| This dissertation studies how people in a household setting learn to use a novel TV-based Internet appliance. Past research on the problem of how users learn to operate computers and software has been rooted in psychology and problem solving (Polson & Lewis, 1990) and the unit of analysis has been the individual learning in isolation. While a framework of exploratory learning has been developed (Rieman, 1996), the present work extends this framework to better account for multiple people learning a user interface together.; The result is called Joint Exploratory Learning (JEL), which includes types of learning events and JEL components. In an analysis 2404 lines of transcription, 154 learning events were identified as locations where household member(s) clearly learned something about how to use the system. These events were coded with JEL components that describe the basic structure of the learning event. They occurred singularly, such as with Try (trying something and learning from it) or in combination with other JEL components, as in Ask-Tell (asking someone and getting an answer). The frequency of JEL components and their combinations in learning events was calculated, and Try accounted for 60% of all observed learning events. Other notable JEL component combinations include Suggest-Try, Try-Breakdown; Tell; Ask-Tell; and Reason. The interrater reliability of the learning event coding scheme used was calculated as 0.732.; One finding is that learning emerges (Suchman, 1987) as a by-product of interaction of household members with the technology or with other people present. Socially oriented JEL combinations include Suggestions followed by a Try or an explicit Ask of a question followed by a Tell of the solution. Another finding is that breakdowns (Winograd & Flores, 1986) play a significant role in joint exploratory learning because they cause a shift of attention from task to tool, corroborating other research on the role of error in learning (Seifert & Hutchins, 1992).; The joint exploratory learning framework developed here could be reliably used to better understand the nature of how people conceive of particular designs and what makes a good design work, potentially leading to specific ways of improving designs. |
