|
Activity Number:
|
479
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Government Statistics
|
| Abstract - #305728 |
|
Title:
|
Multiple-Record Applicants in the Analysis of Hiring Disparity
|
|
Author(s):
|
Charles McGhee*+ and Marika Litras and Michael Sinclair
|
|
Companies:
|
U.S. Department of Labor and Office of Federal Contract Compliance Programs and Office of Federal Contract Compliance Programs
|
|
Address:
|
7600 Oak Knoll Drive, North Richland Hills, TX, 76180,
|
|
Keywords:
|
hiring discrimination ; multiple records ; regression assumptions
|
|
Abstract:
|
The Office of Federal Contract Compliance Programs is responsible for enforcing provisions that prohibit employment discrimination by federal contractors and subcontractors. A significant challenge in analyzing hiring disparity is developing statistical models from data that contain multiple-record applicants (MRAs). This issue is central to hiring evaluations where jobs are seasonal or have high turnover. How multiple-record applicants are analyzed, moreover, has important implications for defining applicants and meeting independence assumptions of logistic regression. This research outlines various settings in which MRAs are common, proposes opportunity- and person-based models for dealing with MRAs, and covers the strengths and weaknesses of these approaches. Case studies in low and high MRA settings are presented.
|
