Feature Requests

Stratifying random assignment in experiments among demographic quotas would be useful. Right now this is only possible by asking participants their demographics in your survey prior to randomly assigning them to condition. This creates a methodological artifcat in that by doing so, you are causing participants to think specifically about their identity before answering survey questions. Research (see Shih et al., 1999, for example) shows that this type of identity priming can alter responses. If you allowed passthrough data for Connect Demographic Info (e.g., political affiliation, gender), you could use survey flow logic to stratify random assignment without priming the social identity. Reference: Shih, M., Pittinsky, T. L., & Ambady, N. (1999). Stereotype susceptibility: Identity salience and shifts in quantitative performance. Psychological science, 10(1), 80-83. https://doi.org/10.1111/1467-9280.00111

The only way to reliably find a decent amount of work is to use an extension like tab reloader to continuously refresh the dashboard page at short intervals. Most tasks end up being gone before a notification can even trigger. A system, like Prolific uses, that automatically refreshes/updates the dashboard every time new tasks are posted would make the platform far less of a hassle to use.

I've returned a few projects due to previous completion and not wanting to take an opportunity or funds from another participant. Can we figure out how to inform if someone has completed a project?

Add a way to access receipts from funding the lab account from the account page.

We have an online game we are sending users to play for psych research. We are using five simultaneous CR experiments, and want to put players in one of five conditions based on the experiment they're in (on the same website). The easiest way to do this would be for us to pass a query string parameter, e.g., "?xyz=study1" that we could then use to assign participants to a group

Currently, participants who opt in to study notifications get such notifications only sometimes (sometimes there are studies upon refreshing Connect that no notifications came through about), and they're completely silent. At least while there are so few studies on the platform (i.e., studies aren't constantly available), participants should get notifications for every study and have the option to get audio alerts with these notifications. Also, the notifications should not group together and prevent participants from seeing the details of one particular notification, which is what currently happens if the participant has at least one study notification they have yet to open.

allow a percentage or N of the sample to pilot. for example, i might want to set the study to collect for 10%, it'll automatically pause collection upon 10%, i'll examine data for any issues, and resume collection when it's done. currently, you need to create two separate projects (one pilot, one the rest), and if you're using the redirect link for qualtrics, you need to change your actual survey (if you're using the same survey) for the second project.

As on MTurk, accepting one study shouldn't preclude a participant from accepting another (with the understanding that the participant would complete the studies separately from one another). This will allow participants to maximize their earnings where they otherwise would have lost out on either or both studies and will prevent participants from canceling participation midway through a study in the hopes of securing a spot in a higher-paying study.

Can we get an extension similar to Prolifics that will show how many jobs are available without having to go to the website and refresh the page.

In AMT we have the option to upload a csv file (called the input file) where we can have multiple HITs ( https://blog.mturk.com/using-csv-files-to-create-multiple-hits-in-the-requester-ui-22a25ec563dc ). In CR we have only one hyper link. This is not convenient for experimentalists who need to run several experiments, each of which with a different link because this means creating a different survey per link. In our case, for instance, this would mean creating 4000 surveys.... So I think it would be a good further development for CR to implement something like this as well.