|
In this example,
the LSAT data of Example
2 are analyzed assuming a one-factor model. The purpose of
the analysis is to compare the goodness-of-fit with that of
the two-factor model, and to use the change in chi-square
between the models as a test of statistical significance of
the second factor. The computation of classical item statistics
is skipped (SKIP = 1), and the factor loadings are not rotated
or saved.
>TITLE
LSAT DATA NON-ADAPTIVE FULL-INFORMATION ITEM FACTOR ANALYSIS
TEST OF FIT
>PROBLEM NITEM=5,RESPONSE=3,SKIP=1;
>NAMES ITEM1,ITEM2,ITEM3,ITEM4,ITEM5;
>RESPONSE '8','0','1';
>KEY 11111;
>TETRACHORIC NDEC=3,LIST;
>FACTOR NFAC=1,NROOT=3;
>FULL CYCLES=16;
>TECHNICAL NOADAPT;
>INPUT NIDW=2,SCORES,WEIGHT=PATTERN;
(2A1,5A1,3X,I3)
100000 3 12
200001 6 19
300010 2 1
400011 11 7
500100 1 3
600101 1 19
700110 3 3
800111 4 17
901000 1 10
1001001 8 5
1101010 0 3
1201011 16 7
1301100 0 7
1401101 3 23
1501110 2 8
1601111 15 28
1710000 10 7
1810001 29 39
1910010 14 11
2010011 81 34
2110100 3 14
2210101 28 51
2310110 15 15
2410111 80 90
2511000 16 6
2611001 56 25
2711010 21 7
2811011173 35
2911100 11 18
3011101 61136
3111110 28 32
3211111298308
|
