*This talk is part of the ENGAGEMENT & MOTIVATION track*
Citizen cyberscience projects like the physics game Quantum Moves have the potential to generate an extraordinary amount of data.
As researchers, we have the opportunity to study both the “core” contribution of players as well as the psychological factors contributing to motivation and performance in-game.
In this talk, we present the core design and educational theories behind our second www.scienceathome.org game PsychLab (working title), made to crowdsource not statistical calculation but causal inferences about relationships within our jumble of data from Quantum Moves.
Many statistical tasks, like finding correlations between player performances and psychological traits, can be carried out near-automatically, but computers will never be able to suggest meaningful causal relationships, or do its own data collection based on new causal hypotheses. Also, the near infinite number of potential relationships make intractable to try out all combinations on a computer and therefore necessitates human selection.
Players build statistical path models by organizing nodes and boxes hierarchically with hypothesized causal connectors. A fun and friendly graphical interface shields users from the mindboggling calculations going on in the engine room, and tickles the human “puzzle instinct”.
"PsychLab" will be an invaluable tool for processing a lot of data, but also an opportunity to teach statistical thinking (linear models, path models, correlation, causation, moderation, mediation, t-tests, ANOVAs, different types of data) at a simple conceptual level, allowing everyone in our community to take part of those scientific processes, that usually go on behind the scenes.
Down the line, we envision PsychLab as an open source tool for our colleagues in the citizen cyberscience community.
**For the record, this is a ridiculously short presentation slot, so please find us during the summit, if you want to hear more.**
20th–22nd February 2014