How NPR’s Stephen Fowler crunches the data on the Epstein story : NPR

SARAH MCCAMMON, HOST:

If you’ve spent any time on the Internet in recent years, you’ve probably heard the name Jeffrey Epstein and the mystery surrounding the records related to his case. Last month, a few days before Christmas, some of these records were officially released. Not a handful of documents, but tens of thousands of pages – court records, emails, photos, all landing at once and instantly lighting up social media, cable news and conspiracy theories.

STEPHEN FOWLER, BYLINE: When the first batch of files came out, well, what I thought was something I couldn’t say on the radio.

MCCAMMON: This is Stephen Fowler. He is a political journalist based in Atlanta. We worked together on NPR’s Washington Desk.

FOWLER: Some of us started going through the documents as soon as they came out. We had to download them. We had to classify them, organize them, sort them and make sure we didn’t miss anything, count the total number of pages and documents to see if anything was missing or needed to be formatted differently.

MCCAMMON: Sorting through this type of data is exactly what Stephen does. He has become our go-to reporter for digging through vast amounts of documents, from government data to the Epstein files, to understand what they actually show and, just as importantly, what they don’t show. And I admit that I got back into this rhythm for a day or two during the holidays, when everyone was throwing themselves into what needed to be done, and that takes a lot of work. So for this week’s Journalist’s Notebook, I wanted to ask Stephen about his process. Where does he start when these files are made public, and how does he navigate through them? It turns out that the work begins long before these documents are published.

FOWLER: We had a story ready to go, saying, hey, the Justice Department started releasing the Epstein files, which we already had before we figured out how we got to this point, what the law said, directing the release, the political fallout. So we didn’t really start from scratch.

MCCAMMON: How do you go about going through those tens of thousands of pages and just figuring out what’s in there and what’s important?

FOWLER: So that prep work that I talked about was important because we didn’t know what was going to be released, when it was going to be released, how it was going to be released – things like what file format it would be, whether there was going to be images or PDFs or videos or anything like that. But we knew there were tens of thousands of other records that had already been made public — court cases, Freedom of Information Act requests from other agencies. There were files in the FBI vault and private records of Epstein that were turned over to Congress through a subpoena. So those of us who were methodically going through these pages had a baseline to look at to understand what was already out there, what it looked like, and to be able to determine if it was new? Oh, it’s redacted. Oh, it’s a copy of the same email chain we saw 10 pages ago, so move on to the next thing.

We had a spreadsheet and Slack chat of notable pages, mentions, and files that were being discovered so we could go back, track, and document. And so everything that you’ve seen online or heard people talking on air about the Epstein files, there’s someone who has personally laid eyes on that document and what it was saying and made sure that that’s what it was saying, and here’s where you can find it.

MCCAMMON: Yeah, I mean, that was one of the challenges, at least during the brief period that I was working on some of these cases. You know, a lot of it – I mean, it was taken – these were files that the Justice Department had received from all sorts of people in all sorts of places, right? And even determining what was legitimate was difficult.

FOWLER: Yeah. This is where there were a lot of challenges because the files that were released were not organized in any order. There was no road map. There was no table of contents. There was no asterisk or disclaimer to say, by the way, we have to publish this by law, but this is actually something that’s not true, and it’s actually something that we’ve encountered. There were files that contained tips from the public that contained conspiracy theories about Donald Trump and Jeffrey Epstein and other things that were completely unverifiable and unverified, but they were part of the file that was supposed to be released by law, so they were released. So it was very difficult to know exactly what we were looking at.

Another thing that made it difficult to review, Sarah, was the sheer number of things that were redacted. I mean, there were pages and pages where all you could see were black boxes. There were no hidden words there or, you know, this part here. There were only complete pages completely blackened and blank. It was therefore difficult to derive information from the black rectangles.

There was, however, an email sent by federal agents after Epstein’s 2019 arrest for alleged sex trafficking of minors. It mentioned 10 possible co-conspirators, including Epstein’s accomplice, Ghislaine Maxwell. Most other names on this list are redacted and the law makes it clear that such information must be disclosed if it is there. So really, the big story coming out of the Epstein files ended up being all the information, files, and possibilities that weren’t disclosed.

MCCAMMON: Now, it wasn’t just about the Epstein files. You were also able to dig into some DOGE files, which I think we all still remember was the so-called Department of Government Effectiveness. You were very active in sorting them out during the first months of Trump’s second term. And you announced news. I mean, tell me more about that. How did you go about browsing these DOGE files?

FOWLER: Well, it’s very different looking at federal government data versus a bunch of disparate documents and records from court cases and the like. I mean, the federal government – I don’t know if people realize this – has one of the richest, most comprehensive, most easily accessible data sets on things that you can imagine. Whatever you think about what the government is doing or thinking, there’s probably a dataset waiting for you to download. The trick is knowing what you’re looking at and how it’s used.

So over the last year, I’ve been tracking federal spending data, tens of thousands of changes to federal contracts under DOGE. There were federal statistics on leasing, other things like that that I looked at to understand how the government was or wasn’t working and whether DOGE was actually doing the things it claimed to be doing. I mean, showing your work is also very different, because with the Epstein files, you can say, hey, on this page with this file number, this thing was said.

With DOGE data, there were so many spreadsheets. I mean, my editors can tell you – spreadsheet after spreadsheet after spreadsheet that had meticulous calculations from taped, in-depth interviews with all these different experts in these fields to make sure I understood what I was looking at. And did I mention there are spreadsheets? So it took a lot of effort for me to be able to show the work and make sure I knew what I was looking at so I could tell the audience what we were looking at.

I think the biggest example of this is the infamous DOGE receipt wall. This is the website where DOGE claims to always track how much taxpayer money it has saved through its changes. The data released by DOGE did not match the source data released by the federal government. There were exaggerations, errors, and inaccuracies, and it took a lot of time and effort to compare these data sources to discover the truth behind things.

MCCAMMON: What else has all this data yielded in this first year of Trump’s second term? What did this teach you about covering this political moment and what it could mean?

FOWLER: I think it’s even more important to show the work, bring receipts and give people the primary sources that I used to do this cover. First, I want people to trust the reporting and what I have to say, and be able to see for themselves how I got here and also pull back the curtain so the public can learn how to examine this information for themselves. And then, the figures are difficult to transmit on the radio. If you put out too much, people’s eyes start to glaze over. If you have too few, you risk underestimating your findings and missing important context.

So spending the last year basing a lot of my reporting on this data – even though not every story starts with one, here’s that number to use here – I mean, it helps contextualize the scope or scale or importance of what’s happening in politics, whether it’s the small fraction of the Epstein files or the 58th lawmaker retiring or federal spending.

MCCAMMON: That’s Stephen Fowler, a political reporter for NPR’s Washington Desk. Thank you, Stéphane.

FOWLER: Thank you.

(SOUNDBITE OF MUSIC)

Copyright © 2026 NPR. All rights reserved. Visit the terms of use and permissions pages on our website at www.npr.org for more information.

The accuracy and availability of NPR transcripts may vary. The text of the transcript may be edited to correct errors or match updates to the audio. Audio on npr.org may be edited after its original broadcast or publication. The authoritative record of NPR’s programming is the audio recording.