Testing Sleep Consolidation in Skill Learning: A Field Study Using an Online Game

Abstract

Using an observational sample of players of a simple online game (n > 1.2 million), we are able to trace the development of skill in that game. Information on playing time, and player location, allows us to estimate time of day during which practice took place. We compare those whose breaks in practice probably contained a night's sleep and those whose breaks in practice probably did not contain a night's sleep. Our analysis confirms experimental evidence showing a benefit of spacing for skill learning, but fails to find any additional benefit of sleeping during a break from practice. We discuss reasons why the well established phenomenon of sleep consolidation might not manifest in an observational study of skill development. We put the spacing effect into the context of the other known influences on skill learning: improvement with practice, and individual differences in initial performance. Analysis of performance data from games allows experimental results to be demonstrated outside of the lab, and for experimental phenomenon to be put in the context of the performance of the whole task.

Game‐XP: Action Games as Experimental Paradigms for Cognitive Science

“…For web‐based or cellphone‐based games, it may be possible to obtain a complete record of skill acquistion by mining Big Data sources. Big Data: Many of the most popular games are web‐based or cellphone‐based with the side effect that much or all of the interactions, decisions, and keystrokes made during the game are available in datafiles on the internet. Hence, researchers can obtain Big Data (Griffiths, ) and/or naturally occurring datasets (NODS, see Goldstone & Lupyan, ), which contain hundreds of thousands or millions of records (e.g., Huang, Yan, Cheung, Nagapan, & Zimmermann, ; Sangster, Mendonca, & Gray, ; Stafford & Haasnoot, ; Thompson, McColeman, Stepanova, & Blair, ). Joint Action and Teams: It seems fair to say that, outside of studies of language, the cognitive revolution has not greatly influenced the study of people in cooperative or competitive settings. In recent years, Joint Action (e.g., Knoblich, Butterfill, & Sebanz, ; Sebanz & Knoblich, ) has emerged primarily as the study of interactions between two humans.…”

“…Unlike other skill development domains—for example, spoken language, playing the violin, soccer—each action taken during a game, is conducted through a computer and so may be unobtrusively recorded. (Stafford & Haasnoot, )…”

Female Chess Players Outperform Expectations When Playing Men

“…Experimental studies encourage us to focus on isolated causal factors. Observational studies encourage us to see all factors in the context of other factors (Stafford & Haasnoot, 2017). Observing a phenomenon 'in the wild' provides a strong validation of the generality and robustness of an effect.…”

Allen Newell's Program of Research: The Video‐Game Test

“…Four of the eight papers (Huang, Yan, Cheung, Nagappan, & Zimmermann, ; Stafford & Haasnoot, ; Thompson, McColeman, Stepanova, & Blair, ; van der Maas & Nyamsuren, ) used video‐game archival data. The analysis of large databases for studying cognition, which was not anticipated by Newell, magnifies some of the problems he identified (in particular, the peril of averaging across strategies and tasks), but also provides means to address them (e.g., using sophisticated statistical and data‐mining techniques for identifying strategies).…”