We academics write a lot — and we cite a lot, too. We cite peer-reviewed articles, books, reports, and other texts. The problem is, academics don’t always read what they cite. This is not only wrong; it can actually do real damage.

Consider an academic article that came out at the height of the COVID-19 pandemic and argues that doctors’ racist biases are a main reason for the higher COVID-19-related hospitalization and mortality rates among African Americans. It says that “there is evidence of medical bias in the testing and treatment of African-Americans with COVID-19” and cites this report as the source. The problem? The report contains no such evidence.

This could have passed for a mere mistake or a mix-up in an article’s citations. But strangely, this scientific article was not the only one that, over the coming months, attributed this very finding to the same report. Others have, too. This is how we, two psychologists who are interested in issues of racism and discrimination, encountered the report and took to reading it ourselves.

The report summarizes data that were collected by a biotech company called Rubix Life Sciences in early 2020. The company tracked billing information pertaining to health care patients in seven U.S. states who had symptoms related to COVID-19. The report includes a list of medical tests for which billing information was located. However, it does not include information about the numbers or percentages of white versus Black patients who were given each medical test. In other words, it is not possible to make inferences about racial differences — let alone racist biases among doctors — based on this report. Furthermore, the report itself doesn’t make any such statements.

How, then, did a dry report by a biotech data company come to be cited by multiple scientists as containing “evidence” it does not contain? It is quite clear the scientists never read it. We believe we know where they got the idea from: An April 2020 NPR story declared that “an African American with symptoms like cough and fever was less likely to be given one of the scarce coronavirus tests.” What source was linked as a reference? You guessed it: the above-mentioned report, which never studied, and never found, any of this.

We don’t know, of course, why the report got featured on NPR the way it did. It should be acknowledged that NPR has since edited the original article, and its current version no longer claims it as a fact that African Americans are less likely to be referred to COVID-19 testing than whites. We appreciate this correction. Nonetheless, academic publications are not as easily correctable, and scientists must in due diligence check their sources. Our aim, as we write this, is not to point fingers. But we do venture to say that writers, academic or otherwise, should properly read what they cite.

This news story is the earliest mention we could find of Rubix Life Science’s report that didn’t come from the company itself. The earliest scientific article we found that mentions the so-called (erroneous) finding came out in May, and it cites the NPR story itself. Based on this timeline, we infer that the scientific community “discovered” the wrong findings when they got featured in the news. And we think it highly problematic that scientists went on to repeat such a claim without actually reading its source.

What we reveal here goes far beyond a simple, easily forgivable mistake. Endorsing this bogus claim translates to a very serious accusation against doctors and health care providers. It implies that, in spite of numerous laws, antidiscrimination policies, and mandatory diversity training programs, doctors out there have neglected their African American patients in the face of a deadly pandemic. If this claim has no leg to stand on, it doesn’t do a great injustice only to the doctors. Far worse, it might exacerbate the existing feelings of anxiety, helplessness, and medical mistrust among African Americans.

When I (J.L.), an African American man, first read in these articles that African American patients with COVID-19 symptoms were not given the COVID-19 tests they needed, my immediate reaction was “Yes, this sounds about right!” This reaction was based on my personal life experiences, but it was also consistent with something bigger — the repeated messages that I and most other Americans get exposed to, by both academia and the news media, regarding systemic racism and its evils. It made the claim easy to believe, and yet it wasn’t true.

Despite the steps taken to correct the falsehood, the original version of the story keeps circulating. The false claim has been repeated by scientific articles even as late as March 2021, in an availability cascade that is virtually impossible to stop. The false claim has also surfaced in another place of importance: the U.S. Congress. It was cited in a Joint Economic Committee report, saying “there have been reports that Blacks showing symptoms of the coronavirus may be less likely to be tested for it.” Politicians, too, ought to read what they cite. But we mustn’t feed them wrong information.

Academic writing has established, time-honored standards. Why would we compromise them by citing a source we never read? One plausible explanation is being under time constraints. Another explanation, potentially more profound, is that we tend to be less critical — and less inclined to check our sources — when we believe that their claim is true. As psychology has taught us, when a statement is consistent with what we are already thinking, we take it at face value. Many in the academic community believe that systemic racism is a prevailing problem across the United States, including in our health care services. This may be why they won’t stop to question a piece that seems to fit in this puzzle.

Here, we anticipate, many well-meaning readers might ask: “What of it? Why are you nitpicking on one anecdote that happens to be false, when there is so much other evidence for racism out there?” Our answer is: “Because it matters.” We believe that dispensing falsehoods does both science and the African American community a bad service.

Take, for example, the COVID-19 vaccination. Although it is becoming increasingly available, recent data show that vaccination rates among African Americans are still very low. Alongside factors like access and availability, surveys show that many Blacks are reluctant to receive it. Medical mistrust, we see, is a prevailing problem in this community. What can be done, then, to encourage African Americans to benefit from the vaccine? One thing is sure: It will not be achieved by making the medical system look bad undeservedly. And — we know this is a tough thing to say these days — it will not be achieved by rushing to affirm claims of systemic biases instead of assessing them critically.

We should know better than this. The country’s long history of medical racism used to rely on, and be justified by, biased science. Attempting to correct this with scientific writing that is biased in the opposite direction will lead to fear mongering and further interracial distrust. Bad science used to create a racial divide, but it takes open, unbiased, and accurate science to resolve it.