ABSTRACT
We start the chapter by presenting the premises from EAST2 and the corresponding research questions motivated by these premises, given the substantive context of the study. We then describe the experimental data strategy design used to collect data to examine those research questions. An experimental data strategy design was used in this study because we were interested in the “process interaction” at a fine level of granularity. We adopted a decision task that concerns fish and wildlife habitat redevelopment site selection in an estuary in the City of Seattle Washington. The chosen task is modeled after a real task performed by a multi-stakeholder group convened by the National Oceanic and Atmospheric Administration. For experimental treatments we varied the task in terms of complexity, i.e. fewer sites and fewer criteria, versus more sites and more criteria. Video-tape recordings of the interaction were made and then coded with three coding systems: decision functions that emulate task functions, decision aid structures that describe the software tools, and group working relations that describe conflict. The coding was performed for group interaction at the level of “group attention” with PGIS-based collaborative decision aids. Based on the raw codings we describe the different ways that data were prepared for analysis. We then present our data analysis and the findings associated with each approach, comparing them on a premise by premise, hence research question by research question basis. In the final section we draw conclusions about experimental designs and analysis strategies used to examine humancomputer-human interaction during PGIS use.
