Factor Loading larger than 1 PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
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!
 Bengt O. Muthen posted on Monday, December 01, 2014 - 5:47 pm
See our FAQ

Standardized coefficient greater than 1
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: