Log File Analysis: Predicting Success From Online Trajectories
Document Type
Presentation
Publisher
American Educational Research Association (AERA)
Publication Date
1-1-2007
Abstract
Log files are a sequence of behavioral data stored in a permanent file (Hulsof in press). They offer time-stamped (sometimes location-stamped) records of human choices made through computer interfaces with learning environments. Such files result from online Internet navigation (stored server-side or as cookies client-side), online and console-based game play, and the use of computer-based modeling and writing tools, and from avatar-based communication and interactions. Log files retain a record of process and thus afford researchers interested in thinking and learning a rich source for assessment and design research decision-making. Log files are used for market research (as when Amazon.com selects other book choices to present to you based on your log file of book-buying choices) and for customization of advertising in many e-commerce settings. Educational research is just starting to tap into this potential.
We propose a symposium in which panel participants from 5 research Universities will be asked to present the challenges faced by their specific form of log file data: Dede (logfile data from River City); Barab and Ingram-Goble (Bot data from Quest Atlantis); Puntambekar (log file data from COMPASS); Young and Judd (log file data from Jason Online and Jasper); Zheng and Penumarthy (China-US chat logs). Compiling multiple examples of the form and content of log files will provide a first step to systematically address the process of log file analysis. Researchers will be asked to exchange data sets in hopes of checking the extent to which analysis procedures can generalize outside their respective projects. This should introduce a discussion of the shared approached to log file analysis that are developing and identify the needs for collaborative development of methods. The chair will encourage audience and participant questions and dialog.
Recommended Citation
Clarke, J. & Dede, C. (2007). Log File Analysis: Predicting Success From Online Trajectories. Paper presented at the American Educational Research Association, Chicago.
