DISFILE=filename produces an output file containing the counts for each distractor or option or category of each item, similar to Table 10.3. See also Distractor Analysis
DISFILE=? opens a Browse window
DISOPTION= Code or DISOPTION= Score summary. These control the contents of the output fields.
With DISOPTION= Code
ITEM CODE VALUE SCORE UNWTD UNWTD % WTD WTD % AVGE MEAS P.SD MEAS S.E. MEAS INFT MNSQ OUTF MNSQ PTMA LABEL
13 * -1 -1 17 56.67 17.0 56.67 1.14 1.99 .50 .00 .00 .46 nh01 supermarket
13 a 0 0 2 15.38 2.0 15.38 -1.07 .12 .12 .68 .61 -.31 nh01 supermarket
13 b 0 0 1 7.69 1.0 7.69 -.13 .00 .00 1.73 1.55 .14 nh01 supermarket
13 c 1 1 8 61.54 8.0 61.54 -.58 .82 .31 1.35 1.40 -.11 nh01 supermarket
13 d 0 0 2 15.38 2.0 15.38 .14 .60 .60 2.26 2.43 .36 nh01 supermarket
With DISOPTION= Score
ITEM CODE VALUE SCORE UNWTD UNWTD % WTD WTD % AVGE MEAS P.SD MEAS S.E. MEAS INFT MNSQ OUTF MNSQ PTMA LABEL
13 -1 17 56.67 17.0 56.67 1.14 1.99 .50 .00 .00 .46 nh01 supermarket
13 0 5 38.46 5.0 38.46 -.40 .68 .34 1.41 1.53 .11 nh01 supermarket
13 1 8 61.54 8.0 61.54 -.58 .82 .31 1.35 1.40 -.11 nh01 supermarket
This file contains 1 heading lines (unless HLINES=N), followed by one line for each CODES= of each item (whether the code is observed for the item or not) containing:
Columns:
|
|
|
Start:
|
End:
|
Heading
|
Description
|
1
|
10
|
ITEM
|
Item entry number
|
11
|
20
|
CODE
(blank for DISOPTION=
S)
|
Response code in CODES= or *** = missing (not subtotal)
|
21
|
25
|
VALUE
(blank for DISOPTION=
S)
|
Numerical value of the response code (directly or with KEYn=, NEWSCORE=, IVALUEx=, MISSCORE= etc.). -1 means "response is ignored"
|
27
|
30
|
SCORE
|
Score for category (DISOPTION=C) or summed across categories (DISOPTION=S) used for estimating the Rasch item measure (after recounting, weighting, etc.). -1 means "response is ignored"
|
31
|
40
|
UNWTD
|
Count of all valid response codes in data set (excludes missing except for "MISSING" lines)
|
41
|
50
|
UNWTD %
|
Percent of UNWTD responses to this item.
|
51
|
60
|
WTD
|
Count of all valid response codes (weighted with PWEIGHT=) in data set (excludes missing except for "MISSING" lines)
|
61
|
70
|
WTD %
|
% of WTD responses to this item
|
71
|
80
|
AVGE MEAS
|
Average person measure for WTD responses
|
81
|
90
|
P.SD MEAS
|
P.SD is the population standard deviation of the person measures for WTD responses
|
91
|
100
|
S.E. MEAS
|
Standard Error of average person measure for WTD responses = sqrt (P.SD (AVGE MEAS) / (WTD - 1) )
|
101
|
101
|
INFT MNSQ
|
Infit Mean-square of weighted responses, omitting responses in extreme scores
|
111
|
120
|
OUTF MNSQ
|
Outfit Mean-square of weighted responses, omitting responses in extreme scores
|
121
|
130
|
PTMA
|
PTMA is the point-correlation between this code or summary score, scored 1, or its non-occurrence, scored 0, and the person raw scores or measures chosen by PTBISERIAL=. The computation is described in Correlations.
|
132
|
-
|
LABEL
|
Item label
|
DISOPTION=Code - Since the DISFILE= has the same number of CODES= and MISSING "***" entries for every item, the repeated fields are filled out with "0" for any unobserved response codes.
DISOPTION=Score - Only observed scores are reported.
Percentages for non-missing response-scores are computed based on total count of non-missing response scores. Percentages for missing response-scores are computed based on total count of all response scores.
When CSV=Y, commas separate the values with quotation marks around the "Item label", response in CODES=, and ***. When CSV=T, the commas are replaced by tab characters.
Example 1: You wish to write a file on disk called "DISCOUNT.TXT" containing the item distractor counts from Table 14.3, for use in constructing your own tables:
DISFILE=DISCOUNT.TXT
Example 2: You want to see if some distractors are more attractive for boys or girls.
1.Do your usual Winsteps analysis. Output the DISFILE= to Excel: this has counts for all persons for all distractors. PFILE=pf.txt
2.Put PSELECT=girls in your Winsteps control file. Do your usual Winsteps analysis. Output the DISFILE= to Excel: this has counts for girls. PAFILE=pf.txt to make the measures in the different analyses comparable.
3.Put PSELECT=boys in your Winsteps control file. Do your usual Winsteps analysis. Output the DISFILE= to Excel: this has counts for girls. PAFILE=pf.txt
4.Combine the DISFILE= Excel worksheets to compute percentages for each distractor. The DISFILE= also shows which option is correct.
