Integrating platforms with one another (Empirica, Dallinger, Pushkin, etc)

[James Houghton]

The different platforms presented today have different offerings - Dallinger for example includes recruitment, and Pushkin gives you a web presence. How easy would it be to use the platforms together? For example, to use Dallinger's recruitment tools to push folks to an Empirica experiment? Or to use Pushkin to give a web presence to an experiment built with o-tree?

[jkhart...@gmail.com]

Some of these will be easier than others. Pushkin is written to assume that the experiment builder (usually jsPsych) is purely front-end, and that any communication with the database runs through Pushkin (specifically, through pushkin-api and pushkin-client). I don't have a good enough sense of how group experiments are handled, but I imagine that oTree and Empirica integrate both front-end and back-end. The most straightforward integration would be to put all the subject-organization logic into the worker (a language-agnostic process that runs in the background). But I don't think this would be plug-and-play. It would probably require extracting specific functions from Empirica and writing new code based on them. Is my guess.

Integration between Pushkin and Dallinger, on the other hand, probably wouldn't be that hard. The strategy would be the same (use functions from Dallinger in the worker), but I think the iterative nature of Dallinger would make that a bit less painful.

But I don't know. I think this is something Abdullah, Jordan, and I have wanted to talk about but haven't found the time.