There is no single best tool for behavioral research — there is the right tool for the study in front of you. Here is where MediaLab fits, and where something else would serve you better.
Almost every comparison on this page comes back to that one question.
MediaLab is built for lab-based sessions on a dedicated Windows PC: a participant sits down, works through a sequence of questionnaires and media, and the data lands on your machine. It excels where you need tight control over what is shown, in what order, and for how long — priming procedures, media presentation, mixed questionnaire-and-stimulus designs — and where you want your data to never leave the building.
Choose a web survey platform instead if your participants are remote, on their own devices, and you are collecting self-report data. That is what tools like Qualtrics are designed for, and MediaLab is not a substitute for them.
Choose DirectRT instead — or alongside — when the dependent variable is a reaction time measured in milliseconds.
Compared on how each tool is built and paid for, rather than on a feature checklist — those change every release, and this does not.
| MediaLab | DirectRT | Qualtrics | Gorilla | Inquisit | PsychoPy | |
|---|---|---|---|---|---|---|
| Where it runs | Lab PC, offline | Lab PC, offline | Browser | Browser | Lab PC or browser | Lab PC or browser |
| Built for | Questionnaires plus media, in session | Reaction time | Web surveys at scale | Online experiments | Cognitive tasks | Experiments, broadly |
| Timing control | Presentation-level | Millisecond | Not a design goal | Good, browser-bound | Millisecond | Millisecond |
| Where data lives | Your machine | Your machine | Vendor cloud | Vendor cloud | Your machine or their cloud | Your machine or Pavlovia |
| Pricing model | Buy once, own it | Buy once, own it | Annual subscription | Per participant | Annual subscription | Free, open source |
| Coding needed | None | None | None | None | Its own script language | Optional, Python |
| Platform | Windows | Windows | Any | Any | Windows, macOS, web | Windows, macOS, Linux |
Every product here is actively developed and we have tried to describe each one as its own makers would recognise. Details change; if something below has gone out of date, tell us and we will correct it. Product names are the trademarks of their respective owners.
Including, in every case, the situation where we would point you elsewhere.
Our own reaction-time program — the comparison researchers ask us about most.
Your dependent variable is how quickly someone responded. DirectRT is built around millisecond-accurate presentation and response capture, for lexical decision, priming, evaluative categorisation and similar latency-based paradigms.
Your dependent variable is what someone answered. Questionnaires, scales, open-ended responses, consent, debriefing and media presentation, with branching between them.
Most labs run both. MediaLab handles the session — consent, measures, instructions — and hands off to DirectRT for the timed task, then carries on afterwards. That is the pairing they were designed for, and the bundle costs less than two separate licenses. See how DirectRT compares →
The dominant survey platform in higher education.
You are surveying people remotely, on their own devices, at scale — and especially if your university already holds a site license, which thousands do. In that case it is effectively free to you, with sophisticated logic, panels and distribution that MediaLab does not attempt to match. Check with your library or IT before you buy anything.
The study happens in a lab session rather than a browser tab. When you need to present video, audio or images under controlled conditions, interleave them with measures, hand off to a reaction-time task, or keep participant data entirely on your own hardware for ethics or data-residency reasons.
General-purpose survey tool, widely used outside academia.
You want a straightforward questionnaire online quickly and cheaply, and you do not need experimental control. It is easy, familiar to participants, and inexpensive at small scale.
The study is an experiment rather than a survey — random assignment to conditions, stimulus presentation, timing, counterbalancing — or when analysis-ready output and local data storage matter more than a shareable link.
Browser-based experiment builder, popular for online data collection.
You need a large or geographically diverse sample and cannot bring people into a lab. It builds real experiments — not just surveys — without code, and independent timing research has found browser-based collection good enough for a wide range of designs.
You run participants in person and would rather own the software outright than pay per participant. A lab machine you control also removes the browser, the operating system and the network connection as sources of variability you cannot see.
The two commercial lab packages most often weighed against ours.
You need one program to do everything including the millisecond work, you rely on a specific published task from their libraries, or you need deep integration with EEG, fMRI or eye-tracking hardware. E-Prime in particular is long established in cognitive neuroscience for exactly that.
Your work is questionnaire-and-media centred, you would rather not learn a scripting language, and you prefer a one-time purchase to a subscription. MediaLab is deliberately narrower and, for the studies it targets, quicker to build in.
Free, open-source, and excellent. We will not pretend otherwise.
Budget is the binding constraint, or you want an open format, a large community, published timing performance among the best available, and Python underneath. If you or your students are comfortable with code, it is hard to argue against.
You want a shorter path from idea to running study for questionnaire-based designs, and a support line staffed by the people who wrote the software. Free is not free if it costs a graduate student three weeks — but if that student enjoys Python, it genuinely is free.
MediaLab is free for 30 days, with no payment details required. Build the first few screens of something you actually plan to run.