Facets handles mixtures of rating scales and partial credit items within the same test easily using Models=.  Datasets containing only MCQ items can be analyzed.

 

Items 1-10 are dichotomies, already scored 0,1

Items 11-14 are 3-category rating scales 1,2,3 - each with its own scale definition, i.e., partial credit items.

Items 15 is a binomial count 0-10 successes

 

Facet 1 is person, Facet 2 is judge (only for items 11-14), Facet 3 is item:

 

Models=

?, X, 1-10, D  ; items 1-10 -dichotomies - no judge

?, ?, 11-14#,R3 ; items 11-14 - each with own scale

?, X, 15, B10 ; item 15 - 10 binomial trials

*

 

Here are typical lines of data for person 23:

data=

23,0,1-10,0,1,1,1,1,0,0,0,1,1 ; person 23, with no judge, responds to items 1-10, scored 0 or 1.

23,3,11-14,3,2,3,1  ; person 23, judge 3, rated on items 11-14 on rating scales with highest category 3.

23,6,12-14,1,,3   ; person 23, judge 6, rates only items 12 and 14

23,0,15,7   ; person 23, no judge, 7 successes out of 10 binomial trials.

 

---

To score multiple-choice MCQ dichotomies:

 

The scoring key for the 10 items is: CADBDCDABB

 

Also shown are items 11-14 which are partial credit or rating scale items.

Item 15 is a binomial trial item giving successes out of 10.

 

Models=

?,X,1,C  ; using scoring key C as the model code

?,X,2,A

?,X,3,D

?,X,4,B

?,X,5,D

?,X,6,C

?,X,7,D

?,X,8,A

?,X,9,B

?,X,10,B

?, ?, 11-14#,R3 ; items 11-14 - each with own scale

?, X, 15, B10 ; item 15 - 10 binomial trials

*

Rating (or partial credit) scale=A,D ; Scale A - dichotomous

0=Wrong,,,B+C+D+ ; add other possible wrong responses here

1=Right,,,A  ; ",,," is to skip over anchor value positions

*

Rating (or partial credit) scale=B,D

0=Wrong,,,A+C+D+.    ; add other possible wrong responses here

1=Right,,,B

*

Rating (or partial credit) scale=C,D

0=Wrong,,,A+B+D+.    ; add other possible wrong responses here

1=Right,,,C

*

Rating (or partial credit) scale=D,D

0=Wrong,,,A+B+C+.    ; add other possible wrong responses here

1=Right,,,D

*

Here are typical lines of data for person 23:

data=

23,0,1-10,D,B,C,A,.,D,A,B,D,C ; person 23, with no judge, responds to items 1-10

23,3,11-14,3,2,3,1  ; person 23, judge 3, rated on items 11-14

23,6,12-14,1,,3   ; person 23, judge 6, rates only items 12 and 14

23,0,15,7   ; person 23, no judge, 7 successes out of 15.