Platforms /


Final Comments

Platform Describe any abilities, qualities or limitations not covered in the previous questions
Amazon MTurk HITS are usually short, a few minutes.. but you can get the same Turker to do more than one, which they will do if they think the experiment is interesting and pays enough.
Django Django is very powerful and general purpose, but we just use the out-of-the-box functionality to control flow across consent, survey, screening, main task pages, etc. which are built with jspsych. This basic use case is made very easy by following their beginner tutorials. Thus, Django + jspsych on a linux machine gives us a fully functioning online platform.
Gorilla One limitation is the lab-wide management of experiment deployment on Gorilla. With the "pay-as-you-go" level of pricing, deployment needs to be done on the lab-central account for payment unless local labs allow reimbursement to students who are individually deploying. We figured out a work-around by establishing "projects" in a central account then sharing them to students in the lab. Students can work on adding/modifying "experiments" under "shared projects" which are synchronized to the central account for final deployment.
hearX Mostly for clinical audiology telehealth practice.
ispring The major strength of this software/platform is that it does not require coding expertise, making it quite accessible past the initial costs. It is also capable of supporting long term projects and managing user data securely for online projects. A major weakness is that it was not designed for psychology experiments and built-in options are structured around elearning practices. It can be repurposed, but that does require programming expertise.
Jacoti Class II FDA listed Medical Device In the USA, designed for consumer and professional use. If used by consumer, meant as a screener. If used under the guidance of an Audiologist, intended to be used as a hearing fitting and hearing diagnostic audiometer. White papers available showing differences between app test results and audiometric testing conducted in a traditional sound booth.
jspsych As with any front-end library, pairing with a backend server is needed for functionality in practice. The documentation in jspsych provides many pointers to how one might do that.
MATLAB The primary advantage of the MATLAB web app server is the ability to convert existing MATLAB code into a format that works in a browser, allowing researchers to rapidly transition an offline study to an online one without much effort spent redeveloping a task. Since experimenters or institutions will be required to maintain their own server it should be easy to ensure that the platform meets or exceeds all oversight guidelines during the set up and configuration process. Using this platform may require more development work than some of the other platforms dedicated to hosting experiments online, especially on the participant management side, but experiments designed here could be linked to other platforms like MTurk to handle recruiting and payment. There is no cost per participant or per experiment run on the app server. The license allows researchers to upload an unlimited number of apps to the server and the only limits on participants come from the hardware’s ability to handle a given number of simultaneous participants, with no limitations other than time on the total number of participants who can complete a task. My own development efforts on bringing the AFC framework into the MATLAB web app environment will hopefully allow those labs familiar with the framework to rapidly adapt any AFC experiments to be run in the browsers. I hope to be ready with a publicly available demo that I can share with the team in the near future.
PART/BGC Science PART was developed as a Portable Automated Rapid Test platform to bring auditory processing testing to patients in clinics and research participants in their own environments. The development of BGC Science was precipitated by the need for remote testing, substituting the ability to use the participant’s own hardware for control over stimulus level and frequency content. Both platforms have proved to be very effective, especially for suprathreshold signals and tasks that have very little learning associated with them. For audiometric testing and more complex tests, the team is partnering with NCRAR researchers who have developed custom hardware and are working on moving the field forward in terms of adaptive testing procedures. For any project involving PART or BGC Science, the first step should be to contact the developers, who are NIH-funded to work with individual clinicians and researchers interested in developing new projects.
Prolific You can anonymously message individual participants and have a chat-stream with them. This is super useful! Support for multipart/longitudinal studies exists, which can be adapted even for non-longitudinal studies (e.g., early termination of study if participants fail certain screening/quality checks).
Qualtrics It is possible to make an experiment with synchronous interactions in Qualtrics using and opensource software, SMARTRIQS:
SHOEBOX Meant for audiology tele-health practices. Quality audiology researchers/clinicians on the scientific board for this company.