Do coursebook writing tasks engender confirmation bias?

Bias is part of human nature. We all have biases, many of which are implicit. One particular form of this is confirmation bias, the “tendency to search for, interpret, favor, and recall information in a way that confirms one’s preexisting beliefs” (Wikipedia). In other words, it is having an opinion and accepting anything that supports it while rejecting anything that does not. Critical thinking is considered a kind of antithesis or antidote to this type of bias, which is why it and related concepts (i.e. evidence-based thinking) have become so popular lately, being a major part of the United States’ Common Core standards and a skill that is constantly being discussed in all circles of education, ELT included.

In an ideal situation, before an opinion is formed, one carefully considers both sides of an issue, weighs and evaluates the evidence, and then makes an informed decision. In a university course, this usually plays out through reading and discussion, followed by writing in order to synthesize and make sense of the ideas. While this may be an exaggeration of what goes on in a university classroom, the underlying process of moving from reading to evaluation to writing is likely not.

While doing my research on EAP writing coursebooks, looking at various aspects of the writing tasks, I noticed a trend: very few coursebooks follow the process outlined above. In fact, few of them deal with evaluation of evidence at all, and if they do, evaluation is usually superficial, dealt with once, included towards the end, or tacked on as an appendix. Some ask students to complete a T-chart of pros and cons, but the thinking doesn’t go much deeper than that. The ones that do try to incorporate it more often or ask students to look at both sides of an issue should certainly be commended. But those that ask students to “Consult one source, either in the library or on the Internet, and use some of the information, including quotations, in your essay” and leave it at that need to re-evaluate what they are asking.

In my research, I looked at 130 writing tasks from 12 popular EAP writing books. In them, I noticed a number several issues that I feel are connected to confirmation bias:

  • More than half of these books require no reading from writing at all.
  • Only 30% of coursebooks had integrated writing tasks integrated with the coursebook readings.
  • More than half of the tasks that did require writing from reading and research often asked students to do “internet research” with nary a critical analysis included.

How is this related to confirmation bias? The way many of these tasks typically work is to first design a thesis (and perhaps even supporting points), and then find evidence that gives support to the thesis. Few of the tasks, if any, required reading about a topic from various perspectives and critically analyzing the readings before forming an opinion. From these findings, it’s clear that the academic writing process of moving from reading (including analysis and discussion) to writing is not well represented by these tasks. By emulating a backwards process of writing followed by optional reading, these books seem to be engendering confirmation bias and leaving students with the idea that supporting their ideas requires finding expert quotations that are in agreement with them. It is giving them a backwards and flawed writing experience, one which disengages critical thinking and does not prepare them for the academic world they are supposedly designed for.

Of course, the teacher is not (or at least should not) be at the mercy of the coursebook. It is up to the teacher to carefully adapt, supplement, and design tasks in a way that teaches a more critical reading and writing process. However, it still must be stated that the coursebook holds enormous power as a syllabus-defining text and the use of it often leaves little time for extension, adaptation, or circumvention.

In addition, another such caveat is the fact that not all university courses likely offer the critical practices that can break confirmation bias. I’m sure there are many courses that are as guilty as coursebooks for playing service to it. Some course simply want a regurgitation of ideas from the textbook. Others want  a mirror image of the professor’s opinion.

One reason confirmation bias is so hard to overcome is that because it is a natural human tendency. There has been quite a bit of research into how to deal with confirmation bias. A recent study by Villarroel et al (2016) state that simply reading evidence from both sides, or reading “disconfirming evidence” is not effective in overcoming confirmation bias. In their research, they found that “argumentative discourse can mitigate the effects of confirmation bias when discourse goals are directed towards consensus-seeking dialogue” rather than simply proving or disproving opinions. They state that speakers are likely to be less biased when they place more value on consensus than simple persuasion: “Depending on the argumentative goal, students may tend to use data in a biased manner to fit their thesis.” Similar findings were reported by Lin, Chiu, Hsu & Wang (2015).

This has clear and broad implications:

Teachers must learn to help students examine and revise prior beliefs through the careful analysis of evidence. This is no easy task. Even under the guidance of experienced teachers, students struggle with making sense of evidence when it challenges their prior beliefs
Our findings also support the view that collaborative argument has the potential to promote a more careful analysis of evidence conducive to conceptual change.
value of consensus goals over persuasion goals in argumentative dialogue

These findings have some interesting implications for discussion and speaking. Debates and other side-taking activities may be counterproductive in helping students challenge the beliefs they have. Rather, cooperative critical discussion of evidence to reach a mutual conclusion may be a more productive and effective use of class time. This could have a radical effect on different speaking activities we use in class.

In terms of writing instruction, these findings echo the sentiments I expressed at the beginning of the post. Readings on important issues need to include various perspectives if there is any chance of true critical engagement of ideas. The reading of competing materials, however, is not enough. Discussion seems to be a vital component of mitigating confirmation bias and, therefore, should become an integral part of the reading/writing process. And, as the study above notes, this discussion should be with the goal of establishing a consensus rather than simple persuasion.

Returning to the original question of whether coursebooks engender confirmation bias, it seems that they do, but perhaps no more than other activities meant to engage evidence-based thinking and critical engagement. Although teaching critical thinking and mitigating confirmation bias is possible, it is no easy task. As more research comes out that shows how to diminish this type of bias, our teaching can be more informed and our students’ learning can be both more critical and more effective.

References

Lin, C. H., Chiu, C. H., Hsu, C. C., & Wang, T. I. (2015). The Influence of Playing a for or Against a Controversial Position on Elementary Students’ Ability to Construct Cogent Arguments. The Asia-Pacific Education Researcher, 24(2), 409-418. Retrieved from here.

Villarroel, C., Felton, M., & Garcia-Mila, M. (2016). Arguing against confirmation bias: The effect of argumentative discourse goals on the use of disconfirming evidence in written argument. International Journal of Educational Research, 79, 167-179. Retrieved from here.

3 thoughts on “Do coursebook writing tasks engender confirmation bias?

  1. Hi Anthony – I was particularly intriguing by the multiple references to consensus-seeking discussion activities rather than side-taking debates as a way to challenge, rather than confirm, bias (I even cited you on it in my recent blog post). What would that look like in practice? I’d love to hear some practical ideas for setting that up.

    And what about debate activities where you’re forced to take a an opinion opposite to your own (in my experience this only sometimes works – and may just produce a half-hearted debate)?

    • Anthony Schmidt says:

      Thanks for commenting (and citing).

      For debate, while may be good practice in building up argumentative language and seeing an issue from another perspecitve, it may be detrimental to critical thinking, or at least confirmation bias (keep in mind, MAY is an important word because this is coming from one study). The authors state that in a debate based on persuasion:

      “Instead, persuasion goals seem to trigger a defense motivation that is antithetical to the task of evaluating data with care and impartiality (Hart et al., 2009), inducing participants to misinterpret data to strengthen their claims.”

      As for practical activities, the original article has the example task they used in the research. Do you have access to the article? I’m not sure I can sure such a large amount of their article here. The example was data presented for and against nuclear power by experts. Participants had access to 4 charts that served as evidence. The goal was to write an argumentative essay. Before essay writing, the participants in the consensus treatment group had to present their opinion and justification to a partner and work together to reach an agreement on the solution.

      I’m sure there are a number of ways to make this work in the classroom. The joint-solution discourse goal can work for almost any topic.

      One final point from the article: “none of the participants in this condition actually changed sides. Given the complexity of the topic and the strength of views, it is unlikely that a 30-minute dialogue and four pieces of evidence would cause individuals to switch sides. Nonetheless, consensus goals did impact the way in which individuals processed and interpreted evidence, showing that reduced confirmation bias came not from reaching a consensus in opinion, but from the active attempt to reconcile divergent views.”

      I am planning to do a mini-debate on the atomic bomb in Hiroshima and Nagasaki soon. I typically to the pro/con debate, but this time I am going to give students as much evidence as I can and put them in a position to decide a solution to present to Truman. If we actually get time to do this (some other things are also coming up), and it goes well, I will follow-up with a blog post.

      • Thanks for thorough reply, Anthony. No, I don’t have access, but your explanation was clear. I look forward to hearing about how your mini-debate goes. I hope it’s the bomb!

Leave a Reply