Q15|t3           -1.737    0.066  -26.317    0.000   -1.737   -1.737
    Q15|t4           -0.137    0.037   -3.718    0.000   -0.137   -0.137
    Q16|t1           -2.257    0.102  -22.096    0.000   -2.257   -2.257
    Q16|t2           -2.008    0.081  -24.639    0.000   -2.008   -2.008
    Q16|t3           -1.690    0.064  -26.461    0.000   -1.690   -1.690
    Q16|t4           -0.233    0.037   -6.291    0.000   -0.233   -0.233
    Q17|t1           -2.285    0.105  -21.762    0.000   -2.285   -2.285
    Q17|t2           -1.962    0.078  -25.018    0.000   -1.962   -1.962
    Q17|t3           -1.655    0.062  -26.538    0.000   -1.655   -1.655
    Q17|t4           -0.181    0.037   -4.888    0.000   -0.181   -0.181
    Q18|t1           -2.285    0.105  -21.762    0.000   -2.285   -2.285
    Q18|t2           -2.008    0.081  -24.639    0.000   -2.008   -2.008
    Q18|t3           -1.655    0.062  -26.538    0.000   -1.655   -1.655
    Q18|t4            0.038    0.037    1.025    0.305    0.038    0.038
    Q19|t1           -2.205    0.097  -22.687    0.000   -2.205   -2.205
    Q19|t2           -2.041    0.084  -24.345    0.000   -2.041   -2.041
    Q19|t3           -1.777    0.068  -26.156    0.000   -1.777   -1.777
    Q19|t4           -0.343    0.038   -9.148    0.000   -0.343   -0.343
    Q20|t1           -2.257    0.102  -22.096    0.000   -2.257   -2.257
    Q20|t2           -2.041    0.084  -24.345    0.000   -2.041   -2.041
    Q20|t3           -1.856    0.072  -25.750    0.000   -1.856   -1.856
    Q20|t4           -0.178    0.037   -4.830    0.000   -0.178   -0.178
    Q21|t1           -2.285    0.105  -21.762    0.000   -2.285   -2.285
    Q21|t2           -2.008    0.081  -24.639    0.000   -2.008   -2.008
    Q21|t3           -0.203    0.037   -5.473    0.000   -0.203   -0.203
    Q22|t1           -2.230    0.100  -22.403    0.000   -2.230   -2.230
    Q22|t2           -2.096    0.088  -23.827    0.000   -2.096   -2.096
    Q22|t3           -1.767    0.067  -26.201    0.000   -1.767   -1.767
    Q22|t4           -0.305    0.037   -8.158    0.000   -0.305   -0.305
    Q23|t1           -2.315    0.108  -21.396    0.000   -2.315   -2.315
    Q23|t2           -2.158    0.093  -23.193    0.000   -2.158   -2.158
    Q23|t3           -1.821    0.070  -25.946    0.000   -1.821   -1.821
    Q23|t4           -0.135    0.037   -3.660    0.000   -0.135   -0.135
    Q24|t1           -2.383    0.116  -20.552    0.000   -2.383   -2.383
    Q24|t2           -1.947    0.077  -25.130    0.000   -1.947   -1.947
    Q24|t3           -1.607    0.060  -26.601    0.000   -1.607   -1.607
    Q24|t4            0.057    0.037    1.552    0.121    0.057    0.057
    Q25|t1           -2.422    0.121  -20.058    0.000   -2.422   -2.422
    Q25|t2           -1.832    0.071  -25.885    0.000   -1.832   -1.832
    Q25|t3           -1.178    0.048  -24.711    0.000   -1.178   -1.178
    Q25|t4            0.294    0.037    7.867    0.000    0.294    0.294
    Q26|t1           -2.422    0.121  -20.058    0.000   -2.422   -2.422
    Q26|t2           -2.041    0.084  -24.345    0.000   -2.041   -2.041
    Q26|t3           -1.329    0.051  -25.891    0.000   -1.329   -1.329
    Q26|t4            0.042    0.037    1.142    0.253    0.042    0.042
    Q27|t1           -2.512    0.133  -18.877    0.000   -2.512   -2.512
    Q27|t2           -2.096    0.088  -23.827    0.000   -2.096   -2.096
    Q27|t3           -1.533    0.058  -26.591    0.000   -1.533   -1.533
    Q27|t4            0.061    0.037    1.669    0.095    0.061    0.061
    Q28|t1           -2.348    0.112  -20.995    0.000   -2.348   -2.348
    Q28|t2           -2.181    0.095  -22.949    0.000   -2.181   -2.181
    Q28|t3           -1.777    0.068  -26.156    0.000   -1.777   -1.777
    Q28|t4           -0.222    0.037   -5.999    0.000   -0.222   -0.222
    Q29|t1           -2.348    0.112  -20.995    0.000   -2.348   -2.348
    Q29|t2           -2.230    0.100  -22.403    0.000   -2.230   -2.230
    Q29|t3           -1.737    0.066  -26.317    0.000   -1.737   -1.737
    Q29|t4           -0.422    0.038  -11.122    0.000   -0.422   -0.422
    Q30|t1           -2.348    0.112  -20.995    0.000   -2.348   -2.348
    Q30|t2           -2.116    0.090  -23.630    0.000   -2.116   -2.116
    Q30|t3           -1.767    0.067  -26.201    0.000   -1.767   -1.767
    Q30|t4           -0.341    0.038   -9.090    0.000   -0.341   -0.341
    Q31|t1           -2.315    0.108  -21.396    0.000   -2.315   -2.315
    Q31|t2           -2.137    0.091  -23.419    0.000   -2.137   -2.137
    Q31|t3           -1.540    0.058  -26.598    0.000   -1.540   -1.540
    Q31|t4           -0.170    0.037   -4.596    0.000   -0.170   -0.170
    Q32|t1           -2.348    0.112  -20.995    0.000   -2.348   -2.348
    Q32|t2           -1.933    0.077  -25.235    0.000   -1.933   -1.933
    Q32|t3           -1.157    0.047  -24.499    0.000   -1.157   -1.157
    Q32|t4            0.343    0.038    9.148    0.000    0.343    0.343
    Q33|t1           -2.512    0.133  -18.877    0.000   -2.512   -2.512
    Q33|t2           -1.767    0.067  -26.201    0.000   -1.767   -1.767
    Q33|t3           -1.032    0.045  -23.029    0.000   -1.032   -1.032
    Q33|t4            0.382    0.038   10.136    0.000    0.382    0.382
    Q34|t1           -2.512    0.133  -18.877    0.000   -2.512   -2.512
    Q34|t2           -1.798    0.069  -26.058    0.000   -1.798   -1.798
    Q34|t3           -0.968    0.044  -22.136    0.000   -0.968   -0.968
    Q34|t4            0.543    0.039   14.005    0.000    0.543    0.543
    Q35|t1           -2.703    0.166  -16.288    0.000   -2.703   -2.703
    Q35|t2           -1.906    0.075  -25.428    0.000   -1.906   -1.906
    Q35|t3           -1.120    0.046  -24.104    0.000   -1.120   -1.120
    Q35|t4            0.446    0.038   11.701    0.000    0.446    0.446
    Q36|t1           -2.464    0.126  -19.505    0.000   -2.464   -2.464
    Q36|t2           -1.933    0.077  -25.235    0.000   -1.933   -1.933
    Q36|t3           -1.335    0.051  -25.922    0.000   -1.335   -1.335
 [ reached getOption("max.print") -- omitted 1 row ]

Variances:
                   Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
   .Q1                0.541                               0.541    0.541
   .Q2                0.463                               0.463    0.463
   .Q3                0.346                               0.346    0.346
   .Q4                0.559                               0.559    0.559
   .Q5                0.362                               0.362    0.362
   .Q6                0.176                               0.176    0.176
   .Q7                0.252                               0.252    0.252
   .Q8                0.351                               0.351    0.351
   .Q9                0.234                               0.234    0.234
   .Q10               0.513                               0.513    0.513
   .Q11               0.187                               0.187    0.187
   .Q12               0.543                               0.543    0.543
   .Q13               0.374                               0.374    0.374
   .Q14               0.122                               0.122    0.122
   .Q15               0.146                               0.146    0.146
   .Q16               0.080                               0.080    0.080
   .Q17               0.069                               0.069    0.069
   .Q18               0.204                               0.204    0.204
   .Q19               0.139                               0.139    0.139
   .Q20               0.075                               0.075    0.075
   .Q21               0.060                               0.060    0.060
   .Q22               0.239                               0.239    0.239
   .Q23               0.158                               0.158    0.158
   .Q24               0.226                               0.226    0.226
   .Q25               0.326                               0.326    0.326
   .Q26               0.160                               0.160    0.160
   .Q27               0.156                               0.156    0.156
   .Q28               0.120                               0.120    0.120
   .Q29               0.237                               0.237    0.237
   .Q30               0.113                               0.113    0.113
   .Q31               0.330                               0.330    0.330
   .Q32               0.242                               0.242    0.242
   .Q33               0.192                               0.192    0.192
   .Q34               0.129                               0.129    0.129
   .Q35               0.121                               0.121    0.121
   .Q36               0.112                               0.112    0.112
    IGA               0.459    0.028   16.667    0.000    1.000    1.000
    KATW              0.766    0.026   29.218    0.000    1.000    1.000
    KATC              0.626    0.022   28.562    0.000    1.000    1.000
    IM                0.774    0.016   48.975    0.000    1.000    1.000
    IS                0.758    0.019   40.180    0.000    1.000    1.000

Scales y*:
                   Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
    Q1                1.000                               1.000    1.000
    Q2                1.000                               1.000    1.000
    Q3                1.000                               1.000    1.000
    Q4                1.000                               1.000    1.000
    Q5                1.000                               1.000    1.000
    Q6                1.000                               1.000    1.000
    Q7                1.000                               1.000    1.000
    Q8                1.000                               1.000    1.000
    Q9                1.000                               1.000    1.000
    Q10               1.000                               1.000    1.000
    Q11               1.000                               1.000    1.000
    Q12               1.000                               1.000    1.000
    Q13               1.000                               1.000    1.000
    Q14               1.000                               1.000    1.000
    Q15               1.000                               1.000    1.000
    Q16               1.000                               1.000    1.000
    Q17               1.000                               1.000    1.000
    Q18               1.000                               1.000    1.000
    Q19               1.000                               1.000    1.000
    Q20               1.000                               1.000    1.000
    Q21               1.000                               1.000    1.000
    Q22               1.000                               1.000    1.000
    Q23               1.000                               1.000    1.000
    Q24               1.000                               1.000    1.000
    Q25               1.000                               1.000    1.000
    Q26               1.000                               1.000    1.000
    Q27               1.000                               1.000    1.000
    Q28               1.000                               1.000    1.000
    Q29               1.000                               1.000    1.000
    Q30               1.000                               1.000    1.000
    Q31               1.000                               1.000    1.000
    Q32               1.000                               1.000    1.000
    Q33               1.000                               1.000    1.000
    Q34               1.000                               1.000    1.000
    Q35               1.000                               1.000    1.000
    Q36               1.000                               1.000    1.000

Patrick (Malone Quantitative)

unread,
Jan 2, 2021, 5:08:00 PM1/2/21
to lav...@googlegroups.com
Erdem,

Nothing jumps out at me. I suspect it reflects the fact that some of your factors are so highly correlated.

Pat

--
You received this message because you are subscribed to the Google Groups "lavaan" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lavaan+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/lavaan/c187de43-cc19-4981-b0c9-9d0c171dab13n%40googlegroups.com.


--
Patrick S. Malone, Ph.D., Malone Quantitative
NEW Service Models: http://malonequantitative.com

He/Him/His

Erdem Uygun

unread,
Jan 3, 2021, 7:48:07 AM1/3/21
to lav...@googlegroups.com

Hi Patrick and thank you for your valuable response.

When I inspect correlations among factors, I get the results below:

> lavInspect(fit,"cor.lv")
     IGA   KATW  KATC  IM    IS  
IGA  1.000                        
KATW 0.683 1.000                  
KATC 0.844 0.720 1.000            
IM   0.879 0.647 0.871 1.000      
IS   0.751 0.711 0.738 0.801 1.000

Factors are indeed highly correlated. I repeated the test with one-factor solution but I am still getting the same warning. By the way, when I use any ML estimator (ML, MLM, MLR), there is no warning but with low global fit indices scores most probably because my data is significantly not normal:

> model<-'
+ GC=~Q1+Q2+Q3+Q4+Q5+Q6+Q7+Q8+Q9+Q10+Q11+Q12+Q13+Q14+Q15+Q16+Q17+Q18+Q19+Q20+Q21+Q22+Q23+Q24+Q25+Q26+Q27+Q28+Q29+Q30+Q31+Q32+Q33+Q34+Q35+Q36
+ '

>
> fit<-cfa(model,data=CFA_full,ordered=c("Q1","Q2","Q3","Q4","Q5","Q6","Q7","Q8","Q9","Q10","Q11","Q12","Q13","Q14","Q15","Q16","Q17","Q18","Q19","Q20","Q21","Q22","Q23","Q24","Q25","Q26","Q27","Q28","Q29","Q30","Q31","Q32","Q33","Q34","Q35","Q36"))
Warning message:
In lav_model_vcov(lavmodel = lavmodel, lavsamplestats = lavsamplestats,  :
 
lavaan WARNING:

    The variance-covariance matrix of the estimated parameters (vcov)
    does not appear to be positive definite! The smallest eigenvalue
    (= -3.130112e-19) is smaller than zero. This may be a symptom that

    the model is not identified.

>
> summary(fit,fit.measures=TRUE,standardized = TRUE)
lavaan 0.6-7 ended normally after 85 iterations


  Estimator                                       DWLS
  Optimization method                           NLMINB
  Number of free parameters                        179

                                                     
  Number of observations                          1165
                                                     
Model Test User Model:
                                               Standard      Robust
  Test Statistic                              12039.349    9221.853
  Degrees of freedom                                594         594
  P-value (Chi-square)                            0.000       0.000
  Scaling correction factor                                   1.366
  Shift parameter                                           409.172

       simple second-order correction                              

Model Test Baseline Model:

  Test statistic                            971449.445  110645.100
  Degrees of freedom                               630         630
  P-value                                        0.000       0.000
  Scaling correction factor                                  8.824

User Model versus Baseline Model:

  Comparative Fit Index (CFI)                    0.988       0.922
  Tucker-Lewis Index (TLI)                       0.987       0.917

                                                                 
  Robust Comparative Fit Index (CFI)                            NA
  Robust Tucker-Lewis Index (TLI)                               NA

Root Mean Square Error of Approximation:

  RMSEA                                          0.129       0.112
  90 Percent confidence interval - lower         0.127       0.110
  90 Percent confidence interval - upper         0.131       0.114

  P-value RMSEA <= 0.05                          0.000       0.000
                                                                 
  Robust RMSEA                                                  NA
  90 Percent confidence interval - lower                        NA
  90 Percent confidence interval - upper                        NA

Standardized Root Mean Square Residual:

  SRMR                                           0.082       0.082


Parameter Estimates:

  Standard errors                           Robust.sem
  Information                                 Expected
  Information saturated (h1) model        Unstructured

Latent Variables:
                   Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all
  GC =~                                                                
    Q1                1.000                               0.618    0.618
    Q2                1.084    0.035   30.806    0.000    0.670    0.670
    Q3                1.194    0.036   33.335    0.000    0.738    0.738
    Q4                0.981    0.037   26.723    0.000    0.606    0.606
    Q5                1.187    0.036   33.116    0.000    0.734    0.734
    Q6                1.343    0.040   33.367    0.000    0.831    0.831
    Q7                1.281    0.041   31.076    0.000    0.792    0.792
    Q8                1.191    0.037   31.917    0.000    0.736    0.736
    Q9                1.099    0.037   30.047    0.000    0.679    0.679
    Q10               0.884    0.042   20.969    0.000    0.547    0.547
    Q11               1.124    0.038   29.694    0.000    0.695    0.695
    Q12               0.850    0.044   19.261    0.000    0.526    0.526
    Q13               1.214    0.040   30.286    0.000    0.751    0.751
    Q14               1.482    0.046   32.151    0.000    0.916    0.916
    Q15               1.465    0.045   32.317    0.000    0.906    0.906
    Q16               1.530    0.048   31.915    0.000    0.946    0.946
    Q17               1.539    0.049   31.565    0.000    0.951    0.951
    Q18               1.401    0.043   32.562    0.000    0.866    0.866
    Q19               1.457    0.045   32.356    0.000    0.901    0.901
    Q20               1.529    0.048   31.842    0.000    0.945    0.945
    Q21               1.548    0.049   31.633    0.000    0.957    0.957
    Q22               1.363    0.042   32.086    0.000    0.843    0.843
    Q23               1.442    0.045   32.177    0.000    0.892    0.892
    Q24               1.349    0.041   32.889    0.000    0.834    0.834
    Q25               1.256    0.041   30.745    0.000    0.777    0.777
    Q26               1.422    0.044   32.085    0.000    0.879    0.879
    Q27               1.424    0.044   32.559    0.000    0.881    0.881
    Q28               1.459    0.046   31.564    0.000    0.902    0.902
    Q29               1.344    0.042   32.013    0.000    0.831    0.831
    Q30               1.465    0.046   31.837    0.000    0.906    0.906
    Q31               1.250    0.040   31.128    0.000    0.773    0.773
    Q32               1.242    0.039   32.003    0.000    0.768    0.768
    Q33               1.328    0.042   31.286    0.000    0.821    0.821
    Q34               1.429    0.045   31.422    0.000    0.883    0.883
    Q35               1.436    0.046   31.200    0.000    0.888    0.888
    Q36               1.386    0.045   30.903    0.000    0.857    0.857
    GC                0.000                               0.000    0.000
   .Q1                0.618                               0.618    0.618
   .Q2                0.551                               0.551    0.551
   .Q3                0.455                               0.455    0.455
   .Q4                0.632                               0.632    0.632
   .Q5                0.461                               0.461    0.461
   .Q6                0.310                               0.310    0.310
   .Q7                0.373                               0.373    0.373
   .Q8                0.458                               0.458    0.458
   .Q9                0.538                               0.538    0.538
   .Q10               0.701                               0.701    0.701
   .Q11               0.517                               0.517    0.517
   .Q12               0.724                               0.724    0.724
   .Q13               0.436                               0.436    0.436
   .Q14               0.160                               0.160    0.160
   .Q15               0.179                               0.179    0.179
   .Q16               0.105                               0.105    0.105
   .Q17               0.095                               0.095    0.095
   .Q18               0.250                               0.250    0.250
   .Q19               0.189                               0.189    0.189
   .Q20               0.107                               0.107    0.107
   .Q21               0.084                               0.084    0.084
   .Q22               0.290                               0.290    0.290
   .Q23               0.205                               0.205    0.205
   .Q24               0.304                               0.304    0.304
   .Q25               0.397                               0.397    0.397
   .Q26               0.227                               0.227    0.227
   .Q27               0.225                               0.225    0.225
   .Q28               0.186                               0.186    0.186
   .Q29               0.309                               0.309    0.309
   .Q30               0.180                               0.180    0.180
   .Q31               0.403                               0.403    0.403
   .Q32               0.410                               0.410    0.410
   .Q33               0.326                               0.326    0.326
   .Q34               0.220                               0.220    0.220
   .Q35               0.212                               0.212    0.212
   .Q36               0.265                               0.265    0.265
    GC                0.382    0.025   15.351    0.000    1.000    1.000
Do you have any idea why this occurs and what to do in such situations? Thank you very much.




Erdem UYGUN, Phd Student

Department of Educational Sciences/Curriculum and Instruction



Patrick (Malone Quantitative) <mal...@malonequantitative.com>, 3 Oca 2021 Paz, 01:08 tarihinde şunu yazdı:

Patrick (Malone Quantitative)

unread,
Jan 3, 2021, 8:15:26 AM1/3/21
to lav...@googlegroups.com
Weird. I have no idea why a one-factor model--where the variance is well significant--would give you that warning. I'll punt to others.

Pat

Erdem Uygun

unread,
Jan 3, 2021, 8:20:14 AM1/3/21
to lav...@googlegroups.com
I really appreciate it because I spent days and hours on that issue. I'm looking forward to seeing others' responses. Thank you.


Erdem UYGUN, Phd Student

Department of Educational Sciences/Curriculum and Instruction


Patrick (Malone Quantitative) <mal...@malonequantitative.com>, 3 Oca 2021 Paz, 16:15 tarihinde şunu yazdı:

Edward Rigdon

unread,
Jan 3, 2021, 10:50:19 AM1/3/21
to lav...@googlegroups.com
With ordinal variables, the set of parameters includes thresholds related parameters. With skewed data, there may be high uncertainty about these additional parameters, perhaps making identification marginal. You might try simulating a much larger dataset to see if sample size might be the problem.

Juan Diego Hernández Lalinde

unread,
Jan 3, 2021, 11:14:18 AM1/3/21
to lav...@googlegroups.com
Hello guys,

I have a question (it could be a suggestion) regarding this problem. Due to the eigenvalue is quite small, perhaps you could try by using different starting values and compare the results. If the chi-square statistic and fit indices are the same (or very similar), you could ignore this warning message.

What do you think about this? Again, I'm not sure if this is a good practice, but I have read this recommendation in several forums and it sounds logical to me.

Good luck.

Erdem Uygun

unread,
Jan 3, 2021, 1:21:23 PM1/3/21
to lav...@googlegroups.com

Hi all, I want to share a finding of mine with you. When I specifically write estimator = “DWSL”, I receive no warning with better results. So additionally I want to ask that what is the difference between “DWSL” and “WLSMV” estimators? I see in books referring WLSMV mostly.


At lavaan website it says:

 

In both cases, lavaan will automatically switch to the WLSMV estimator: it will use diagonally weighted least squares (DWLS) to estimate the model parameters, but it will use the full weight matrix to compute robust standard errors, and a mean- and variance-adjusted test stastistic.

 

I understand that lavaan actually uses DWSL when we put “ordered” command. Is it because there is a lavaan-related problem that generates previous warning with WLSMV that does not exist when specifically writing DWSL as the estimator for bypassing default WLSMV option, or do I miss something else?

 

Note: My sample size is 1165.


Thank you.

 

model<-'

+ IGA=~Q1+Q2+Q3+Q4+Q5+Q6+Q7+Q8

+ KATW=~Q9+Q10+Q11+Q12

+ KATC=~Q13+Q14+Q15+Q16+Q17+Q18+Q19+Q20+Q21+Q22+Q23

+ IM=~Q24+Q25+Q26+Q27+Q28+Q29+Q30+Q31

+ IS=~Q32+Q33+Q34+Q35+Q36

+ '

>

> fit<-cfa(model,data=CFA_full,estimator = "DWLS", ordered=c("Q1","Q2","Q3","Q4","Q5","Q6","Q7","Q8","Q9","Q10","Q11","Q12","Q13","Q14","Q15","Q16","Q17","Q18","Q19","Q20","Q21","Q22","Q23","Q24","Q25","Q26","Q27","Q28","Q29","Q30","Q31","Q32","Q33","Q34","Q35","Q36"))

>

> summary(fit,fit.measures=TRUE,standardized = TRUE)

lavaan 0.6-7 ended normally after 125 iterations

 

  Estimator                                       DWLS

  Optimization method                           NLMINB

  Number of free parameters                        189

                                                      

  Number of observations                          1165

                                                      

Model Test User Model:

                                                      

  Test statistic                              3339.255

  Degrees of freedom                               584

  P-value (Chi-square)                           0.000

 

Model Test Baseline Model:

 

  Test statistic                            971449.445

  Degrees of freedom                               630

  P-value                                        0.000

 

User Model versus Baseline Model:

 

  Comparative Fit Index (CFI)                    0.997

  Tucker-Lewis Index (TLI)                       0.997

 

Root Mean Square Error of Approximation:

 

  RMSEA                                          0.064

  90 Percent confidence interval - lower         0.062

  90 Percent confidence interval - upper         0.066

  P-value RMSEA <= 0.05                          0.000

 

Standardized Root Mean Square Residual:

 

  SRMR                                           0.044

 

Parameter Estimates:

 

  Standard errors                             Standard

  Information                                 Expected

  Information saturated (h1) model        Unstructured

 

Latent Variables:

                   Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all

  IGA =~                                                                

    Q1                1.000                               0.677    0.677

    Q2                1.082    0.013   84.772    0.000    0.733    0.733

    Q3                1.194    0.013   91.334    0.000    0.809    0.809

    Q4                0.980    0.012   81.989    0.000    0.664    0.664

    Q5                1.180    0.013   91.794    0.000    0.799    0.799

    Q6                1.340    0.014   98.539    0.000    0.908    0.908

    Q7                1.277    0.013   95.960    0.000    0.865    0.865

    Q8                1.189    0.013   92.767    0.000    0.805    0.805

  KATW =~                                                               

    Q9                1.000                               0.875    0.875

    Q10               0.797    0.010   78.080    0.000    0.698    0.698

    Q11               1.030    0.011   96.872    0.000    0.901    0.901

    Q12               0.772    0.010   76.490    0.000    0.676    0.676

  KATC =~                                                               

    Q13               1.000                               0.791    0.791

    Q14               1.184    0.008  145.214    0.000    0.937    0.937

    Q15               1.168    0.008  145.083    0.000    0.924    0.924

    Q16               1.212    0.008  150.084    0.000    0.959    0.959

    Q17               1.220    0.008  150.581    0.000    0.965    0.965

    Q18               1.128    0.008  140.263    0.000    0.892    0.892

    Q19               1.172    0.008  144.850    0.000    0.928    0.928

    Q20               1.215    0.008  151.737    0.000    0.962    0.962

    Q21               1.226    0.008  152.827    0.000    0.970    0.970

    Q22               1.102    0.008  135.024    0.000    0.872    0.872

    Q23               1.160    0.008  144.441    0.000    0.918    0.918

  IM =~                                                                 

    Q24               1.000                               0.880    0.880

    Q25               0.933    0.007  139.945    0.000    0.821    0.821

    Q26               1.042    0.006  160.321    0.000    0.916    0.916

    Q27               1.044    0.006  163.322    0.000    0.919    0.919

    Q28               1.067    0.006  166.126    0.000    0.938    0.938

    Q29               0.993    0.007  147.823    0.000    0.873    0.873

    Q30               1.070    0.007  162.402    0.000    0.942    0.942

    Q31               0.931    0.007  138.048    0.000    0.819    0.819

  IS =~                                                                 

    Q32               1.000                               0.871    0.871

    Q33               1.032    0.008  132.329    0.000    0.899    0.899

    Q34               1.072    0.008  140.234    0.000    0.933    0.933

    Q35               1.077    0.008  142.589    0.000    0.938    0.938

    Q36               1.082    0.008  140.853    0.000    0.942    0.942

 

Covariances:

                   Estimate  Std.Err  z-value  P(>|z|)   Std.lv  Std.all

  IGA ~~                                                                

    KATW              0.405    0.006   66.042    0.000    0.683    0.683

    KATC              0.452    0.005   91.751    0.000    0.844    0.844

    IM                0.524    0.006   94.672    0.000    0.879    0.879

    IS                0.443    0.005   83.496    0.000    0.751    0.751

  KATW ~~                                                               

    KATC              0.498    0.005   92.114    0.000    0.720    0.720

    IM                0.499    0.006   78.905    0.000    0.647    0.647

    IS                0.541    0.007   81.675    0.000    0.711    0.711

  KATC ~~                                                               

    IM                0.606    0.005  130.906    0.000    0.871    0.871

    IS                0.509    0.004  113.757    0.000    0.738    0.738

  IM ~~                                                                 

    IS                0.613    0.005  122.415    0.000    0.801    0.801

    IGA               0.459    0.008   55.252    0.000    1.000    1.000

    KATW              0.766    0.014   56.289    0.000    1.000    1.000

    KATC              0.626    0.007   85.688    0.000    1.000    1.000

    IM                0.774    0.007  105.932    0.000    1.000    1.000

    IS                0.758    0.009   86.681    0.000    1.000    1.000


Erdem UYGUN, Phd Student

Department of Educational Sciences/Curriculum and Instruction




Juan Diego Hernández Lalinde <hernandez...@gmail.com>, 3 Oca 2021 Paz, 19:14 tarihinde şunu yazdı:

Terrence Jorgensen

unread,
Jan 5, 2021, 10:39:54 PM1/5/21
to lavaan

what is the difference between “DWLS” and “WLSMV” estimators?


"WLSMV" is a keyword, not an estimator.  As the text states (which you pasted in your last post; see also the ?lavOptions help page), WLSMV implies DWLS.  Specifically, when you set (or rely on the default setting when declaring any outcomes as ordered=), estimator = "WLSMV" is a shortcut for the following three arguments:
  1. estimator = "DWLS"
  2. se = "robust.sem"
  3. test = "scaled.shifted"
So if you explicitly only declare (1) without also declaring (2) and (3), that might be the source of your discrepancy.  You can always check your model results to verify:
lavInspect(fit, "options")[c("estimator","se","test")]

You need (2) and (3) to obtain robust results; otherwise, you cannot trust your test statistics or CIs to have nominal Type I error rates or CI coverage rates.

Terrence D. Jorgensen
Assistant Professor, Methods and Statistics
Research Institute for Child Development and Education, the University of Amsterdam

Reply all
Reply to author
Forward
0 new messages