Dr. Edward L. Fink of the Communication Department at the University of Maryland recently discussed a 2008 study that he described as having "practical implications for advertising, marketing, health campaigns, and political campaigns."
The study, published in the journal Communication Monographs, is titled "The Effects of Message Discrepancy and Source Credibility." It was completed by Dr. Fink and associates Sungeun Chung and Stan A. Kaplowitz.
In it, the research team concluded that persuasive messages from sources with high levels of credibility induced much greater differences in belief change than messages from sources with low credibility, which is to be expected. More interestingly, the team discovered that in the short term, messages that advocated extreme views brought about a greater belief change than messages that advocated only slight change.
STUDENTS WERE FORCED TO MAKE TOUGH DECISIONS BEFORE AND AFTER BEING PERSUADED
The experiment used 95 Michigan State students, who sat in front of computers and read two different presentations. The first presentation was a fake sentencing guideline for criminal court cases involving armed robbery. The second presentation was supposedly taken from a meeting of the board of trustees at their university. In that presentation, students discovered that the board planned to raise the tuition at Michigan State the following year.
After reading the first presentation, the students were each given an argument from a judge's opinion about how long sentencing for armed robbery cases really should be. The judge's opinions varied in recommended sentencing length, and the students were told before they read whether the judge was respected or not by his peers. Using the computer-mouse technique, the research team tracked how persuaded or not the students were becoming until they made a final decision on how long they believed the sentencing should be - usually about 50 seconds later.
The second study followed suit - after reading the presentation, students were given a different argument from a member of the board of trustees who they were told was either pro-student or anti-student (thus determining his credibility to the students). Their beliefs were tracked.
THE MODEL USED WAS CONFIRMED MORE STRONGLY THAN ANTICIPATED
Using previous persuasion models, the team knew that source credibility and message discrepancy - how different the message being pushed was from the original beliefs held by the students - were the two biggest factors in persuasion. The researchers sought to determine whether they could find an equation that roughly tracked the participants' beliefs from the beginning of the study until they made a new decision; it turned out that the equation they decided upon was near perfect.
"I was surprised the model worked so well," admitted Dr. Fink. "It wasn't a smooth curve because of the short time period, so we had to compromise a bit, but it lined up the way we hoped it would."
The team discovered that beliefs progress over an L-shaped curve until they settle on a new equilibrium, or state of rest. Two of the curves - belief changes charted from students after being persuaded from high credibility sources with either an extremely different message (blue curve) or only a marginally different message (red curve) are shown below. The graph indicates that the extreme position changed the students' beliefs much more.
THE EXPERIMENT CAN BE DUPLICATED OUTSIDE OF THE LABORATORY
Dr. Fink believes that the results of the study have real-life implications. He was most interested in the students' belief changes on the topic of the tuition raise, because that decision would affect the students directly. He thus described that particular study as having a level of ego involvement, which is important with advertising.
"If somebody is not interested in cars, or if somebody is not interested in purchasing a car, then an advertisement about cars wouldn't have the same effect, it wouldn't have ego involvement," explained Dr. Fink. "You have to have the combination of the two for a car advertisement to actually have a level of involvement."
Dr. Fink also believes the study should show similar results across different demographics of age. "I might say a joke that most college kids wouldn't laugh at but that most 40-year-old adults would find funny, and vice versa," he said.
"As long as the content is equal amongst groups, then the process is the same as well. For example, if you were doing a study on criminal sentencing for trial lawyers, because they're a group, just like college students are a group, it would have the same effect as the study done on college tuition for the college kids."
THE RESULTS COULD BE SEEN APPLIED AROUND ELECTION TIME
One thing that Dr. Fink emphasized was that this study's endpoint was when the participants reached a new decision. In this way, the study was different from previous research that examined the impact of persuasive messages for days and weeks after the participants had been persuaded. The results were markedly different as well.
"Messages with moderate message discrepancy [in the previous study] showed to be fairly ineffective initially, like in our study. Over a course of weeks, however, they become extremely effective. The opposite holds true, as well - over time, they flip."
Dr. Fink advised that this phenomenon could be utilized by political campaigns as the next election nears.
"Election day advertisements should be completely different from ads that were run two weeks prior," he said. "The ads run on election day should be much more extreme than the others."
THIS STUDY HAS BEEN ANALYZED MULTIPLE TIMES
This was not the first time that Dr. Fink and his team analyzed data from this type of study. According to Dr. Fink, this was actually the third trial of a study originally conducted in 1996.
"We had a lot of different questions that weren't answered in previous studies, and the data was rich," Dr. Fink said. "There were still a lot of things to do, so we analyzed [the data] a third time."
He said that the focus of this research was different from the previous times, because there wasn't such an end-result focus.
"Most studies are just question and answer - yes or no - and they take up to 40 seconds. This was a study about the thinking process that goes into it."
