gameswithwords

The model doesn't do a good job of telling apart the various Chinese languages (yet), so we aren't including that distinction in the feedback (yet).

That's right. In our data, Standard American, Canadian, & AAVE are all pretty similar. Partly that's a function of the questions we ask. But partly that's because there's a clear divide between the north american dialects and the UK-based ones.

This has to be best best comment I've seen about the project yet! You mind if I use this as an example for my students? (I mean, I will anyway. I just want to know if you mind.)

When we're done with the project, we'll put up a final version that actually gives you a breakdown of why you get the score you do, what parts of the world certain grammar rules are particularly common -- all the things people (including me!) want to know.

We're actually getting some of that up bit-by-bit as data comes in (you can watch for it on our site blog). Thank you to everyone who has been participating so far.