How the five-paragraph essay fails students

Photo by Erik Mclean on Unsplash

Once, while teaching undergraduate rhetoric/composition, a student turned in a final paper where he defended the existence of a high-tech ancient civilization which was informed, to some extent, by aliens. This civilization, with the help of aliens, was responsible for building pyramids in Egypt, Mexico, Peru, Eastern Europe, and Antarctica. Throughout the essay, he cited the work of the hack-archaeologist Graham Hancock. Although his bibliography included an article on how “Ancient Aliens” theories are racist (and I searched the paper to see how this student grappled with the ideas of that article), I couldn’t find that article mentioned a single time in the paper.

When the student asked why he received a low grade for that paper, I told him, “This isn’t research. This is confirmation bias. You had the idea that ancient aliens built the pyramids. You found a researcher who believes that. You cited him almost exclusively. You can rewrite the paper, but grapple with research that contests that of Graham Hancock.”

So, he rewrote the paper. His rewrite defended the idea that ancient aliens did not build the pyramids. Although I agree with his second paper more than I agree with the first, his grade was unchanged. He missed the point. I wasn’t asking for a paper I agreed with. I was asking for a research paper.

After speaking with colleagues, I’ve stumbled on the idea that part of the problem is the way students are taught to write “research” papers in high school. I’m sure this isn’t universally true, but many of my students have told me that they were taught to write five paragraphs in this order: introduction/thesis, evidence 1, evidence 2, evidence 3, conclusion/restate thesis. Each paragraph of evidence should support the claim made by the author. When I’ve asked how this paper should confront other opinions, my students have told me that they weren’t taught that.

If that’s what they were taught, it’s no wonder students equate research with finding support for their beliefs. I hope that’s changing because that is not research. That is an exercise in confirmation bias.

True research is done when someone has a question and seeks an answer. In this process, we often find contradicting evidence. That’s what students really need to learn: how do you confront contradicting evidence?

Let me propose a hypothetical.

I am a high school student or college freshman. I am writing a paper on whether religion is helpful or harmful to the human psyche. My personal belief, my bias, as a religious person, is that it is helpful to the human psyche. Although I might want to write an essay on why religion is helpful, this is a research paper, so instead, I propose the question, is religion helpful or harmful to the human psyche? I can be blunt about my bias. Yes. I could even phrase it as a hypothesis. “It seems to me that religion would be helpful because….”

Then, I sit down to investigate. I find two peer-reviewed articles published in reputable journals. One claims that religious people coped better with isolation during the pandemic, the other claims that religious women are more likely to be victims of domestic violence. What should I do now?

If I were looking for evidence that supports my claim, I would cite the first paper and ignore the second one. But that doesn’t answer my question. Research, an authentic desire to find an answer to my question and write a well-written paper, forces me to explore both papers.

So, I read through imaginary-paper 1, and I find that the research took 300 participants of different religious backgrounds, equal numbers self-described “religious” and “nonreligious,” and found that the religious participants self-reported lower rates of depression and anxiety, and higher rates of positive emotions during the pandemic.

Imaginary-paper 2 asked 400 women about their romantic relationships over the course of five years. All women were in their early twenties; approximately 200 described themselves as “religious” and 200 as “non-religious.” The religious women self-reported being in abusive relationships over the five years at a much higher rate than the nonreligious women.

I could draw many conclusions from this. Maybe religion is more helpful to regulating negative emotions, but it also makes one more tolerant of harmful experiences. Maybe both papers ignored other factors like class, i.e., in imaginary-paper 1, the people of religious backgrounds might have had more income stability than those who self-described as non-religious, and vice versa for imaginary-paper 2. I can criticize the limitations of both research papers.

Or, maybe I’m not fully knowledgeable of research methods and controlled variables, so I can’t criticize the methods of either paper. In which case, I can look for other papers and scientific consensus.

Let’s say that when I look into scientific consensus, I find overwhelmingly that religious people are more likely to underreport negative feelings and equally or more likely to be diagnosed with depression or dementia later in life than non-religious people, leading one to believe that religious people are just lying to themselves or researchers about their emotions. Shoot, that contradicts my hypothesis. How would I grapple with that?

It’s a sweet lie, I tell myself. It’s a lie that’s added richness to my life and created beautiful works of literature like The Divine Comedy and The Ascent of Mount Carmel. However, I should acknowledge that my religious convictions might encourage me to lie to myself about my emotions, and I should examine how I can use my religion to be more honest with myself about my emotions. I can turn to the writings of Santa Teresa de Ávila, someone who was brutally honest about her emotions, and see how examination of conscience and prayer can be beneficial in my interior life. There’s hundreds of directions I can take this hypothetical paper, but none of those directions ask me to restate my thesis as a conclusion.

There are two points I want to make with this: the first point is that real research is more interesting than finding evidence that agrees with a thesis. It creates conflict that provokes plot and gives an essay the form of a story. It invites the writer to be honest about her biases and grow as an academic and as a person. The second point, real research guarantees that the conclusion will be more than just a repetition of my introduction. Rather than restating a thesis, it presents a new idea or the question in a new light.

Maybe in my essay about religion and psychology, I do find an answer that could be a thesis: “Research suggests that religion is often associated with a repression of emotions that could be harmful to long-term psychological health.” But that answer is new and discovered, and it doesn’t have to end there. I can continue, “However, with certain measures, religion can also be helpful toward regulating emotions.”

If my student who wrote his ancient-aliens essay ended it with, “All research I have found in peer-reviewed journals of merit supports the idea that there is no solid evidence suggesting the existence of a high-tech ancient civilization informed by aliens, but I want to believe. Maybe the more interesting question is why people are drawn to this idea? I know I’m drawn to it because…” that would have been excellent! I’m not asking my students to agree with me. I’m asking them to write an interesting essay and to grapple with research in the field.

There’s a reason why you’ve never read a published paper that’s written as a thesis with supporting claims. Yes, it’s bad research, but it’s also boring. It’s boring to read a paper that could be summarized in the first paragraph. Good writing moves the reader’s eye down the page. Real research allows the writer to discover a structure rather than starting with “I need this number of paragraphs with each making this kind of point.”

This leads me to my biggest problem with the five-paragraph essay. Aside from being bad research, it’s bad writing. It’s unpublishable. It’s boring. Its form is rigid, and it discourages the kind of creativity that good writing demands. I see its utility for high school — teachers want to make sure their students understand course content, so they ask them to write what they know. That has a place in early learning. However, there are consequences to believing that finding evidence to support a claim is research. Either they go to college and have to unlearn what their teachers taught them, or they don’t go to college and continue thinking that confirmation bias is “research.”

Today, we live in a post-fact world. You can have any opinion on reality and find someone online who agrees with you. You can believe that the Cookie Monster killed Obama, who was replaced by a body double during his first term, and you can find “evidence” to support your claim. This isn’t a failure of intelligence, it’s a failure of research literacy. Again, I’m not asking that people who believe what I think is nonsense stop believing it. I’m not looking for agreement. I’m looking for good writing and intellectual sincerity. I’m asking for people to grapple with their biases and reckon with potential problems in their beliefs, rather than alienate themselves from their communities because they’ve found an online niche that echoes them.

I’m not saying the five-paragraph essay is the reason conspiracy theories exist, but unteaching the five-paragraph essay—and teaching what real research looks like—is a good start. It’s a way to invite our conspiracy-minded friends and students to engage with complexity, grow as thinkers, and stay in relationship with the world instead of retreating from it.