Behavioral Science Dictionary

Fear appeal

Also known as: Fear-arousing communication

Social Influence

Scare people about a threat, then show them a way out.

What it means

A persuasive message that arouses fear by stressing the severity of a threat and the audience's vulnerability to it, in order to motivate protective behavior. On average such messages produce a small-to-moderate gain in attitudes, intentions and behavior, and those that also state that the recommended action works are stronger still, while the one synthesis to separate the two kinds of efficacy information found that self-efficacy statements added nothing reliable. The extended parallel process model predicts that high threat without efficacy diverts people into fear control - denial and avoidance - rather than the danger control that produces action, but the largest meta-analyses find outright backfire rare and the predicted interaction only partly supported. The handful of studies built to test that interaction do recover it, on a base their authors called thin. A pooled average cannot show what happened to the least confident recipients, who are often the target of public-health, safety and security campaigns.

The original demonstration

In 1953 Irving Janis and Seymour Feshbach gave the entering class of a Connecticut high school a lecture on dental hygiene. About two hundred students were split into four groups of fifty: three heard illustrated slide lectures differing only in how frightening they were, the fourth a talk on an unrelated subject. The strong version showed photographs of diseased gums and badly decayed teeth; the moderate version softened the imagery; the minimal version kept the slides but substituted X-rays, diagrams and healthy teeth, and described the consequences of decay in impersonal terms rather than omitting them. Questionnaires a week before and a week after recorded what changed. The strong version drew the most reported worry, so the manipulation worked on feeling. It worked least on conduct. Net change in self-reported conformity to the lecture's dental-care recommendations ran at about 36 percent in the minimal condition, 22 percent in the moderate and 8 percent in the strong. That reversal was repeated in textbooks for decades as a rule of thumb: fear defeats itself past some optimum. Repetition is not replication. It was one experiment, self-reported at a single follow-up, and no meta-analysis since 1982 has recovered the inverted U.

How it works

Three generations of theory tried to explain why arousing fear might fail to produce protection. Howard Leventhal proposed in 1970 that a threatening message sets off two reactions in parallel: an effort to control the danger, directed at the world, and an effort to control the fear, directed at the feeling. Only the first yields protective action. Ronald Rogers gave the idea a cognitive structure in protection motivation theory, though the version usually taught is not the one first published. The 1975 statement had three components: how bad the harm would be, how likely it was, and whether the recommended response would avert it - response efficacy. The revision Rogers published with James Maddux in 1983 added self-efficacy, whether the recipient believes they can carry the action out, and regrouped the components into threat appraisal and coping appraisal. Kim Witte's extended parallel process model made the relation conditional rather than additive. Threat sets how strongly a person responds at all; efficacy sets which way they go. Past a threshold, high efficacy channels the arousal into danger control - accepting the message and acting on it - and low efficacy into fear control: denial, avoidance, dismissing the source as alarmist. Because these are alternatives, raising threat without a usable response does not merely waste effort but makes matters worse. That last claim is the one most often quoted in practice, and the one carrying the least evidence.

What the evidence shows

Kim Witte and Mike Allen's 2000 meta-analysis is the most cited of several, not the first. Sutton in 1982, Boster and Mongeau in 1984 and Mongeau in 1998 had pooled the literature before them, and none of them found the inverted U either. Witte and Allen confirmed it, reporting a linear relation in which stronger appeals were more persuasive. What was new was considering threat and efficacy together: most behaviour change under strong threat with high efficacy, most defensive responding under strong threat with low efficacy - though from comparing separate subsets of studies, not a direct test of the interaction. Gjalt-Jorn Peters, Robert Ruiter and Gerjo Kok argued in 2013 that most of the literature could not test the model, since a proper test manipulates threat and efficacy independently and measures behaviour. That filter left eight studies for the main effects and six for the interaction. Within those six the crossover appeared: threat helped only when efficacy was high (d = 0.31), efficacy only when threat was high (d = 0.71). Under low efficacy the effect of threat was negative and borderline (d = -0.31, p = .06), which they read as a sign that threat can depress behaviour when people cannot act. They were candid that six studies is a thin base for a widely taught rule and that the analysis may be underpowered; their own funnel plots also showed a publication-bias signal on the efficacy main effect, the larger of the two. Larger syntheses push the other way. Paschal Sheeran, Peter Harris and Tracy Epton pooled experiments that had successfully raised risk appraisal - a broader category than fear appeals, taking in perceived severity as well - and found about d = .31 on intention across 217 tests and d = .23 on behaviour across 93. Melanie Tannenbaum and colleagues assembled 127 articles, 248 samples and 27,372 participants for an average of d = 0.29, with very few conditions under which fear appeals were ineffective and none under which they backfired. A 2022 meta-analysis by Elisabeth Bigsby and Dolores Albarracín narrowed that question. Across 158 effect sizes and 19,736 participants, fear appeals moved intentions and behaviour more when they carried positive response-efficacy information - a statement that the recommended action works - while self-efficacy information made no reliable difference. Kok and colleagues replied in 2018 that a positive average is compatible with real harm to subgroups who are often the intended audience. The defensible summary is a small-to-moderate average benefit, improved by showing the recommended action works, with routine boomerang unsupported in aggregate - though the one estimate pointing that way came from the studies built to detect it.

Where it shows up

Public health remains the main arena - pictorial warnings on tobacco packaging, drink-driving films for school leavers, sun-protection campaigns, screening reminders. Commercial versions are more numerous: a home-security vendor recites break-in rates for a postcode before quoting a fee; an antivirus renewal notice counts the threats it blocked last month. Corporate security-awareness training is the best-studied non-health case. A 2023 meta-analysis pooling several hundred security studies found efficacy a stronger predictor of protective motivation than threat, and fear itself more important than threat perception. On that evidence the work is done by the part that shows staff what to do, not the reconstructed breach that opens it. Advocacy has drifted the other way. Messaging that details an unliveable climate future while offering nothing an individual can plausibly do supplies threat without a response, the configuration the models warn about, and the one least tested. Where a seller manufactures the vulnerability it then charges to remove, the appeal shades into a dark pattern whether or not it moves the metric.

Limits and caveats

An average effect is not a safety guarantee. The people least confident of their own capability - long-term smokers, patients already carrying a diagnosis - are both the likeliest to respond defensively and the usual target, and a pooled estimate cannot say what happened to them. The construct is looser than it looks. Most studies labelled fear appeals manipulate threatening content without verifying that fear was aroused, so what is measured is often severity and susceptibility rather than emotion. Outcomes skew toward intentions and short-horizon self-report; behavioural effects are smaller, and there is little evidence on anything sustained across months. Habituation to the same graphic material is assumed more than tested. Effectiveness and legitimacy are separate questions. Fear is an unpleasant state imposed without consent, and campaigns that frighten people about a condition often stigmatise the people who have it. A message can move the outcome measure and still be the wrong thing to send.

Examples

A graphic anti-speeding ad changes driving only if it also convinces viewers that slowing down is easy and effective.

A health agency sends two versions of a screening reminder. One describes late-stage outcomes in graphic detail and stops there; the other pairs the same imagery with a pre-booked appointment slot and a number to call. Both lift attendance over sending nothing, but the second lifts it substantially more.

A staff cybersecurity module opens with a reconstruction of a ransomware incident, then closes without showing what a suspicious message looks like or who to forward it to. Employees afterwards rate the risk as higher and reporting ticks up slightly, but nothing like the sister module that ends with a one-click report button and a screenshot of a real lure.

An insurer's direct-mail piece leads with the cost of an uninsured house fire, and the action it asks for is a single cheap step - a quote that takes five minutes. It performs markedly better than the same insurer's campaign urging households to change their fire-safety habits week after week.

A road-safety film shown each year to the same cohort of school leavers loses measurable effect by its third showing. The footage has not changed, the audience has seen it before, and the shock no longer registers.

First described in Hovland, Janis & Kelley (1953); Witte's EPPM (1992).

Key references

  1. Janis, I. L., & Feshbach, S. (1953). Effects of fear-arousing communications. The Journal of Abnormal and Social Psychology, 48(1), 78-92. doi.org/10.1037/h0060732
  2. Leventhal, H. (1970). Findings and theory in the study of fear communications. Advances in Experimental Social Psychology, 5, 119-186. doi.org/10.1016/S0065-2601(08)60091-X
  3. Rogers, R. W. (1975). A protection motivation theory of fear appeals and attitude change. The Journal of Psychology, 91(1), 93-114. doi.org/10.1080/00223980.1975.9915803
  4. Maddux, J. E., & Rogers, R. W. (1983). Protection motivation and self-efficacy: A revised theory of fear appeals and attitude change. Journal of Experimental Social Psychology, 19(5), 469-479. doi.org/10.1016/0022-1031(83)90023-9
  5. Witte, K. (1992). Putting the fear back into fear appeals: The extended parallel process model. Communication Monographs, 59(4), 329-349. doi.org/10.1080/03637759209376276
  6. Witte, K., & Allen, M. (2000). A meta-analysis of fear appeals: Implications for effective public health campaigns. Health Education & Behavior, 27(5), 591-615. doi.org/10.1177/109019810002700506
  7. Peters, G.-J. Y., Ruiter, R. A. C., & Kok, G. (2013). Threatening communication: A critical re-analysis and a revised meta-analytic test of fear appeal theory. Health Psychology Review, 7(sup1), S8-S31. doi.org/10.1080/17437199.2012.703527
  8. Sheeran, P., Harris, P. R., & Epton, T. (2014). Does heightening risk appraisals change people's intentions and behavior? A meta-analysis of experimental studies. Psychological Bulletin, 140(2), 511-543. doi.org/10.1037/a0033065
  9. Tannenbaum, M. B., Hepler, J., Zimmerman, R. S., Saul, L., Jacobs, S., Wilson, K., & Albarracín, D. (2015). Appealing to fear: A meta-analysis of fear appeal effectiveness and theories. Psychological Bulletin, 141(6), 1178-1204. doi.org/10.1037/a0039729
  10. Kok, G., Peters, G.-J. Y., Kessels, L. T. E., ten Hoor, G. A., & Ruiter, R. A. C. (2018). Ignoring theory and misinterpreting evidence: The false belief in fear appeals. Health Psychology Review, 12(2), 111-125. doi.org/10.1080/17437199.2017.1415767
  11. Bigsby, E., & Albarracín, D. (2022). Self- and response efficacy information in fear appeals: A meta-analysis. Journal of Communication, 72(2), 241-263. doi.org/10.1093/joc/jqab048
  12. Lowry, P. B., Moody, G. D., Parameswaran, S., & Brown, N. J. (2023). Examining the differential effectiveness of fear appeals in information security management using two-stage meta-analysis. Journal of Management Information Systems, 40(4), 1099-1138. doi.org/10.1080/07421222.2023.2267318

Related terms

← All 1001 terms