Where we are
You have a sentence and three supports from session 1, and an opening and a last line from session 2. Noor Haddad, our constructed team lead, has the same, and in four weeks she walks on at her first company all-hands. Her talk is ready on paper. What she dreads is the two minutes before it. Tonight is about that body, honestly: what the evidence says, where it disagrees with itself, and what you can do that does not depend on any single study being right. A short self-study card, "Nerves, honestly", in Communication, session 5 ↗ covers the basics for meetings; this page goes deeper into the evidence.
★ What you walk out with tonight. A written pre-talk routine for your own talk: the physical signs you get, a reappraisal sentence in your own words with an honest reason to use it, your first sentence rehearsed to automatic, a ten-minute checklist, a plan for the moment you blank, and one line to write afterwards. You also leave knowing which part of the famous "get excited" finding held up when it was repeated, and which part did not.
A racing heart is arousal, and arousal has two names
5 min liveBefore a talk the body is in high arousal: heart up, breathing up, attention narrowed. Anxiety is one name for that state. Excitement is another. The psychologist Alison Wood Brooks argued in 2014 that the two are close relatives, both high in arousal and different mainly in how they feel, and that the standard advice asks for the hardest move of all.
LiveTwo states, one level of arousal3 min▶
Brooks's paper, "Get excited: Reappraising pre-performance anxiety as excitement" (2014), rests on one observation. Anxiety and excitement both come with high arousal: the racing heart is there in both. They differ in whether the coming moment feels like a threat or an opportunity. She calls the two states "arousal congruent".
From that she draws the argument in the figure. "Calm down" asks you to travel from high arousal to low arousal in the minutes before you stand up, which is the direction your body is least likely to go while a room is waiting. "I am excited" asks you to keep the arousal and change its label, a shorter move.
Treat this as Brooks's argument, carefully made and tested, not as settled physiology. The figure is a picture of her reasoning, and one of her own measurements is a useful reminder of its limit: in her third study, a maths task, heart rate did not differ between the conditions. Saying "I am excited" changed the label people gave the state; it did not slow the pulse.
Self-studyWhy almost everyone gives the opposite advice2 min▶
Before her main studies Brooks ran a pilot on Amazon Mechanical Turk with 300 people. The scenario will sound familiar: tomorrow you, or a coworker, give a thirty-minute keynote to a whole company of about five hundred people, CEO included, and feel extremely anxious. Of the 153 asked about themselves, 90.97% chose "Try to relax and calm down" as the best advice, and 7.74% chose "Try to be excited instead of anxious". The pattern was the same when people were asked about a coworker.
That is the starting point the paper argues against, and it is a fair description of most well-meant advice before a talk. Notice what the pilot measured: what people believe helps, not what helps. It tells you the advice is near universal. It does not tell you the advice fails; that is what the studies in Part 2 try to test.
What none of this covers: clinical speech anxiety. If fear stops you speaking at all, that is worth a professional's help, and nothing on this page is a substitute for it.
The advice you were given before your first presentation. Think of the last time a colleague saw you were nervous before a meeting or a pitch. The likely words were "relax", "breathe", "calm down". In Brooks's pilot, asked about their own all-company keynote, 90.97% of people picked it as the best advice. It is kind advice, and Brooks's point is only that it asks for a bigger change than the alternative. Whether the alternative makes the talk better is a separate question, and Part 2 is where it gets hard.
Get excited: what one study found, and what happened when it was repeated
7 min liveThe "say I am excited" tip travels on one study about public speaking. It is a good study, and it was repeated in 2025 by researchers who copied its method. Part of the finding came back. Part of it did not. Both halves matter, and the honest position is to hold them side by side.
LiveBrooks 2014, Study 2, in full3 min▶
Who and what. 140 participants were given two minutes to prepare a speech on why they would be a good work partner, then delivered it on camera, told it would be judged by a committee of their peers. Just before speaking, each said one sentence aloud: "I am excited" or "I am calm".
What others saw. Three raters who did not know which sentence a speaker had said scored the videos. Speakers who said "I am excited" were rated more persuasive (d = .681), more competent (d = .458), more confident (d = .705) and more persistent (d = .505). They also spoke for longer: 167 seconds against 132.
What the speakers felt. The excited group reported more excitement (4.75 against 4.09). In Brooks's words, "Self-reported anxiety did not differ significantly across conditions" (5.06 against 5.27). That is the detail most retellings drop: the sentence did not reduce anxiety. People still felt as anxious; they also felt more excited.
Its scope. A two-minute prepared speech, in a study setting, judged by a committee they were told about, with ratings from three blind raters. The paper's other studies cover karaoke singing and a maths test, with a fourth on threat versus opportunity mind-sets. It is a strong paper, and it is one paper.
LiveThe 2025 replication, and what it did not find3 min▶
Poynter and Pasqualini repeated Study 2 directly, copying its method, and published it in 2025 in Psychology of Language and Communication. The study was preregistered on OSF, per the authors. They report finding no earlier published attempt to replicate it.
What came back. The self-report finding reproduced: speakers who said "I am excited" felt more excited, and their anxiety was unchanged, just as in 2014.
What did not. The finding that matters most to a speaker, that other people rate the talk as better, did not. In the authors' words: "None of the observer ratings differed significantly between conditions."
Its limits, which are real. 74 people gave self-reports and 71 were rated by observers. Data collection stopped early because of the pandemic. Most participants were women, and they took part for course credit rather than pay. The authors estimate their sample had between .60 and .90 power to detect effects the size of Brooks's, so a real but smaller effect could have been missed. One small, early-stopped replication does not overturn a finding. It does mean the performance claim now rests on one study that found it and one that did not.
Self-studyThe wider literature: a small effect, and signs of bias3 min▶
Bosshard and Gomez (2024), Scientific Reports. A meta-analysis of stress-arousal reappraisal and stress-mindset interventions: 35 studies, 44 effects, 5,638 participants. The pooled effect was small, d = 0.23. After trim-and-fill, a correction for studies that may be missing from the published record, it fell to d = 0.14 (confidence interval .04 to .24). Egger's test for funnel asymmetry was significant (p = .042), a sign of publication bias. Two cautions: this is a broader family of interventions than Brooks's one-sentence self-talk, and it could not be confirmed that Brooks's paper is one of the included studies.
Ganley and colleagues (2021), Journal of Numerical Cognition. Reported from the abstract only, as the full text was not read. 300 participants tried four brief interventions before a maths test, one of them reappraising anxiety as excitement. The abstract reports that none had an effect on state anxiety or on performance. A different task, not a speech, so it is a conceptual test rather than a replication.
The honest reading. Three claims, three different standings. Feeling more excited while anxiety stays the same: reproduced. Giving a talk that others rate as better: one study found it, one small direct replication did not. Reappraisal in general: a small effect that shrinks when you allow for missing studies. Nobody has settled this, and this page does not either.
The replication is the real-world lesson. For eleven years "say I am excited" circulated as a performance tip on the strength of one public-speaking study, and the authors of the 2025 replication found no earlier published attempt to repeat it. When someone did, the part about how speakers feel came back and the part about how audiences rate them did not. Whenever a nerves tip comes with "studies show", the first useful question is how many studies, and whether anyone has repeated the one it rests on.
Fear above death, traced
3 min live"People fear public speaking more than death" is said at the start of half the talks about talks. It is meant kindly: you are not alone. Followed back to where it starts, the line rests on one market-research survey from 1973 whose original report could not be read, and current data tell a more interesting story.
LiveWhere the line starts2 min▶
The usual source is a 1973 survey by the market-research firm Bruskin Associates, reported in the Sunday Times (London) on 7 October 1973 under the headline "What people usually fear". As reported by later researchers, 41% listed "speaking before a group". The original survey report could not be read for this page.
One detail decides what the number means, and it could not be settled: whether respondents ticked as many items as they liked from a list, or named their single greatest fear. Secondary accounts disagree. If it was a list, "41% listed speaking" and "people fear speaking more than death" are very different claims. The ranking was then reported in The Book of Lists (1977), which was not opened here either, and has been repeated as fact ever since.
Self-studyWhat current data, read at source, show3 min▶
Chapman University Survey of American Fears, 2024 (Wave 10). 1,008 adults on the SSRS web panel, fielded 21 March to 1 April 2024. The share saying they were afraid or very afraid: public speaking 29.0%, dying 31.6%, people they love dying 57.8%. Public speaking and dying are close, with dying slightly ahead, and fear for the people you love is far above both.
Dwyer and Davidson (2021), Journal of Education and Educational Development. 374 students. Before a public speaking course, the fear of public speaking and the fear of death "were not significantly different". After the course, death was chosen more often (170 against 129). A course in speaking moved the ranking.
The lesson. The fear is common and real: nearly three in ten adults in the 2024 survey. The ranking against death depends on how the question is asked, and in one study it shifted after people practised. You do not need the dramatic version to be taken seriously, and the less dramatic version comes with a hopeful detail.
An opening line worth retiring. A familiar scene at an internal training day: the facilitator opens with "you know people fear this more than death", the room laughs, and the line sets up speaking as something terrifying. The 2024 Chapman numbers give a truer opening, and a kinder one: about 29% of adults say they are afraid of public speaking, so in a room of twenty, several people feel what you feel, and in one study a speaking course changed how students ranked it.
Your pre-talk routine, written down
★ 22 min · everyone buildsSix steps that turn tonight into something you can do on the day. Each produces a line or two on paper, and Noor's constructed versions sit alongside as the worked example. Nothing here depends on the reappraisal finding being right: it is one line in the routine, and the rest is preparation you control.
Name the physical signs you get. Write down what your body does before you speak, in plain words, in the order it usually happens. Naming them in advance turns "something is wrong" into "this is the thing I expected".
Write the reappraisal sentence in your own words, and the honest reason you might use it. Brooks used "I am excited". Yours can be different, as long as it keeps the arousal and changes the label. Then write one line on why: it costs nothing and may make the state feel more like excitement; the evidence that it makes the talk better is mixed.
Rehearse the first sentence of your talk to automatic. Take the opening you wrote in session 2. Say it out loud until you could say it while thinking about something else. When everything else wobbles, this is the sentence that gets you moving.
Write your ten-minute checklist. Four items at minimum: water, the notes card (throughline, three supports, the ask), arrive early, and stand where you will speak before anyone else is in the room. Put each against a time.
Plan what you do when you blank. Blanking is common and recoverable, and the recovery is mechanical: pause (it looks deliberate from the room), look at the card, say the throughline sentence from session 1 again, then go to the next support. The card is your map back.
Write a one-line plan for afterwards. Within an hour of speaking, write down what actually happened next to what you feared would happen. Over several talks this record is the most useful evidence you will have about your own nerves.
One line, said plainly. If fear stops you speaking at all, not just makes it uncomfortable but stops it, that is worth a professional's help. A routine is for nerves you can carry; it is not a treatment.
Before session 4
30 min- Run the routine once, for real, before your next meeting where you speak, even a short one. Afterwards write the one line from step 6: what you feared, what happened.
- Say your first sentence out loud ten times on two separate days, standing up. Note whether it still feels automatic on the second day.
- Find one nerves tip that says "studies show" in an article, a post or a training. Write down what it claims, and whether it names the study, how many people were in it, and whether anyone has repeated it.
- Find the recording from session 1 (your sentence and three supports, under two minutes). Session 4 measures it, so bring it.
Three questions
Clear all three and session 3 gets stamped in your passport.
1 · In Brooks's 2014 Study 2, what happened to the self-reports of speakers who said "I am excited" before their speech?
Excitement went up (4.75 against 4.09) while anxiety stayed about the same (5.06 against 5.27). The sentence added excitement; it did not remove anxiety. Heart rate was measured in Study 3, the maths task, and it did not differ by condition.
2 · What did the 2025 direct replication by Poynter and Pasqualini find for observer ratings, and what is the honest reading?
The self-report half reproduced; the observer-rated half did not. With 71 rated speakers and an early stop, one replication cannot settle it, and the wider meta-analysis shows a small effect with signs of publication bias. Contested is the accurate word.
3 · Where does "people fear public speaking more than death" come from, and what does the 2024 Chapman survey show?
The line traces to a 1973 survey known only through later reports, and whether people picked from a list or named one fear could not be settled. Chapman 2024 puts the two close, with dying slightly ahead, and fear of people you love dying at 57.8%.
What this session teaches, and where it came from
Every fact on this page, and its verification tier, is recorded in the course's source map.
Session 3 cheat sheet · pin this
Session 4 is the pace bench. You will measure how fast you actually speak, how long your sentences run and how often you pause, against 2,299 TED talks, and see what speaking faster does to a talk when nerves push you to rush. Bring the recording from session 1.