Rating (or partial credit) scale output is easier to understand when the categories are shown with their substantive meanings. Use CFILE= to label categories using their scored values, i.e., after rescoring. Use CLFILE= to label categories using their original codes, i.e., before any rescoring. Labels for the original categories in the data can be specified using CLFILE= and a file name, or CLFILE=* and placing the labels in the control file. Each category number is listed (one per line), followed by its descriptive label. Original category values are used. There are several options:

 

 

 

Its format is:

CLFILE=*

item entry number + category code   category label

%item label + category code  category label

*

 

XWIDE=2 ;observations are two columns wide

CODES = " 0 1 299"       ; codes are 0, 1 , 2, 99

CLFILE=*

99    Strongly Agree ; original code of 99 has the label "Strongly Agree"

2  Agree  ; original code of blank+2 (or 2+blank) has the label "Agree"

2+99 Heartily Agree ; for item 2, code 99 has the label "Heartily Agree"

3+0 Disagree  ; for item 3, code 0 means "Disagree"

*

 

Example 1: Identify the three LFS categories, D=Dislike, N=Don't know, L=Like.

CODES        =DNL

NEWSCORE=012

CLFILE=*

D Dislike

N Neutral

L Like

*

 

The labels are shown in Table 3.2 as:

-------------------------------------------------------

 CATEGORY OBSERVED AVGE  INFIT OUTFIT  STRUCTURE

  LABEL    COUNT   MEASURE MNSQ  MNSQ  MEASURE

-------------------------------------------------------

    0        378     -.87  1.08  1.19   NONE     Dislike

    1        620      .13   .85   .69     -.85   Don't know

    2        852     2.23  1.00  1.46      .85   Like

-------------------------------------------------------

 

Example 2: Items 1-10 (Grouping 1) are "Strong Disagree, Disagree, Agree, Strongly Agree". Items 11-20 (Grouping 2) are "Never, Sometimes, Often, Always".

NI=20

CODES        =ABCD

NEWSCORE=1234

ISGROUPS=11111111112222222222

CLFILE=*

7+A Strongly Disagree ; 7 is any item in Grouping 1

7+B Disagree

7+C Agree

7+D Strong Agree

13+A Never   ; 13 is any item in Grouping 2

13+B Sometimes

13+C Often

13+D Always

*

 

Example 3: To enter CLFILE= information on the DOS Prompt or Extra Specifications lines, using commas instead of blanks as separators:

C:>WINSTEPS SF.TXT SFO.TXT  CLFILE=* D,Dislike N,Don't-know L,Like *

 

Example 4: One grouping of items has a unique response format, but the other groupings all have the same format. Here, each grouping has only one item, i.e., ISGROUPS=0

NI=20

CODES=1234

ISGROUPS=0

CLFILE=*

1 Strongly Disagree ; This rating scale is used by most items

2 Disagree

3 Agree

4 Strong Agree

16+1 Never  ; 16 is the one item using this rating scale

16+2 Sometimes

16+3 Often

16+4 Always

*

Example 5: Several categories are collapsed into one category. The original codes are A-H. After rescoring there is only a dichotomy: 0, 1.

NI=30

CODES = ABCDEFGH

NEWSCORE=00011110

CFILE=*

0 Fail  Specify the categories as recoded

1 Pass

*

; or

CLFILE=*

A Fail

B Fail

C Fail

D Pass

E Pass

F Pass

G Pass

H Pass

*

 

Example 6: Identifying the distractors for a multiple-choice MCQ item.

Here is item 5:

5. The shape of the relationship between raw score and measures on the latent variable is

 

CODES = abcd

 

CLFILE=*

5+a diatonic

5+b harmonic

5+c monotonic

5+d synchronic

*

 

Example 7: Using the item labels

CLFILE=*

%APMED+1 Attentive

%APMED+2 Interested

%APMED+3 Residual

%APMED+9 NOT ASKED THAT YEAR

*