Date Approved
2025
Degree Type
Open Access Senior Honors Thesis
Department or School
Economics
First Advisor
Jenni Putz, Ph.D.
Second Advisor
Amanda Stype, Ph.D.
Third Advisor
Barbara Patrick, Ph.D.
Abstract
Ghost jobs are advertised positions where companies have no intent to hire which create significant frustration for job seekers. This study explores how to automatically detect these fake listings using public data. Analyzing 849 LinkedIn job postings, I developed a two-step detection method. First, I calculated a ghost job score based on warning signs like listing duration and vague descriptions. Second, I used a BERT neural network to analyze the text patterns. The results demonstrate that machine learning can identify deceptive hiring practices, offering a new way to measure labor market inefficiencies.
Recommended Citation
Mathison, Caleb, "Classification of ghost job listings: A two step BERT approach" (2025). Senior Honors Theses and Projects. 881.
https://commons.emich.edu/honors/881