How do differing discourse goals affect students’ abilities to process evidence? Does the act of argument and persuasion mean they read evidence from a biased perspective? If they argue from the opposite side’s perspective, will that change their own opinion? What if they had to come to a mutual decision? Would that affect their opinion?
All of these ideas are inspired by an article I wrote about last month when I discussed whether coursebooks promote or reinforce confirmation bias – the “tendency to search for, interpret, favor, and recall information in a way that confirms one’s preexisting beliefs” (Wikipedia). The study which looked at the effect of persuasion and consensus building on dealing with possibly disconfirming evidence (Villarroel et al., 2016). The authors found:
- Persuasion-based dialogue triggers confirmation bias in writing.
- Consensus-based dialogue mitigates confirmation bias in writing.
Basically, this experimental study found students using evidence to come to a mutual decision were “less likely to read disconfirming evidence incorrectly, partially or subjectively in favor of their prior beliefs”. The authors state that an emphasis on evidence-based argumentation (the trend in most education circles) “can only be successful if teachers create an environment conducive to the open and impartial examination of evidence.” Here, debate and “cognitive conflict” and simply being exposed to disconfirming evidence is not seen as useful. They feel that “collaborative argument has the potential to promote a more careful analysis of evidence conducive to conceptual change”.
These are some big ideas, and I wanted to play with them to understand how they could be applied to my own teaching situation. However, I did not follow the experimental design of the original article. In that article, students were given information about nuclear energy (for, against, and others that can be interpreted) and had to write an argumentative essay. After that, they were assigned to someone with the opposite view and then had to come to a mutual solution to a related problem (all done via text chat). This is easily adaptable to written and oral tasks, but, unfortunately, I did not have enough time to do so. Therefore, my activities are inspired by but not necessarily based on the original research.
For my class, which was an advanced language class with the theme of US history, I saved the last two weeks to discuss and debate the dropping of the atomic bomb on Hiroshima and Nagasaki. We did a number of activities, including research gathering, analyzing the structure of a debate, crafting logical arguments and refutations, etc.. I then set up two activities to try Villarroel et al’s ideas:
- a whole-class semi-formal debate arguing from the opposite side’s perspective: for or against dropping the bomb
- a brief consensus activity in which they must advise Harry Truman to drop the bomb, not drop the bomb, or do something else
From just observing my students I saw that arguing the opposite side’s perspective (from their own perspective) quite passionately did not change minds, even though it seemed that it had. Students were fiercely in debate and it really seemed they were convinced of their position. But after the debate, a show of hands indicated that they had not actually changed their opinions. From the next activity, I (and my students) realized just how hard reaching a consensus was. This activity, where students were paired off with someone from the opposite team, often devolved into debate rather than mutual decision. However, with some redirection, students were able to eventually able to come to an agreement. Some recommended dropping the bomb elsewhere, some recommended changing the conditions of surrender for the Japanese, and others recommended waiting. If they had been given more time, more evidence, and perhaps the task to not only orally discuss but produce in writing a joint statement, this activity may have been better. Nevertheless, I was duly impressed with my student’s engagement with evidence throughout both activities.
In the end, opinions did not change much. This is similar to what the authors found: “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 emphasized much above because opinions did change somewhat. I surveyed students afterwords and was able to see how exactly their opinions moved after each debate. I’ve reproduced the results in the table below:
Note: an opinion of 1 equals “Yes, definitely drop the bomb” and an opinion of 10 = “No, definitely do not drop the bomb”.
|Student||Opinion before debate||Opinion after debate||Opinion after consensus|
All opinions changed after the debate, most likely because students had a chance to examine more evidence and discuss both sides of the issue. Interestingly, most opinions trended downward towards dropping the bomb. Two students (B, F) who started somewhat neutral (6) ended more in favor of dropping the bomb (4). After the consensus, most opinions did not change. One (G) changed from neutral (after the debate, originally starting against the bomb), to a slightly weaker anti-bomb opinion (8). Student E started against the bomb (3), became neutral after the debate, and then became even more against the bomb (2) after the consensus activity. The consensus activity results were mixed, likely due to time and activity constraints, but also possibly because the argumentation (the activity did include argument, as noted above) was informal, non-structured, and between two people rather than a formal, 4-person team debate,
All students said the consensus-building activity had some impact on their opinion, but overall, doing the research itself had the greatest impact. The research was relatively simple (Wikipeda, Debatepedia) and presented both sides of the argument. If they had to conduct research on their own, confirmation bias and faulty evidence may have played a larger role in this activity.
Interestingly, when asked if students’ final opinion were based more on evidence or feeling, the response was overwhelmingly “Mostly evidence, some feeling”. Only student E responded 100% on evidence. This shows that evidence plays a large role, but prior beliefs and “gut feelings” cannot be discounted or easily mitigated.
In conclusion, I found that doing both activities were worthwhile, fun, and productive. I neither confirmed nor refuted Villarroel et al’s ideas but rather used them to inform an interesting activity that dealt with evidence and argument, two staples of critical thinking and higher education. Running this “pilot” has given me a number of ideas for future activities, including spending more time on consensus building, which is perhaps a more important task than simple argument, as this seems “to trigger a defense motivation that is antithetical to the task of evaluating data with care and impartiality”. If we are trying to change hearts and minds, we need to all be on the same page.
(Thanks to Kyle Dugan for making me think about how the research article’s ideas can be applied to the classroom.)