Why all the coefficients are negative?

[A.Faruk KILIÇ]

Hi everybody;

i am newbie about R. I estimate in IRTPRO and R's MIRT package for a two-dimensional and dichotomous scoring test. 

In R, all the a1 coefficients are negative. However, all of the IRTPRO's estimates are positive and their estimates are different. How can I resolve this situation?

I use mirt packaged via this code:

b23PLM=mirt(df,2, technical = list(NCYCLES=10000), itemtype = "3PL", method = 'EM', SE=1)

coef(b23PLM)

Results for MIRT Package

$ï..M1

             a1     a2       d     g  u

par      -8.422  2.883 -12.581 0.173  1

CI_2.5  -13.028 -0.794 -19.349 0.155 NA

CI_97.5  -3.817  6.560  -5.813 0.192 NA

$M3

             a1     a2       d     g  u

par      -7.106  3.327 -13.336 0.189  1

CI_2.5  -12.287 -0.489 -22.744 0.172 NA

CI_97.5  -1.926  7.143  -3.928 0.208 NA

$M5

             a1     a2       d     g  u

par      -7.270  1.833  -8.137 0.247  1

CI_2.5  -10.506 -1.245 -12.108 0.219 NA

CI_97.5  -4.034  4.912  -4.165 0.278 NA

$M6

             a1     a2       d     g  u

par      -4.505 -1.577  -7.411 0.222  1

CI_2.5  -10.638 -5.013 -17.409 0.192 NA

CI_97.5   1.628  1.858   2.586 0.256 NA

$M7

            a1     a2      d     g  u

par     -3.336  1.046 -5.343 0.200  1

CI_2.5  -4.686 -0.383 -7.439 0.175 NA

CI_97.5 -1.985  2.475 -3.247 0.227 NA

$M8

            a1    a2      d     g  u

par     -2.418 1.471 -5.240 0.193  1

CI_2.5  -3.874 0.091 -8.042 0.166 NA

CI_97.5 -0.963 2.850 -2.439 0.224 NA

$M9

             a1     a2       d     g  u

par      -9.957  3.604 -13.903 0.208  1

CI_2.5  -17.718 -1.308 -24.858 0.188 NA

CI_97.5  -2.195  8.517  -2.948 0.230 NA

$M10

             a1     a2       d     g  u

par      -7.466  3.022 -12.476 0.094  1

CI_2.5  -12.009 -0.355 -19.658 0.080 NA

CI_97.5  -2.922  6.399  -5.293 0.110 NA

$M11

             a1     a2       d     g  u

par      -9.155  0.358 -10.356 0.273  1

CI_2.5  -26.171 -3.258 -30.386 0.242 NA

CI_97.5   7.860  3.973   9.673 0.307 NA

$M12

            a1     a2      d     g  u

par     -1.372 -0.191 -1.839 0.282  1

CI_2.5  -2.524 -0.915 -4.011 0.141 NA

CI_97.5 -0.220  0.533  0.333 0.483 NA

$M13

            a1     a2      d     g  u

par     -1.405 -0.053 -1.214 0.107  1

CI_2.5  -2.463 -0.738 -3.268 0.004 NA

CI_97.5 -0.347  0.631  0.839 0.782 NA

$M14

            a1    a2      d     g  u

par     -2.539 1.378 -4.032 0.147  1

CI_2.5  -3.810 0.157 -5.890 0.108 NA

CI_97.5 -1.267 2.600 -2.173 0.196 NA

$M15

            a1     a2      d     g  u

par     -3.901 -0.382 -5.890 0.262  1

CI_2.5  -6.261 -2.062 -9.543 0.234 NA

CI_97.5 -1.541  1.298 -2.237 0.293 NA

$M16

             a1     a2       d     g  u

par      -6.103  2.994 -11.481 0.346  1

CI_2.5  -11.176 -0.574 -20.806 0.324 NA

CI_97.5  -1.029  6.562  -2.156 0.368 NA

$M17

            a1     a2       d     g  u

par     -3.670  0.843  -7.482 0.214  1

CI_2.5  -5.682 -0.678 -11.294 0.195 NA

CI_97.5 -1.657  2.364  -3.670 0.235 NA

$M19

            a1     a2      d     g  u

par     -2.822  1.088 -3.629 0.196  1

CI_2.5  -4.012 -0.175 -5.383 0.151 NA

CI_97.5 -1.631  2.352 -1.875 0.251 NA

$M20

             a1 a2       d     g  u

par      -5.321  0 -12.742 0.209  1

CI_2.5  -14.011 NA -32.378 0.192 NA

CI_97.5   3.368 NA   6.893 0.227 NA

Result for IRTPRO

Item	Label	a1		a2		c		logit g		g
1	VAR1		4.22		-0.12		-6.01		-1.79	0.14
2	VAR2		0.73		-0.35		-1.68		-1.45	0.19
3	VAR3		4.29		-0.24		-8.05		-1.51	0.18
4	VAR4		3.66		0.81		-6.30		-1.08	0.25
5	VAR5		2.93		0.14		-2.48		-1.80	0.14
6	VAR6		1.43		1.42		-3.06		-1.49	0.18
7	VAR7		2.22		-0.50		-3.44		-1.57	0.17
8	VAR8		2.43		-0.62		-5.06		-1.44	0.19
9	VAR9		7.49		0.47		-10.29		-1.39	0.20
10	VAR10		5.28		-0.10		-9.04		-2.32	0.09
11	VAR11		3.82		1.43		-4.29		-1.09	0.25
12	VAR12		2.97		0.99		-5.15		-0.51	0.37
13	VAR13		1.44		0.79		-1.98		-1.45	0.19
14	VAR14		1.88		0.02		-3.10		-1.74	0.15
15	VAR15		1.90		1.44		-3.66		-1.13	0.24
16	VAR16		2.75		-0.83		-5.14		-0.80	0.31
17	VAR17		3.00		1.06		-6.87		-1.32	0.21
18	VAR18		1.37		0.42		-2.74		-1.41	0.20
19	VAR19		2.62		0.01		-3.20		-1.46	0.19
20	VAR20		1.35		0.17		-4.98		-1.41	0.20

[Phil Chalmers]

This is an exploratory model, therefore the raw slopes are meaningless and simply indicate the arbitrary orientation which the model last converged to. You want to use something like summary() to rotate the orientation. If you are familiar with EFA with continuous variables (via the psych package, for example) then this type of interpretation is essentially the same. 

Phil

--
You received this message because you are subscribed to the Google Groups "mirt-package" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mirt-package+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.