Forensic Research Archive of Manipulation and Exploitation
Illustration: a redacted Cold War intelligence-dossier collage centred on a world map of influence operations, with the U.S. Capitol among the evidence.

Techniques · Classical

Testimonial

Illustration, not an archive object

Also called Appeal to authority, Celebrity endorsement

A named person or institution vouches; the vouch is meant to end checking.

The tell

Celebrity, clergy, or 'experts say' without the underlying data.

Not this

A primary witness describing what they saw.

History and effect

IPA 1937; grew out of print advertising's use of celebrity endorsement and was carried wholesale into political and wartime messaging.

Trust in a named person substitutes for verification of the claim, shortcutting the audience's own scrutiny.

Lineage: IPA 1937

How it works

A testimonial pairs a claim with a name. The sender finds a person or institution the audience already trusts or admires (a general, a doctor, a film star, a scientist, a church, a famous visitor) and attaches their approval to a product, policy, war effort or government. The claim itself may be vague or never stated in detail. The endorsement is the argument.

The mechanics turn on a transfer of trust across a boundary. The endorser's reputation was earned in one field, and the testimonial spends it in another: a general vouching for supply work, a physician's white coat selling a brand of cigarette, an actor's fame recommending a candidate, a novelist praising a foreign country's economy after a guided tour. The audience is invited not to ask whether the endorser knows anything about the question, or whether the endorser was paid, briefed or selected, and not to ask for the data the endorser saw, if any.

Several features make a testimonial more effective. The endorser is named and recognizable. The link between the endorser and the sender (payment, employment, political alliance, a hosted trip) is hidden or not mentioned. Only favorable endorsers are shown, and the ones who declined or disagreed are left out. In the institutional form, a vague appeal ("experts agree", "doctors recommend", "scientists say") stands in for a named person, which removes even the possibility of checking who said what.

The modern variant removes the endorser's consent entirely. Cloned voices and recut video can make a real, trusted person appear to endorse something they never said. Matryoshka's midterm clips: borrowed faces and a borrowed logo documents clips of real actors recut with cloned audio to attack U.S. politicians; the actors had said nothing of the kind. Here testimonial overlaps with Synthetic media.

Why it works

The foundational experiment is Hovland and Weiss (1951). They gave people identical articles attributed either to a high-credibility source or a low-credibility one, and found more immediate opinion change with the high-credibility source, even though the content was the same. They also found that the difference shrank over the following weeks, a pattern that became central to later research on the "sleeper effect". A later meta-analysis by Kumkale and Albarracín (2004) found that the sleeper effect appears under specific conditions (for example, when the message is strong and the discounting cue comes after it) and is not a general rule. FRAME has a separate technique page on Sleeper effect.

Petty, Cacioppo and Goldman (1981) showed when a source's prestige matters most. Students read a proposal for a new comprehensive exam policy attributed either to a commission chaired by a Princeton professor or to a local high school class. When the policy would not affect them personally, the prestigious source increased agreement regardless of whether the arguments were strong. When it would affect them, argument quality mattered more than the source. This is the core prediction of the Elaboration Likelihood Model (Petty and Cacioppo 1986): endorsements work best on audiences who are not motivated or able to examine the claim.

Obedience research is often cited in discussions of authority. Stanley Milgram's experiments (1974) found high rates of compliance with an authority figure's instructions. The findings have been partly replicated under modified, safer conditions (Burger 2009), but archival work has raised questions about how the experiments were run and reported (Perry 2013), and researchers disagree on how far they generalize. They are best treated as evidence that authority cues are powerful in some settings, not as a precise measure. Cialdini (2001) lists authority as a core compliance principle and notes how often its symbols (titles, uniforms, credentials) do the work on their own.

Celebrity endorsement has its own account. McCracken (1989) argued that celebrities carry cultural meanings from their roles and public lives, and that endorsement moves those meanings onto the product. In politics, Garthwaite and Moore (2013) estimated that Oprah Winfrey's 2008 endorsement of Barack Obama gained him votes in the Democratic primary; the size of such effects is debated, and most research finds celebrity endorsements matter less in politics than in advertising.

Examples across eras and sides

  • "Pershing says" (United States, 1918). An American Expeditionary Forces poster, Pershing Says: The S.O.S. Shares the Success With Us, quotes General John J. Pershing thanking the Services of Supply for its prompt response to every demand. The commanding general's name is the whole message; there is no account of what the supply service delivered (Sources: NARA 1918).
  • Western visitors to the Soviet Union (1930s). Soviet authorities hosted prominent Western writers and intellectuals on guided visits, and the favorable accounts some of them published abroad were used as testimony for Soviet achievements. Paul Hollander documented the same pattern of hosted "political pilgrims" in later visits to China and Cuba. The visitors were real and often sincere, but they saw what their hosts arranged (Sources: Hollander 1981).
  • "More doctors smoke Camels" (United States, R. J. Reynolds, 1940s). Cigarette advertising from the 1930s to the early 1950s used images of physicians, and R. J. Reynolds's Camel campaign claimed that more doctors smoked Camels than any other brand, based on surveys the company's agency arranged. The physician stood in for any claim about health that the ads did not make directly. The same industry later ran A Frank Statement to Cigarette Smokers, which leaned on unnamed "authorities" in a similar way (Sources: Gardner and Brandt 2006).
  • The Pentagon's military analysts (United States, 2002 to 2008). The New York Times reported that the Department of Defense cultivated retired military officers who appeared on television as independent analysts, giving them briefings and access while many had business ties to defense contractors. Viewers saw experienced officers; they were not told about the briefings or the business ties (Sources: Barstow 2008).
  • Celebrity endorsements in both parties (United States, 2007 to 2008). In the same election cycle, Oprah Winfrey campaigned for Barack Obama in the Democratic primary and the actor Chuck Norris appeared in a campaign ad for Mike Huckabee in the Republican primary. Celebrity endorsement is legal and usually disclosed; it fits the technique when fame substitutes for any argument on the issues. Research on the Winfrey endorsement suggests a measurable effect, though the estimate is contested (Sources: Garthwaite and Moore 2013).
  • Fabricated endorsers (United States and Russia-aligned operations, 2024 to 2026). Before the January 2024 New Hampshire primary, robocalls used an AI-generated imitation of President Joe Biden's voice to tell Democrats not to vote; the Federal Communications Commission later fined the political consultant responsible. The Matryoshka network recut real celebrity videos with cloned audio to attack Democratic candidates (Matryoshka's midterm clips: borrowed faces and a borrowed logo). In both cases the "endorser" never said the words (Sources: FCC 2024).

How to spot it

  • The main reason offered to believe the claim is who said it, not what they showed.
  • The endorser's expertise is in a different field from the claim (an actor on medicine, a general on economics, a novelist on agriculture).
  • The endorser's relationship to the sender (payment, employment, a hosted trip, a party role) is missing or in small print.
  • Vague sources stand in for names: "experts say", "doctors recommend", "scientists agree", with no study or person you can check.
  • Only supporters are quoted, and you cannot tell whether other qualified people disagree.
  • A clip or quote of a famous person endorsing something surprising has no original source, no date and no matching statement on the person's own channels.

Where it ends: edge cases and legitimate persuasion

Named variants include the celebrity endorsement, the expert or credential endorsement ("four out of five doctors"), the institutional seal, the ordinary-person testimonial (which overlaps with Plain folks), the negative testimonial (showing that a disliked figure supports the other side, a form of Transfer), and the fabricated endorsement (Synthetic media). When many endorsers are stacked together to suggest everyone agrees, the testimonial shades into Bandwagon or Manufactured consensus. When an endorsement is routed through an apparently independent outlet to hide its origin, it overlaps with Narrative laundering.

Not every appeal to a source is propaganda. A primary witness describing what they saw is evidence. An expert explaining a finding while showing the data, the method and the uncertainty is giving an argument, not a testimonial. Relying on a qualified, independent authority is often reasonable when you cannot check a question yourself; everyone does it with doctors and engineers. U.S. law also draws a line: the Federal Trade Commission's endorsement guides require advertisers to disclose material connections, such as payment, between an endorser and the company (FTC 2023).

The test a reader can apply has three questions. Is the endorser speaking within their own competence? Is the relationship between the endorser and the sender disclosed? And would the claim still stand, with its evidence visible, if the name were removed? If the answer to the last question is no, the name was doing the work.

Sources

  • Barstow, David. 2008. "Behind TV Analysts, Pentagon's Hidden Hand." The New York Times, 20 April 2008.
  • Burger, Jerry M. 2009. "Replicating Milgram: Would People Still Obey Today?" American Psychologist 64(1): 1–11.
  • Cialdini, Robert B. 2001. Influence: Science and Practice. 4th ed. Boston: Allyn & Bacon.
  • Federal Communications Commission (FCC). 2024. Forfeiture order against political consultant Steve Kramer concerning AI-generated robocalls before the New Hampshire presidential primary. Washington, DC: FCC.
  • Federal Trade Commission (FTC). 2023. Guides Concerning the Use of Endorsements and Testimonials in Advertising. 16 C.F.R. Part 255.
  • Gardner, Martha N., and Allan M. Brandt. 2006. "'The Doctors' Choice Is America's Choice': The Physician in US Cigarette Advertisements, 1930–1953." American Journal of Public Health 96(2): 222–232.
  • Garthwaite, Craig, and Timothy J. Moore. 2013. "Can Celebrity Endorsements Affect Political Outcomes? Evidence from the 2008 US Democratic Presidential Primary." Journal of Law, Economics, and Organization 29(2): 355–384.
  • Hollander, Paul. 1981. Political Pilgrims: Travels of Western Intellectuals to the Soviet Union, China, and Cuba, 1928–1978. New York: Oxford University Press.
  • Hovland, Carl I., and Walter Weiss. 1951. "The Influence of Source Credibility on Communication Effectiveness." Public Opinion Quarterly 15(4): 635–650.
  • Kumkale, G. Tarcan, and Dolores Albarracín. 2004. "The Sleeper Effect in Persuasion: A Meta-Analytic Review." Psychological Bulletin 130(1): 143–172.
  • McCracken, Grant. 1989. "Who Is the Celebrity Endorser? Cultural Foundations of the Endorsement Process." Journal of Consumer Research 16(3): 310–321.
  • Milgram, Stanley. 1974. Obedience to Authority: An Experimental View. New York: Harper & Row.
  • National Archives and Records Administration (NARA). 1918. "Pershing says, 'I want S.O.S. to know how the First Army appreciates the prompt response made to every demand...'" Poster, National Archives Identifier 512615.
  • Perry, Gina. 2013. Behind the Shock Machine: The Untold Story of the Notorious Milgram Psychology Experiments. New York: The New Press.
  • Petty, Richard E., and John T. Cacioppo. 1986. Communication and Persuasion: Central and Peripheral Routes to Attitude Change. New York: Springer-Verlag.
  • Petty, Richard E., John T. Cacioppo, and Rachel Goldman. 1981. "Personal Involvement as a Determinant of Argument-Based Persuasion." Journal of Personality and Social Psychology 41(5): 847–855.

Case studies

In the collection

All 18 testimonial examples
Australia's imperishable record. Australians! The empire needs you (Farmer's, Sydney)Australia's imperishable record. Australians! The empire needs you (Farmer's, Sydney)1915 · Testimonial, Glittering generalitiesMr. John Dillon, M.P., on recruiting in Ireland. Join an Irish regiment to-dayMr. John Dillon, M.P., on recruiting in Ireland. Join an Irish regiment to-day1915 · Testimonial, Semantic hijackingReinforcements. Our scrap of paper, honor it. Vote Yes. Referendum (Australia, 1917)Reinforcements. Our scrap of paper, honor it. Vote Yes. Referendum (Australia, 1917)1917 · Testimonial, TransferIs Ireland a part of England? Vote for Sinn Fein (Sinn Féin publicity sheet, 1918 general election)Is Ireland a part of England? Vote for Sinn Fein (Sinn Féin publicity sheet, 1918 general election)1918 · TestimonialAmerica When FeminizedAmerica When Feminized1920 · Fear appeal, TestimonialAttention American Soldiers! I Cease Resistance (surrender pass, leaflet 17-J-1)Attention American Soldiers! I Cease Resistance (surrender pass, leaflet 17-J-1)1945 · Euphemism / doublespeak, TestimonialChúng tôi đã rời bỏ hàng ngũ V.C... Các bạn còn chần chờ gì nữa? (We Left the VC Ranks... What Are You Still Waiting For?)Chúng tôi đã rời bỏ hàng ngũ V.C... Các bạn còn chần chờ gì nữa? (We Left the VC Ranks... What Are You Still Waiting For?)1967 · Testimonial, BandwagonNancy Reagan holds up a 'Just Say No' sign for press photographersNancy Reagan holds up a 'Just Say No' sign for press photographersSep 6, 1986 · Testimonial

In the watch briefs

All 11 briefs

In campaigns

Related techniques

TransferPlain folksBandwagon