MFORMS= supports the reformatting of input data records, and also equating multiple input files in different formats, such as alternate forms of the same test. Data after END NAMES or END LABELS is processed first, as is data specified by DATA= in the core control file.
MFORMS= file name |
file containing details |
MFORMS= file name + file name + ... |
multiple file names |
MFORMS = * |
in-line list |
MFORMS = ? |
opens a Browser window to find the file |
Question: I've been trying to analyze two test forms, sharing 1-50 as common and 51-70 as unique. The common items also have IWEIGHT=. I am trying to estimate the difficulties for each form. I want to retain the KEY1= so I can have option information (so using CODES=ABCD), but also enable test form comparisons. I've tried building the combined control and data file using MFORMS=, but no success yet. Any advice would be great.
Reply: The items become
1-50 common
51-70 form 1
71-90 form2
and the Item labels, KEY1= and IWEIGHT= match these numbers
Combine the data files by
a) copying your form 1 data into the combined file
followed by
b) items1-50 of your form 2 data
c) and then items 51-70 of your form2 data into the position for items 71-90 of the combined data.
To do this copying, please use "rectangular copying" (alt+Mouse). This is available in Microsoft Word and also the freeware NotePad++ https://notepad-plus-plus.org/
Data reformatted by MFORMS= can be accessed, viewed, edited and "saved as" permanently using the "Edit" pull-down menu. It has a file name of the form: ZMF.....txt
Procedure:
1. Make a list of all your different items. Usually we list the common items first.
2. For each item, list its entry number (sequence number) on each form.
3. Use MFORMS=. Transform the list of item numbers into MFORMS= instructions.
Here is the layout:
MFORMS=* |
|
DATA=forma.txt |
; the name of an input data file |
L=2 |
; there are 2 lines in input data file for each data record |
I1 = 20 |
; response to item 1 of the combined test is in column 20 of the input data file |
I3-5 = 21 |
; items 3, 4, 5 are in columns 21, 22, 23 of the input data file |
I16-20=11 |
; items 16, 17,18, 19, 20 are in columns 11, 12, 13, 14, 15 |
P1=9 |
; the first character of person label is in column 9 of the input data file |
P3-8=1 |
; person label characters 3 through 8 start in column 1 |
C20-24="FORMA" |
; put in columns 20-24 the letters FORMA |
C40-90 = 2:1 |
; put in columns 40-90 the characters in the second line of the data record |
# |
; end of definition - start of next file reformat |
DATA=formb.txt |
; name of input data file |
P3-7=1 |
; information for columns 3-7 of person label starts in column 1 of data record |
..... |
|
* |
; end of mforms= command |
Details: |
|
MFORMS=* or filename |
* instructions follow in control file, and end with another * or instructions are in a file. |
data=filename |
name of input file name to be reformatted. The reformatted records are placed in a temporary work file. This may be accessed from the Edit pull-down menu, and saved into a permanent file. This temporary file is processed after any Data Files specified with the master data= instruction and in the same way, e.g., any FORMAT= command will be applied also to the temporary work file. |
L=nnn |
nnn is the count of lines in each input data record. If L=1 this can be omitted. L=4 means that 4 input data lines are processed for each data record output. |
Cnnn= column in original data record |
nnn is the column number in the formatted data record. XWIDE= does not apply. C10-12 means columns 10, 11, 12 of the formatted record. C1= refers to column 1 of the formatted data record. This can also be used to move item and person information. |
Innn= column in original data record |
nnn is the starting item number in the formatted data record. nnn-mmm are the starting to ending item numbers in the formatted data record. XWIDE= is applied, so that I3-5= with XWIDE=2 means 6 characters. I1= points to column Item1= in the formatted data record. |
Pnnn= column in original data record |
nnn is the starting column number in the person label in the formatted person label. XWIDE= is not applied. P6-8= always means 3 columns starting in column 6. P1= points to column Name1= in the formatted data record. |
.....=nnn |
nnn is the starting column of the only, or the first, line in the input data record. |
....=m:nnn |
m is the line number in each data record nnn is the starting column number of that line |
....="xxxx" |
xxxx is a character constant to be placed in the formatted data record. Note: for I18-20="abc" with XWIDE=2, then response to Item 18 is "ab", 19 is "c ", 20 is " ". |
# |
end of processing of one file, start of the next |
* |
end of Mforms= processing |
Item bank |
Datafile1.txt |
Datafile2.txt |
Datafile3.txt |
1-3 |
1-3 |
1-3 |
1-3 |
4-6 |
4-6 |
- |
- |
7-9 |
- |
4-6 |
- |
10-12 |
- |
- |
4-6 |
Datafile1.txt: (Items 1-6)
TOMY ABCDAB
BILL BCDADD
Datafile2.txt (Items 1-3 and 7-9)
TOTO BBADAB
MOULA BADADD
Datafile3.txt (Items 1-3 and 10-12)
IHSANI ACCDAB
MALIK CBDDCD
Control file:
TITLE="Multiple MCQ forms with one scoring key"
NI=12 ; 12 ITEMS IN TOTAL
ITEM1=11
NAME1=1
CODES="ABCD"
KEY1=BACCADACADDA
mforms=*
data=datafile1.txt ; name of data file
L=1 ; one line per person
P1-10=1 ; person label in columns 1-10
I1-3=11 ; items 1-3 in columns 11-13
I4-6=14 ; items 4-6 in columns 14-16
#
data=datafile2.txt
L=1
P1-10=1
I1-3=11
I7-9=14 ; items 8-9 in columns 14-16
#
data=datafile3.txt
L=1
P1-10=1
I1-3=11
I10-12=14 ; items 10-12 in columns 14-16
*
&END
; item identification here
END NAMES
Here is how the data appear to Winsteps for analysis:
TOMY ABCDAB
BILL BCDADD
TOTO BBA DAB
MOULA BAD ADD
IHSANI ACC DAB
MALIK CBD DCD
Example 3: Test 1 is a 4-item survey. Test 2 is a 4-item survey its first two items are the last two items on Test 1. Those items share the same rating scale, called "C". The other 4 items also share another rating scale, called "U".
Item bank |
Datafile1.txt |
Datafile2.txt |
1-2 |
1-2 |
- |
3-4 |
3-4 |
1-2 |
5-6 |
- |
3-4 |
Datafile1.txt: (Items 1-4)
TOMY 3241
BILL 3352
Datafile2.txt (Items 3-4 and 5-6)
TOTO 3325
MOULA 2143
TITLE="Combing two surveys"
NI=6 ; 6 ITEMS IN TOTAL
ITEM1=11
NAME1=1
CODES="12345" ; all valid codes on all the survey items
ISGROUPS = UUCCUU ; assigns items to rating scales
mforms=*
data=datafile1.txt ; name of data file
L=1 ; one line per person
P1-10=1 ; person label in columns 1-10
I1-4=11 ; items 1-4 in columns 11-14
#
data=datafile2.txt
L=1
P1-10=1
I1-2=13 ; items 1-2 on the second survey are items 3 and 4 on the first survey
I3-4=15 ; items 3-4 in columns 15-16
#
&END
; item identification here
END NAMES
Here is how the data appear to Winsteps for analysis:
TOMY 3241 ; items 1-2 unique, items 3-4 common
BILL 3352
TOTO 3325 ; items 3-4 common, items 5-6 unique
MOULA 2143
Example 4: Test 1 is a 4-item survey. Test 2 is a 4-item survey with two items in common with Test 1 which are to be anchored to their Test 1 values.
Item bank |
data1.txt |
data2.txt |
1 |
1 |
- |
2 |
2 |
3 |
3 |
3 |
- |
4 |
4 |
2 |
5 |
- |
1 |
6 |
- |
4 |
Test 1 has 4 rating scale items. Each item has its own partial-credit structure:
title = "Test 1"
item1 = 1 ; items start in column 1
ni = 4 ; 4 items
name1 = 5 ; person label starts in column 5
namlen = 14 ; length of person name
codes = 01234 ; rating scale
ISGROUPS = 0 ; each item has its own rating scale structure
stkeep = YES ; this is probably what you want for these type of data
data = data1.txt
ifile = items1if.txt ; item calibrations from Test 1 for Test 2 (output)
sfile = items1sf.txt ; structure calibrations from Test 1 for Test 2 (output)
&END
Test 1 item 1
Test 1 item 2
Test 1 item 3
Test 1 item 4
END NAMES
data1.txt is:
1234Person 1-1
3212Person 1-2
......
Test 2 has 4 items. 1 and 4 are new - we will call these items 5 and 6 of the combined Test 1 and 2. Item 2 is Test 1 item 4, and item 3 is Test 1 item 2.
title = "Test 2 (formatted to match Test 1)"
item1 = 1 ; items start in column 1
ni = 6 ; 4 items in Test 1 + 2 more in Test 2
name1 = 7 ; person label starts in column 7
namlen =14 ; length of person name
codes = 01234 ; rating scale
stkeep = YES ; this is probably what you want for these type of data
ISGROUPS = 0 ; each item has its own rating scale structure
iafile = items1if.txt ; item calibrations from Test 1 (input - unchanged)
safile = items1sf.txt ; structure calibrations from Test 1 (input - unchanged)
MFORMS = * ; reformat the Test 2 data to align with Test 1
data = data2.txt ; the name of an input data file
L = 1 ; there is 1 line in input data file for each data record
I2 = 3 ; response to item 2 of Test 1 is in 3 of the data2.txt file
I4 = 2 ; response to item 4 of Test 1 is in 2 of the data2.txt file
I5 = 1 ; item 5 is in column 1 of data2.txt
I6 = 4 ; item 6 is in column 4 of data2.txt
P1-14 = 5 ; the first character of person label is in column 5 of data2.txt for 14 columns.
* ; end of mforms= command
&END
Test 1 item 1 (blank in Test 2)
Test 1 item 2 (Test 2 item 3)
Test 1 item 3 (blank in Test 2)
Test 1 item 4 (Test 2 item 2)
Item 5 (not in Test 1, Test 2 item 1)
Item 6 (not in Test 1, Test 2 item 4)
END NAMES
data2.txt is:
5426Person 2-1
1234Person 2-2
....
The formatted file (see Edit pull-down menu MFORMS==) is
2 456Person 2-1
3 214Person 2-2
....
Example 5. For my computer-adaptive test (CAT), I want a new MFORMS= specification for each respondent.
In your Winsteps control file, put
MFORMS=myrespondents.txt (or any other file name)
Then myrespondents.txt can contain an unlimited number of MFORMS instructions
DATA=respondent1.txt
....
#
DATA=respondent2.txt
....
#
DATA=respondent3.txt
....
#
......
......
#
Example 6: Constructing a combined scoring key.
1. Construct MFORMS= as above.
2. Put each data files scoring key in the format of that data file in a separate data file.
3. In your control file, include:
CODES= list of valid data codes, e.g., CODES=ABCDE
IREFER=*
1-number of items A ; all items will be scored by IVALUEA=
*
IVALUEA=10000
IVALUEB=01000
IVALUEC=00100
IVALUED=00010
IVALUEE=00001
Run your control file with MFORMS=, using the data files containing only the keys.
Look at Edit menu, MFORMS=. It will have a data record for each files key, in the combined file format.
Now, Output Files menu, XFILE=,
clear all settings
check: item entry number
response code in data file
OK
Output to text file space separated.
You will see something like:
; RESIDUAL FILE FOR YOUR TEST
ITEM C
11 A
12 B
13 E
14 A
Delete the two heading lines.
The other lines are the scoring key.
In IREFER=, replace
1-number of items A
with all these lines, or you can also put them in a separate file,, and specify IREFER=theirfilename
IREFER is acting like a key, by referring each observation to the IVALUE which matches its scoring key
Now run your Winsteps control file with the correct data files (delete the scoring keys in the data files).
Winsteps should produce the correct output.
Help for Winsteps Rasch Measurement and Rasch Analysis Software: www.winsteps.com. Author: John Michael Linacre
Facets Rasch measurement software.
Buy for $149. & site licenses.
Freeware student/evaluation Minifac download Winsteps Rasch measurement software. Buy for $149. & site licenses. Freeware student/evaluation Ministep download |
---|
Forum: | Rasch Measurement Forum to discuss any Rasch-related topic |
---|
Questions, Suggestions? Want to update Winsteps or Facets? Please email Mike Linacre, author of Winsteps mike@winsteps.com |
---|
State-of-the-art : single-user and site licenses : free student/evaluation versions : download immediately : instructional PDFs : user forum : assistance by email : bugs fixed fast : free update eligibility : backwards compatible : money back if not satisfied Rasch, Winsteps, Facets online Tutorials |
---|
Coming Rasch-related Events: Winsteps and Facets | |
---|---|
Oct 21 - 22 2024, Mon.-Tues. | In person workshop: Facets and Winsteps in expert judgement test validity - UNAM (México) y Universidad Católica de Colombia. capardo@ucatolica.edu.co, benildegar@gmail.com |
Oct. 4 - Nov. 8, 2024, Fri.-Fri. | On-line workshop: Rasch Measurement - Core Topics (E. Smith, Winsteps), www.statistics.com |
Jan. 17 - Feb. 21, 2025, Fri.-Fri. | On-line workshop: Rasch Measurement - Core Topics (E. Smith, Winsteps), www.statistics.com |
May 16 - June 20, 2025, Fri.-Fri. | On-line workshop: Rasch Measurement - Core Topics (E. Smith, Winsteps), www.statistics.com |
June 20 - July 18, 2025, Fri.-Fri. | On-line workshop: Rasch Measurement - Further Topics (E. Smith, Facets), www.statistics.com |
Oct. 3 - Nov. 7, 2025, Fri.-Fri. | On-line workshop: Rasch Measurement - Core Topics (E. Smith, Winsteps), www.statistics.com |
Our current URL is www.winsteps.com
Winsteps® is a registered trademark