The “Friday Night (data) Massacre”

When government data disappear, how do we make evidence-based, informed policy decisions?

The evening of Friday, January 31, dozens of government websites suddenly went dark. Agency websites from the Centers for Disease Control to the Census Bureau and the Federal Aviation Administration were suddenly unavailable or reported an error due to “maintenance.” When websites went back up, suddenly whole pages and datasets—largely those related to “gender ideology”—were missing, like they’d never been there at all. The cuts were sweeping and over-broad, including words like equity, even if they refer to home equity.

Most Americans don’t spend their Friday nights scrolling .gov webpages and perusing massive microdatasets. But many in the academic and policy communities—researchers at universities and colleges, nonprofits, think tanks, and, yes, the government—began to panic at what began to smell like our own “Friday night massacre.”

Cutting off the flow of large-scale data matters.

It’s been two weeks now, and it’s still unclear the extent of the damage that has been done. We do know already that dozens of databases and info sheets are gone. Individual researchers are trying to stockpile data while it’s posted or secure files on personal servers. But those data sources we used to trust are also now suspicious—if they’ve been scrubbed of all transgender people, what else has been scrubbed out? Our normally safe and comfortable number-crunching now feels uncertain.

As of February 1, the CDC ordered a mass retraction of articles under review or accepted but not yet published at scientific and medical journals. Articles using language such as “Gender, transgender, pregnant person, pregnant people, LGBT, transsexual, non-binary, nonbinary, assigned male at birth, assigned female at birth, biologically male, biologically female” are to be retracted and revised to remove this language. This expands the limit on data and information from just .gov websites all the way to independent peer review.

Who benefits from research using federal data?

All of us.

Data are the foundation of modern policy making and policy evaluation, and federal data in particular inform the entirety of the US health, education, economic, and welfare system. We rely on information to administer the massive social safety net programs—Medicaid and Medicare, SNAP, TANF, Section 8—that serve millions of Americans each year.

Research informs what works and why and how and for whom—and what doesn’t. With a presidential administration aiming to reduce “government excess,” this is particularly frightening. Cutting off access to large-scale data prevents us from being able to effectively assess the costs or benefits of different programs (putting aside the many valid critiques about the ineffectiveness of CBA for social policymaking). This means that our ability to advocate for keeping certain programs funded is hamstrung.

When the Trump administration is trying to make sweeping cuts to federal spending and the budget deficit, our best defense is the broad base of evidence showing that these programs work.

What is at stake?

Put plainly, data is knowledge is money is power. Now, more than ever.

Without effective data infrastructure, those looking to protect safety net programming cannot effectively advocate for their program’s efficacy, meaning they cannot effectively advocate for their funding. By controlling the data spigot, the administration can control the knowledge, meaning they control the money and they control the power—regardless of what the people who democratically elected them actually want.

This is going to hurt the most marginalized among us the most. Data specifically under threat are those related to queer and trans folks—especially trans youth—and on topics like abortion. When data were re-posted to CDC websites, many quickly noticed that references to transgender and LGBTQ+ people had been scrubbed.

The erasure of trans folks in particular has been seen across government websites. The National Parks service removed references to transgender people on the official web page for the Stonewall National Monument website. The State Department now only provides guidance for “LGB” travelers abroad. Yet, deleting the letter T or removing data from a website does not stop trans people and queer folks from existing.

It does, however, limit our ability to effectively advocate for these people. Making whole swaths of people invisible in the data further prevents us from understanding things like how queer and trans folks interact with the health care system, participate and contribute to the economy, and more.

A lack of data and data infrastructure also makes it increasingly difficult, if not nearly impossible, to identify programs that are effective and might merit scaling up versus those that are ineffective and require reform or sunsetting. Without data, we cannot have evidence-based policy.

And this is precisely what that current administration has in mind in gutting the Department of Education, the Institute of Education Sciences, and the What Works Clearinghouse. After all, it will be easier (and more lucrative) to privatize and profit off of the public education system without pesky evidence about effectiveness (or ineffectiveness) hanging around.

What we all stand to lose if we don’t get data for 2025.

A lot of attention has been paid to the scrubbing of historical data from the Census, CDC, NIH, and more. But the current anti-data energy suggests that new data collection is also under threat. This is, in large part, a philosophical and political decision—Musk and Trump have demonstrated they do not care about protecting massive datasets.

But importantly, it’s also a practical decision. Terminating probationary federal employees (on Friday, the CDC announced it would lose 10% of its workforce) will substantially reduce the data collecting power of these agencies going forward.

This will leave researchers unable to answer important questions like:

• How many families slipped into poverty due to the Trump administration’s tax cuts for the rich and tax increases for the rest of us?

• How many trans kids died by suicide after being denied life-saving treatment?

• How did veteran health suffer when the VA laid off over a thousand workers Friday?

• How did civilian health suffer from the pending NIH funding cuts that slashed individual hospital systems’ budgets by up to $100 million?

• How much forest and how many homes were lost due to worsened wildfire management after federal layoffs?

• How are these impacts experienced differently by people of color, when the administration wants to defund NSF projects that dare to use words like “black,” “latinx,” or “race and ethnicity”?

• How many children in the world’s most vulnerable regions suffered from malnutrition when programs like USAID Feed the Future fell to sweeping DOGE cuts?

If the goal is obfuscation—making it difficult or even impossible to assess the damage done by defunding the social safety net—then this will put the new administration firmly in the end zone. Without data and research, it becomes considerably easier to reframe and refute “hard truths” as “fake news.”

What else can be done?

Researchers and community members have power here. Here are a few ideas of how you can take action today.

1. Help us build our own resources.

This is not the time for data hoarding. Fortunately for us, there is a considerable infrastructure already for sourcing and sharing data. If you have data stored in some recessed DropBox folder or on an old hard drive, consider contributing to a public data sharing effort like those described below.

• The Internet Archive stores a mirror of federal websites at the end of every presidential term at GovWayback, pushing back against such moves as removing the Spanish-language version of whitehouse.gov.

• The Data Rescue Project is tracking data that have been saved and stored. ICPSR, the data repository at University of Michigan, houses DataLumos, an online archive for federal data to which anyone can contribute.

• Boston University’s School of Public Health has started up FindLostData.org, which crawls major databases regularly to store up-to-date copies.

• The Data Foundation, which champions the use of open data and evidence-informed public policy to make democratic society better for everyone, provides secure anonymous federal evidence, data and analysis tracking through SAFE-Track.

• The Integrated Public Use Microdata Series, commonly known as IPUMS, is a rich data source hosted at the University of Minnesota. It collects, preserves, and provides access to census and survey data from around the world, ranging from the US Census and the Integrated Health Demographic and Health (DHS) Surveys from around the globe. IPUMS is more valuable than ever in the context of a federal administration that disappears data to hide truths and control citizens.

2. Keep doing research—and bringing it to public audiences.

The stories that we tell with data—and the nuance and complexity and context that trained researchers can bring to otherwise sterile numbers—are powerful. Many folks don’t immediately think of the ways they themselves are impacted by federal policy and federal funding cuts. Academic and policy-oriented research can help shine a light on how we are all affected by what’s happening in Washington, D.C. But this requires us to make strong, evidence-backed, nuanced cases and to spread what we know far and wide. We can spend our energies bringing research out of the academic journals and evaluation reports into the public eye via op-eds, social media posts, legislative testimony, and, yes, even on blogs like this one.

3. Organize round tables/panels at convenings.

The more we put our heads together, the more we can effectively advocate, figure out workarounds, and support each other. Conference roundtables and panels are a great way to bring folks together to brainstorm and collaborate. But, don’t wait for in-person events—we saw during the pandemic how effectively zoom can be leveraged for research. Check with disciplinary associations and enterprising academic units/individuals to learn about what is already being organized and how you can pitch in.

4. But also: be safe.

The present moment is understandably scary. It’s important that we do this work and share these resources with each other, but we need to also consider how we can resource without adding new risk. The data repositories we’ve listed above are all public about their work, but we don’t want to put targets on data rescuers’ backs. Keep in mind that it is likely that many of the data repositories we have shared will also come under attack. These organizations will need our support and protection, whenever we can bear it. Be smart, and stay safe. The work that you do is important, and we need to prepare to run an ultra-marathon, not a sprint.

Note: an earlier version of this post incorrectly named the IES the Institute of Education Services.