Over the last four years of my life, my number one priority was getting into graduate school. There’s nothing I enjoy doing more than learning, reading, and researching, and graduate school seemed like the best place to do that. As a psychology undergraduate, I worked in multiple research labs, took online courses in programming and statistics, and read a lot of extra books and papers to supplement what I was learning in class. In my fourth year, I applied to a few graduate programs and got into my #1 choice; a 2-year master’s in political communication research at the University of Amsterdam.
As soon as I started school in September, I felt something was off. Talking to the other students, I realized all of us shared a common story: we were smart kids who did well in our undergrad, and graduate school just seemed like the “natural next step.” None of us could explain why we were in this particular program, or give an account of how it would advance our careers. Some, including me, hinted at the possibility of a PhD, but this was just deferring the question of “why are you here,” not answering it.
I was expecting the “school” portion of graduate school to be exciting and interesting, but I quickly became frustrated. Roughly half of the assigned papers were pre-replication crisis, so I didn’t know whether I should believe the results or not. The other half were post-replication crisis, and they were boring; just table after table of trivial correlations wedged between a lengthy literature review and an over-generalizing discussion section. I seriously doubted that studying these papers would get me closer to the truth on important matters rather than further away from it.
Within the first month, I managed to secure a research assistantship with a professor. But in the same private meeting where I was offered the position, the professor confided in me, saying that like me, they were “also very skeptical” of a lot of the research in the field, including some pieces they taught in their class. Now, one way to take this would be to think to myself, “wow, this professor is being really honest with me about the limitations of the research; that’s great.” Instead I thought to myself, “jeez, in class the professors are teaching these papers but in private they’re skeptical of them; that’s not good.” More conversations with other students and faculty confirmed this for me. In class these theories and findings were treated seriously. In private they were ridiculed.
During this period, I got along very well with the other students and faculty, and I never felt uncomfortable expressing my opinions in class. The people in my program expressed a wide range of political views and I’d say the median viewpoint in any given classroom was old-school liberalism. But the fact that I wasn’t distracted by concerns over “radicals on campus” or “leftwing bias” only made me more cognizant of how much I actually disliked the graduate school experience and how little faith I had in the value of the research.
So two months into graduate school, I dropped out. It was an easy decision.
Having spent some time reflecting on my experience in graduate school and within academia more generally, I’ve come up with what I believe were my three main reasons for leaving. The first is that I didn’t believe the academic literature I was studying was true, useful, or insightful. The second is that most of the students and faculty had beliefs about people and society that I felt didn’t square with the empirical evidence. The third is that I was much more interested in the social science research happening outside of academia. In the remainder of this post, I’ll discuss these reasons further in-depth, treating them not just as personal reasons for leaving academia but as good reasons for anyone to leave, or never to begin graduate school. The reasons I discuss aren’t unique to my program, department, or sub-discipline. They pervade social science academia, with some fields being worse than others.1I think it’s fair to say that economics and quantitative sociology (particularly demography) suffer from these problems less than other social sciences. They also tend to replicate well compared to other social sciences, as can be seen here: https://fantasticanachronism.com/2020/09/11/whats-wrong-with-social-science-and-how-to-fix-it/ That’s why I’m comfortable generalizing from my experience studying communications, psychology, and political science to the social sciences more broadly.
We’ve all read pieces that criticize academia for being too woke or too bureaucratic, and while I touch upon those issues, my main criticism is just that it’s incredibly underwhelming.
1. Most of What You Study in Social Science Academia is False, Useless, or Obvious
I learned about the replication crisis a quarter of the way through my undergrad. I had the uneasy feeling that most of what I was learning was probably fake; that much of the social-scientific literature was one big house of cards that should have fallen years ago. But my focus on the replication crisis blinded me to a larger truth: that most demonstrably true social science research findings are useless or obvious.
Useless findings sound interesting, but for all practical purposes they don’t matter. You can tell a nice story about the findings, and they might even be statistically significant, but the effects are so tiny as to render them inapplicable to any real-world prediction or application. This is the fate of many social-scientific studies, which tell a nice story about the difference between groups – Democrats and Republicans, single and married, control and treatment – but gloss over how tiny the differences are. A classic example is an influential political psychology paper called “The Secret Lives of Liberals and Conservatives.”2Carney, Dana R., John T. Jost, Samuel D. Gosling, and Jeff Potter. 2008. “The Secret Lives of Liberals and Conservatives: Personality Profiles, Interaction Styles, and the Things They Leave Behind.” Political Psychology 29(6): 807-840. The authors found several liberal-conservative differences in personality, behavior, and even what items people had in their bedrooms and offices. But as one psychologist critical of the paper pointed out, “the ‘secret lives’ title is apt” given how small the differences are and how little of their variance can actually be attributed to political ideology.3Kalmoe, Nathan P. 2020. “Uses and Abuses of Ideology in Political Psychology.” Political Psychology 41(4): 774.
Obvious findings are those you probably could have known about without empirically researching them. They’re more useful than useless findings (obviously), but you can’t tell a good story about them because the story would be boring. For example, people who are more interested in politics are more likely to know a lot about politics and to vote, or people who post about protests on social media are more likely to attend them.4Prior, Markus. 2019. Hooked: How Politics Captures People’s Interest. Cambridge University Press: p. 6; Boulianne, Shelley, Karolina Koc-Michalska, and Bruce Bimber. 2020. “Mobilizing Media: Comparing TV and Social Media Effects on Protest Mobilization.” Information, Communication, and Society. 23(5): 642-666 (Note: the specific finding I’m referring to here is on page 653). I think social science is heading in this direction, and while the studies are getting more scientifically precise, the results are getting less interesting and surprising as a consequence.
There are social science findings that are true, useful, and counterintuitive, but they are rare, probably less than 1% of published research. Most of the findings that appear to fall under this category are ‘too good to be true,’ and are eventually revealed to be p-hacks, flukes, or fabrications, though they still get cited a lot!5Serra-Garcia, Marta, and Uri Gneezy. 2021. “Nonreplicable Publications are Cited More than Replicable Ones.” Science Advances 7(21). More often than not, the true, useful, and counterintuitive findings are null results: you think something mattered, and it turns out not to matter at all. Unfortunately, null results rarely get the attention they deserve due to publication bias.
Many professors and graduate students are aware of these issues but cope with them by saying, “we’re getting better now that we have open science research practices.” What they don’t realize, or prefer not to think about, is that as social science methodology has improved the findings have become more useless and obvious. Most counterintuitive results fail to replicate or are disproven, and effect sizes shrink with bigger samples and more control variables. What’s left is often so banal or self-evident that it’s not clear why it needed studying in the first place. I’m not suggesting we go back to how we used to do social science, where bad statistics and small samples artificially inflated effect sizes. But I am suggesting we grapple with the fact that social-scientific advancement has revealed most social science research to be of little to no value.
2. Most People Studying Social Science Have the Wrong Priors
Social science students and academics frequently hold a set of views that other scientists have refuted. They often think that gender is a social construct, or that parenting matters more than genetics in determining social and psychological outcomes. Many believe they can fix complex social issues like poverty, racial disparities, or political polarization by reminding people of their biases or ‘nudging’ them towards pro-social behaviors.6Singal, Jesse. 2021. The Quick Fix: Why Fad Psychology Can’t Cure Our Social Ills. Farrar, Straus and Giroux.
I would say that many who study social science overweight the influence of social factors and behavioral interventions and underweight the impact of biological factors, random variance, and culture. Rather than adjust their priors, they tend to double-down on their belief that tiny effects matter or that obvious results are surprising.
If they were to adjust their priors, expecting most social factors and interventions to have little impact if any, two things would happen. First, a literature chock-full of tiny effects would be taken as evidence that a phenomenon isn’t worth studying in-depth. Second, consolatory statements like “humans are complicated” or retreats into philosophy of science would be seen for what they are: attempts to cope with the disparity between social reality and social reality as perceived by social scientists.
There is value in studying and reporting null results and small effects, but only if we’re honest about what they mean. The value comes from slashing false lay-beliefs about social phenomena (i.e., “advertisements are highly effective”) rather than constructing a story about the importance of those phenomena when the data argue the opposite. However, most people who choose to study social science don’t go into it expecting small effects and null results. They go into it thinking the effects will be big and the results will be socially significant. Were they not under this impression, who would knowingly pay money to study a subject?
3. The Most Innovative and Impactful Social Research is Happening Outside of Academia
Good social research often requires large, high-quality datasets. The biggest social/behavioral datasets are owned by companies like Facebook, Nielsen, and Palantir, and governments that collect a lot of data on their citizens, such as China. Unlike academia, the businesses and institutions that collect this data have massive budgets and often collect data on real behavior instead of approximating it in a laboratory setting or survey. They also operate with fewer ethical constraints than in academia, where even the simplest, most inoffensive survey must be cleared by an ethics committee. This can take anywhere from several days to several weeks, unlike in business research or media where you can put out a survey instantly. For example, a presidential approval poll would require an ethics review in academia, but not if done for a newspaper.
I’m not suggesting the problems described above, like small effects or non-replicable results, are absent from non-academic research. These issues exist in all research environments, public or private. But in a non-academic setting, you’ll have the opportunity to work on projects that matter; that help a business identify a market and sell a product or help a government understand its citizens and predict their behavior. The feeling that you are doing work that “matters” is incredibly important to the kinds of people who get involved in social science, and venturing outside of academia to work on projects with tangible impacts is one way of recapturing that feeling. Projects also get done a lot faster outside of academia, which helps you quickly determine whether the research you’re conducting has value or not.
But let’s say you don’t care about the nitty-gritty empirical data and are more interested in social science theory. It’s reasonable to suspect that academia might be a better place to study theory than a private company or government, but is it where the best social theory is emerging?
I’d argue that the best modern social theory comes from three non-academic sources. The first is blogs and newsletters, where top journalists, political actors, and social theorists are sharing their research and insights for free or a small monthly fee. The second is the tech world, which makes sense considering our thoughts, feelings, and actions are increasingly expressed online or through technology. The third is social media platforms like Twitter and Reddit, where novel social and political theory is frequently proposed, discussed, and dissected.
I don’t deny that interesting theoretical work comes out of academia, but academia holds no monopoly over the study of people and society. And if your goal is to propose new social theory, you’re better off being entrepreneurial with your ideas instead of getting bogged down in academia’s incentive structure, which discourages innovation and favors incrementalism.
Many people on the right and heterodox left are concerned about the lack of conservative voices on campus, particularly in the social sciences and humanities. There are far fewer right-leaning faculty, and right-leaning students are less likely to enroll in an undergraduate program or pursue post-graduate education than their left-wing classmates.7Kaufmann, Eric. 2021. “Academic Freedom in Crisis: Punishment, Political Discrimination, and Self-Censorship.” Center for the Study of Partisanship and Ideology. Available at https://cspicenter.org/reports/academicfreedom/.
These are troubling issues that deserve to be addressed, but I think moderates and conservatives’ desire to make academia a more politically tolerant place may have blinded them to how unremarkable contemporary academia actually is. Even if 50% of social science faculty were right-wing and moderate or conservative students felt perfectly comfortable expressing their views and values, I don’t think it would justify their spending more time in academia, splitting hairs and boiling the ocean.
There’s never been a better time to study people and society empirically. Books and articles are easily accessible, publicly available datasets are abundant, and conducting survey research is cheap with services like MTurk or Prolific. You can learn statistics, programming, and survey design skills through online courses or private tutoring, which cost a fraction of graduate school tuition fees. Or you can join a company or government with access to big datasets and go crazy with multiple regression.
Life is too short to spend reading outdated articles and writing pointless papers, and people and society are too interesting to be studied in such a myopic manner. So if you’re thinking about graduate school or are already enrolled but not enjoying yourself, do something else instead.
Jonah Davids is the director of communications at CSPI.
- 1I think it’s fair to say that economics and quantitative sociology (particularly demography) suffer from these problems less than other social sciences. They also tend to replicate well compared to other social sciences, as can be seen here: https://fantasticanachronism.com/2020/09/11/whats-wrong-with-social-science-and-how-to-fix-it/
- 2Carney, Dana R., John T. Jost, Samuel D. Gosling, and Jeff Potter. 2008. “The Secret Lives of Liberals and Conservatives: Personality Profiles, Interaction Styles, and the Things They Leave Behind.” Political Psychology 29(6): 807-840.
- 3Kalmoe, Nathan P. 2020. “Uses and Abuses of Ideology in Political Psychology.” Political Psychology 41(4): 774.
- 4Prior, Markus. 2019. Hooked: How Politics Captures People’s Interest. Cambridge University Press: p. 6; Boulianne, Shelley, Karolina Koc-Michalska, and Bruce Bimber. 2020. “Mobilizing Media: Comparing TV and Social Media Effects on Protest Mobilization.” Information, Communication, and Society. 23(5): 642-666 (Note: the specific finding I’m referring to here is on page 653).
- 5Serra-Garcia, Marta, and Uri Gneezy. 2021. “Nonreplicable Publications are Cited More than Replicable Ones.” Science Advances 7(21).
- 6Singal, Jesse. 2021. The Quick Fix: Why Fad Psychology Can’t Cure Our Social Ills. Farrar, Straus and Giroux.
- 7Kaufmann, Eric. 2021. “Academic Freedom in Crisis: Punishment, Political Discrimination, and Self-Censorship.” Center for the Study of Partisanship and Ideology. Available at https://cspicenter.org/reports/academicfreedom/.