I began by researching the causes of discrimination in recruiting and analyzing three HR software tools. Using insights from inclusive recruiting methods, I created series of interface interventions that incorporate anonymized profiles, positive discrimination, and AI assisted evaluations to promote a more equitable digital recruiting experience.

Problem

In Germany, discrimination in job application procedures still happens frequently, with ethnic origin being one of the most common causes.

This is why I investigated for my master’s thesis project, how HR software can contribute to supporting inclusion. My aim was to develop interface interventions that reduce discrimination and support the equitable treatment of all applicants.

Process

First, I analyzed three recruiting software solutions to find out how discrimination occurs in the interfaces and whether it is triggered rather by the technology or by human behavior. The main findings are:

Problems in the analyzed recruiting softwares

1. Too much personal data provokes prejudice
2. Lack of standardization leads to undifferentiated assessments
3. Screening based solely on application documents reproduces structural discrimination
4. Filters for educational qualifications enable direct discrimination

I then researched methods from literature in the fields of consulting and social sciences that can contribute to more equitable recruiting processes. The following methods are applicable to the software:

Based on these findings, I developed my interventions for the improvement of recruiting interfaces. To test their potential for reducing discrimination, I discussed them in a small testing with the same three recruiting experts.

Result

I developed a set of interface interventions for recruiting software to help reduce discrimination in application processes. They are based on findings from scientific literature and my own analysis of best practice softwares. In the following, I explore different approaches designed to help recruiters treat all candidates and resumes equally, and make more objective hiring decisions.

For a more detailed explanation of my concept, feel free to read my master thesis below (in German).

Overview

List of Applicants

Applied Methods

• Anonymization
• 2 ranking modes
• Highlighted diversity attributes
• Positive discrimination

Ranking Mode 1: by Criteria from the Job Description

Ranking Mode 2: by Tests & Evaluations

Profile

Applied Methods

• Anonymization
• Standardized layout of the CV for all applicants
• Adaptation to requirements for the application process

Anonymized State

Visible State

Evaluation

Applied Methods

• More objective evaluation through AI transcripts
• Standardized evaluation forms

Feedback from Expert Testing

To evaluate my concepts, I conducted a mini testing with a focus group of the three recruiting experts from the beginning of my research. I first used the Thinking Aloud method to check whether the screens were plausible and then asked them about the potential they see in the interventions.

Project Outlook

The next step should be to verify in a larger testing group whether the interventions produce the desired effects. This would require a detailed prototype in which recruiters can go through the entire process for an application file.

At the same time, the effects on applicants should be investigated. Specifically, it should be tested whether applicants are actually open to stating diversity attributes.

‍