Software

Delibera 1.0

General Aims

The software product created specifically with NSF funding to support the experimental side of the VAP was called Delibera.  Its production was managed by Carnegie Mellon University’s Institute for the Study of Information Technology and Society (InSITeS), under the technical leadership of Stuart Easterling, with primary supervision by Robert Cavalier.  The philosophical aim of VAP software creation was helping people understand informational materials through rich visual language and the opportunity for well-structured conversation. This was to be accomplished both through carefully designed computer mediated communication and through document design and access. The first goal was to create an immersive online environment utilizing good human-computer interaction principles and audio /video capability in order to make the program accessible to all participants, even those who are not tech savvy.  If successful, this interface would help place the user in a “virtual agora,” a marketplace of ideas where they would experience conversations with others in a high telepresence setting. These goals have been pursued throughout the development stages of all our software efforts.

Features

Supporting the experimental goals of the VAP required a largely intuitive and user-friendly online environment for both synchronous and asynchronous online deliberation.  At a high level, Delibera’s features can be understood under six headings:

Delibera was also programmed to permit researcher collection of clickstream data, permitting analysis of such phenomena as reaction time, emoticons used, sequence of pages visited, and the like.

Performance

As explained under Social Science, the VAP used Delibera in two very different environments.  During Phase I, it was used in a tightly controlled on-campus environment, where all users participated with broadband ethernet connections and students were available to tutor participants in the use of both hardware and software.  Delibera gave satisfactory and reliable performance under these conditions.

During Phase II, Delibera was deployed in a far more challenging environment.  The eight-month experiment in at-home online deliberations was intended to include 410 of the original 568 participants who were selected to receive a computer. These users differed substantially in technical experience and in their at-home communications environment, with some using broadband connections and others using much slower dial-up. 

At this stage, Delibera did not perform as seamlessly in supporting synchronous on-line audio conferencing.  It had to be modified repeatedly for increased robustness against user error, and had some difficulty managing audio traffic over different bandwidths within a single conference.  Unfortunately, these problems coincided with occasional short-term electrical outages at Carnegie Mellon and subsequent server problems unrelated to the software, all of which required occasional modifications to the planned sequence of real-time online meetings.

During Phase II, the VAP experienced a significant drop-off in participation.  Response rates to questionnaires dropped to 230 in early stages. Nearly two-thirds of the participants who withdrew formally contacted us, with roughly half stating specific reasons for their withdrawal.  Of those participants who offered reasons, only one specifically named “software problems,” although the performance of Delibera could have been a factor for 18 other participants who mentioned “equipment problems” or “inadequate support.”  We do not have any more systematic evidence that would enable us to attribute the overall drop-off more definitively to any particular factor, whether related or unrelated to the quality of software, experimental design, or the level of incentives provided for participation.