skip to main content skip to footer

A Method for Estimating from Speeded Test Data the Power Condition Scores and Item Difficulties

Author(s):
Lord, Frederic M.
Publication Year:
1950
Report Number:
RB-50-24
Source:
ETS Research Bulletin
Document Type:
Report
Page Count:
8
Subject/Key Words:
Difficulty Level, Item Analysis, Power Tests, Test Items, Timed Tests

Abstract

A method for estimating from speeded test data is presented. The procedure produces estimates of scores and item difficulty from data on ordinary time limit tests that would have been attained if the test had been given with no time limit. The data are presented in matrix form. The matrix of data was rearranged. It is supposed that there are 100 examinees and 100 test items. The first r items are reached by all testees; the last 100-r items are not. Also, the first c examinees reached all 100 items; the remaining 100-c did not. The number of correct answers of each testee on the first r items is the first estimate (x') of the true power score of that examinee. The number of right answers attained by the first c testees is the first estimate (d') of the power-condition item difficulty. x" is the second estimate of the testee's power score. d" is an improved estimate of the power condition item difficulty. (SGK).

Read More