[R] Partial Credit Models using eRm

Michael Ranopa ranopy at hotmail.com
Mon Mar 15 13:19:54 CET 2010


Hello all,
 
 
I have just started using fitting the PCM (Partial Credit Model) using eRm and have 2 problems which I cannot solve, I have checked everywhere on the net, but no joy:
 
Firstly,
 
I have fit a PCM model to 10 variables with differing response categories (3 for the first 6 items, 4 for the following 2 and 2 for the last two items).
 
mat1 <- matrix(c(rasch_bart$bart_bowel, rasch_bart$bart_blad, rasch_bart$bart_toil, rasch_bart$bart_feed, rasch_bart$bart_dress, rasch_bart$bart_stairs,  rasch_bart$bart_trans, rasch_bart$bart_mob, rasch_bart$bart_groom, rasch_bart$bart_bath), nrow=100, ncol=10)
 
bart_pcm <- PCM(mat1)
summary(bart_pcm)

 
# Item location map
plotPImap(bart_pcm, item.subset=c(1,2))
 
but I can't get the person-item plot to work. I receive an error from R saying:
 
Error in xb[, dim(xb)[2]] : incorrect number of dimensions.
 
 
 
Secondly,
 
How would you interpret the following parameter estimates i.e. the eta and beta estimates?:
 
> summary(bart_pcm)
Results of PCM estimation: 
Call:  PCM(X = mat1) 
Conditional log-likelihood: -206.1496 
Number of iterations: 40 
Number of parameters: 19 
 
Basic Parameters (eta) with 0.95 CI:
       Estimate Std. Error lower CI upper CI
eta 1     0.786      0.565   -0.320    1.893
eta 2    -0.440      0.480   -1.382    0.501
eta 3    -0.709      0.565   -1.817    0.399
eta 4    -0.547      0.428   -1.387    0.292
eta 5    -2.161      0.596   -3.330   -0.992
eta 6     5.036      0.674    3.715    6.358
eta 7     4.250      0.692    2.894    5.606
eta 8    -0.813      0.410   -1.615   -0.010
eta 9    -3.735      0.657   -5.023   -2.447
eta 10   -0.727      0.397   -1.506    0.051
eta 11   -4.671      0.709   -6.061   -3.281
eta 12   -2.417      0.463   -3.324   -1.510
eta 13    3.750      0.585    2.604    4.897
eta 14    2.499      0.613    1.297    3.700
eta 15   -0.589      0.707   -1.975    0.797
eta 16    1.637      0.515    0.627    2.647
eta 17    2.164      0.561    1.065    3.264
eta 18   -0.062      0.656   -1.347    1.223
eta 19   -2.989      0.494   -3.959   -2.020
 
Item Parameters (beta) with 0.95 CI:
            Estimate Std. Error lower CI upper CI
beta I1.c1    -0.262      0.560   -1.360    0.835
beta I1.c2     0.786      0.565   -0.320    1.893
beta I2.c1    -0.440      0.480   -1.382    0.501
beta I2.c2    -0.709      0.565   -1.817    0.399
beta I3.c1    -0.547      0.428   -1.387    0.292
beta I3.c2    -2.161      0.596   -3.330   -0.992
beta I4.c1     5.036      0.674    3.715    6.358
beta I4.c2     4.250      0.692    2.894    5.606
beta I5.c1    -0.813      0.410   -1.615   -0.010
beta I5.c2    -3.735      0.657   -5.023   -2.447
beta I6.c1    -0.727      0.397   -1.506    0.051
beta I6.c2    -4.671      0.709   -6.061   -3.281
beta I7.c1    -2.417      0.463   -3.324   -1.510
beta I8.c1     3.750      0.585    2.604    4.897
beta I8.c2     2.499      0.613    1.297    3.700
beta I8.c3    -0.589      0.707   -1.975    0.797
beta I9.c1     1.637      0.515    0.627    2.647
beta I9.c2     2.164      0.561    1.065    3.264
beta I9.c3    -0.062      0.656   -1.347    1.223
beta I10.c1   -2.989      0.494   -3.959   -2.020
 
Thank you in advance for any help which is much appreciated.
 
Michael
 
 
 
 
 
  		 	   		  
_________________________________________________________________
Tell us your greatest, weirdest and funniest Hotmail stories



More information about the R-help mailing list