Universities are supposed to be places where all students can learn, free from discrimination.
A key part of this ideal is academics welcoming all students to study and research, regardless of their racial background.
But as our new research shows, Australian academics responded differently to potential PhD students, depending on whether they were called “Melissa” or “Rahul”.
Racism on campus
Many overseas and Australian studies have shown racism is both a historical and ongoing problem for universities.
A 2020 Australian study showed universities tend to be run by older, white men. A 2021 UK study showed academics from different cultural backgrounds face racism at work.
But there has been less specific attention paid to those trying to become academics.
The main way people start an academic career is via a doctoral degree. In the Australian system, before a student is accepted they usually require an established academic to agree to supervise them. So a student’s initial communication with a potential supervisor is very important.
How we set up our research
To investigate whether racism is playing a role at the entrance point to PhD study, in 2017 we sent about 7,000 emails from fictitious students to academics based at the main campuses of Australia’s Group of Eight universities (billed as Australia’s top research universities).
These are the Australian National University, Monash University, University of Adelaide, University of New South Wales, University of Melbourne, University of Sydney, University of Western Australia and University of Queensland.
We emailed staff ranked senior lecturer or above, as these are the levels most likely to be supervising PhD students. Academics were identified by university websites, and we sent emails to everyone who fit our rank criteria across all disciplines.
In this process, we found 70% of relevant academics were male and 84% were white. This did not improve in the more senior ranks – more than 68% of professors were white men.
What did the email say
The emails asked for an meeting to talk about potential PhD supervision.
They were identical apart from the senders’ names. These names were tested to be associated with male and female and with white-European, Indigenous, South Asian, Chinese and Arab identities. Recipients were randomly allocated to different name groups.
The emails indicated the sender was an Australia-based student with fluent English. It conveyed an interest in the recipient’s research and urgency in meeting because the sender was only on campus for several days. It also noted “I have recently finished my honours degree” (a common path into a PhD in Australia) and was sent from a University of Sydney email address.
What did we find
Responses agreeing to a meeting or requesting further information were categorised as “positive”. Those who declined a meeting were “non-positive”. Automated replies and those who did not reply were “non-responses”.
Of 6,928 emails sent, 2,986 (43.1%) received a reply within 24 hours and 2,469 (35.6%) received a positive reply. There were 3,942 (56.9%) non-responses and 517 (7.5%) non-positive responses (declining a meeting).
We initially planned to give academics a week to respond, but after IT at one university noticed several staff had received emails with identical text, we ended the experiment after 24 hours.
From here, the results were stark: emails from names associated with non-white racial groups received significantly fewer responses and positive replies than those from names typically associated with white individuals.
An email from “Melissa Smith” was far more likely to get a positive response than an identical email from “Grace Chen Jinyan” (six percentage points lower) or “Omar al-Haddad” (nine percentage points lower).
The most dramatic gap was in the positive response rates to Melissa Smith, compared with “Rahul Kumar”. The rate of positive responses to Melissa was 12 percentage points higher than for Rahul.
Overall, our statistical analysis showed the white-sounding names averaged a 7% higher reply rate and a 9% higher positive response rate than the non-white sounding names. Both these findings were highly statistically significant, meaning we can be very confident the results were not due to chance.
Of course, some faculty members may simply have been unable to meet with the student, or may have missed the email. However, given the randomisation used, it is reasonable to assume bias explains the gap in responses to students with different names.
This is alarming because it suggests racial bias is quietly influencing who gets a foot in the door of academia even before formal admissions processes begin.
Silver linings
One seemingly positive finding was academics at the more junior end of our study group appeared to show less bias towards students of different backgrounds.
For academics at senior lecturer or associate professor levels, Melissa was 10.5% more likely to receive a positive response than Rahul, while the corresponding figure for full professors was 14.7%.
However, junior academics often have little institutional power or much of a say on hiring. More research is needed to explore whether generational change is achievable (albeit painfully slow).
We also found that, unlike similar US studies, there was no significant bias against female students. In fact, there was some evidence of positive bias, or preference, for female students.
Backlash to our study
We based our study on a peer-reviewed study carried out in the United States, and followed a research ethics protocol approved by our university.
However, minutes after academics received our follow-up email telling them they had been part of a research study (part of our ethics protocol), the backlash began.
The University of Sydney, our home institution at the time, received more than 500 inquiries about the study. While some were curious or supportive, the majority were complaints. These were primarily about our use of deception (a well-researched and supported method of studying bias). Megan MacKenzie, the more junior author (at the time a senior lecturer), received calls threatening her with consequences for her career.
Although unpleasant, the reaction was revealing. It reinforces other research on how defensive racial majorities can be when they believe they are suspected of bias. It also complements work showing internal resistance to diversity efforts in higher education.
What can we do
Universities pride themselves on being meritocracies, where the best ideas and brightest minds rise to the top. But our study suggests racial bias is undermining this principle by influencing who is even considered for an academic career.
There is growing acknowledgement racism is a significant problem on Australian university campuses (as well as in broader society). In May, the federal government asked the Australian Human Rights Commission to study the prevalence and impact of racism at Australian universities.
But this study is not due to deliver its final report until June 2025, and any ensuing action will be further away still.
What can be done now to tackle this issue?
First, universities need to acknowledge academia remains overwhelmingly white and male, in spite of efforts to increase diversity.
Second, universities also need to acknowledge the existence of racial bias, the need for ongoing research into how it operates in higher education and the most effective strategies to tackle it.
Megan MacKenzie is Professor and Simons Chair in International Law and Human Security, Simon Fraser University.
Benjamin E Goldsmith is Professor of Political Science, Australian National University
This article was first published on The Conversation.
