As the political world anxiously awaits the release of Robert Mueller’s report Thursday morning, much of the focus has been on what we won’t see. It’s quite likely the version that’s released to the public will be heavily redacted. Though legitimate reasons exist for the government to excise sensitive information from a public document, any omissions threaten to inspire conspiracy theories about why parts of the report was suppressed, particularly after Attorney General William Barr rushed out his own interpretation of Mueller’s findings — which favored President Trump — in a letter to Congress within 48 hours of receiving the document.

Barr said the delay in the report’s public release could be attributed to the tedious process of redacting key information. He specifically said government lawyers will omit secret grand jury testimony, information that could compromise intelligence sources and methods or ongoing investigations, and information about private citizens who are peripheral to the case.

The problem is that Barr cannot credibly commit to only redact information that is truly sensitive, especially after he already said the evidence “is not sufficient to establish that the President committed an obstruction-of-justice offense.” (A conclusion, it should be noted, that Mueller himself never reached.) The power of redaction, which necessarily withholds information, is vulnerable to exploitation. As New York’s Adam K. Raymond notes, “Some Democrats are worried that Barr will use his powers of redaction to block information just because it’s embarrassing to Trump or his allies.” Indeed, Democrats have already authorized Jerry Nadler, the chair of the House Judiciary Committee, to subpoena the full report, though it is not clear if they will be successful or, for that matter, if the full document will ever be released publicly.

The time it has taken to release the report may be the result of slow bureaucratic review processes, not malevolence. Similarly, any overuse of redactions may be the result of risk aversion by overly cautious government lawyers rather than partisan efforts to prevent the release of damaging information. But in this charged context, suspicions are high, and even innocuous delays and redactions could be seen as efforts to protect Trump.

When people were exposed to redacted information rather than unredacted information, they were significantly more likely to endorse conspiracy beliefs.

In their research into the origins of conspiracy theories, the social psychologists Albert A. Harrison and James Moulton Thomas found that belief in government cover-ups often stems from attributions like these, which frequently blame government actors for intentionally withholding or suppressing information and neglect the organizational factors that contribute to such delays and omissions. Harrison and Thomas note how time pressures and jurisdictional disputes between government bodies produced anomalies and conflicting reports about the assassination of John F. Kennedy. Skeptics seized on these discrepancies, helping to fuel lasting conspiracy theories about a cover-up.

Redactions have always been particularly vulnerable to conspiracy theories; they literally cover up information in documents. That’s why redactions inspire so much suspicion. When, for example, lawmakers decided to redact a section in a congressional report into the September 11 terrorist attacks that discussed possible Saudi involvement in the plot, government skeptics seized on the opportunity. The so-called “28 pages” quickly became a focal point for critics of the government. As Sharon Premoli, the co-chair of 9/11 Families United for Justice Against Terrorism, said in 2014, “The redaction of the 28 pages has become a cover-up by two presidents, and cover-up implies complicity.” These suspicions helped force the release of the 28 pages in 2016, but despite their “near-mythic status” among lawmakers and victims’ families, the Washington Post concluded that the redacted pages “do not appear to add significantly to information collected in subsequent investigations.”

In 2014, my students and I conducted an experimental study to measure the effects of redactions on conspiracy beliefs. We expected redactions to make people more likely to infer that the government was engaged in some sort of malfeasance — and therefore more likely to endorse a conspiracy theory about the events or information in question. We tested this hypothesis in the context of the crash of TWA Flight 800, which exploded after takeoff from John F. Kennedy International Airport in 1996. According to the National Transportation Safety Board (NTSB), which conducted an extensive investigation, “the probable cause… was an explosion of the center wing fuel tank, resulting from ignition of the flammable fuel/air mixture in the tank.” Still, conspiracy theories have repeatedly been offered about the crash — including that the plane was shot down — though the NTSB found no evidence to support them.

To test whether redactions inspire conspiracy theories, my students and I presented respondents in an online survey with government documents about the crash that we altered to include or omit apparent redactions. The documents were otherwise identical and included identical visible text, allowing us to isolate the effect of exposure to redactions alone, rather than the combined effects of seeing redactions and of not seeing the redacted information.

The results were clear. When people were exposed to redacted information rather than unredacted information, they were significantly more likely to endorse conspiracy beliefs about the TWA Flight 800 crash: that it was shot down by a missile, that the government was involved, or that the government was covering up the true cause. They were more skeptical about the NTSB’s explanation for the crash and the credibility of the government investigation.

We also found that the presence of redactions undermined the effectiveness of the evidence presented to counter conspiracy theories. Exposure to documents from the investigation reduced conspiracy beliefs and increased belief in the official account among respondents, but only among those who saw unredacted documents. The very presence of redactions, in other words, prevented the evidence from helping to change people’s minds about the cause of the crash.

We should worry that the redactions in Mueller’s report will have the same effect on Trump’s most sensationalistic critics, who are committed to the worldview that he and his campaign were acting on Russia’s behalf. In reality, though Trump’s associates had extensive Russian contacts, Mueller concluded — at least according to Barr — that “the investigation did not establish that members of the Trump campaign conspired or coordinated with the Russian government in its election interference activities.”

The Twitter conspiracy entrepreneurs who have found a large audience since Trump’s victory are already downplaying Mueller’s reported conclusions, which they alternately attribute to an ever-growing web of deceit or dismiss as a prelude to future investigations to come. Any redactions may provide an escape hatch that allow them, and the Democrats who follow them, to continue to believe what they want to believe, evidence be damned.