NotesFAQContact Us
Search Tips
ERIC Number: ED186444
Record Type: Non-Journal
Publication Date: 1979-Sep-21
Pages: 41
Abstractor: N/A
Reference Count: N/A
An Empirical Study of the Use of Item-Item Curves to Detect Item Bias: Preliminary Report.
Goldberg, Jill; Hulin, Charles L.
A Job Descriptive Index was studied for possible bias by examining item-item curves for various pairs of items computed for contrasting samples. The two pairs of contrasting samples were 669 part-time workers versus 920 full-time workers and 486 male workers versus 1,146 female workers in a large merchandising company. Item-item curves were computed by obtaining the scores for all the workers on the 18-item, 3-choice scale; then stratifying each sample on the basis of total score into groups of about 50 each; finding the percent of yes responses in each group for each item; and finally plotting percents of yes responses on various pairs of items for the two contrasting samples on the same pair of item axes. In general, the items had closely similar curves for the contrasting samples except for one item that was chosen much more frequently by part-time workers than by full-time workers. (CTM)
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A