If GRFILE=filename is specified, a file is output which contains a list of measures relative to item difficulty (x-axis coordinates) and corresponding expected scores and category probabilities (y-axis coordinates) to enable you to use your own plotting program to produce item-category plots like those in the Graphs menu.
GRFILE=? opens a Browse window
Column Heading
|
Column Contents
|
Fixed Columns
|
ITEM (or ITEM= value)
|
Item number
|
6
|
MEAS (Measure without item difficulty added)
MEAS+ (Measure with item difficulty added)
|
IADD=No (default): The measure relative to item difficulty (user-rescaled by USCALE=) with UDECIMALS= decimal places.
The plotted range is at least MRANGE= away from its center.
IADD=Yes adds the item difficulty to MEAS for absolute measure relative to the latent variable.
|
7
|
SCOR (Score)
|
Expected score for the Modeled Item Characteristic Curve (ICC) or Item Response Function (IRF).
|
7
|
INFO (Information)
|
Statistical information for Item Information Function. Sum these for all items for the Test Information Function (TIF).
|
7
|
0 (Category number)
|
Probability of observing lowest category
|
7
|
1, 2, etc. (Category numbers)
|
Probability of observing higher categories
|
7
|
If CSV=Y, values are separated by commas. When CSV=T, values are separated by tab characters.
|
For the numbers plotted for empirical ICCs, see the "Empirical ICC" plot, "Copy data to Clipboard"
Example 1: You wish to write a file on disk called "MYDATAGR.txt" containing x- and y-coordinates for plotting a generic item's category response curves and its item information function. Item difficulty is not added.
IADD=No
GRFILE=MYDATAGR.txt
With CSV=Y
"PROBABILITY CURVES FOR LIKING FOR SCIENCE (Wright & Masters p.18) Aug 13 2023 12:40"
"ITEM","MEASURE","SCORE","INFORMATION",0,1,2
1,-4.00,.0420,.0409,.9583,.0414,.0003
1,-3.99,.0423,.0412,.9580,.0417,.0003
1,-3.98,.0427,.0415,.9577,.0420,.0003
1,-3.98,.0430,.0418,.9573,.0423,.0003
1,-3.97,.0433,.0421,.9570,.0427,.0003
With CSV=N (fixed spacing)
; PROBABILITY CURVES FOR LIKING FOR SCIENCE (Wright & Masters p.18) Aug 13 2023 12:40
ACT MEAS SCOR INFO 0 1 2
1 -4.00 .0420 .0409 .9583 .0414 .0003
1 -3.99 .0423 .0412 .9580 .0417 .0003
1 -3.98 .0427 .0415 .9577 .0420 .0003
1 -3.98 .0430 .0418 .9573 .0423 .0003
1 -3.97 .0433 .0421 .9570 .0427 .0003
Example 2: I want the information function for every item in my test relative to the latent variable (absolute x-axis).
IADD=Yes ; item difficulty is added
GRFILE=MYDATAGR.txt
Columns: ITEM, MEAS, INFO
; PROBABILITY CURVES FOR LIKING FOR SCIENCE (Wright & Masters p.18) Aug 13 2023 12:40
ACT MEAS+ SCOR INFO 0 1 2
1 -7.40 .0021 .0021 .9979 .0021 .0000
1 -7.38 .0022 .0022 .9978 .0022 .0000
1 -7.37 .0022 .0022 .9978 .0022 .0000
1 -7.36 .0022 .0022 .9978 .0022 .0000
1 -7.34 .0023 .0023 .9977 .0023 .0000
Example 3:The Category Information Functions are the item information function multiplied by the category probability for each category for each measure. Use Excel for the calculation.
IADD=Yes
GRFILE=MYDATAGR.xls
