Responses are revalued according to the matching codes in IVALUE=. If this implies that the items may have different rating (or partial credit) scale structures, so ISGROUPS= may also be required.
IREFER= *file name
|
file containing details
|
IREFER = *
|
in-line list
|
IREFER = codes
|
codes for item groups
|
IREFER= has three forms: IREFER=AABBCDAAD and IREFER=* list * and IREFER=*filename
Valid one-character IREFER= codes include: !#$%&-./123456789<>@ABCDEFGHIJKLMNOPQRSTUVWXYZ^_|~
A-Z are the same as a-z.
Characters with ASCII codes from 129-255 can also be used, but display peculiarly: ÎÏÐÑÒÓÔÕÖ×ØÙÚÛÜÝÞßàáâã etc.
When XWIDE=2 or more, then
either (a) Use one character per XWIDE and blanks,
NI=8
XWIDE=2
IREFER=' A B C D D C B A'
or (b) Use one character per item with no blanks
NI=8
XWIDE=2
NEWSCORE='ABCDDCBA'
Item identifying codes can be letters or numbers. "A" is the same as "a", etc.
Example 1. There are 3 item types. Items are to rescored according to Type A and Type B. Other items to keep original scoring.
CODES = 1234
IREFER = AAAAAAAABBBBBBBB******* 3 item types: ("a" is the same as "A" in these codes)
IVALUEA = 1223 Recode Type A items
IVALUEB = 1123 Recode Type B items
IVALUE* = 1234 Recode Type * item. Can be omitted
or
IREFER=*
1-8 A
9-16 B
17-23 *
*
or
IREFER=*filename.txt
in filename.txt:
1-8 A
9-16 B
17-23 *
Example 2. There are 3 item types. Responses are 2 characters wide. Items are to rescored according to Type A and Type B. Other items to keep original scoring.
XWIDE=2
CODES = '1 2 3 4 '
IREFER = AAAAAAAABBBBBBBB******* 3 item types
IVALUEA = '1 2 2 3 ' Recode Type A items
IVALUEB = '1 1 2 3 ' Recode Type B items
IVALUE* = 1234 Recode Type * item. Can be omitted
Example 3: All items are to be rescored the same way
NI = 100 100 ITEMS
IREFER=*
1-100 X FOR ALL 100 ITEMS, reference is X
*
Codes = 12345 rescore 12345
IVALUEX = 12223 into 12223
Example 4: Items are to be rescored in 3 different ways, forward scoring, reversed scoring and dichotomized scoring. Then the items are to be divided into 4 rating scale structures.
ISGROUPS=11112223312444 ; 4 RATING SCALE GROUPINGS
IREFER =AAAABBBCCABBBB ; 3 RECODINGS
CODES =01234 ; ORIGINAL CODES IN DATA
IVALUEA =01234 ; ORIGINAL CODING MAINTAINED - THIS LINE CAN BE OMITTED
IVALUEB =43210 ; CODING IS REVERSED
IVALUEC =*112* ; DICHOTOMIZED WITH EXTREME CATEGORIES MARKED MISSING
Example 5: Multiple-choice test with 4 options, ABCD
IREFER=ABCDDABADCDA ; SCORING KEY
CODES =ABCD ; VALID DATA CODES
IVALUEA=1000 ; A SCORED 1
IVALUEB=0100 ; B SCORED 1
IVALUEC=0010 ; C SCORED 1
IVALUED=0001 ; D SCORED 1
MISSCORE=0 ; EVERYTHING ELSE IN THE DATA SCORED 0
Example 6: A 10 item multiple choice test with 4 partial credit performance items scores 0,1,2,3.
NI = 14 ; 14 items
; 12345678901234
ISGROUPS=DDDDDDDDDD0000 ; 10 Dichotomies in Group "D". 4 Partial Credit items each in its own group "0"
IREFER =ABCDDABADC1111 ; ITEM REFERENCE FOR THE 14 ITEMS
CODES =ABCD0123 ; VALID DATA CODES
IVALUEA=10000000 ; A SCORED 1
IVALUEB=01000000 ; B SCORED 1
IVALUEC=00100000 ; C SCORED 1
IVALUED=00010000 ; D SCORED 1
IVALUE1=00000123 ; PERFORMANCE ITEMS: 0,1,2,3
MISSING-SCORED=0 ; EVERYTHING ELSE IN THE DATA SCORED 0
Example 7: Partial credit items are to be rescored:
IREFER = ABCAAB ; there are six items items 1, 4 and 5 have one rescoring. Items 2 and 6 another rescoring. Item 3 a third rescoring
CODES = 1234 ; the original scoring
IVALUEA = 1123 ; the new scoring for A-type items
IVALUEB = 1334 ; for B-type items
IVALUEC = 1234 ; original scoring maintained
Now there is a decision:
does 134 mean the same as 123, so that 2 is a structural zero, if so STKEEP=NO
does 134 mean the same as 1(2)34, so that 2 is an incidental or sampling zero, if so STKEEP=YES
Example 8: An MCQ item where two boxes can be checked for one item, e.g., choose 2 options out of 5.
Load the data into Excel. Each one-box item has a separate column. The two-box item is converted into 5 columns, one for each option. For example, 2 out 5 becomes 5 dichotomies, one for each option:, A-E,
Each dichotomy is scored 1 if the response is right, 0 if wrong, so
the two correct answers are scored 1 if selected, 0 if not
the three incorrect answers are scored 0 if selected, 1 if not
then
the two correct dichotomies are weighted 1
and the three incorrect dichotomies are weighted 0
Use Winsteps Excel/RSSST menu to create the Winsteps data file.
