WHETHER RECOMMENDING A RESTAURANT, endorsing a brand or promoting a policy, it seems reasonable to think that the more arguments you can provide, the more persuaded your audience will be. Indeed, past research suggests that, assuming one has compelling points, providing many rather than few arguments generally promotes greater persuasion.
Laypeople share this belief. When I asked 249 online participants whether giving a low or high number of arguments would be more persuasive, 59 per cent indicated that using a high number of arguments would be the better strategy.
In contrast to this view, in recent research with Mohamed Hussein and Zakary Tormala of Stanford University’s Graduate School of Business, we posited that a high rather than low number of arguments might have conflicting effects on persuasion. And further, these competing effects might actually cancel each other out, resulting in minimal or no overall benefit to persuasion from using more arguments. In this article I will summarize our research and the key takeaways.
Less Can Be More
Past research suggests that the more arguments one provides for a position, the more persuasive one tends to be. In one notable study, participants were asked to imagine being trial jurors and read arguments for both the prosecution and the defence. Then, they rated how guilty the defendant was. The researchers varied the number of arguments given by the prosecution and the defence to be one, four or seven. Results indicated that the more arguments provided by the prosecution, the more guilty participants found the defendant to be — and the same went for the defence. In other words, more arguments led to more persuasion.
In another example, undergraduates read an editorial promoting a campus policy. The researchers varied argument quantity (three versus nine), argument strength (weak versus strong) and participants’ involvement (low versus high). They found that the effect of argument quantity depended on involvement. Under low involvement, nine arguments led to more persuasion than three, regardless of argument strength. This effect was thought to be driven by a simple ‘more-is-better’ heuristic. Under high involvement, the argument quantity effect depended on argument strength. Because high-involvement participants processed more carefully, nine (versus three) arguments promoted persuasion when arguments were strong, but undermined persuasion when arguments were weak. These results suggest that as long as one’s arguments are strong, providing more of them fosters more persuasion.
In our research, we hypothesized that people actually draw competing inferences about sources who provide many arguments. On the one hand, they might come across as more expert on the topic, which boosts persuasion; on the other, they might come across as more biased, undermining persuasion. These inferences, we felt, could suppress each other, producing minimal or no overall benefit to providing many arguments.
It seems reasonable to suspect that a source who provides many arguments might be seen as having more expertise on the topic at hand. ‘Source expertise’ refers to the extent to which a source is knowledgeable, competent and capable of making accurate claims. Compared to someone who offers few compelling arguments (e.g. one or two), a source who provides many (e.g. nine or 10) might seem to possess greater knowledge and competence and, thus, more expertise. As indicated, prior work suggests that experts are more persuasive than nonexperts. Thus, to the extent that using more arguments increases perceived expertise, we expected it to enhance persuasion.
At the same time, however, we felt that a source who uses many arguments might come across as more biased on the issue at hand. ‘Source bias’ refers to the extent to which a source has a slanted or skewed perspective on an issue or a vested interest in a specific outcome. Compared to someone who provides few arguments, a source who provides many might elicit questions such as, ‘Why is she providing so many arguments? Does she have a vested interest?’ or ‘Does he have a hidden agenda here?’ In other words, message recipients might question whether sources providing many arguments are biased.
Although there is not a wealth of research on source bias, the nascent literature suggests that perceived bias undermines persuasion. For example, one researcher found that under-justifying one’s position (e.g. providing weak arguments) can lead to perceptions of greater bias, which reduces persuasion.
My colleagues and I felt that just as under-justifying one’s position with weak arguments could trigger perceived bias, overjustifying one’s position with too many arguments could, as well. Specifically, we predicted that providing more arguments than is typical in a given setting increases perceived bias. For example, if the norm on a customer review platform such as Yelp is to provide three reasons for your positive review, providing 10 might lead people to infer that you are biased. If this proved to be true, we hypothesized that providing many arguments might undermine persuasion, as well.
As predicted, there was both a positive effect of over-arguing (through perceived expertise) and a negative effect (through perceived bias).
We tested our hypotheses across three experiments, using different topics and settings, including brand endorsements on social media and restaurant reviews. In our first experiment, participants viewed a social media post containing either three or nine arguments. We predicted that providing more arguments would increase perceived expertise and perceived bias, and that these perceptions would have a ‘cancelling out’ effect on persuasion.
Four hundred one participants completed our study, and 369 were included in the final analysis. Participants imagined that they were browsing Instagram when they came across a story from someone they followed. Next, they viewed an Instagram story that appeared over two pages. On the first page, they saw a picture of cold-brew iced coffee. The picture contained text that endorsed the brand of cold-brew (‘Love this cold-brew coffee!’ and ‘Highly recommend that you give it a try!’). We made an effort to mimic a typical Instagram story, but did not specify the brand.
The second page differed by condition. Participants in the three-arguments condition saw three arguments advocating for the coffee brand (e.g. ‘In terms of flavour, this one definitely stands out. Not too sweet and has a smooth and creamy texture.’) In both conditions, the arguments were randomly selected from a larger list of 10 and all appeared on the same page.
Participants indicated their overall attitude towards the brand using a star-rating similar to those commonly employed online. They could report any star-rating between .5 and 5, with half-star increments (e.g. .5, 1, 1.5), making this a 10-point scale. They reported their interest in trying the brand of coffee on a 7-point scale (1 = not at all interested, 7 = very interested). Then they completed a three-item ‘bias index’:
- How much do you believe that this user has a biased perspective on this brand of cold-brew coffee?
- How one-sided does this user seem on the subject of this coffee brand?
- To what extent do you believe this user has a vested interest or a hidden agenda in promoting this brand?
Responses, provided on scales ranging from 1 (not at all) to 7 (extremely), were averaged. Participants also completed a two-item ‘expertise index’:
- How knowledgeable do you think this user is about this brand of coffee?
- How much of an expert on this brand of coffee do you think this user is? Responses, provided on scales ranging from 1 (not at all) to 7 (extremely) and once again, were averaged.
RESULTS: Participants perceived the source as ‘more expert’ in the nine-argument rather than three-argument setting and more biased in the nine-argument rather than three-argument setting. In terms of persuasion, attitudes were slightly more favourable in the nine-argument condition, but there was no difference in interest in trying the brand between the two settings.
While there was no overall effect of the number of arguments on interest, we did find evidence for suppression: As predicted, there was a positive indirect effect through perceived expertise and a negative indirect effect through perceived bias.
For attitudes, we found a slight-but-significant positive effect of number of arguments. Importantly, though, a negative indirect effect through bias attenuated what would have been a stronger overall effect. Indeed, we found a positive indirect effect through perceived expertise and a negative indirect effect through perceived bias.
In short, our first study had conflicting effects on persuasion. Offering nine rather than three arguments increased perceived expertise, which promoted persuasion, but also increased perceived bias, which undermined persuasion. Combined, these effects produced no or minimal benefit to increased argument quantity.
Our second experiment was similar to the first but used a different context. Participants read either one or 10 positive reviews of a restaurant, all written by the same reviewer and ostensibly taken from Yelp.com. We predicted that participants would perceive the source as more expert and more biased in the 10- compared to one-review condition. Furthermore, we predicted that these perceptions would have countervailing effects on downstream consequences, including participants’ attitudes towards the restaurant and interest in trying it. We also measured an additional consequence — interest in receiving future information from the source — to assess whether the suppression effect would generalize to outcomes beyond the attitude object.
Four hundred and four participants completed our study on Prolific Academic, and 374 were included in the final analysis. All participants read reviews of a restaurant called HappyGreens. Participants in the first condition saw one positive review, whereas those in the second condition saw 10 positive reviews, all written by the same source. They then answered questions about their attitudes, interest in trying the restaurant, desire to receive future information from it, and perceived source bias and source expertise.
RESULTS: Participants perceived the source as more expert in the 10-review rather than one-review condition; and also, more biased in the 10-review rather than one-review condition. There was no difference in any of the persuasion outcomes. Attitudes did not differ following 10 versus one review; interest in trying the restaurant did not differ following 10 versus one review; and interest in receiving future information from the source did not differ. For all three persuasion outcomes (attitudes, interest in the restaurant, interest in receiving further information), we found evidence for suppression. That is, we found a null overall effect of number of arguments on these persuasion outcomes. In addition, there was a positive indirect effect through perceived expertise and a negative indirect effect through perceived bias.
Experiment 2 replicated the ‘dual effect’ of argument quantity. First, compared to one positive review, providing 10 increased perceived expertise, which had a positive impact on persuasion. However, 10 positive reviews also triggered perceived bias, which had a negative effect on persuasion. These two competing effects suppressed each other, producing no overall benefit to offering many arguments compared to just one.
Together, Experiments 1 and 2 showed that increasing argument quantity can have conflicting effects on persuasion. Using many rather than few arguments increased perceived expertise and bias, and these perceptions had a cancelling effect on persuasion-relevant outcomes.
Our outstanding question was this: At what point does adding more arguments introduce perceived bias and offset the benefit of conveying expertise? Our third experiment addressed this question. In an online review context, we presented participants with anywhere from one to 10 reviews and tracked perceptions of expertise and bias across all ten conditions. We also assessed participants’ perceptions of norms around posting online reviews — specifically, how many reviews are typical and how many reviews would be too many. We explored the possibility that when argument quantity exceeds the norm in a given context, the countervailing effects of source bias and expertise emerge.
One thousand and twelve participants completed our study, and 969 were included in the final analysis. Similar to Experiment 2, participants read positive reviews of a restaurant called HappyGreens. Here, the number of reviews varied from one to 10. Following the review(s), participants reported attitudes, interest in trying the restaurant, and perceived bias and expertise.
Finally, we assessed perceived norms. First, participants reported the number of reviews people generally leave on Yelp (‘In general, when someone reviews a restaurant on Yelp, how many reviews on the restaurant’s page do they leave?’). Second, they reported how many reviews would be too many (‘Sometimes people leave more than one review of the same restaurant on Yelp. How many reviews from the same reviewer seem like too many? For example, how many reviews from the same reviewer would make you feel doubt or suspicion?’). These questions, we reasoned, could help us identify the point at which adding arguments might trigger perceived bias and suppress the expertise effect on persuasion.
RESULTS: Participants perceived the source as more expert in the 10-review rather than the one-review condition, and as more biased in the 10-review condition. In terms of persuasion, there was no difference in attitudes between the ten-review and onereview condition, nor in interest in trying the restaurant.
For attitudes, while there was no overall effect of number of arguments, we did find a positive indirect effect through perceived expertise and a negative indirect effect through perceived bias. For interest in trying the restaurant, there was no overall effect of number of arguments, but there was a positive indirect effect through perceived expertise and a negative indirect effect through perceived bias.
Put simply, including additional arguments — even cogent ones — often yields no benefit.
Next, we examined perceived norms to pinpoint when we might expect the countervailing forces of bias and expertise to produce suppression. Most participants (84 per cent) reported that they would expect to see only one review from the same reviewer on a restaurant’s Yelp page. The average number of reviews that would seem like ‘too many’ was 3.32. Together, these results suggest that as the number of reviews from the same reviewer passes one and approaches three, we start to observe suppression. Thus, we conducted the same analyses described above, but instead of comparing one- and 10-review conditions, we compared one- and two-review conditions.
Here, we observed an immediate rise in both perceived expertise and perceived bias when comparing one to two reviews. Consistent with this finding, there was no difference in attitudes or interest in trying the restaurant between the one- and two-review conditions, and we found suppression effects for both outcomes. For attitudes, there was a positive indirect effect through perceived expertise and a negative indirect effect through perceived bias. For interest too, there was a positive indirect effect through perceived and a negative indirect effect through perceived bias.
Our third experiment replicated the earlier studies. Most notably, it provided a simple norms-based approach to detect the point at which adding arguments might be perceived as over-arguing. It helps establish the role of norms in shaping the inferences people draw from argument quantity. Practically, it helps persuaders gauge norms and modify their messages to avoid over-arguing their positions.
Of course, argument-quantity norms will vary across contexts. With online reviews in Experiment 3, the perceived norm was one review; providing more triggered perceived bias. In other contexts, such as writing an editorial or pitching a new idea in a meeting, norms might be higher. Normative differences should influence when increasing argument quantity no longer boosts persuasion.
Key Takeaways: Our three experiments show that providing a high rather than low number of arguments can have conflicting effects on persuasion. Presenting many arguments boosts perceived expertise, which has a positive effect on persuasion; but at the same time it builds perceived bias, which has a negative effect on persuasion. These forces can cancel each other out, providing minimal or no benefit to offering many arguments to support one’s position. We observed these effects in the context of online endorsements of products and restaurants, but we suspect that the same phenomenon occurs in other settings as well (e.g. when a colleague speaks up at a team meeting). Put simply, including additional arguments — even cogent ones — often yields no benefit.
Although it can be tempting to add arguments to your message to increase persuasion, as indicated herein, additional arguments can introduce perceived bias. Adding arguments comes with costs, both in terms of time (e.g. the time it takes to generate arguments) and money (e.g. adding words in a print ad can be expensive). Our findings act as a reminder to think twice before including additional arguments in any message, even if the arguments are compelling. The trade-offs between expertise and bias can actually undo any persuasive advantage of further argumentation.
Rhia Catapano is an Assistant Professor of Marketing at the Rotman School of Management. This article has been adapted from her working paper, “15 Reasons You Should Read This Paper: How over-Arguing Increases Perceptions of Both Expertise and Bias.”
Share this article: