EFILE=filename produces an output file containing the Rasch-model expected score on each item for the measures in the operational range of the set of items.
Subsets of items can be reported by using IDELETE= or ISELECT= from the Specification menu.
Options using ETYPE=
ETYPE=E
|
Expected score on the item for the ability measure according to the Rasch model. For dichotomous items scored 0/1, this is the probability of success.
|
ETYPE=P
|
Most probable category on the item for the ability measure according to the Rasch model. If more than one category, then middle or immediately higher category is shown.
|
ETYPE=T
|
Thurstonian category on the item for the ability measure according to the Rasch model. If more than one category, then middle or immediately higher category is shown.
|
Example 1: ETYPE=E with dichotomous 0-1 items.
; EXPECTED SCORES FOR Knox Cube Test (Best Test Design p.31)
MEASURE SCORE 4 5 14 15 16 17
-7.00 .11 .07 .04 .00 .00 .00 .00
-6.85 .12 .08 .04 .00 .00 .00 .00
-6.70 .14 .09 .05 .00 .00 .00 .00
-6.55 .16 .10 .06 .00 .00 .00 .00
.....
7.70 5.74 1.00 1.00 .98 .92 .92 .92
7.85 5.78 1.00 1.00 .98 .93 .93 .93
8.00 5.81 1.00 1.00 .99 .94 .94 .94
MEASURE -7.00
|
ability measure in logits or user-scaled units (UIMEAN=, USCALE=)
|
SCORE .11
|
expected raw score = sum of expected scores on the items with item weighting if IWEIGHT=
|
4, 5, 14, 15, ...
|
item entry numbers
|
.07, .04, .00, ...
|
Rasch-model expected score or category on the item for the MEASURE
|
Example 2: Reckase Chart of Expected Scores on All Items for All Measures.
Reckase, M. D. (2001). Innovative methods for helping standard-setting participants to perform their task: The role of feedback regarding consistency, accuracy, and impact. In: Gregory J. Cizek (Ed.), Setting performance standards: Concepts methods and perspectives (pp. 159-173). Mahwah, NJ: Erlbaum.
MacCann & Stanley, "Rasch Standard Setting", Practical Assessment Research & Evaluation, 11, 2, 2, Table 1.
Example 3: Angoff standard-setting ratings to measures to cut-scores
In https://scholarworks.umas.edu/cgi/viewcontent.cgi?article=1188&context=par "The Use of Rasch Modeling To Improve Standard Setting", their Table 1 is the EFILE= output
Their Table 2 is deduced from the EFILE= output
Expected Score using Excel
Example of 5-category rating scale:
Column A: item difficulty
Column B: 1
Column C: exp(-item difficulty - threshold 1)
Column D: Column C * exp(-item difficulty - threshold 2)
Column E: Column D * exp(-item difficulty - threshold 3)
Column F: Column E * exp(-item difficulty - threshold 4)
Column G: Column B + C + D + E + F
Column H: ( Column C + 2 * Column D + 3 * Column E + 4 * Column F ) / Column G = Expected score
