McConnell Library Scholars Repository

Are We Trading One Bias for Another?: Consideration of Computer and Human Evaluation of Resumes

Hardin, Hannah (2016) Are We Trading One Bias for Another?: Consideration of Computer and Human Evaluation of Resumes. Masters thesis, Radford University.

This is the latest version of this item.

[img]
Preview
PDF - Accepted Version
Available under License Creative Commons Attribution No Derivatives.

Download (2MB) | Preview

Abstract

Research on resumes has largely focused on biases concerning applicant characteristics, largely ignoring how resumes are analyzed by humans. Additionally, technological advances in resume screening including the use of computer aided text analysis presents a gap in research, which the current study addresses through the comparison of human evaluation and computer aided text analysis of resumes. Researchers predicted that human raters would be more accurate than computer systems when hiring applicants from resume ratings, that computer systems using synonyms would be less accurate than those that used single words when assessing resumes with more ambiguity (i.e. of average quality), and that computer systems would be less accurate than human raters when assessing resumes with ambiguity. Using signal detection theory, results demonstrated that computer systems were more accurate than human raters when ambiguity is introduced, but equally as accurate as human raters for high quality resumes (low ambiguity), regardless of using synonym or single word systems. Additionally, research found that human intuition based hiring was the least accurate method, as well as the most liberal. Human hiring decisions made from ratings and subsequent rankings (logic based) including ambiguous resumes were more accurate than intuition-based methods but less accurate than hiring decisions for non-ambiguous resumes. Human and computer ratings were equally as hiring high quality (not ambiguous) resumes. The current study provides initial evidence that computer systems used in resume screening provide a valid, reliable alternative to human-based manual scoring of resumes.

Item Type: Thesis (Masters)
Subjects: B Philosophy. Psychology. Religion > BF Psychology
H Social Sciences > H Social Sciences (General)
H Social Sciences > HA Statistics
Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: Radford University > College of Graduate and Professional Studies
Depositing User: Hannah Hardin
Date Deposited: 31 May 2019 15:15
Last Modified: 31 May 2019 15:15
URI: http://wagner.radford.edu/id/eprint/290

Available Versions of this Item

  • Are We Trading One Bias for Another?: Consideration of Computer and Human Evaluation of Resumes. (deposited 31 May 2019 15:15) [Currently Displayed]

Actions (login required)

View Item View Item