For Science!
Oh, man. Has anybody else noticed that we’ve had a truly excellent run of MiniLD themes lately? I’m in, and on time too — I started thinking seriously about things this morning, but didn’t make a post until now. I’ve got something on the order of seven million ideas for this one, ranging from the totally reasonable to the downright Vista-esque, but instead of actually fitting the theme properly I’m going to engage in a little quick-and-sly reinterpretation. It may not be educational, but it will be for education! I present a ludicrously rough version of PopCat!, a collaborative offline-multiplayer wordgame:
A bit of backstory (warning — academic-speak to follow): I’m working on my PhD in cognitive science, and my research revolves around something called categorization — the process by which people recognize objects and group them into logical categories. I’m trying to build a computational model of this using only words, but there just isn’t any data to be found. So far I’ve been getting folks on Mechanical Turk to fill in dumb little surveys, but maybe, just maybe, I can make the task fun enough (and competitive!) that people will produce my data for free. Maybe.
/end academic speak
Anyway, the game is something like Scattergories played in reverse; you’re given a bunch of words and have to come up with a category for each one. There’s a timer, so you have to think fast, and you get points for each category proportional to how many other people came up with the same category. The idea is to make it something of a mind game — do more people thing “tomato” is a fruit or a vegetable? Which is worth more?
The game is in Flash and Flixel (well, B-Flixel specifically), while the data server is in PHP. Joy.
Here’s something that might be interesting to keep in mind if you want to use the data academically:
Generally I’d think of tomatoes as vegetables, but if I’m typing and there’s a timer I’d probably say it’s a fruit.
Snap decisions are exactly what I’m after — in ivory-tower-speak this is a Speeded Categorization Task. If I’ve thought things through correctly, the conditions of the game should be the same as in a lab study. Well, more or less — few cogsci labs feature Kongregate integration. But hey.
The nice thing is that the core (show a word, get a word, give feedback, repeat) is incredibly easy, leaving me plenty of time to polish it all up in a nice little package. And if a mere 400 people play just one round before quitting (suckered in by pretty visuals) I’ll have collected about $35 worth of data.