Support for surveydown in the OpenScales project

[stmueller]

Description

Hi,
I've recently launched the OpenScales project--a open format and repository for psychological/health scales:
https://github.com/stmueller/OpenScales
http://openscales.net

The format is intended to support most of the critical factors scale developers incorporate into their personality, health, usability, and other scales. It include common question types, scale-level logic like common likert forms and question heads, scoring and coding logic, translation, visibility logic, and some randomization capabilities. As you are probably aware, the typical way scales are distributed is via pdf files and appendices of journal papers with basic instructions, which can often lead to mistakes and errors.

I provide scale running tools in several ways, as well as conversion scripts that can help support common survey systems that have ability to import surveys. I was excited to find surveydown, and added a minimal conversion utility to export a surveydown-compatible package:

https://github.com/stmueller/OpenScales/blob/main/tools/convert_to_surveydown.py

This is also available by navigating to any scale on openscales.net and using the surveydown conversion download.

This does not implement all of the features of osd into a surveydown survey, but it generates sort of a built package that can be run. Although I'm an R user since the 1990s, I haven't wrapped my head around the architecture of surveydown completely, so there may be issues and ways to improve it that you might more easily see.

This should give easy access to use surveydown in any of the 800+ CC, public domain, and otherwise redistributable scales I've implemented. Obviously, a user would have to have some familiarity with setting up the surveydown system, and although I was able to do it as I've used R and shiny extensively, this is probably an obstacle for the median survey researcher.

I'm hoping that scale developers will choose to implement and distribute their scales in this format regardless, and am trying to build tools to support this more broadly, and supporting surveydown is just one of the ways I'm doing this. I just wanted to make you aware of this, and note that I'd be happy to accept any updates to the conversion script (You can fork the script and include it in your own repo), and if you would find it useful for your users to link to the project/website. Also, since the .osd format is documented and fairly simple, it may be possible for surveydown to work directly with .osd; that is, have a generic runner that loads the .osd (a json format) file and implements a survey, instead of building/converting into a surveydown-based survey. This might give more control than the conversion approach I've taken. I'd be happy to discuss this or other possibilities if you would find it useful.

Shane Mueller
Professor of Psychology
Michigan Technological University

[jhelvy]

Shane, this is amazing, it's super exciting to see this level of integration with fields like yours emerging. I've got a few conferences coming up that are going to take my attention, but sometime over the summer I'll try and take a closer look at this to see what I could possibly do to further improve the integration.