Insights

Research Dissemination: Methods, Examples and How to Plan It

November 13, 2026 · @Owen · 9 min read

Since I started working in marketing, my fascination with human choices and psychology started to take shape. As TwentyFun became an outlet for running my own experiments based on previous studies, I found it to be a very real way of showing experiments in an interactive way.

Research dissemination is the work of getting findings to the people who can use them: other researchers, practitioners, policymakers and the public. It is no longer optional. Horizon Europe requires most applications to include a plan for dissemination and exploitation of results, including communication activities. The US National Science Foundation scores every proposal on “broader impacts” alongside intellectual merit.

The EU separates three related ideas. Communication tells a wide audience that the project exists and why it matters. Dissemination shares the results with people who can use them. Exploitation is the actual use of those results, in further research, policy or products.

This guide covers the common dissemination methods, when each one fits, and how to plan and budget for it. It also looks at one method that gets less attention than it deserves: interactive experiments people can try for themselves.

Common dissemination methods

There is no single right channel. Most projects use a mix, chosen by who needs to hear about the results and what they should do with them.

Journal article (open access)

Who it reaches
Other researchers
Best for
The full method and results, peer reviewed
One-way or interactive
One-way

Preprint

Who it reaches
Researchers, journalists
Best for
Getting results out before peer review
One-way or interactive
One-way

Conference talk or poster

Who it reaches
Your research field
Best for
Feedback and new contacts
One-way or interactive
Some dialogue

Policy brief

Who it reaches
Policymakers, public sector
Best for
Short, actionable recommendations
One-way or interactive
One-way

Press release and media

Who it reaches
The general public
Best for
A clear, newsworthy finding
One-way or interactive
One-way

Plain-language summary or blog

Who it reaches
Practitioners, the public
Best for
Explaining what the results mean
One-way or interactive
One-way

Social media

Who it reaches
Broad, mixed audiences
Best for
Short highlights and links
One-way or interactive
Some dialogue

Podcast or video

Who it reaches
The general public
Best for
Telling the story behind the research
One-way or interactive
One-way

Open data and code

Who it reaches
Other researchers
Best for
Reuse and replication
One-way or interactive
Hands-on, for experts

Interactive experiment or game

Who it reaches
The public, students, other researchers
Best for
Letting people feel a finding, and collecting data
One-way or interactive
Interactive

Most of these share one thing: the audience reads, listens or watches. That works well for explaining a result. It works less well when the point is something people only really get once they have tried it themselves.

The case for interactive formats

Most dissemination is one-way: the researcher explains, the audience reads or listens. Interactive formats flip that. The reader makes a choice, sees the result, and then learns what the research found. There are three good reasons to take this seriously.

People understand more when they do something. The strongest evidence comes from education, not dissemination, but the principle carries over. A meta-analysis of 225 studies of university science teaching found that students in traditional lectures were 1.5 times more likely to fail than students in classes with active learning (Freeman et al., PNAS 2014).

An interactive experiment can reach people a paper never will. In 2014, researchers at the MIT Media Lab launched the Moral Machine, a game-like site where visitors decided how a self-driving car should act in unavoidable accidents. It gathered nearly 40 million decisions from people in 233 countries and territories, and the results were published in Nature (Awad et al., Nature 2018). The researchers did not predict how widely it would spread (MIT Technology Review).

Dissemination and data collection can be the same thing. The Moral Machine was both outreach and the study itself. Foldit, a protein-folding game from the University of Washington, went further. For over a decade, scientists had failed to solve the structure of a protease from a virus that causes AIDS in monkeys. In a three-week challenge, Foldit players produced models good enough to solve it (Nature Reviews Microbiology).

All kids learn by doing. In my head, this is the fastest and simplest way to learn something new: a concept, a technology or an experience someone else has had. By making my own online experiments based on others, I've come to see that some of the results are pretty close to the original studies, even though people play at home behind a screen with no intention of taking part in an actual experiment. Doing something yourself outweighs a PDF, and you can actually feel and get your mind around something in a much shorter time than by digesting a heavy article.

Not every study needs a game, and a few clicks are not a controlled lab setting. Samples from a public website are self-selected and rarely representative. But when the goal is to make a finding felt, reach a broad audience, or collect a large exploratory dataset, an interactive format often does what a PDF cannot.

Examples from TwentyFun

TwentyFun is a small collection of browser experiments based on classic studies. The numbers below are early and come from self-selected visitors, so they are illustrations, not research findings. They do show what happens when a study becomes something people play.

Mind Lab: 32 classic experiments

Players make the choice from a classic experiment, guess what most people chose, and then see the original study. So far: more than 1,700 anonymous choices from 428 unique visits, with an average session of 4 minutes and 4 seconds.

  • Trolley problem: 76% refused to pull the lever (44 of 58). The historical version is in Real Trolley Problems.
  • Prisoner's dilemma: 67% cooperated (44 of 66), even though defecting pays more against a trusting partner.
  • Monty Hall: 70% stayed with their first door (43 of 61), although switching wins two times out of three. Stay or Switch lets you test it a hundred times.
  • Dictator game: asked to share $10 with a stranger, 32% gave nothing and 23% gave exactly $5. Very few gave something in between.
  • Ultimatum game: 38% rejected an unfair offer, giving up free money to punish a greedy partner.

Gut Check: hidden A/B tests

Players are randomly split into two groups that see slightly different wording, and only learn afterwards how they were nudged. So far: 362 answers and 32 complete runs through all eight questions, with an average of 3 minutes and 29 seconds. The effects point the same way as the original studies.

  • Anchoring (after Strack and Mussweiler, 1997): asked whether Gandhi died before or after the age of 9, group A's median guess was 67. Asked about the age of 140, group B's median was 86. An obviously absurd number moved the estimate by 19 years.
  • The calculator (after Tversky and Kahneman, 1981): would you drive 20 minutes to save $5? For a $15 calculator, 70% said yes. For a $125 calculator, only 36% did. In the original study, the figures were 68% and 29% (Tversky and Kahneman, Science 1981).
  • Leading questions (after Loftus and Palmer, 1974): how fast were the cars going when they “hit” each other, or “smashed into” each other? The median guess went from 50 mph to 60 mph. In the original, the average estimate went from 34 to 41 mph (Loftus and Palmer, 1974).
  • Beer on the beach (after Richard Thaler): the same cold beer, from a run-down kiosk or a resort hotel. The median maximum price went from $12.50 to $20.
  • The blind spot: before seeing their results, players guessed they had been nudged on 3.7 of the 8 questions on average. The real number was 4.7. People underestimated how much they had been influenced by almost one full question.

The Hurry: the Good Samaritan, replayed

In Darley and Batson's 1973 study, seminary students with time to spare stopped for a man slumped in a doorway 63% of the time. Those told they were late stopped 10% of the time. The Hurry puts players in the same situation and randomly assigns them to a “hurry” or a “relaxed” condition.

Of the first 20 players, none stopped for the coughing man on their first run, in either condition. One relaxed player crossed the whole campus in 43 seconds without helping anyone. On second and third runs, every player stopped.

This is not a replication, and it is too small to conclude anything about helping. The original study was done on theology students. Mine was played by random people online, most likely not theology students. I was actually not surprised that nobody stopped to help. People treat it like a game and try to beat the clock. I was more surprised that people stopped when they played again. It's almost a reversed version of the original experiment. It shows that you can gamify an experiment, flip what you measure, and still be surprised. I've learned that an experiment done online is not always the same as the original, but you can still get interesting results.

More examples: Eye Duel recreates Asch's conformity experiment, Flicker is a change blindness test, and How Fast Do You Forget? is built on Ebbinghaus's forgetting curve. See them all in psychology games, or browse interactive learning.

How to budget for interactive dissemination

If you want an interactive experiment, the easiest time to pay for it is before the project starts: put it in the grant application. Here is where it usually fits. Always check the details with your grants office, since rules vary between calls and institutions.

Horizon Europe. Describe the activity in Part B under “Measures to maximise impact – Dissemination, exploitation and communication” (General Annexes 2026–2027). A service bought from an outside provider is usually budgeted as “other goods, works and services”, a purchase cost paid at market price (FFG).

NSF. Explain in your Broader Impacts section how the experiment helps your research benefit society. NSF says broader impacts can come through the research itself or through activities directly related to it (NSF). In the budget, it belongs under other direct costs, for example publication, documentation and dissemination or consultant services (NSF proposal budget).

What to include in the budget line

  • Design and development of the experiment
  • Data collection: consent, random assignment to conditions, export to CSV
  • Hosting and maintenance for the full project period
  • Launch: getting the experiment in front of people
  • A short summary of how many played and what they chose

A sample line for your budget justification

Interactive online experiment based on [study], allowing the public to take part in the main task and see how their choices compare with the research findings. Includes design, development, consent and data export, and hosting for the duration of the project. Expected outcome: [number] participants and a shareable public version of the study.

Ask for a fixed price before you submit. Then the number in your application is real, and you know exactly what you are getting.

A simple dissemination plan template

A good plan answers five questions for each audience: who they are, what they should take away, how you reach them, when, and how you will know it worked. Copy the table and fill in one row per audience.

Other researchers

Key message
The full results and method
Format
Open access article, preprint
Channel
Journal, conferences
When
End of project
How you measure it
Citations, downloads

Policymakers

Key message
What the findings mean for policy
Format
Policy brief
Channel
Direct contact, events
When
After main results
How you measure it
Meetings, references in policy

The general public

Key message
One finding they can feel for themselves
Format
Interactive experiment
Channel
Website, social media, press
When
Mid-project or at publication
How you measure it
Players, completion rate, shares

Students and teachers

Key message
How the study works
Format
Interactive experiment, plain-language summary
Channel
Course pages, teacher networks
When
Ongoing
How you measure it
Use in classes, return visits

Two tips. Start with the audience, not the format: decide who needs to hear about the results before choosing how. And plan the interactive part early. If it launches mid-project, the data it collects can feed back into your research.

Have a study people should experience?

I would love to work with behavioural psychology. I find the topic personally fascinating, and studies with a surprising effect on how we think make for really fun interactive experiences.

Giving a study a platform where it can be played and enjoyed gives it a reach beyond scientific circles, from students to people who just enjoy playing and learning online. In my opinion, that is very valuable. Impact is a good word for it: an interactive explanation or game gives the study a personal impact on whoever plays it.

If you have a study you'd like more people to experience, I would love to take a look at it and your conclusions, and find a fun way for people to get that aha moment. I'm especially looking for researchers who think this medium could work for their study and who want a stronger impact.

Send me a message through the contact page →

Tell me a little about your study.