IFILE= item output file

Up  Previous  Next

IFILE=filename produces an output file containing the information for each item. This file contains 4 heading lines (unless HLINES=N), followed by one line for each item containing, for fixed fields and the standard field selection:

 

Columns:

 

 

 

 

Start

End

Format

Description

1

1

A1

Blank or ";" if HLINES=Y and there are no responses or deleted or extreme (status =0,-1, -2, -3)

2

6

I5

1. The item sequence number (ENTRY)

7

14

F8.2

2. Item's calibration (user-rescaled by UMEAN=, USCALE=, UDECIM) (MEASURE)

15

17

I3

3. The item's status (STATUS)

 

 

 

 

2 = Anchored (fixed) calibration

 

 

 

 

1 = Estimated calibration

 

 

 

 

0 = Extreme minimum (estimated using EXTRSC=)

 

 

 

 

-1 = Extreme maximum (estimated using EXTRSC=)

 

 

 

 

-2 = No responses available for calibration

 

 

 

 

-3 = Deleted by user

 

 

 

 

-4 = Inestimable: high (all responses in the same category with ISGROUPS=0 or CUTHI=)

 

 

 

 

-5 = Inestimable: low (all responses in the same category with ISGROUPS=0 or CUTLO=)

 

 

 

 

-6 = Deselected

18

25

F8.1

4. The number of responses used in calibrating (COUNT) or the observed count (TOTAL=Y)

26

34

F8.1

5. The raw score used in calibrating (SCORE) or the observed score (TOTAL=Y)

35

41

F7.2

6. Item calibration's standard error (user-rescaled by USCALE=, UDECIM=) (ERROR)

42

48

F7.2

7. Item infit: mean square infit (IN.MSQ)

49

55

F7.2

8. Item infit: t standardized (ZSTD), locally t standardized (ZEMP) or log-scaled (LOG)

56

62

F7.2

9. Item outfit: mean square outfit (OUT.MS)

63

69

F7.2

10. Item outfit: t standardized (ZSTD), locally t standardized (ZEMP) or log-scaled (LOG)

70

76

F7.2

11. Item displacement (user-rescaled by USCALE=, UDECIM=) (DISPLACE)

77

83

F7.2

12. Item by test-score correlation: point-biserial (PTBS) or point-measure (PTME)

84

90

F7.2

13. Item weight (WEIGHT)

91

96

F6.1

14. Observed percent of observations matching prediction (OBSMA)

97

102

F6.1

15. Expected percent of observations matching prediction (EXPMA)

103

109

F7.2

16. Item discrimination (this is not a parameter, merely a descriptive statistic)

 

 

 

 

 

If ASYMPTOTE=Yes:

110

115

F6.2

17. Item lower asymptote (LOWER)

116

121

F6.2

18. Item upper asymptote(UPPER)

... add 12 to the next column locations

 

 

 

 

 

 

If PVALUE=Yes:

110

115

F6.2

19. Item p-value or average rating (PVALUE)

... add 6 to the next column locations

 

 

 

 

 

 

110

110

1X

 Blank

 

111

111

A1

20. Grouping to which item belongs (G)

112

112

1X

 Blank

 

113

113

A1

21. Model used for analysis (R=Rating, S=Success, F=Failure) (M)

114

114

1X

 Blank

 

115

144+

A30+

22. Item name (NAME)

 

 

 

 

 

The format descriptors are:

 

In = Integer field width n columns

 

Fn.m = Numeric field, n columns wide including n-m-1 integral places, a decimal point and m decimal places

 

An = Alphabetic field, n columns wide

 

nX = n blank columns.

 

 

When CSV=Y, commas separate the values, which are squeezed together without spaces between. Quotation marks surround the "Item name", e.g., 1,2,3,4,"Name". When CSV=T, the commas are replaced by tab characters.

 

Example:        You wish to write a file on disk called "ITEM.CAL" containing the item statistics for use in updating your item bank, with values separated by commas:

                       IFILE=ITEM.CAL

                       CSV=Y

 

When W300=Yes, then this is produced in Winsteps 3.00, 1/1/2000, format:

 

Columns:

 

 

 

 

Start

End

Format

Description

1

1

A1

Blank or ";" if HLINES=Y and there are no responses or deleted (status = -2, -3)

2

6

I5

1. The item sequence number (ENTRY)

7

14

F8.2

2. Item's calibration (user-rescaled by UMEAN=, USCALE=, UDECIM) (MEASURE)

15

17

I3

3. The item's status (STATUS)

 

 

 

 

2 = Anchored (fixed) calibration

 

 

 

 

1 = Estimated calibration

 

 

 

 

0 = Extreme minimum (estimated using EXTRSC=)

 

 

 

 

-1 = Extreme maximum (estimated using EXTRSC=)

 

 

 

 

-2 = No responses available for calibration

 

 

 

 

-3 = Deleted by user

18

23

I6

4. The number of responses used in calibrating (COUNT) or the observed count (TOTAL=Y)

24

30

I6

5. The raw score used in calibrating (SCORE) or the observed score (TOTAL=Y)

31

37

F7.2

6. Item calibration's standard error (user-rescaled by USCALE=, UDECIM=) (ERROR)

38

44

F7.2

7. Item mean square infit (IN.MSQ)

45

51

F7.2

8. Item infit: t standardized (ZSTD), locally t standardized (ZEMP) or log-scaled (LOG)

52

58

F7.2

9. Item mean square outfit (OUT.MS)

59

65

F7.2

10. Item outfit: t standardized (ZSTD), locally t standardized (ZEMP) or log-scaled (LOG)

66

72

F7.2

11. Item displacement (user-rescaled by USCALE=, UDECIM=) (DISPLACE)

73

79

F7.2

12. Item by test-score correlation: point-biserial (PTBS) or point-measure (PTME)

80

80

1X

15. Blank

 

81

81

A1

16. Grouping to which item belongs (G)

82

82

1X

17. Blank

 

83

83

A1

18. Model used for analysis (R=Rating, S=Success, F=Failure) (M)

84

84

1X

19. Blank

 

85

132+

A30+

18. Item name (NAME)

 

Example of standard IFILE=

 

; ITEM  Knox Cube Test (Best Test Design p.31)  Nov 10 15:40 2005

;ENTRY MEASURE STTS COUNT    SCORE  ERROR IN.MSQ IN.ZSTD OUT.MS OUT.ZSTD DISPL PTME WEIGHT OBSMA EXPMA  DISCR G M NAME

;    1   -6.74 -1    34.0     34.0   1.87   1.00    .00   1.00    .00    .00    .00   1.00    .0    .0   1.00 1 R 1= 1-4

    4   -4.54  1    34.0     32.0    .82    .92   -.16    .35   -.24    .00    .55   1.00  94.1  94.1   1.07 1 R 4= 1-3-4


Help for WINSTEPS® Rasch Measurement Software: www.winsteps.com.