Actually, facts do care about your feelings

Pursuing objectivity requires being cognizant of your biases, not pretending they don’t exist.

“Facts don’t care about your feelings” is one of the most well-known phrases among young conservatives today — and it’s easy to understand why. When Ben Shapiro first popularized the phrase, he did so in his perceived context of a college culture gone mad. He was pushing back against the rise of safe spaces, trigger warnings, and microaggressions emerging across college campuses.

The phrase empowers a basic statement: college isn’t about hiding from scary facts or silencing opposite opinions, and no matter what you do, life isn’t going to reshape reality around your delicate feelings. It’s embracing these uncomfortable facts that allows us to understand the complex realities of the world we live in, which is essential to being a functioning human being.

It’s a powerful message and one rational people should struggle to disagree with. In the context, it’s also a correct message. When the question is “should we avoid facts because of our feelings?”, the answer can be a very simple “no”. It’s a simple argument to point out that avoiding facts doesn’t change them, and this simple statement clearly demonstrates the simple truth.

But what happens when things get more complex? This is where I’ve seen this phrase abused to death — and the reason behind this critique. When questions get more complex, feelings tend to shape how we navigate an unclear topic.

The example of police brutality.

Let me give you an example topic that encapsulates the complexities of science: What is the role of racial bias in police shootings?

A conservative friend might start with the fact that black citizens disproportionately commit crime at higher rates than white citizens. A whole lot of people stop there — they conclude that this datapoint alone solves the issue. Black people are shot more than white people merely because they commit more crimes.

A liberal friend comes by and points out findings from implicit bias studies, pointing out psychological research that shows that both college students and police officers are more likely to notice crime-related objects after seeing a black face. The same research shows that the more stereotypically black a face is, the more likely police are to label it as the face of a criminal.

In response, your conservative pal points out that a recent paper by a Harvard economist found that there’s no racial bias in police shootings. I guess that solves that, right?

But what if it turns out this study isn’t peer-reviewed and has been criticized for methodological issues, including a lack of familiarity with criminological analysis? Would you be able to recognize that the paper wasn’t peer-reviewed just by looking at it? Would you have considered that the paper could be flawed?

Maybe your liberal friend isn’t able to either — and responds that there are papers (1, 2, 3, 4) that show that police officers are more likely to shoot black people and that they also shoot faster. Your conservative friend reads Heather Mac Donald though, and counters with new research showing that police officers are less likely to shoot black people and that they also shoot more slowly.

Now we’ve got conflicting data, and no clear way of weighing them. Your liberal pal may point out that there’s multiple papers all reaching the same conclusion, whereas your conservative friend could argue that not only is their research more recent, but it uses a different method that may be more realistic. Welcome to scientific hell. This is the normal weather.

If someone is especially thorough, they might check to see what papers have cited that new paper with anomalous finding — and you’d find that in the same journal, a paper was published simultaneously arguing that the conflicting findings can both be true under a new framework.

Or the discussion could go an entirely different route: in response to the first claim that black people commit more crimes, perhaps the liberal friend points out that we don’t actually know what true crime rates are. What we usually refer to as “crime rates” are police records — which are not a random sampling of crimes, and are known to have significant racial biases (1, 2, 3). To quote Lum & Isaac : “..it is clear that police records do not measure crime. They measure some complex interaction between criminality, policing strategy, and community–police relations”.

Getting the picture? This is the web of scientific knowledge that can exist for any topic.

Your feelings affect what facts you use

Any individual fact (singular) is unlikely to be affected by your feelings, as long as it is a clear-cut statement. However, when you start talking about facts (plural), things get messy.

The reason for this is that very few people have a complete understanding of the facts on any issue. If you’re not a Ph.D in a relevant field, or haven’t spent hundreds of hours studying this topic, you almost certainly don’t know all the facts.

This isn’t hyperbole. Multiple new papers on a topic can come out daily. Some topics have over a half-century of accumulated research that you have to wade through. To even begin to start to comprehend this stuff at a necessary level one has to understand scientific methodology, how to appropriately weigh conflicting pieces of evidence, and how to detect pseudoscience masquerading as science. To fully comprehend this stuff, it usually requires advanced understanding of statistical analysis, a well-trained ability to break down and rip apart underlying assumptions and flaws in the experimental design and conclusions, and an encyclopedic recall of the details and findings of similar experiments. Ask someone with a Ph.D how long it took for them just to learn the required background of a new subject they wanted to broach.

The bottom line is this: if you aren’t an expert in a field (and you probably aren’t), you don’t know the facts. What you possess is knowledge of a subset of the facts, and how you decide what facts are included in that subset is of paramount importance. This filter is going to be determined by your background, your experiences, your social circle, your interests, and yes, by your feelings.

How you feel about the political slant of news sources is going to determine which ones you prefer. How you feel about the underlying conclusion of research is going to determine what findings you emphasize and what findings you discard — as well as what research you seek out. You can’t separate the influence your feelings have on the facts that you know.

The bastardization of #FactsNotFeelings.

Here’s how I usually see #FactsNotFeelings being used now:

1. Person A decides to opine on a particularly complex question, say, the existence of implicit bias among law enforcement officers and how it affects interactions police have with civilians.
2. To showcase their deep understanding of this topic, Person A highlights a specific datapoint or chart that they believe validates their stance. This usually is a single piece of evidence, not placed in meaningful context.
3. Person A then finishes their exhibition by stating #FactsNotFeelings. This has the dual purpose of screaming “I’m not biased. I care only about the evidence” as well as “If you disagree it’s because you’re overly emotional”.
4. If Person B comes along and questions the conclusion, even slightly, their argument is usually ignored. Often Person B won’t address the datapoint directly, but instead adds another datapoint.
5. Person A usually views this as a sneaky trick, and says Person B is trying to avoid the facts due to their conflicting feelings. Person B is labeled as hopelessly biased due to their emotional baggage and discarded into the wind.
6. I scream at my computer screen in response.

This is incredibly wrong, and it’s an incredibly dangerous mindset to settle into. Using this statement as an end-all “I’m right, you’re wrong” foments intellectual laziness and a holier-than-thou (more-objective-than-thou?) mindset. It suppresses introspection of underlying biases, and lulls the user into a complacency that all of their stances are grounded in thoroughly vetted facts.

In other words, this creates an intellectual safe space for the user. It boils complex realities down into simple datapoints and declares victory. Misuse of this phrase is antithetical to its “embrace all knowledge” origins.

Embracing the complexity of an unclear world

One thing you shouldn’t do is stop trying to earnestly discuss topics. I’m also not saying that you should avoid saying “here are the facts” — because anyone who takes that literally in a conversation deserves a click of the mute button.

What you should do is realize that data is messy. Science is hard. But the most important thing is to recognize when your feelings are affecting how you’re treating your data.

Be aware of your biases and confront them head-on. If you can manage that, you might be able to say “facts, not feelings” — and be telling the truth.

Author’s note: Please also see a similarly titled piece by John-Pierre Maeli, that examines this topic from a different angle.

Edits as of July 2020: Partially in response to the Black Lives Matter movement, I wanted to revisit this piece. I added another paragraph to shine light on the complexities around measuring “crime rate”, and softened some language in the introduction.