The report from FBI special counsel Robert Mueller was released to the public Thursday — and, as expected, large chunks of his findings are redacted.
In theory, the government has to release as much as it possibly can. But as former White House lawyer Andy Wright told Vox’s Alex Ward, “A redaction is sort of a compromise between withholding a document entirely versus releasing the full document.” For what it’s worth, Attorney General William Barr said the White House did not play a part in the redaction process — and that instead, it was DOJ lawyers, Mueller’s team, and members of the intelligence community.
So what was redacted?
We can’t see behind the black bars. But we can learn some things from where they’re placed — and how many there are.
The most redacted sections were on Russia’s hacking and election interference
One way to see which sections are blacked out most is by just looking at the pages.
But to get a little more insight, we tried to approximate how much each page was redacted by section.
We wrote a computer program to look at the millions of pixels on each page and find how many of them were black. (More about methodology at the bottom.) This analysis found that about 7 percent of the entire report was redacted — but some sections were withheld much more than others.
In the chart below, the full length of the bar represents how long each section is. The black portion is how much each section was redacted.
Here, it’s pretty clear that the section titled “Russian ‘Active Measures’ Social Media Campaign” was among the most redacted. From the unredacted portions, we can tell that it’s about Mueller’s findings on Russia’s efforts to sow discord around the election using social media.
Another heavily redacted section is the one called “Russian Hacking and Dumping Operations,” which explains Mueller’s findings on Russia’s role in hacking top Democrats’ emails and releasing them to the public.
The primary reason for these redactions was that it could cause a “harm to ongoing matter,” with a faraway second being information that would reveal investigative techniques. One of these matters is likely the trial of Roger Stone, the Trump associate who allegedly sought out those emails and lied to Congress about it. It also appears the details of investigations into Russian election interference have been withheld. For example, a section describing the structure of the Internet Research Agency, which tried to sow political discord on social media, was almost entirely redacted.
Why parts of the Trump-Russia collusion section were redacted
There are four reasons Barr gave for why certain sections would be redacted, and good reasons for them:
- Revealing it would harm an ongoing matter. Ward writes, “Some redactions in the report will exist mainly to allow other related cases to continue unimpeded.”
- It reveals investigative techniques. Ward writes that “it’s highly likely the US government had to use sensitive spying methods — like, say, an undercover agent or top-secret surveillance technology — to investigate.”
- It reveals private information of third-party individuals. Ward writes the “Justice Department has long had a policy of not divulging people’s names during an investigation unless they are indicted.”
- It was obtained via grand jury testimony. This is secret by law.
We ran a similar pixel analysis (and double-checked it manually) to see what kind of redaction rationale was most common in each section.
This reveals that most of the ongoing matter redactions were in sections about Russian election interference and hacking.
There’s also a decent portion of material redacted because the “Justice Department has long had a policy of not divulging people’s names during an investigation unless they are indicted,” Ward writes.
But when it came to the section on the Trump campaign’s links and contacts with Russia, the primary redaction rationale was because the information was obtained through grand jury testimony.
Many of these redactions hide details about Trump associates’ interaction with Russian officials, like the section about the infamous Trump Tower meeting with Russia lawyer Natalia Veselnitskaya.
There’s a decent amount of redactions, but we can read enough to know Trump and his associates were open to Russia’s help
Barr, the Trump-nominated attorney general, initially released a summary of Mueller’s findings. And then he tried to characterize the report as one that says Trump did not “collude” with Russia.
But it turns out Barr left out some crucial information.
Vox’s Zack Beauchamp lays out some of the damning things the report reveals about Trump’s action during the election. This includes Trump directing an adviser to find the hacked Hillary Clinton emails, and approving another adviser’s efforts to set up a meeting between Trump and Russian President Vladimir Putin. And there may well be more under these black bars.
It’s easy to joke about entire pages that are redacted. After all, 40 percent of pages have redactions!
But that shouldn’t be the focus of this report. Much of it — including 96 percent of the section on Trump and his associates’ connection to Russia — is unredacted. From what we can gather from surrounding context, there don’t appear to be any egregious, politically motivated redactions. (House Democrats have subpoenaed the unredacted report.) But even if there were, there is more than enough information to see that Trump and his associates knew about Russia’s attempts to help him win the election — and were willing to receive that help.
Methodology: We wrote a computer program to turn each page of the report into an image. Then we analyzed each pixel of the image to see whether it was black. Obviously, both the text and the redactions are black, which makes this exercise a bit hard. But we found that a page that is unredacted but full of text is about 3.5 percent black, and a page that is entirely redacted is about 34.5 percent black (because of border and imperfections in the redaction). So we were able to use these baselines to estimate how much each page was redacted. To determine the rationale for redactions, we used a similar method to count the number of times each rationale was used by counting pixel colors — and then double-checked it by going to each page by hand and recounting.