Consider the Source: Persuasion of Implicit Evaluations is Moderated by Source Credibility
By Colin Tucker Smith (University of Florida and Ghent University) *
Consider a string of web searches I recently embarked upon. Is the new locally-sourced restaurant in town any good? What is the definition of “locally-sourced?” Is genetically-modified food safe? Is Monsanto evil? Is President Obama pro-Monsanto? The Internet is awash with opinions, and it is easy to get inadvertently swept away in it. In any case, when I reported the results of my search to my partner, she had an additional question: Where did I get my information? This question feels increasingly relevant in the “age of the internet” and what is at least subjectively a more and more fractured media environment. Of course it does matter who the information came from. Indeed, when it comes to political issues, the inability of political partisans to agree on simple facts would be comical if it wasn’t so dangerous.
Social psychologists have long known that the evaluations that people directly self-report (i.e., “explicit” evaluations) are impacted by the credibility of the source of the information (e.g., Briñol & Petty, 2009; Hovland, Janis, & Kelly, 1953). In the current work (conducted with co-authors Jan De Houwer and Brian Nosek), we were interested in whether this is also true for a different type of evaluations – “implicit” evaluations. Implicit evaluations stand in contrast to explicit evaluations, which people can directly self-report and which are thought of as relatively conscious summaries of mental contents which people experience as being “theirs”. Implicit evaluations, on the other hand, are activated automatically, occur relatively outside of conscious awareness, do not come with a sense of ownership, and are most commonly measured via computerized tasks that do not alert the participant as to what is being measured. Because implicit evaluations have proven to be an important predictor of behavior (e.g., physicians’ treatment decisions, police shooting decisions), researchers have been interested in ways to change them. Recently we noticed that researchers were almost exclusively using one of two methods to change implicit evaluations. One way is by pairing an attitude object with something positive many times in an attempt to retrain the underlying information. The second way is by manipulating what information is activated (e.g., by having people focus on positive examples of a negatively-stereotyped group). While both of these methods have had great success (see Lai, Hoffman, & Nosek, in press for a review regarding reducing implicit prejudice), in the current work, we wanted to expand the range of methods for researchers interested in changing implicit evaluations. As a start, we took inspiration from the long tradition of researchers studying persuasion of explicit evaluations to see if their methods would also be effective with implicit measures.
To test this, we presented participants with information about a fictitious laundry detergent company named Soltate (Smith, De Houwer, & Nosek, 2013). Every participant read the same information, which was quite positive. For example, participants were told details such as “The laundry detergents made by Soltate are some of the best detergents available today” which “over the past few years they have topped the charts in customer satisfaction in the regions where they are sold.” Importantly, although all participants read the same product information, we varied the source of the information across participants. We tested whether participants’ implicit evaluations were impacted by the credibility of the source, which we operationalized through ‘expertise’ and ‘trustworthiness’. The source’s ‘expertise’ was manipulated by telling participants in the ‘High Expertise’ condition that the information about the detergent was collected by a man named Jonathan Brower who “is a Yale-educated chemist who works for his state’s Department of Water Resources” and “volunteers on a bi-partisan Consumer Protection Board for his local government.” In contrast, in the ‘Low Expertise’ condition, Jonathan Brower was a 14-year old who “spends one Saturday each month meeting with his middle school’s Consumer Club” which is “a small group of students ranging from 4th to 8th grade, and is usually supervised by Jonathan’s Physical Education teacher”. Similarly, we manipulated ‘trustworthiness’ by telling participants in the ‘High’ Trustworthiness’ condition that the information about the laundry detergent was found on the homepage of a local Consumer Protection Board, while those in the ‘Low Trustworthiness’ condition were told that the information was found on the homepage of the company that made the laundry detergent in question.
In five studies, using three different implicit measures, we found that who presented the information was an important factor in the strength of the resulting implicit evaluations. Specifically, when the source was high in expertise (Studies 1, 2, and 5) or trustworthiness (Studies 3 and 4), implicit evaluations were stronger than when the source was low in those qualities. This is clear evidence that even those parts of our minds that operate relatively outside of conscious awareness are impacted by who tells us information. In addition, it potentially greatly expands the types of manipulations which researchers can use to change implicit evaluations if the long history of persuasion research also proves relevant for implicit evaluations.
Back to my Internet search reported at the beginning, these findings lead me to believe that we need to be more careful of how we come by our information, because it is currently unclear whether we are able to correct for the potentially biasing effects of the source of information we read. Even though we may be able to correct our relatively conscious evaluations in the short term, we may be less likely to correct our less conscious evaluations and over time these implicit evaluations may drive the evaluations we stand behind as our own. As one example, implicit evaluations of a political proposal are more positive when presented by a member of one’s own political party even though participants report being entirely unaffected by that information (Smith, Ratliff, & Nosek, 2012). Crucially, these implicit evaluations mediate self-reported evaluations gathered up to three weeks later. More research is needed to investigate how long-lasting the effects of source information are on implicit evaluations, whether people can correct for this influence, and the conditions under which implicit evaluations predict future self-reports and behavior. By the way, the restaurant was great, just as Yelp predicted.
Colin Tucker Smith is a post-doctoral researcher in the Department of Psychology at the University of Florida. He studies implicit evaluations with an emphasis on understanding where such evaluations come from, whether they have conscious consequences for the individual, and how they can be changed. Recently, he has been engaged in work on how implicit evaluations interact with political and religious ideologies to predict behavior.
Briñol, P., & Petty, R. E. (2009). Source factors in persuasion: A self-validation approach. European Review of Social Psychology, 20, 49-96.
Hovland, C. I., Janis, I. L., & Kelly, H. H. (1953). Communications and persuasion: Psychological studies in opinion change. New Haven, CT: Yale University Press.
Lai, C. K., Hoffman, K. M., & Nosek, B. A. (in press). Reducing implicit prejudice. Social and Personality Psychology Compass.
Smith, C. T., De Houwer, J., & Nosek, B. A. (2013). Consider the source: Persuasion of implicit evaluations is moderated by source credibility. Personality and Social Psychology Bulletin, 39, 193-205.
Smith, C. T., Ratliff, K. A., & Nosek, B. A. (2012). Rapid assimilation: Automatically integrating new information with existing beliefs. Social Cognition, 30, 199-219.
* Author’s note: This research was supported by Ghent University Grants BOF/GOA2006/001, BOF/01M00209, and Project Implicit. The author is currently at the University of Florida; this research was conducted while he was a post-doctoral researcher at Ghent University.