Example 10: Combining tests with common items

Top Up Down  A A

This uses MFORMS=, but it can also be done, more awkwardly, with FORMAT=

 

Test A, in file EXAM10A.TXT, and TEST B, in EXAM10B.TXT, are both 20 item tests. They have 5 items in common, but the distractors are not necessarily in the same order. The responses must be scored on an individual test basis. Also the validity of each test is to be examined separately. Then one combined analysis is wanted to equate the tests and obtain bankable item measures. For each file of original test responses, the person information is in columns 1-11, the item responses in 41-60.

 

The combined data file specified in EXAM10C.TXT, is to be in RFILE= format. It contains

 

Person information of 11 characters: Columns 1-30 (always), but only 1-11 are active.

Item responses to 35 items: Columns 31-64

 

The identification of the common items is:

 

Test Item Number (=Location in item string)

Bank

1

2

3

4

5

6-20

21-35

A:

3

1

7

8

9

2, 4-6, 10-20

 

B:

4

5

6

2

11

 

1, 3, 7-10, 12-20

 

I. From Test A, make a response (RFILE=) file..

 

; This file is EXAM10A.TXT

TITLE="Analysis of Test A"

RFILE=EXAM10AR.TXT ; The constructed response file for Test A

NI=20   ; 20 items

ITEM1=41  ; Items start in column 41 of data record

NAME1=1  ; Start of person label

NAMELEN=11  ; Length of person label

CODES="ABCD# " ; Beware of blanks and # meaning wrong!

   ; Blanks are included in CODES=, but they are scored incorrect, because never keyed correct

KEY1=CCBDACABDADCBDCABBCA   ; the MCQ key

&END

BANK 2   TEST A 1 ; first item name

BANK 6   TEST A 2

BANK 1   TEST A 3

 .

BANK 20  TEST A 20

END NAMES

Person 01 A                             BDABCDBDDACDBCACBDBA

 .

Person 12 A                             BADCACADCDABDDDCBACA

 

The RFILE= file, EXAM10AR.TXT, is:

 

0        1         2         3         4         5

12345678901234567890123456789012345678901234567890

Person 01 A                   00000000110010001001

.........

Person 12 A                   00001110000001001011

 

II. From Test B, make a response (RFILE=) file.

 

; This file is EXAM10B.TXT

TITLE="Analysis of Test B"

RFILE=EXAM10BR.TXT ; The constructed response file for Test B

NI=20

ITEM1=41  ; Items start in column 26 of reformatted record

NAME1=1  ; Start of person label

NAMELEN=11  ; Length of person label

CODES="ABCD# " ; Beware of blanks meaning wrong!

KEY1=CDABCDBDABCADCBDBCAD  ; Key in data record format

&END

BANK 21  TEST B 1

BANK 4   TEST B 2

BANK 22  TEST B 3

  .

BANK 35  TEST B 20

END NAMES

Person 01 B                             BDABDDCDBBCCCCDAACBC

  .

Person 12 B                             BADABBADCBADBDBBBBBB

 

The RFILE= file, EXAM10BR.TXT, is:

 

Person 01 B                   01110101011001000100

...

Person 12 B                   00000001010000101000

 

III. Analyze Test A's and Test B's RFILE='s together:

 

; This file is EXAM10C.TXT

TITLE="Analysis of Tests A & B (already scored)"

NI=35   ; 35 items in total

ITEM1=31  ; reformatted data record

CODES=01  ; scored right-wrong. 

   ; Blanks ignored as "not administered"

MFORMS=*        ; multiple data files in different formats

DATA=EXAM10AR.txt ; first file

L=1             ; one line per data record

P1-11 = 1       ; person id starts in column 1 of input file

I1=33           ; item 1 is Test A's item 3 in Test A's column 33

I2=31           ; item 2 is Test A's item 1 in column 31

I3-5=37         ; items 3-5 are Test A's items 7-9 starting in column 37

I6=32

I7-9=34

I10-20=40

#

DATA=EXAM10BR.txt ; second data file

L=1

P1-11 = 1       ; person id starts in column 1 of input file

I1-3 = 34       ; items 1-3 are Test B's items 4-6 starting in Test B's column 34

I4 = 32         ; item 4 is Test B's item 2 in column 32

I5 = 41         ; items 5 in Test B's item 11 in column 41

I21 = 31

I22 = 33

I23-26 = 37

I27-35 = 42

*

&END

BANK 1   TEST A 3 B 4

...

BANK 35  TEST B 20

END NAMES

 

The combined data file (which can be accessed from the Edit pull-down menu) is:

 

Person 01 A                   00001000010010001001

...

Person 12 A                   00100001100001001011

Person 01 B                   10111               010101001000100

...

Person 12 B                   00000               000101000101000

 

After running EXAM10C, I want to see two ICCs: One for test A and another Test B. How do I do this?

 

This graph is not produced directly by Winsteps, but can be produced in Excel.

 

 

After the Exam10C analysis,

use the "Specification" pull-down menu to delete items not in Test A: Use IDELETE=21-35

Display the Test Characteristic Curve

Select "Copy data to clipboard".

Paste into an Excel worksheet

 

Use the "Specification" pull-down menu to reinstate all items: IDELETE=+1-35

 

Use the "Specification" pull-down menu to delete items not in Test B: Use IDELETE=6-20

Display the Test Characteristic Curve

Select "Copy data to clipboard".

Paste into an Excel worksheet

 

In Excel, scatterplot the pasted columns.