16

We are currently using Inquisit to conduct Internet experiments.

Ideally, we would like to implement something that is strictly browser-based, without any extra plug-ins or downloads. We also require reaction time measurements.

Is there anything open source out there like this?

This question has outlined various methods that don’t fulfill all our requirements.

Jeromy Anglim
  • 30,741
  • 11
  • 93
  • 221
Allison Yan
  • 163
  • 5
  • Welcome to cogsci.SE! Can you expand a little bit in your question on why the answers on the earlier question don't fulfill your needs? It is not clear to me from the question, but do you still want to collect timing data? Also, I feel like the last paragraph is distinct enough from the ones above it to merit a separate question. – Artem Kaznatcheev Apr 07 '12 at 18:17
  • I would like a more concrete answer on how to implement an experiment that is browser based, but that does not require users to download any extra plug-ins as Java does. And yes, I want to collect timing data at the same time. I'll create a new question for that last bit. Thanks! – Allison Yan Apr 07 '12 at 19:14
  • 2
    Okay, after you ask the new question, you can click the 'edit' button below this question to remove the last paragraph and incorporate the content of your comment (i.e. that you want timing data and that the answers in the other question demand plug-ins). Also, take a look at this answer, it seems to come pretty close, but technically does require the flash plug-in, but most browsers have that by default. Hopefully @AndyDeSoto will notice this question and give more insights. – Artem Kaznatcheev Apr 07 '12 at 19:23
  • 2
    Flash and Javascript are two options but I don't have first hand experience. Also, Inquisit version 4 has mac support – Jeromy Anglim Apr 08 '12 at 06:49
  • I wouldn't trust flash to run smoothly enough for accurate measurements personally, but JS can be pretty accurate – Ben Brocka Apr 08 '12 at 15:13
  • @Jeromy Anglim: Thank you! I had no idea Inquisit was in beta 4. – Allison Yan Apr 09 '12 at 18:41
  • @BenBrocka why would JS be more reliable than flash? – Artem Kaznatcheev Apr 09 '12 at 22:43
  • @ArtemKaznatcheev Flash has more significant performance issues at the micro level, I wouldn't trust Flash to time clicks within a millisecond. – Ben Brocka Apr 09 '12 at 22:59
  • 2
    General question to those who are providing helpful links: how good is the timing on these packages? I know some providers of experimental software still recommend XP as the best for experiments as it won't misbehave and make your experiment a low priority in terms of resources, thereby screwing up the timing. – vizzero Apr 14 '12 at 15:26
  • @vizzero What do you expect? The OP wants something that runs in the browser, so timing will be comparatively horrible. The users will have all different kind of operating systems (probably most Windows) and a lot of different configurations (background programs such as antivirus, viruses (!), ...). So even if your program is good the shear amount of possible differences in user configuration will screw any reliable estimate. – Henrik Apr 14 '12 at 19:27
  • 1
    @Henrik : I am entirely aware of that stuff - hence my comment :) The user said "We also require reaction time measurements." I wanted them to be clear that they are in dangerous territory for that kind of thing. – vizzero Apr 15 '12 at 11:19

8 Answers8

6

WebExp is a client/server based psychology/linguistics experiment creation/running system written in Java. It is freely available.

A subject types in the appropriate web address and they see the experiment pages that have been created; obviously you have to have access to a server on which the experiment software+configuration files are running. It supports timing and works quiet well.

Derek Jones
  • 76
  • 2
4

I think ProjectImplicit will be what you want. It is also Java based and runs fully in the browser. It is by the Harvard guys that did run the IAT via web and collected ten thousand datasets this way.

See here for their services (I am not sure if it is free but seems so at least for non-commerical research). If you like it and use it perhaps you can post some infos afterwards here.

Henrik
  • 318
  • 2
  • 8
3

For an open source JavaScript/HTML/CSS solution, check out jsPsych: http://www.jspsych.org. It can be used for reaction time measurement and interactive designs. An article describing the library was recently published in Behavior Research Methods.

de Leeuw, J. R. (2014). jsPsych: A JavaScript library for creating behavioral experiments in a Web browser. Behavior research methods, 1-12.

Josh
  • 5,904
  • 2
  • 25
  • 47
2

Also, consider lab.js. It's Javascript-based, so no downloads.

lab.js is a tool for constructing browser-based studies built for (but not limited to) the social and cognitive sciences. Studies can be assembled using a graphical builder, or through code. In either case, lab.js takes care of study presentation, response collection and data storage, while providing the greatest possible flexibility with regard to study design.

Ruben
  • 472
  • 3
  • 11
2

If you want to deliver visual stimuli with accurate timing, don't use this method as the general consensus is that flat screen monitors can not yet be fully trusted for timing (though tests are being done, and some monitors may be ok, the jury is very much out on this one). For example, if you want a stimulus up for, say 500ms, or some smaller number like 50ms, delivering experiments via the web will not work - or at least, you'll have to throw loads of people out. I could see one method would be to let vast numbers of people take part and then ditch their data if their machine didn't meet certain requirements.

vizzero
  • 1,135
  • 8
  • 11
  • this isn't an answer, but it is a very interesting question/comment. I would encourage you to ask this as a separate question since answers are not supposed to be used for further questions or discussion. – Artem Kaznatcheev Apr 14 '12 at 15:15
  • Can do - though would it be better as a comment on the original question? I added it as an answer given that there were several separate answers with different software. – vizzero Apr 14 '12 at 15:20
  • 1
    I would comment on the original question, unless you have enough of a follow up question to make a separate one. I know that the OP wanted to ask another question about the reliability of timing data, and the second part of your answer would be great for that; but she has yet to ask the follow up. – Artem Kaznatcheev Apr 14 '12 at 15:22
  • @ArtemKaznatcheev Agreed. The first paragraph should have been a comment, but I find the second paragraph to be a suitable 'answer'. Perhaps it's best if vizzero edits the post as such. – Steven Jeuris Apr 14 '12 at 15:25
  • 1
    We have done some serious test (using photoreceptors and other crazy equipment) at our lab using a 120hz LG Flatron W2363D (3D LCD monitors). The results were surprising. Almost all measures indicated advantages for the LCD (lesser dropped frames, ...). Unfortuanetly, the guy who did it does not want to write a paper about it. But I could ask him which measures he took. – Henrik Apr 14 '12 at 20:18
1

"We are currently using Inquisit to conduct Internet experiments. Since it is not compatible with non-Windows systems"

As of this time, Inquisit runs on Windows, Mac, iOS, and Android devices for precision-timed psychological testing. It also integrates with Qualtrics, Sona, TurkPrime, and other online participant recruiting platforms.

It does still require that participants download the app, either from directly millisecond.com, the App Store, or the Google Play store.

Details available at https://www.millisecond.com.

seandr
  • 11
  • 2
1

I've been developing an online platform to run HTML5/Javascript experiments, recruit participants via email, Facebook, or Twitter, and collect and evaluate results in real time. Neither you nor your participants require anything but a web browser.

I agree with @vizzero that timing behavior is a challenge. I optimized timing by preloading all assets and generally recommend Google Chrome, as its internal clock has the highest resolution of all modern web browsers. However, you have to verify that browser based experiments meet your needs.

Please see stato.de for a demo; it does not require signup.

Nick Stauner
  • 9,414
  • 2
  • 26
  • 58
stato
  • 51
  • 2
0

You should have a look at otree, "a Django-based framework for implementing multiplayer decision strategy games".

It is manly designed for experimental economics, but it is a very neat and versatile piece of software, and I am sure you could actually run most psychological experiment with it.

Once the experiment is coded, it only requires internet access and a browser on the part of the participant.

Big plus : otree comprises functionalities to connect with Amazon Mechanical Turk.

Martin Van der Linden
  • 313
  • 1
  • 6
  • Essentially the same answer as the one I posted for http://cogsci.stackexchange.com/questions/699/software-for-online-psychological-experiments-that-dont-require-users-to-downlo. Don't know if that's acceptable by CogSci's guidelines. – Martin Van der Linden Sep 09 '15 at 16:01