Online cognitive tasks for teaching psychology

What a class needs from a task tool

A lab session with the classic tasks has the same shape whatever the tool. Students do the task, the data comes back to you, and the next session looks at the effect the lecture introduced. Five things decide whether a tool fits that shape.

• Set-up on the student's side. Anything a student has to download or configure costs you the first ten minutes of the session and a few of the students.
• Who pays. Some tools charge per participant, some charge the institution, some charge nobody.
• Data. You need the trial-level data of the whole class in one file, soon enough to use it.
• Ready-made tasks. Building a Stroop is a fine exercise in a programming course and a detour in a cognitive psychology one.
• Language. If you teach in Spanish, the instructions and the stimuli have to be in Spanish, not only the menus.

The sections below take each tool through those five points. Prices and plans change: the details were checked in October 2026 against each tool's own site, and are worth confirming there before you plan a term around them.

PsyToolkit

PsyToolkit is an online environment for building and running reaction-time experiments and questionnaires, written and maintained by Gijsbert Stoet (Stoet, 2010; Stoet, 2017). It is the closest thing teaching has to a standard: the 2017 paper describing it appeared in Teaching of Psychology, and the site says it suits teaching from secondary school upwards.

Set-up

The instructor edits experiments in a web editor under an account. Students open a link. An offline Linux version exists, but at the time of writing the site marks it as unavailable.

Cost

No charge for university students and staff, and classroom use at not-for-profit institutions is explicitly allowed. Commercial use is not, and citing the two papers is a condition of use.

Data

Collected on PsyToolkit's server and downloaded as text, CSV or Excel. The records sit under your account.

Tasks

A large library of ready-made experiments, including Stroop, flanker, digit span, reverse digit span, 2-back and the psychomotor vigilance task, each with its code open so a class can read how it works.

Language

The library tasks are in English, with a German Stroop as the exception, and are written to be easy to translate. The survey side lists Spanish among its interface languages; the experiments you translate yourself.

Scores come back raw: reaction times and accuracy per trial. There are no norms, so a student sees their own numbers rather than a comparison with a reference group. For a lab on the effect itself, that is exactly right.

PsychoPy and Pavlovia

PsychoPy is the open-source desktop application for building experiments, with a visual Builder and a Python Coder (Peirce et al., 2019). Pavlovia is its hosting service: a Builder experiment is exported to JavaScript and served from Pavlovia.

Set-up

The instructor installs PsychoPy and holds a Pavlovia account. Students open a link; the usual reading of the user forum is that they need no account, although the official documentation does not say so in as many words.

Cost

PsychoPy costs nothing. Pavlovia charges a credit for each participant whose data is saved, or an annual site licence that covers everyone at an institution. Many universities already hold one, so ask your department before assuming a per-student fee.

Data

Each run saves a file to the experiment's project on Pavlovia, which you download: PsychoPy's own CSV, one row per trial.

Tasks

The Builder ships with demos, Stroop among them, and Pavlovia hosts public experiments you can fork. You still assemble, check and pilot each task.

Language

Anything you type. Stimuli and instructions are yours, so a Spanish task is as easy as an English one. The Pavlovia interface itself is in English.

PsychoPy is the right choice when the course is also about building experiments, and the timing work behind it is unusually well documented (Bridges et al., 2020). For a session where the task is a means to an effect, the build is the overhead.

Gorilla

Gorilla is a commercial online experiment builder with a visual task builder and a questionnaire builder (Anwyl-Irvine et al., 2020).

Set-up

Nothing on either side. The instructor needs an account; students open a link.

Cost

Building is at no charge. The company charges for the data collected, one token per participant per experiment, at academic and public-sector rates with a minimum purchase. A department subscription includes teaching use, and Gorilla's teaching page describes university courses run that way. Students added as collaborators can preview a task without a token, but that data is not saved.

Data

Downloaded from the Gorilla project as spreadsheets once the tokens are in place.

Tasks

An open-materials page lists classic tasks, including Stroop, flanker, Simon, digit span and n-back, that you clone into your own project.

Language

Not stated by Gorilla. The builder accepts any text you enter, so stimuli can be in Spanish.

Gorilla's strength is the polished builder and the support around it. The cost model is the thing to check: without a departmental subscription, each student run costs a token.

lab.js, jsPsych and the open-source stack

lab.js is an open-source builder with a visual interface that works from a tab, so the person building the study installs nothing (Henninger et al., 2022). The catch is hosting: lab.js builds the experiment but does not collect the data for you. You export it to your own web space, to JATOS (an open-source server your university runs, or the hosted MindProbe instance), to Open Lab, or to a survey platform.

Cost

The software costs nothing. Hosting is whatever your host charges: nothing on your own server or a JATOS instance, a plan with an academic discount on Open Lab.

Data

Wherever the host puts it, usually CSV. On your own server you write the receiving script.

Tasks

Templates for common paradigms; a Stroop or a flanker is built from them in an afternoon.

Language

Yours to type.

jsPsych is the JavaScript library most of the open-source world builds on, this site included (de Leeuw, 2015; de Leeuw et al., 2023). It has plugins for the classic paradigms and an active community, and it needs a programmer and a host. Cognition.run hosts jsPsych tasks with a no-cost tier capped by number of runs. OpenSesame with OSWeb and JATOS is the same idea with a desktop builder (Mathôt et al., 2012). PEBL is a desktop battery with many classic tests (Mueller and Piper, 2014) and a newer build that opens in a tab.

Neuropsych

Neuropsych is this site: standardized cognitive tasks, built on jsPsych, in English and Spanish, scored against published normative tables. It is not an experiment builder. The tasks are fixed, documented and the same for everyone, which is the point for a class and the limitation for a study that needs its own design.

Set-up

The instructor creates a class link at neuropsych.io/teach/ in about a minute and pastes it into the course page. Students open it on their own laptop or phone and sign in with an email code or a Google account to keep their results. There is nothing to configure on either side.

Cost

Nothing is charged to students, instructors or the institution, for any class size.

Data

A class page with one row per consenting student under a code, live as they finish: scores, percentiles against the published norms and response-time summaries, with a CSV, a trial-level export where the task logs it, and a PDF class report that shows whether the class reproduced the effect. Names and email addresses never reach the instructor.

Tasks

The classic ones a syllabus covers, among them Stroop, flanker, go/no-go, stop-signal, digit span, Corsi blocks, n-back, operation span, simple and choice reaction time, trail making, Wisconsin card sorting, Tower of London, number comparison and a verbal list-learning task. Each has a page with its scoring rules and reference values, and every rule is open on the scoring engine page.

Language

The tasks a lab needs, the class page and the report exist in Spanish, with Spanish instructions and dictation audio, so a bilingual department can run the same lab in either language. A few verbal tasks are English only.

What you do not get: your own paradigm, your own stimulus timing, or a within-subject design you invented. For that, use a builder.

The comparison at a glance

Which one, for which course

• Cognitive psychology, the effect as the point of the session. Neuropsych or PsyToolkit. Both give you the classic task tomorrow. Ours adds the norms and the class report; PsyToolkit adds the task's open code and a larger library.
• Research methods or statistics, where the class needs its own dataset. Neuropsych for a dataset in one file with percentiles, trial data and scoring rules the class can audit; PsyToolkit for raw reaction times the class processes from scratch.
• Experimental psychology, where students design a study. PsychoPy, lab.js or Gorilla. Building is the learning.
• A Spanish-language course. Neuropsych if the standard tasks are enough; any builder if you are willing to translate.
• A single demonstration in a lecture. Neuropsych. A student does the Stroop on a phone during the break and sees the effect in their own numbers.

A note on timing and what online data can show

Online reaction-time tasks are less precise than a lab machine with a response box, and the loss is smaller than instructors often fear. The classic effects replicate online: Stroop, flanker, Simon, task switching and the attentional blink were all reproduced with web participants in early work (Crump et al., 2013), performance on a range of tasks matched the lab (Germine et al., 2012), and within-subject effects survive the added noise (Hilbig, 2016). Measurements of the platforms themselves put the variability of online presentation and response timing between a few and a few tens of milliseconds, depending on the browser and the device (Bridges et al., 2020; Anwyl-Irvine et al., 2021). For a class demonstration of an effect of a hundred milliseconds, every tool here is good enough. For a study that needs millisecond precision, none is a substitute for the lab, and those two papers are the ones to read before choosing.

A percentile is a separate matter. Only a tool with normative tables can place a student's score against a reference sample, and the comparison is with a published norm collected on a different version of the task, which is why this site says so on every result.

Questions instructors ask

What is the closest PsyToolkit alternative for a class?

For the classic tasks, this site: students open a class link, the tasks are fixed and scored, and the class data arrives with percentiles and a PDF report. PsyToolkit remains the better choice when you want to edit the task's code or need a paradigm outside the standard set.

Do students need an account?

PsyToolkit, Pavlovia and Gorilla collect data under the instructor's account, and a participant normally opens the link without signing in. Neuropsych asks students to sign in with an email code or a Google account so that each keeps their results, and shares them with the class only after an explicit yes. The instructor sees a code per student, never a name.

Can a class on a Spanish-speaking campus use these?

With the builders you translate the instructions and stimuli yourself, which is quick for a Stroop and slower for a task with recorded audio. On Neuropsych the Spanish version is maintained as a first-class twin, written in the informal register students expect, and the class page and the report follow the language of the link.

References

1. Stoet G. PsyToolkit: a software package for programming psychological experiments using Linux. Behavior Research Methods. 2010;42(4):1096-1104. https://doi.org/10.3758/BRM.42.4.1096
2. Stoet G. PsyToolkit: a novel web-based method for running online questionnaires and reaction-time experiments. Teaching of Psychology. 2017;44(1):24-31. https://doi.org/10.1177/0098628316677643
3. Peirce J, Gray JR, Simpson S, MacAskill M, Höchenberger R, Sogo H, Kastman E, Lindeløv JK. PsychoPy2: experiments in behavior made easy. Behavior Research Methods. 2019;51(1):195-203. https://doi.org/10.3758/s13428-018-01193-y
4. Anwyl-Irvine AL, Massonnié J, Flitton A, Kirkham N, Evershed JK. Gorilla in our midst: an online behavioral experiment builder. Behavior Research Methods. 2020;52(1):388-407. https://doi.org/10.3758/s13428-019-01237-x
5. Henninger F, Shevchenko Y, Mertens UK, Kieslich PJ, Hilbig BE. lab.js: a free, open, online study builder. Behavior Research Methods. 2022;54(2):556-573. https://doi.org/10.3758/s13428-019-01283-5
6. de Leeuw JR. jsPsych: a JavaScript library for creating behavioral experiments in a web browser. Behavior Research Methods. 2015;47(1):1-12. https://doi.org/10.3758/s13428-014-0458-y
7. de Leeuw JR, Gilbert RA, Luchterhandt B. jsPsych: enabling an open-source collaborative ecosystem of behavioral experiments. Journal of Open Source Software. 2023;8(85):5351. https://doi.org/10.21105/joss.05351
8. Mathôt S, Schreij D, Theeuwes J. OpenSesame: an open-source, graphical experiment builder for the social sciences. Behavior Research Methods. 2012;44(2):314-324. https://doi.org/10.3758/s13428-011-0168-7
9. Mueller ST, Piper BJ. The Psychology Experiment Building Language (PEBL) and PEBL Test Battery. Journal of Neuroscience Methods. 2014;222:250-259. https://doi.org/10.1016/j.jneumeth.2013.10.024
10. Bridges D, Pitiot A, MacAskill MR, Peirce JW. The timing mega-study: comparing a range of experiment generators, both lab-based and online. PeerJ. 2020;8:e9414. https://doi.org/10.7717/peerj.9414
11. Anwyl-Irvine A, Dalmaijer ES, Hodges N, Evershed JK. Realistic precision and accuracy of online experiment platforms, web browsers, and devices. Behavior Research Methods. 2021;53(4):1407-1425. https://doi.org/10.3758/s13428-020-01501-5
12. Germine L, Nakayama K, Duchaine BC, Chabris CF, Chatterjee G, Wilmer JB. Is the Web as good as the lab? Comparable performance from Web and lab in cognitive/perceptual experiments. Psychonomic Bulletin and Review. 2012;19(5):847-857. https://doi.org/10.3758/s13423-012-0296-9
13. Crump MJC, McDonnell JV, Gureckis TM. Evaluating Amazon's Mechanical Turk as a tool for experimental behavioral research. PLoS ONE. 2013;8(3):e57410. https://doi.org/10.1371/journal.pone.0057410
14. Hilbig BE. Reaction time effects in lab- versus web-based research: experimental evidence. Behavior Research Methods. 2016;48(4):1718-1724. https://doi.org/10.3758/s13428-015-0678-9