|
 |
Factor Loading larger than 1 |
 
|
Message/Author |
|
Hanno Petras posted on Thursday, September 15, 2005 - 12:13 pm
|
|
|
Dear Linda and Bengt, I have run an EFA analysis using the self esteem items from the LSAY data. The problem I encountered that a one and three factor solution converged but not the 2 factor solution. I increased the number of interation and the 2 factor solution conseuqently converged. However, one of the items shows a loading of larger than 1 and has a large negative residual variance. Given that the same problem is also encountered in the 3 factor model, I am wondering if this shows that the last item is not useful and/or a one factor solution is preferrable. Any advice would be greatly appreciated. Below I have posten the converged output. Best, Hanno Mplus VERSION 3.13 MUTHEN & MUTHEN 09/15/2005 11:14 AM INPUT INSTRUCTIONS Title: The variable names are for the data set lsay.dat DATA: File is "j:\ccjs699\dataset\lsay.dat"; VARIABLE: Names are lsayid schcode classize urban tracking ntracks mthlvl female mthflg7-mthflg12 mothed fathed mothsei fathsei homeres race expect parapsh parcpsh parmpsh peerapsh peermpsh bas7 basse7 alg7 algse7 geo7 geose7 qlt7 qltse7 mth7 mthse7 mtha7 mthase7 bas8 basse8 alg8 algse8 geo8 geose8 qlt8 qltse8 mth8 mthse8 mtha8 mthase8 bas9 basse9 alg9 algse9 geo9 geose9 qlt9 qltse9 mth9 mthse9 mtha9 mthase9 bas10 basse10 alg10 algse10 geo10 geose10 qlt10 qltse10 mth10 mthse10 mtha10 mthase10 bas11 basse11 alg11 algse11 geo11 geose11 qlt11 qltse11 mth11 mthse11 mtha11 mthase11 bas12 basse12 alg12 algse12 geo12 geose12 qlt12 qltse12 mth12 mthse12 mtha12 mthase12 mthcrs7-mthcrs12 mtrk10-mtrk12 totstud lchfull lchpart parvis mcirr mclub strat mstrat comp mcomp african hispan asian expel arrest dropot self worth other satisf respect failure esteem problem cloctn dloctn eloctn floctn gloctn hloctn iloctn jloctn kloctn lloctn drink runawa suicid alc7 alc10 alc11 alc12 arest7 runa8 runa9 runa10 runa11 run12 suic8 suic9 suic10 suic11 suic12 drop7 drop8 drop9 drop10 drop11 drop12 fdrop8 fdrop9 fdrop10 fdrop11 fdrop12 enj7 good7 und7 useboy7 nerv7 wor7 scar7 use7 logic7 boybet7 job7 often7 enj8 good8 und8 useboy8 nerv8 wor8 scar8 use8 logic8 boybet8 job8 often8 enj9 good9 und9 useboy9 nerv9 wor9 scar9 use9 logic9 boybet9 job9 often9 enj10 good10 und10 useboy10 nerv10 wor10 scar10 use10 logic10 boybet10 job10 often10; Missing are all(9999); IDVARIABLE=lsayid; USEVAR = self worth other satisf respect failure ; USEOBSERVATIONS = (female EQ 1) ; Analysis: Type is efa 1 3 missing; iterations=5000; Output: sampstat patterns mod (3.84) tech1 ; Plot: Type is plot1 plot2 plot3; *** WARNING in Output command SAMPSTAT option for analysis types MISSING and MCOHORT requires H1. Analysis type H1 is turned on automatically. *** WARNING in Output command MODINDICES option is available only for Analysis types GENERAL and MIXTURE. Request for MODINDICES is ignored. *** WARNING Data set contains cases with missing on all variables. These cases were not included in the analysis. Number of cases with missing on all variables: 14 3 WARNING(S) FOUND IN THE INPUT INSTRUCTIONS The variable names are for the data set lsay.dat SUMMARY OF ANALYSIS Number of groups 1 Number of observations 1476 Number of dependent variables 6 Number of independent variables 0 Number of continuous latent variables 0 Observed dependent variables Continuous SELF WORTH OTHER SATISF RESPECT FAILURE Variables with special functions ID variable LSAYID Estimator ML Information matrix OBSERVED Maximum number of iterations 5000 Convergence criterion 0.500D-04 Maximum number of steepest descent iterations 20 Maximum number of iterations for H1 2000 Convergence criterion for H1 0.100D-03 Input data file(s) j:\ccjs699\dataset\lsay.dat Input data format FREE SUMMARY OF DATA Number of patterns 23 SUMMARY OF MISSING DATA PATTERNS MISSING DATA PATTERNS 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 SELF x x x x x x x x x x x x x x x x x WORTH x x x x x x x x x x x x OTHER x x x x x x x x x x x x x x x SATISF x x x x x x x x x x x RESPECT x x x x x x x x x x x x FAILURE x x x x x x x x x x x x x 21 22 23 SELF WORTH OTHER x SATISF x x RESPECT x x FAILURE x MISSING DATA PATTERN FREQUENCIES Pattern Frequency Pattern Frequency Pattern Frequency 1 1307 9 2 17 2 2 35 10 2 18 3 3 22 11 2 19 1 4 32 12 28 20 1 5 1 13 1 21 1 6 3 14 1 22 1 7 4 15 3 23 1 8 21 16 2 COVARIANCE COVERAGE OF DATA Minimum covariance coverage value 0.100 PROPORTION OF DATA PRESENT Covariance Coverage SELF WORTH OTHER SATISF RESPECT ________ ________ ________ ________ ________ SELF 0.995 WORTH 0.970 0.972 OTHER 0.975 0.953 0.979 SATISF 0.960 0.942 0.949 0.965 RESPECT 0.969 0.949 0.957 0.947 0.974 FAILURE 0.964 0.943 0.950 0.939 0.947 Covariance Coverage FAILURE ________ FAILURE 0.967 SAMPLE STATISTICS ESTIMATED SAMPLE STATISTICS Means SELF WORTH OTHER SATISF RESPECT ________ ________ ________ ________ ________ 1 1.933 2.164 2.063 2.236 2.887 Means FAILURE ________ 1 3.902 Covariances SELF WORTH OTHER SATISF RESPECT ________ ________ ________ ________ ________ SELF 0.729 WORTH 0.295 0.800 OTHER 0.239 0.310 0.770 SATISF 0.383 0.351 0.317 1.018 RESPECT -0.154 -0.141 -0.106 -0.191 1.389 FAILURE -0.184 -0.180 -0.202 -0.291 0.368 Covariances FAILURE ________ FAILURE 1.224 Correlations SELF WORTH OTHER SATISF RESPECT ________ ________ ________ ________ ________ SELF 1.000 WORTH 0.387 1.000 OTHER 0.319 0.395 1.000 SATISF 0.445 0.389 0.358 1.000 RESPECT -0.153 -0.134 -0.102 -0.161 1.000 FAILURE -0.195 -0.182 -0.208 -0.261 0.282 Correlations FAILURE ________ FAILURE 1.000 MAXIMUM LOG-LIKELIHOOD VALUE FOR THE UNRESTRICTED (H1) MODEL IS -11432.745 RESULTS FOR EXPLORATORY FACTOR ANALYSIS EIGENVALUES FOR SAMPLE CORRELATION MATRIX 1 2 3 4 5 ________ ________ ________ ________ ________ 1 2.378 1.069 0.730 0.694 0.585 EIGENVALUES FOR SAMPLE CORRELATION MATRIX 6 ________ 1 0.544 EXPLORATORY ANALYSIS WITH 1 FACTOR(S) : CHI-SQUARE VALUE 93.582 DEGREES OF FREEDOM 9 PROBABILITY VALUE 0.0000 RMSEA (ROOT MEAN SQUARE ERROR OF APPROXIMATION) : ESTIMATE (90 PERCENT C.I.) IS 0.080 ( 0.066 0.095) PROBABILITY RMSEA LE 0.05 IS 0.000 ROOT MEAN SQUARE RESIDUAL IS 0.0554 ESTIMATED FACTOR LOADINGS 1 ________ SELF -0.623 WORTH -0.612 OTHER -0.555 SATISF -0.669 RESPECT 0.259 FAILURE 0.367 ESTIMATED RESIDUAL VARIANCES SELF WORTH OTHER SATISF RESPECT ________ ________ ________ ________ ________ 1 0.611 0.625 0.692 0.552 0.933 ESTIMATED RESIDUAL VARIANCES FAILURE ________ 1 0.865 EXPLORATORY ANALYSIS WITH 2 FACTOR(S) : CHI-SQUARE VALUE 18.600 DEGREES OF FREEDOM 4 PROBABILITY VALUE 0.0009 RMSEA (ROOT MEAN SQUARE ERROR OF APPROXIMATION) : ESTIMATE (90 PERCENT C.I.) IS 0.050 ( 0.028 0.074) PROBABILITY RMSEA LE 0.05 IS 0.461 ROOT MEAN SQUARE RESIDUAL IS 0.0195 VARIMAX ROTATED LOADINGS 1 2 ________ ________ SELF 0.632 0.003 WORTH 0.625 0.002 OTHER 0.555 0.007 SATISF 0.661 0.008 RESPECT -0.225 -0.021 FAILURE -0.259 -10.844 PROMAX ROTATED LOADINGS 1 2 ________ ________ SELF 0.632 -0.004 WORTH 0.625 -0.005 OTHER 0.555 0.001 SATISF 0.661 0.001 RESPECT -0.225 -0.018 FAILURE -0.045 -10.846 PROMAX FACTOR CORRELATIONS 1 2 ________ ________ 1 1.000 2 0.030 1.000 ESTIMATED RESIDUAL VARIANCES SELF WORTH OTHER SATISF RESPECT ________ ________ ________ ________ ________ 1 0.601 0.609 0.692 0.563 0.949 ESTIMATED RESIDUAL VARIANCES FAILURE ________ 1 -116.659 EXPLORATORY ANALYSIS WITH 3 FACTOR(S) : CHI-SQUARE VALUE 0.000 DEGREES OF FREEDOM 0 PROBABILITY VALUE 0.0000 ROOT MEAN SQUARE RESIDUAL IS 0.0001 VARIMAX ROTATED LOADINGS 1 2 3 ________ ________ ________ SELF 0.658 0.086 0.073 WORTH 0.558 0.166 0.070 OTHER 0.305 1.331 0.065 SATISF 0.648 0.115 0.121 RESPECT -0.210 -0.020 -0.185 FAILURE -0.139 -0.057 -1.362 PROMAX ROTATED LOADINGS 1 2 3 ________ ________ ________ SELF 0.689 -0.038 -0.037 WORTH 0.566 0.066 -0.024 OTHER 0.072 1.340 0.003 SATISF 0.667 -0.006 0.014 RESPECT -0.203 0.024 -0.154 FAILURE 0.013 -0.003 -1.373 PROMAX FACTOR CORRELATIONS 1 2 3 ________ ________ ________ 1 1.000 2 0.353 1.000 3 0.274 0.097 1.000 ESTIMATED RESIDUAL VARIANCES SELF WORTH OTHER SATISF RESPECT ________ ________ ________ ________ ________ 1 0.554 0.656 -0.869 0.553 0.921 ESTIMATED RESIDUAL VARIANCES FAILURE ________ 1 -0.878 PLOT INFORMATION The following plots are available: Histograms (sample values) Scatterplots (sample values) Eigenvalues for exploratory factor analysis Beginning Time: 11:14:56 Ending Time: 11:15:00 Elapsed Time: 00:00:04 MUTHEN & MUTHEN 3463 Stoner Ave. Los Angeles, CA 90066 Tel: (310) 391-9971 Fax: (310) 391-8971 Web: www.StatModel.com Support: Support@StatModel.com Copyright (c) 1998-2005 Muthen & Muthen |
|
BMuthen posted on Thursday, September 15, 2005 - 2:32 pm
|
|
|
A negative residual variance may indicate overextraction of factors. This does not imply that one factor is sufficient. Perhaps the best model has one factor and some minor factors that can be captured by correlated residuals. Please do not post full outputs in Mplus Discussion. |
|
Tracy Witte posted on Monday, December 01, 2014 - 11:52 am
|
|
|
I have a question about factor loadings larger than 1.0 when doing an EFA with ordinal variables, using the WLSMV estimator. In one of my factor solutions, one item has a geomin rotated loading of 1.203 on one of the factors. However, none of the items have negative residual variances. Additionally, there are no other warnings in my output, and the solution appears to have converged properly. My question is this: do loadings greater than 1.0 with the WLSMV estimator in EFA necessarily constitute a Heywood case? Should I not be concerned since the residual variances are all positive? Thank you very much for your time! |
|
|
See our FAQ Standardized coefficient greater than 1 |
|
Back to top |
|
|