Interactional analysis in anthropology: A videodisc.

The visual data on the accompanying videodisc are the data from which social scientists have generated published research.

We undertook this videodisc archive project for a number of related reasons:

1. Published research reports that have used visual data cannot include the data on which the research was based. This makes it impossible to replicate the research using the same body of data.
2. Since the data is not visible to the reader, the researcher's manipulations of the data are often confusing or even incomprehensible. Ideally the reader of a research report should be able to "see" the methods applied to the data.
3. A behavioral sequence or relationship may be simple to recognize when it is pointed out on film or videotape. But when the reader must reconstruct that sequence or relationship from a written description, his reconstruction is necessarily personal and subjective. This seriously impairs the "hard science" requirement that a unit of data be the same to everyone--researchers and readers.
4. This concern is even greater when the sequences or relationships are below a threshold of explicit awareness and require some form of "magnification." Also, the reader may be understandably confused when the unit of concern is the organization of relationships and not the familiar "things" or reified abstractions of other kinds of research.

We came to believe that making this collection of visual data available, along with the published research, would move our kind of research away from its neglected step-child status and toward the recognition of this research as conforming to the canons of hard science in which the data is recorded and shared and available for replication. We hope that the availability of the data alongside the methods and findings will help students and colleagues to recognize the value of research in which our data "realities" are (hard) visible behavior and not (soft) socially constructed abstractions.

There have been two methodological approaches to the use of recorded visual data in behavioral research. The first and older approach looked for occurrences of labelled behaviors or events that the researcher understood to have significance in human affairs. Such behaviors for example, be expressions of named emotions in the face or body. the occurrence of words or semantic configurations in speech, or even interpersonal or relational events such as smiling-at, touching, responding to, agreeing with, interrupting, etc. These behavioral might be might, ingeniously conceived but they were socially constructed or scientistconstructed units that were "validated" by consensual agreement of panels of colleagues and then identified in the data. This older form of research found the labelled behaviors it looked for whereas the research related to this disc labelled what it found.

This earlier form of research studied human affairs in terms of the prevailing "common sense" or scientific assumptions about humans and their interpersonal or social behavior and in terms of the prevailing methods and scientific interests. This kind of research described complex webs or chains of causation. Because human behavior is always somewhat unpredictable (in contrast to the assumed predictability in the non-living universe), this form of behavioral research was necessarily "soft."

After World War II a new and different form of behavioral research began to emerge. Instead of studying events or behaviors as units with either causative or effect significance, the search turned to the study of the organization of relationshi~ and the processes by which such relationships were implemented, maintained, or changed. This research did not test theories nor develop theory but it described a relational infrastructure on which to develop theory in areas of special social concern. This kind of research does not begin with a "problem" but it offers descriptions or maps of human interpersonal processes that are useful in clarifying the "problems" that arise in other fields of research.

Fred Ericson's "Gatekeeper" research (part of his data record is on the accompanying disc) is one example of how the "microanalysis" of visual data can illuminate an important social Ifproblem" by making explicit an unintended and unwitting process that emerges, in social terms, as discrimination.

The emerging sophistication of both recording and analytic technology made it possible to discover that human behavior and human relationships are organized more subtly (and more precisely) than our unaided observations had recognized. This is analogous to earlier discoveries about the organization and the processes underlying human physiology or the processes we now study in the atomic or sub-atomic realms.

In the behavioral science that used visual and sound records the methods were often called "microanalysis." Although this "microanalysis" did not study preconceived event-packages and was concerned only with observable relationships that were unrecognized without the equivalent of magnification, the explanatory power of this kind of research was easily recognized by those who set it alongside important social questions.

All the research that emerged from the analysis of the data archived on the accompanying disc was carried out in terms of the second, systemscybernetic-structuralist paradigm that begins without a priori assumptions, no labelled abstractions, and describes only patterns of relationship--the organization--of interpersonal human behavior in specified contexts. The value of this non-content research emerges when it is subsequently applied to particular "real lf human questions. And the value of this kind of research has another dimension: it will be applicable to questions that we will come to ask in the future.

When this project began, we envisioned a fully "interactive" disc with explanatory text overlays, multiple choices for individual management of the data stream, split screens for comparisons--all in designer colors. We proposed to develop a new form of science report literacy and publication. We imagined a form of "science report" that would allow the reader to grasp the methodological procedures and to try his own data manipulations. The technology is available for this and it could surely change the relation of the "scientific" to the "non-scientific" worlds by bringing the non-scientific world into the laboratories--the kitchens of science--without the barriers that language erects for the non-scientist.

There already are elaborately constructed interactive videodiscs designed to teach science subjects. We envisioned discs that would allow interactivity, i.e. participation in the research process.

When we explored our visionary ideas in the real world we found three insurmountable problems. The cost of assembling and affording the multiple skills (scientists, presentation designers, programmers, technicians, publisher-distributors) was beyond even our dreams of support. Hardware compatibility problems separated the scientist from the producers and the end product from the users, severely limiting the usefulness and/or availability. And, most of all, most science reviewers did not share (or understand) our vision, believed that the present forms of scientific communication were best, and were not friendly to the idea that outsiders would snoop through their arcane procedures.

So we settled for a simple straightforward data archive that would fit within the SO-minute frame of a single double-sided disc. Almost any videodisc player will do frame searches and simple forward-backward, slowfast, stop or frame-by-frame play. The appropriate frame codes and timing codes were included in the disc to permit these manipulations.

The choice of the data we included is not ideal. We would have liked a wider disciplinary representation and a sampling that would recognize much more of the excellent work using visual data. Nothing of Gregory Bateson's work is here, for example, simply because of the time and procurement logistics. Our selection, in the end, is personally biassed. I, for example, insisted on including two sets of still photographs because of my own early use of still photography and because I believe that still photography is a much-neglected and little understood source of visual data. The idea of being representative or fair to the field gave way to the press of time and convenience. In one instance (the Rizzo-Tolk data) we included the data of a graduate student who produced a fine dissertation using the videodisc and player for her analysis. Three others of us also intend to do our own forms of analysis on the same data--as a demonstration.

There are awkward mistakes on the disc. We discovered too late the unpleasant jitter on Adam Kendon's Australian aborigine woman telling a story in sign language. Even so the lab still can't explain it. One of my still segments is out of sequence. One original reel-to-reel tape lost contrast through improper storage. The transfer from 24-fps film to the 30-fps of the disc works fine for viewing purposes but the intermittent frame duplication ("pull down") tends to confuse frame-by-frame work.

We believe that this disc and the accompanying "documentation" will be quite useful as an example of a new form of data availability that will eventually take on more importance when there is more standardization in hardware and when microcomputers can be linked to erasable CD-ROM and laser encoding.to those many instructors who include something of "film analysis." This videodisc should also be useful for instructors of communication or human interaction courses who have not, themselves, done such research. It will also be useful to those of us who would like to show students examples of research with data other than our own.

We also believe that this data disk will allow students to discover that this kind of research is not as difficult and complicated to do as it is to read about. And I hope that those who use this disc in classroom teaching will allow students to explore the data individually. Given the opportunity (after "getting the hang of it. ") students almost always see patterns or relationships that we didn't see.

We hope that this disc will encourage a new generation of students to become interested in the organized underlayers of the world they live in.

Our interest and enthusiasm for the idea of this disc would have come to nothing without the understanding, interest. and support of Lita Osmundsen and the Wennner-Gren Foundation for Anthropological Research. We are grateful to them and hope that the use of this disc will rekindle the research community's interest in this form of behavioral research.

We are also grateful to those friends who agreed to support this idea by sending us copies of their data and doing the tedious work of keying their published research reports to the frame numbers on this disc.

An unexpected discovery we made when we received our first discs from the lab: a videodisc is not only our best state of the art archival medium but with the hand-held remote control, the videodisc player far surpasses either the multi-speed film analyzer or the VCR as an analysis tool.

I personally believe that behavioral science research and its significance to the extra-academic world will be revolutionized when the capabilities of the fully interactive videodisc are as available and affordable as personal computers. At that point we might expect to find high school students doing significant behavioral science research.

Paul Byers