If a facet entry in the data file has the same value for all data lines, then it can be omitted from the data, and replaced by a Dvalue= statement, e.g., a set of items administered to all persons.

 

If a facet entry in the data file has its value in the label of another element, then it can be omitted from the data, and replaced by a Dvalue= statement, e.g., a demographic indicator in the person element label.

 

I. Elements numbers the same for every data line. This occurs when, for instance, every examinee is rated on items 1-5, or when every patient is observed on items 1-13.

 

The format for one Dvalue= is:

Dvalue = facet entry number in data, constant value

 

The format for several Dvalues= is:

Dvalue=

facet entry number in data, constant value

facet entry number in data, constant value

....

*

 

Example 1: In the Guilford.txt example, the standard layout is:

Facets = 3

Data =

1, 1, 1-5, 5,5,3,5,3

1, 2, 1-5, 9,7,5,8,5

....

3, 7, 1-5, 7,7,7,5,7

 

This is the same as:

Facets = 3

Dvalue = 3, 1-5

Data =

1, 1, 5,5,3,5,3

1, 2, 9,7,5,8,5

....

3, 7, 7,7,7,5,7

 

II. Elements numbers based on line number. This occurs when, for instance, the examinee is the line number in the data

 

The format for one Dvalue= is:

Dvalue = facet entry number in data, program variable

 

The format for several Dvalues= is:

Dvalue=

facet entry number in data, program variable

facet entry number in data, program variables

....

*

 

Program variables are:

$Line = Data-line number in current data file (excluding comment lines and blank lines)

$Cumulative = Data-line number in current data file (including previous data files, if any)

 

Example: for the Knox Cube Test:

dvalues=

1, $Line  ; facet 1: person elements are same as data line number

2, 1-18   ; facet 2: every person responds to 18 items

*

Data =              

1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0 ; child 1, items 1-18

1,1,1,1,1,1,0,0,1,1,1,0,0,1,0,0,0,0

1,1,1,1,1,1,1,1,1,0,1,0,0,0,0,0,0,0

1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0

 

II. An element identifier for one facet (the target) is in the element label of another facet (the reference) for the same observation. If the segment in the reference element label is a number, that becomes the element number for the target facet. If non-numeric, then the segment is matched to the start of the elements labels given for the target facets, up to :. If ":" is omitted, then a match is made to the left-end of the target element label.

The format is:

Dvalue= facet entry number for target facet, reference facet location in data, start column in the label, length of field in the label

The format for several Davalues= is:

Dvalue=

facet entry number for target facet, reference facet location in data, start column in the label, length of field in the label

facet entry number for target facet, reference facet location in data, start column in the label, length of field in the label

........

*

 

Example 1. The Knox Cube Test with interactions example data set contains replacements for facets 3 and 4:

 

 

Example 2. Ethnicity is indicated by a numeric value in column 1 of the student label. Gender is in column 2.

Facets=4  ; persons, item, ethnicity, gender

Models = ?,?B,?,?B,D ; we want to look for item-gender DIF

Labels=

1, Student

234 = 61 Jose  ; Jose is a Hispanic (6 in column 1), male (column 2 of label=1)

235 = 32 Mary  ; Mary is a Caucasian (3 in column 1), female (column 2 of label =2)

......

*

2, Items

1-4   ; 4 dichotomous items, scored 0-1.

*

3, Ethnicity, A  ; elements anchored at 0 to avoid distorting student measures

3, Caucasian, 0

6, Hispanic, 0

*

4, Gender, A  ; elements anchored at 0 to avoid distorting student measures

1, Male, 0

2, Female, 0

*

Dvalues=

2, 1-4   ; all students administered items 1-4

3, 1, 1, 1  ; element number for facet 3 (ethnicity) is in element label for facet 1, column 1 with a length of 1

4, 1, 2, 1  ; element number for facet 4 (gender) is in element label for facet 1, column 2 with a length of 1

*

Data=

234, 0,1,1,0  ; facet 1 element 234 then 4 observations. would be 234, 1-4, 6, 1, 0,1,1,0 without Dvalues =

235, 1,1,0,0  ; facet 1 element 235 then 4 observations. would be 235, 1-4, 3, 2, 1,1,0,0 without Dvalues =

 

Example 3. Gender is indicated by a non-numeric value in column 2 of the student label

Facets=3 ; persons, item, DIF indicator

Models = ?,?B,?B ; we want to look for gender-related item DIF

Labels=

1, Student

234 = 6M Jose  ; Jose is a male (column 2 of label=1)

235 = 3F Mary  ; Mary is a female (column 2 of label =2)

......

*

2, Items

1-4   ; 4 dichotomous items, scored 0-1.

*

3, Gender, A  ; elements anchored at 0 to avoid distorting student measures; only for DIF analysis

1, M:Male, 0  ; M: - colon, so M is used to match the code in the student label

2, Female, 0  ; no :, so match is made to left-end of label for length of dvalues= field (width 1 in this example).

*

Dvalues=

2, 1-4   ; all students administered items 1-4

3, 1, 2, 1  ; element number for facet 3 is in column 2 of element label for facet 1.

*

Data=

234, 0,1,1,0  ; facet 1 element 234 with 4 observations. would be 234, 1-4, 1, 0,1,1,0 without Dvalues =

235, 1,1,0,0  ; facet 1 element 235 with 4 observations. would be 235, 1-4, 2, 1,1,0,0 without Dvalues =

 

Example 4: To test that Dvalues= is functioning correctly:

title= "dvalue test"

facets = 2

residualfile=dvalueresiduals.txt

models=?,?,R

labels=

1, Facet 1

1=Fred

2=Mary

*

2, Facet 2

1=M:supervisor

2=F:janitor

*

dvalues=

2,1,1,1 ; facet 2 is the first character of the element for facet 1

*

data=

1-2, 0,1

1-2, 1,0

 

; this analyzes as (see residualfile=)

; 1 1 2 2 Obs

; 1 Fred 2 F:janitor 0

; 2 Mary 1 M:supervi 1

; 1 Fred 2 F:janitor 1

; 2 Mary 1 M:supervi 0

 

Example 5: Demonstrates element numbers in element labels:

facets = 3

model=?,?,?,R

labels=

1

1-10

*

2

1-10

*

3

1,1

2,1

3,1

4,2

5,2

6,2

7,2

*

dvalue=*

2, 3,1,1

*

data=

1,1-7,1,0,1,0,1,0,1

2,1-7,0,1,0,1,0,1,0