Our research examines the perceptions and lived experiences of women, Black and Latinx undergraduate students completing degrees in computing majors who encounter algorithmic systems during the hiring process for entry-level technology positions. Specifically, we use Gilliland's procedural justice rules as a theoretical model for understanding applicants' reactions to Applicant Tracking System (ATS) software employers use to automate recruiting and hiring tasks. Procedural justice rules include formal characteristics like the consistency of the administration of interview questions, the opportunity to perform a pre-employment task, and the job-relatedness of the task. Also included are honest and informative explanations of hiring decisions and being treated respectfully by the hiring organization.

Using surveys and workshops, we found that job seekers' perceptions about the fairness of the testing, interviewing, and other procedures applied during candidate selection process are inextricably linked with perceptions of equity in the outcomes of the selection process. Consistent with prior research on selection systems, women and racially minoritized job seekers perceived hiring procedures to be fairer when they are able to demonstrate their technical skills through a test that consistently and accurately evaluated performance. Fairness is also perceived when job seekers are offered two-way communication with an interviewer who does not exhibit personal bias based on the applicants' race, gender and ethnicity. 

When considering the fairness of the hiring decision, situational factors like the characteristics of the decision-maker (human or algorithm), type of task (programming vs. interviewing), explanation of the decision-making rationale (feedback on why I was/was not hired), and the possibility for human oversight (can a hiring decision be modified if a fault in the algorithm was identified?) become important. The salience of discrimination and performance expectations based on racial and gender stereotypesis critical in determining fairness in the hiring process. Thus, ATS software may reproduce human bias when algorithmic methods for detecting, classifying, and selecting human beings are built without considering the broader historical and social context of racial and gender inequality. 

The broader impacts of our work are immediately found in the demographic diversity of our research team, which reflects the under-representation found in our study participants. In addition, using a solid theoretical framework to empirically assess the reactions of job seekers from historically excluded groups in computing offers complex viewpoints highlighting both the positive and negative sides of ATS software, which provides novel insights for diversifying the technology workforce. From an ethical perspective, organizations should be concerned with the effects of automated recruiting and hiring procedures on the psychological well-being of job candidates. Moreover, organizations should be concerned about how perceived unfairness and algorithmic bias may lead applicants to pursue discrimination cases from a legal perspective.