# How to explain your research in two minutes, with 3MT tips

**How do I explain my research simply to a non-specialist in two minutes?** To explain your research in two minutes, say the problem, what you did, what you found and why it matters, in everyday words, with one study or number the listener can hold on to. At 120 to 160 words a minute that is roughly 240 to 320 words. The Three Minute Thesis allows three minutes, one static slide and no props.

Published 2026-09-26 by EdCitation. https://edcitation.com/newsletter/how-to-talk-about-your-research-in-two-minutes

To explain your research in two minutes, give four things in order: the problem, what you did, what you found, and why it matters, put simply enough for a friend from another subject. Name one study or one number they can remember, and stop. That is the elevator pitch for research, the opening answer at a viva, and the core of a Three Minute Thesis talk.

This guide, published by EdCitation, gives a structure of our own, a worked example, the Three Minute Thesis (3MT) rules, what seven universities advise and where they disagree, two lay summary limits, and one experiment on jargon, all read on 27 September 2026. Delivery is in [how to give a presentation without reading your slides](https://edcitation.com/newsletter/how-to-give-a-presentation-without-reading-your-slides); the viva in [how to prepare for a viva or dissertation defence](https://edcitation.com/newsletter/how-to-present-your-dissertation-and-prepare-for-a-viva).

A two-minute explanation often leans on one earlier study that shows the problem is real. EdCitation's free [Find sources](https://edcitation.com/) finds it by topic or by the claim it must support, and [Cite a source](https://edcitation.com/cite) builds its reference for the slide or handout. EdCitation never writes any part of a talk or thesis, so every word you say stays yours.

## How do you explain your research in two minutes?

Use four moves, in order, each with a time. The structure below is ours, built from the university guidance cited later; Queensland's 3MT judging criteria ask for similar content, and VCU Libraries (2026) give a similar four-line skeleton.

### Our four-part structure

| Move | What it answers | Time | Words |
| --- | --- | --- | --- |
| 1. The problem | What is wrong, missing or unknown, and who it affects | 25 seconds | 50 to 65 |
| 2. What you did | Your question and how you answered it | 30 seconds | 60 to 80 |
| 3. What you found | The main result, or what you expect if unfinished | 35 seconds | 70 to 95 |
| 4. Why it matters | Who could use it, and what changes if you are right | 20 seconds | 40 to 55 |
| Close | One sentence the listener could repeat | 10 seconds | 20 to 25 |

The word counts use the 120 to 160 words a minute in our [guide to writing a speech](https://edcitation.com/newsletter/how-to-write-a-speech). Write to the low end: pauses and questions take time.

### Three rules we hold to

1. **One number, one study.** A listener keeps one figure. Choose the one that proves the problem is real, and know where it came from: [Find sources](https://edcitation.com/) finds it.
2. **Swap the term rather than define it.** The experiment below suggests a definition does not undo the effort a technical word costs.
3. **End on the listener.** "Which means your adviser could…" lands better than "and that is my thesis".

## A worked example: one invented project

The project and its result are invented for this guide, to show the structure. The one real study in it, and its numbers, come from the published trial.

Suppose a master's student has tested whether a text message the day before cuts missed appointments at a university careers service.

### The two-minute version

> Every week, students book careers appointments and do not turn up. The adviser sits in an empty room, and a student on the waiting list loses a slot. [The problem]
>
> Hospitals have tried reminders for years. A trial in Geneva with 6,450 patients found a text the day before worked about as well as a phone call, and cost less. So I asked: would the same text work for students? For one term, half the students who booked got a text 24 hours before; half got nothing. I counted who turned up. [What I did]
>
> Students who got a text missed fewer appointments, and the text cost almost nothing to send. [What I found; a real pitch gives the figure]
>
> Advisers are few, and every empty slot is one another student could have used. If the result holds, every booking system on campus could send the same message. [Why it matters]
>
> One text, sent the day before, gives students back appointments that would otherwise be wasted. [Close]

That runs to about 150 words, leaving room for your real figure and a breath between moves. It leaves out the statistical test, the literature review and the ethics approval: those are for the questions afterwards.

### The same project in 20 seconds

MIT's Communication Lab sets about 20 seconds and five parts: introduce yourself, a hook, your idea, why it matters, and an ask (Prost & Larson, n.d.). Cut down: "I'm testing whether a text the day before stops students missing careers appointments. Empty slots waste advisers' time. Early results look promising. Do you run a booking system?"

## What are the Three Minute Thesis rules?

The University of Queensland's rules allow at most three minutes, one static slide, spoken word only, and no props or other media; anyone over three minutes is disqualified (University of Queensland, n.d.-b). 3MT began at Queensland in 2008 with 160 research students, and the university says it now runs at over 900 universities in more than 85 countries (University of Queensland, n.d.-a).

### The rules and the judging

- **Time:** three minutes at most, starting when you first speak or move.
- **Slide:** one static slide from the start; no transitions, animations or movement.
- **Media and props:** no sound or video, no costumes, instruments or laboratory equipment; no poems, raps or songs.

Queensland's FAQs allow palm cards or notes, though they can weaken delivery, and a plainer title than your thesis (University of Queensland, n.d.-c). Two criteria carry equal weight (University of Queensland, n.d.-d): comprehension and content (motivation, strategy and results, conclusions and impact) and engagement and communication (language for a non-specialist, a slide that helps, enthusiasm that holds the room).

### Who can enter: universities differ

- **Queensland:** doctoral candidates past their confirmation milestone, usually about 12 months in.
- **York:** PhD candidates past their first formal progress review and before their viva (University of York, n.d.).
- **Ohio State:** master's and doctoral students, master's generally in their second year; it also tells speakers not to turn to or point at the slide (The Ohio State University Graduate School, n.d.).
- **Washington:** master's and doctoral students completing in 2025 to 2026, with a draft that includes findings (University of Washington Graduate School, n.d.).

In the UK, Vitae has hosted the national final since 2014; only winners of member universities' heats enter, via a video quarter-final (Vitae, n.d.). A master's student could enter at Ohio State but not at York, so read your own graduate school's page first.

## What do universities advise on explaining research to a non-specialist?

They agree on the core: cut jargon, tell a story with a clear point, test it on someone outside your field, and practise aloud. They differ on length, technical terms and the ending.

### Where they agree

Sussex asks you to write for your audience, tell a story with a beginning, middle and end, and imagine a friend from another field (University of Sussex, n.d.). VCU Libraries (2026) suggest plain language and analogies. Ohio State wants attention grabbed early and a slide "not text-heavy"; Washington suggests practising on friends and colleagues.

### Where they differ

- **Length.** MIT says about 20 seconds; VCU allows 30 seconds to two minutes; 3MT gives three minutes. MIT's research pitch opens a conversation, a 3MT is a whole talk, and two minutes suits a viva opener.
- **Technical terms.** Notre Dame's graduate school suggests keeping a needed term and explaining it in the same sentence, in apposition (Capdevielle et al., n.d.). MIT warns against piling up definitions and against words with a different everyday meaning. York's judges want terms explained without the research trivialised.
- **The ending.** MIT ends with an ask; Sussex with a take-home message. Notre Dame adds what the others omit: how your work joins your discipline's wider conversation, which matters when the listener is in your field.

Our reading: at a conference reception, end with MIT's ask; in a 3MT or a viva, end with the take-home sentence.

## What is a lay summary, and how long should it be?

A lay summary is a short written account of research for readers outside the field; funders and journals call it a plain English or plain language summary, and each sets its own length. The same four moves fit it.

### A funder and a journal

The UK's National Institute for Health and Care Research requires a plain English summary in its standard application form; non-specialist reviewers, including members of the public, use it to judge the bid (National Institute for Health and Care Research, 2021). Some programmes ask for 300 words, others up to 750. NIHR wants aims, background, methods, public involvement and dissemination, at about the level of a newspaper article, and not the abstract pasted in.

The American Geophysical Union requires a Plain Language Summary for 13 of its journals: one paragraph of at most 200 words, plus up to three Key Points of 140 characters each, in four parts close to our four moves (American Geophysical Union, n.d.).

### Where they disagree

NIHR writes for members of the public. AGU says its summaries are not for the general public, and pitches them at undergraduate science, for journalists and researchers in other fields. Follow the body you are writing for. A conference abstract is different again: see [how to write a conference abstract](https://edcitation.com/newsletter/how-to-write-a-conference-abstract).

## How long is each format? A table

| Format | Time or length | Rules | Source |
| --- | --- | --- | --- |
| Elevator pitch (MIT) | About 20 seconds | Intro, hook, idea, why it matters, an ask | Prost and Larson (n.d.) |
| Research elevator pitch (VCU) | 30 seconds to 2 minutes | Problem, gap, your work, how it helps | VCU Libraries (2026) |
| Two-minute explanation (ours) | About 240 to 320 words | Four moves; the key study found with [Find sources](https://edcitation.com/) | This guide |
| Three Minute Thesis | 3 minutes at most | One static slide; spoken word; no props or media | University of Queensland (n.d.-b) |
| NIHR plain English summary | 300 to 750 words, by programme | For the public; not the abstract | National Institute for Health and Care Research (2021) |
| AGU Plain Language Summary | One paragraph, 200 words at most | Undergraduate level; acronyms explained | American Geophysical Union (n.d.) |
| AGU Key Points | Up to 3, 140 characters each | No special characters or acronyms | American Geophysical Union (n.d.) |

## Does jargon really lose a listener?

The best evidence we read says yes, and that a definition does not fully repair it. Bullock et al. (2019) gave 650 US adults from an online panel three short paragraphs on self-driving cars, surgical robots and 3-D bioprinting, with or without ten technical terms each, and with or without definitions shown on hovering.

Jargon readers rated the text harder to process: 4.57 against 5.27 on a seven-point scale. Offering definitions made no measurable difference. Harder processing went with more resistance to the message, higher perceived risk and less support for the technologies. The authors name the limits: a non-representative sample, text with no images or context, comprehension not measured directly, and data from one point in time, so the chain of effects is not proven cause.

Our caution: participants read; they did not listen, and a listener cannot reread a sentence. One experiment is not a law, but its lesson matches the university advice: replace the term where you can. A related experiment is summarised in our [prose guide for clear writing](https://edcitation.com/newsletter/a-prose-guide-for-clear-writing).

## How can EdCitation help with the one study your pitch names?

EdCitation finds and cites the study your pitch leans on, and leaves the pitch to you: the argument must be yours, and the reference behind it must be real. EdCitation is the best tool for the reference, because it looks each one up in the publisher's record rather than composing one, as a chatbot does.

### Finding and citing the study

[Find sources](https://edcitation.com/) searches about 300 million published works, by topic or by claim, with filters for kind of source, years and open access, free with no account. On 27 September 2026 we searched "text message reminders missed appointments". The first request gave no answer within 60 seconds; the second returned 421,187 works from Crossref. Second in the list was the Geneva trial: Junod Perron et al. (2013), "Text-messaging versus telephone reminders to reduce missed appointments in an academic primary care clinic: a randomized controlled trial". We read its abstract at source before using its numbers: 6,450 patients, 11.7% missed with a text and 10.2% with a call, judged equivalent, with the text more cost-effective.

[Cite a source](https://edcitation.com/cite) builds the reference from a DOI, title, PubMed ID, ISBN or web address in six styles, and screens for retraction. For 10.1186/1472-6963-13-125 it read Crossref, found no retraction, and returned this [APA 7](https://edcitation.com/cite/apa) entry:

> Junod Perron, N., Dao, M. D., Righini, N. C., Humair, J.-P., Broers, B., Narring, F., Haller, D. M., & Gaspoz, J.-M. (2013). Text-messaging versus telephone reminders to reduce missed appointments in an academic primary care clinic: A randomized controlled trial. *BMC Health Services Research*, *13*(1), Article 125. https://doi.org/10.1186/1472-6963-13-125

On a 3MT slide, "Junod Perron et al., 2013" in small type is usually enough; the full entry goes in the handout. What to say aloud is in [how to cite sources in a speech or debate](https://edcitation.com/newsletter/how-to-cite-sources-in-a-speech-or-debate). If a pitch comes with a written brief, [Check your paper](https://edcitation.com/check) reads it into a checklist. These tools and [Verify references](https://edcitation.com/verify-references) are free; paid plans are on the [pricing page](https://edcitation.com/pricing).

## Quick questions

### How long is an elevator pitch for research?

MIT's Communication Lab says about 20 seconds; VCU Libraries allow 30 seconds to two minutes. Two minutes suits a viva opener.

### What are the 3MT rules?

Three minutes at most, one static slide, spoken word only, and no props, sound or video, under the University of Queensland's rules. Going over means disqualification.

### What are the best 3MT tips?

Write for a friend in another field, tell a story with one take-home sentence, keep the slide uncluttered, and rehearse aloud to a timer in front of people.

### How long is a lay summary?

Whoever asks sets it: NIHR programmes 300 to 750 words, AGU journals one paragraph of at most 200.

### Can EdCitation write my research pitch?

No. EdCitation never writes any part of a talk or thesis. Its free [Find sources](https://edcitation.com/) finds the study your pitch names, and Cite a source builds its reference from the publisher's record.

## References

- American Geophysical Union. (n.d.). *Plain language summary*. AGU Publications. [https://www.agu.org/publications/authors/journals/plain-language-summary](https://www.agu.org/publications/authors/journals/plain-language-summary)
- Bullock, O. M., Colón Amill, D., Shulman, H. C., & Dixon, G. N. (2019). Jargon as a barrier to effective science communication: Evidence from metacognition. *Public Understanding of Science*, *28*(7), 845–853. [https://doi.org/10.1177/0963662519865687](https://doi.org/10.1177/0963662519865687)
- Capdevielle, M., Bendlin, R., & Busl, G. (n.d.). *The elevator pitch: Presenting your research in conversation* [PDF]. University of Notre Dame Graduate School. [https://graduateschool.nd.edu/assets/32665/elevator_pitch_presentation.pdf](https://graduateschool.nd.edu/assets/32665/elevator_pitch_presentation.pdf)
- Junod Perron, N., Dao, M. D., Righini, N. C., Humair, J.-P., Broers, B., Narring, F., Haller, D. M., & Gaspoz, J.-M. (2013). Text-messaging versus telephone reminders to reduce missed appointments in an academic primary care clinic: A randomized controlled trial. *BMC Health Services Research*, *13*(1), Article 125. [https://doi.org/10.1186/1472-6963-13-125](https://doi.org/10.1186/1472-6963-13-125)
- National Institute for Health and Care Research. (2021). *Plain English summaries* [PDF]. Learning for Involvement. [https://www.learningforinvolvement.org.uk/wp-content/uploads/2021/04/Plain-English-Summaries-2022.pdf](https://www.learningforinvolvement.org.uk/wp-content/uploads/2021/04/Plain-English-Summaries-2022.pdf)
- The Ohio State University Graduate School. (n.d.). *Three Minute Thesis (3MT®)*. [https://gradsch.osu.edu/current-students/graduate-student-experience/career-building-strategies-and-skill-development/three](https://gradsch.osu.edu/current-students/graduate-student-experience/career-building-strategies-and-skill-development/three)
- Prost, V., & Larson, D. (n.d.). *Elevator pitch*. MIT Communication Lab, Mechanical Engineering CommKit. [https://mitcommlab.mit.edu/meche/commkit/elevator-pitch/](https://mitcommlab.mit.edu/meche/commkit/elevator-pitch/)
- University of Queensland. (n.d.-a). *About 3MT*. Three Minute Thesis. [https://threeminutethesis.uq.edu.au/about](https://threeminutethesis.uq.edu.au/about)
- University of Queensland. (n.d.-b). *Competition rules*. Three Minute Thesis. [https://threeminutethesis.uq.edu.au/resources/competition-rules](https://threeminutethesis.uq.edu.au/resources/competition-rules)
- University of Queensland. (n.d.-c). *FAQs for competitors*. Three Minute Thesis. [https://threeminutethesis.uq.edu.au/resources/faqs-competitors](https://threeminutethesis.uq.edu.au/resources/faqs-competitors)
- University of Queensland. (n.d.-d). *Judging criteria and panel*. Three Minute Thesis. [https://threeminutethesis.uq.edu.au/resources/judging-criteria-and-panel](https://threeminutethesis.uq.edu.au/resources/judging-criteria-and-panel)
- University of Sussex. (n.d.). *Preparing your 3MT presentation*. Sussex Researcher School. [https://www.sussex.ac.uk/internal/sussex-researcher-school/researcherdev/researchercommunity/threeminthesis/preparing3mt](https://www.sussex.ac.uk/internal/sussex-researcher-school/researcherdev/researchercommunity/threeminthesis/preparing3mt)
- University of Washington Graduate School. (n.d.). *2026 UW Three Minute Thesis (3MT)*. [https://grad.uw.edu/current-students/student-success/3mt/](https://grad.uw.edu/current-students/student-success/3mt/)
- University of York. (n.d.). *Three Minute Thesis (3MT)*. York Graduate Research School. [https://www.york.ac.uk/research/graduate-school/public-engagement/three-minute-thesis/](https://www.york.ac.uk/research/graduate-school/public-engagement/three-minute-thesis/)
- VCU Libraries. (2026, July 14). *Develop your research elevator pitch: Getting started*. Virginia Commonwealth University. [https://guides.library.vcu.edu/research-elevator-pitch](https://guides.library.vcu.edu/research-elevator-pitch)
- Vitae. (n.d.). *Three Minute Thesis competition*. [https://www.vitae.ac.uk/events/three-minute-thesis-competition](https://www.vitae.ac.uk/events/three-minute-thesis-competition)