When is credibility less important

All, while sipping gas like an economy hybrid? But I still doubt the claim of 91 mpg. Below is a chapter excerpt from The Persona Principle book. When first published in the s this chapter was widely excerpted because of its challenging notion that credibility in marketing has more value than truth. It is no less valid today, than when we first wrote these words two decades ago. Your truth must be credible. You are rarely given the opportunity to offer proof once your incredible claim is exposed to your audience.

Many of our clients have credentials that are too good. Cynicism is a normal human reaction to bold claims and innovations. Regardless of what media you use to promote your image, you have seconds, not minutes, to sell your claim. Unless you can confidently and conclusively prove your claim in five words or less, you have no hope of gaining credibility. It is usually impossible to gain credibility when you are making bold claims.

The same is true of clever claims. Smart or overly cheeky headlines and ad styles often destroy credibility, thereby damaging Image-Equity. They apply to any industry. Only one type of venture can benefit from clever and exciting campaigns: those with credible messages and dull images. You can use a more exciting message if you can easily prove your claim. Otherwise, let the facts shine through with cautious persona structures.

This is because they are brilliant, clever, beautiful, well crafted—and totally inappropriate for the company being advertised. Both your agency and your internal marketing staff must address your audience, understand your offering, reach your audience, and captivate. They must produce credible, memorable, appropriate-impact advertising.

Participants were recruited through SONA, an online participant pool system. After giving informed consent, participants received instructions to judge the statements according to their truth and were informed that the experimenters were interested in their spontaneous responses.

If a source was displayed, it appeared below the statement e. Pictures if present also appeared below the respective statement or below the source. After the 60 statements, the participants were thanked, debriefed, and received credit. A 3 source: expert vs. Figure 1 depicts the descriptive results.

Mean truth ratings in Experiment 1 as a function of source and picture conditions. Error bars represent standard error of the means. Due to the incomplete counterbalancing of one statement, we additionally analyzed our data with a linear mixed model that accounts for random variance in participants and items.

We used the R-packages lme4 Bates et al. The tested model included the fixed factors source condition and picture condition and the interaction between the two. Furthermore, the model included random intercepts for participants and statements. The results of Experiment 1 show that people consider source information when forming truth judgments. Importantly, however, they do not merely consider the availability or absence of source information but take the credibility of the source into account.

Specifically, we found that expert sources increased the perceived truth of statements whereas lay sources decreased the perceived truth compared to a control condition without source information. Surprisingly, we did not find an effect of non-probative pictures on truth judgments. We will elaborate on this finding in further detail in Experiment 2. According to Unkelbach and Greifeneder , people integrate declarative judgment cues e.

In Experiment 1, however, participants exclusively relied on source credibility as a cue. In contrast, non-probative pictures did not affect truth judgments, even though they presumably enhance processing fluency.

Because Experiment 1 is the first study that has investigated joint effects of source credibility and non-probative pictures on judged truth, we can only speculate why participants solely focused on source information does the statement come from an expert, a layperson, or is the source unknown?

Possibly, the integration of declarative cues and experiential cues requires specific conditions. Because repetition is one of the strongest and most popular fluency manipulations, we included statement repetition as an additional factor in Experiment 2.

In fact, Unkelbach and Greifeneder had observed additive effects of repetition and declarative advice on truth judgments. Thus, including repetition as a further factor allowed us to test the replicability and generalizability of their findings. Moreover, as stated earlier, investigating joint effects of statement repetition, source credibility, and presentation format is also important for applied reasons, because the three variables typically co-occur in the context of social media.

Ninety-one students 71 female, 16 male, 4 not specified from the University of Oslo, Norway, completed the experiment online. Two participants did not indicate their age. Participants were proficient in Norwegian and received course credit for participation.

Mean truth ratings served as the dependent variable. The materials were the same as in Experiment 1. In order to counterbalance repeated statements versus new statements across participants, we divided the 60 statements into two sets of 30 statements each. Half of the participants received the first set in the exposure phase, the other half received the second set.

Participants received all 60 statements in the judgment phase. Half of these statements were old i. We counterbalanced source types and picture conditions across repetition conditions. Footnote 4. The procedure of Experiment 2 was similar to Experiment 1, except for the following changes. First, there was an exposure phase and a judgment phase. In the exposure phase, we presented only the statements, without sources and pictures.

Second, participants completed a item non-verbal filler task between the exposure phase and the judgment phase in order to minimize the impact of short-term memory for the last statements see Postman and Phillips Each item consisted of an initial matrix identical across trials and a rearranged matrix, in which one symbol was missing. The pattern of this source effect was also identical to Experiment 1. This means that the effects of source and repetition were additive see Fig. Mean truth ratings in Experiment 2 as a function of picture, source, and repetition conditions.

We additionally analyzed our data with a linear mixed model because, as in Experiment 1, one statement was unbalanced across experimental conditions. The tested model included the fixed factors source, repetition, and picture and all possible interactions between these factors as well as random intercepts for participants and statements. In Experiment 2, we found strong, additive effects of source credibility and statement repetition on judged truth.

Expert sources increased the perceived truth of statements whereas lay sources decreased the perceived truth compared to a control condition without source information.

Moreover, truth judgments were higher for repeated compared to new statements. These findings are consistent with the general pattern of results reported by Unkelbach and Greifeneder and thus provide evidence for their replicability and generalizability.

Accordingly, our findings support the assumption that people integrate declarative cues e. Unlike repetition, however, non-probative pictures did not show any influences on truth judgments.

That means, as in Experiment 1, the truthiness effect did not replicate. We will come back to this point in a later section. Experiment 3 tested the generalizability of the previous findings using a different sample population German instead of Norwegian participants and a different set of materials. The experiment was similar to Experiment 2, except that statements were framed as news headlines and the presentation layout was similar to the one on the Facebook social media platform.

We implemented these changes in order to simulate a social media news context. Within this framework source credibility was manipulated by either presenting a statement together with the name and logo of a trustworthy, real news source real source , together with the name and logo of an untrustworthy, made-up news source fake source , or without any source information no source.

As in Experiment 2, we also varied repeated exposure to statements yes vs. Despite the mentioned changes, we expected to replicate the findings of Experiment 2. Participants were recruited at the University of Mannheim, Germany. Sixty-four participants 51 female, 13 male completed the experiment.

We collected statements, most of them from the Internet. We rephrased the statements so that they reflect the typical style of a news headline. That is, all statements included a buzzword or short introduction followed by the actual message e.

Additionally, we assembled 32 real German news sources and made up 16 names for news sources that factually do not exist i. Based on the pre-test, we selected 60 statements truth ratings: 3. Moreover, we selected ten real news sources that were judged as familiar by at least 20 of the 25 pre-testers and ten fake sources that were judged as familiar by no more than 1 of the 25 pre-testers. Presenting the sources together with logos helped participants to differentiate between familiar sources and unfamiliar sources.

For this reason, and for the sake of ecological validity, we decided to present the sources together with their logos in the experiment. We then searched for two thematically associated, non-probative pictures for each of the 60 selected statements.

If both pictures met this criterion, we chose the picture with the higher thematic-fit rating. The two sets served to counterbalance which statements were repeatedly presented and which statements were not. Next, we created pairs of real and fake news sources e. These source pairs were then assigned to three statements within each set based on their thematic fit to the real news source. For example, statements on an economic topic were assigned to a source that typically reports on economic affairs.

Each statement was counterbalanced across all cells of the experimental design according to a Latin square. The procedure of Experiment 3 was similar to Experiment 2, except for the following changes. First, this time the judgment phase directly followed the exposure phase. Second, we told participants at the start of the judgment phase that the statements would appear together with corresponding pictures and source information, if available. In fact, the assignment of sources and pictures was completely under experimental control.

This deception served to unravel source effects from statement effects. If a source was present, it was displayed in the upper left corner of the simulated news posting see Fig. Finally, the experiment involved a source judgment phase in which the 20 sources including their logos were successively displayed in random order.

For each source, participants had to indicate whether they had been familiar with the source prior to the experiment yes vs. Exemplary statement in the fake-source condition and picture condition. Note that in the original study the sources and statements were presented in German. This indicates that participants were able to distinguish between credible and dubious news sources.

A 3 source: real vs. We also replicated the truth effect obtained in Experiment 2, i. Hence, as in Experiment 2, the source effect and the truth effect were additive see Fig. Mean truth ratings in Experiment 3 as a function of picture, source, and repetition conditions.

We did so by running a linear mixed-model analysis. The model included the fixed factors rated source trustworthiness as a continuous variable , statement repetition yes vs. Moreover, the model included random intercepts for participants and statements. Truth judgments from the no-source condition were discarded from the following analysis. In line with the ANOVA findings, truth judgments were significantly affected by perceived source trustworthiness, F 1, Moreover, again, there was no effect of non-probative pictures F 1, Despite a different sample population, a modified presentation layout, and other materials, the results of Experiment 3 were essentially the same as in Experiment 2.

The only notable difference was in the pattern of the source effect. In contrast to the lay sources, fake sources did not decrease the perceived truth of statements compared to the no-source condition. This is remarkable, particularly in light of the low explicit credibility ratings for the fake sources.

Different news sources often report on the same information. For this reason, it is likely to come across the same statement repeatedly when skimming social media news headlines. Experiments 2 and 3 demonstrated that previous exposure increases the perceived truth of a statement regardless of its source.

However, what happens when encountering a news headline that is inconsistent with or even contradictory to another headline seen before e. Ibuprofen prevents severe courses of Covid? Previous studies suggest that people tend to disbelieve statements that are semantically incongruent with previously seen or heard statements Bacon ; Garcia-Marques et al. This illusion of falseness —i. But it is an open question whether the illusion of falseness also replicates in a simulated social media news context that includes source information.

To our knowledge, however, no study has yet investigated whether source credibility affects truth judgments differently in the case of incongruently repeated statements than in the case of congruently i. For this reason, the aim of Experiment 4 was to investigate joint effects of type of repetition congruent repetition, incongruent repetition, no repetition and source credibility real source, fake source on the perceived truth of alleged social media headlines.

Please note that for the sake of ecological validity, we no longer included a no-source condition. Moreover, because pictures had not affected truth judgments in our previous experiments, we also dropped this experimental factor. Instead, all statements appeared together with a picture and a source reference in the truth judgment phase to keep the study as naturalistic as possible. Participants were recruited at the University of Mannheim. Eighty participants 66 female, 14 male completed the experiment.

The experimental design was a 2 source: real vs. We constructed pairs of statements. Each pair consisted of two statements that were semantically incongruent with one another but differed in only one term e. Some of the statements were adapted from Experiment 3, other were newly created based on contents found on the Internet.

As in the previous experiment, all statements had a typical news-headline format. Although we were no longer interested in the effect of pictures, we decided to present each statement with a picture, because this is the typical presentation format of social media news headlines.

For this reason, we selected a non-probative picture for each statement-pair that fitted the content of both statements. Importantly, pre-testers never judged two statements of the same pair as these appeared in separate between-subject conditions.

Based on the pre-test, we selected 54 statement-pairs truth ratings of the individual statements: 3. The sources were the same as in Experiment 3. Therefore, we only used nine of the ten source pairs real—fake pairings of Experiment 3, which we assigned to the 54 statement pairs based on their thematic fit. Statements of the same pair were assigned to two different sets. Each statement was then counterbalanced across all cells of the experimental design according to a Latin square.

The procedure was the same as in Experiment 3, except for the following changes. In the exposure phase, participants judged the interestingness of 36 statements. Eighteen of these statements reappeared as verbatim repetitions in the judgment phase, i. The other 18 statements contained a semantic modification compared to the exposure phase, i. In addition, the judgment phase involved 18 new statements. In the final source judgment phase, participants again provided binary familiarity judgments as well as trustworthiness ratings for each of the 18 sources in random order.

A 2 source: real vs. Likewise, statement repetition affected truth judgments, F 1. Simple contrasts confirmed the expected pattern of statement repetition. Interestingly, this time, the truth effect i. Mean truth ratings in Experiment 4 as a function of source condition and repetition condition.

As for Experiment 3, we ran a linear mixed-model analysis to explore whether the above findings replicate when predicting truth judgments as a function of source trustworthiness ratings instead of real sources vs.

The model included the fixed factors rated source trustworthiness as a continuous variable , statement repetition congruent vs. Once again, we found significant effects of source credibility and statement repetition on the perceived truth of statements, but no interaction between the two factors.

This was the case even though the experiment contained semantically incongruent statement repetitions in addition to congruent repetitions. In accordance with previous research, congruent repetition led to a truth effect whereas incongruent repetition led to an illusion of falseness. That is, statements that diverged from previously presented statements were perceived as less true than unfamiliar statements although participants had no factual knowledge of their validity.

Moreover, source credibility did not moderate this illusion of falseness. This means that participants did not increase their attention to source information when encountering information that was incongruent to earlier encountered information.

Hence, even for incongruent statements allegedly presented by real news sources, there was still a decrease in rated truth both in comparison to congruently repeated statements and to new statements. To our knowledge, this is the first demonstration of the illusion of falseness in a social media news context. We will outline the real-world implications of this finding in the general discussion section. For this reason, we investigated effects of the following variables that, according to previous research, demonstrably affect truth judgments when studied in isolation: source credibility, statement repetition, and non-probative pictures.

As these variables typically co-occur in the context of social media, we were interested in their effects when being jointly manipulated.

Given the support for a basic truthiness effect in the literature e. At present, we can only speculate why the effect did not occur. As the effect is generally smaller than effects of source credibility and repetition, the other effects might have simply overshadowed truthiness.

In line with this idea, there were at least descriptive trends in all experiments that included a picture condition for truth judgments to be higher in the picture compared to the no-picture condition.

Alternatively, the failed replication of truthiness could also be based on our selection of pictures for which we had taken great care to ensure that they are non-probative. A truthiness study with our materials, but without a source and repetition manipulation would allow testing the above explanations against each other.

The effects of seemingly probative pictures either for or against a statement , in contrast, remain an interesting topic for future research. Unlike non-probative pictures, source credibility and repeated exposure turned out to be reliable determinants of judged truth.

Truth judgments were higher for statements presented with credible sources than non-credible sources and statements presented without source information. Moreover, congruent statement repetition increased perceived truth whereas semantically incongruent repetition decreased perceived truth. Importantly, the effects of source credibility and repetition were additive, in line with previous work by Unkelbach and Greifeneder , who had also found additive effects of declarative and experiential cues on judged truth.

When comparing the size of the source effect with the size of the truth effect—the credibility enhancing effect of congruent statement repetition—the former was larger in Experiments 2 and 3 whereas the latter was larger in Experiment 4.

Hence, the weighting of source credibility and fluency seems to be context-dependent. It is possible that differences in the experimental designs affected cue salience and thus resulted in a stronger weighting of source credibility in Experiments 2 and 3.

In addition to the varying size of the source credibility effect, the pattern of this effect also differed between experiments. In Experiment 1 and 2, we manipulated source credibility by source expertise, i. In the other two experiments, by contrast, we manipulated source credibility by source trustworthiness.

In Experiment 3, for example, alleged news headlines were either presented by a trustworthy news source, a fake source i. Interestingly, although expert sources and trustworthy news sources both increased the perceived truth of statements compared to the no-source condition, we observed different results for the lay sources and fake sources.

Statements in the lay-source condition were rated as less true than statements without source information, which suggests that participants discounted the information provided by laypersons. In contrast, this did not happen for the fake sources. Headlines in the fake-source condition received similar truth ratings as headlines presented without source information.

Based on this discrepancy, it would be interesting to investigate whether information presented by lay sources is perceived as less credible than information from unknown news sources. As this goes beyond the scope of our article, it remains an interesting point for future studies.

Based on Lazer et al. Thus, it is good news that source credibility proved to be a reliable determinant of judged truth in our experiments, replicating earlier findings see Wilson and Sherrell In addition, however, participants were susceptible to the truth effect, i. In the context of social media, however, statement familiarity is not a valid cue for truth, as fake news may even spread faster and more broadly than real news Vosoughi et al.

Given the additive nature of the source credibility effect and the repetition effect, participants always provided the highest truth judgments for congruently repeated information presented by credible sources. It is of particular interest that in Experiment 4, in which all statements appeared with source information, the truth effect for congruently repeated statements was even stronger than the source credibility effect.

Moreover, headlines that were incongruent with previously read information produced an illusion of falseness within each of the source conditions; these headlines were rated as less true than new headlines.

People tend to believe information they have encountered before and to distrust information that is inconsistent with previously encountered information. Hence, whatever information comes first has a higher chance of being believed. In contexts in which sources are particularly distinctive Experiments 2 and 3 , however, source credibility seems to have a stronger impact on judged truth than repetition does.

In line with Rapp and Salovich , we suggest that this could be promoted by specific educational programs that train people how to distinguish trustworthy from untrustworthy sources and how to use source cues to evaluate und compartmentalize information.

Of course, the accreditation as a trusted source would have to come from acknowledged and politically independent institutions. It remains to be seen, however, whether the source-tagging strategy we propose is a more promising approach to combat fake news on social media than tagging of individual postings see Pennycook et al.

Although our experiments provide clear evidence that source credibility and repeated exposure have strong, additive effects on judged truth, follow-up studies are warranted. Whitehead identified expertise, dynamism, and trustworthiness as factors impacting credibility [4].

In Web-based communication, the closest link to dynamism would be factors such as design and aesthetics. They were able to identify key factors that impact credibility. Those factors include: age, gender, country of origin, and education and income levels. Robins and Holmes explored the possible link between the concepts of aesthetics and credibility [7]. Their research focused on how users perceived two types i. Low aesthetic treatment is where content is simply placed on a Web site without professional graphic design.

There may be some crudely implemented elements and layout. A high aesthetic treatment presents a professional look and feel. Color and graphics are used to build brand and concept and to enhance communication. In general, results were consistent with the hypothesis that high aesthetic treatment would produce high judgments of credibility. A number of studies have shown that the criteria used in evaluating Web site information is similar to those used in the print environment Brandt, ; McMurdo, [8, 9].

Fonts can also be considered part of the design elements of a Web site and are often credited with creating first impressions. Most empirical research focuses on font and legibility or readability but not on the perceived personality of fonts Shaikh, et al. In , The Stanford Persuasive Technology Lab analyzed comments from 2, Internet users regarding factors that promote credibility of a Web site. Visual design included: layout, typography, font size, and color schemes.

Research has shown that the typeface i. Shaikh, et al. Users consistently attributed personalities to fonts. Serif e. Shaikh manipulated the Web page fonts and presented them in either a neutral or inappropriate typeface [11].

Typeface appropriateness was determined by the author in a previous study using a paired comparison methodology. Appropriateness was categorized as high i.

The Web pages presented in either a neutral or inappropriate typeface resulted in lower ratings for trust, professionalism, and believability. Even a typeface that was seen as neutral resulted in decreased trust, professionalism, and credibility. Research results regarding credibility as it is related gender differences between men and women are inconsistent. Burkhart found that women were considered better writers, more accurate, more trustworthy, more credible and intelligent then their male counterparts [12].

Sweger lists the author of the information being evaluated as a potential target of credibility attributions [16]. This research suggests that a salient factor in assessing credibility of web site presenters may be the gender of the page author.

Gender may serve as a significant cue to users that signals a common group identity and promotes behavior consistent with that identification. Results suggest that gender is a significant factor when looking at perceived Web site credibility measures. Overall, the study shows that gender differences are meaningful in the Web just as they are in other venues.

Males rated both the message credibility and site credibility significantly higher than women, i. There is some evidence to suggest that Web users are becoming more skeptical of online information. The first credibility clues are perceived very quickly and are critical to continued use of the Web site. Designers face increasing pressure to enhance the credibility of Web sites. Skip to main content.