Interaction coding systems are used to assist in the recovery of data that describe decision aid use. An interaction coding system is essentially a set of keywords that summarizes the character of a process and the mechanisms used in that process from a particular perspective. A suite of coding systems is described in this paper.
The coding and analysis approach to research goes by several names in various disciplines. In social science research it is generally known as content analysis (Krippendorf 1980). In psychology, cognitive science and computer science it has been called pr otocol analysis (Ericsson and Simon 1984). In social psychology, sociology and speech communication this process of analysis is called interaction coding analysis (Poole and Roth 1989a). Commonalties across disciplines is indicative of the general applicability for social-behavioral research. Coding systems have been used in research on advanced technology to summarize research data captured about small-group decision interaction (Poole and Roth 1989b). The data is used to identify patterns and the impact s of decision aids on these patterns of decision making. Software tools are available to assist with coding of textual transcripts, and have even been developer for coding data directly from audio-video tapes (Sanderson, James, and Seidler 1989, Roschelle and Goldman 1991). Analysis of the data can be performed using qualitative and quantitative statistical techniques.
The interaction coding systems discussed in this paper are being developed as part of a study of the impacts of computer-assisted decision aids on collaborative spatial decision making (Nyerges and Jankowski 1994). The research design for this study makes use of a laboratory setting in which geographic information systems and multicriteria decision models are the core decision aid technologies being used by small groups of decision makers.
The decision making task has been taken from a realistic decision context dealing with wildlife habitat site selection in the Duwamish Waterway estuary of Seattle, Washington (National Oceanic and Atmospheric Administration 1994). Habitat site selection is growing in importance as environmental analysts are faced with choosing an appropriate balance of financial and human resources. The habitat site selection problem is viewed as consisting of three types of tasks -- a creativity task, a preference task and a cognitive conflict task (McGrath 1984). The creativity task involves sorting out basic issues about what attributes are important as part of the site character, and forming a list of potential sites. The preference task involves selection of a set of attributes that form the criteria for choosing; these criteria becoming the basis of the site selection process. The cognitive conflict task involves group members sorting through the collection of criteria, each member presumably having a different pers pective for wanting particular criteria prioritized over others, and for selecting certain sites over others.
Site selection can be viewed as a decision process that involves conflict management (interaction), due to the different perspectives inherent in the views of participating members. A conflict management activity is composed of two subactivities called id ea differentiation and idea integration. Both idea differentiation and idea integration are essential to resolving differences in viewpoint and reaching a consensus (Walton 1969, Sambamurthy and Poole 1992, Poole, Holmes and DeSanctis 1991).
Idea differentiation involves extracting information on characteristics of sites, and making distinctions among the attributes that characterize sites, hence differentiating the sites themselves. Idea differentiation also involves a contrast of the decisi on strategies that are used, one being favored over another to derive the solution in site selection.
Idea integration involves a process of synthesizing the attributes to establish the criteria to be used in making the decision. Idea integration involves following through in the making of the decision by applying a decision strategy. Such a decision stra tegy could be carried out in either a manual or computer-assisted mode, and should not affect the basic development of a suite of coding systems used to capture the character of the process.
1) decision functions coding system (DFCS) - for summarizing decision phases,
2) group working relations coding system (GWRCS) - for summarizing group social interaction,
3) decision aid coding system (DACS) - for summarizing the kinds of decision aids invoked,
4) aid appropriation coding system (AACS) - for summarizing the manner in which decision aids are invoked during the decision process, and
5) decision aid use coding system - for summarizing how each aid is being used for information interaction.
Each of these is explained in more detail in the following sections.
The DFCS consists of several activity categories. The keywords with their associated phase type codes (in parentheses) are presented below.
1. Problem Activity
1a. Problem Analysis (PA) : Statements that define or analyze the problem
facing the group 1b. Problem Critique (PC, with + or -): Statements that
support or criticize a problem analysis
2. Execute Activity
2a. Orientation (OO): Statements that direct the group's process or help
the group to do its work.
2b. Process Reflection (PR): Statements that comment on the group's
process or progress
3. Solution Activity
3a. Solution Analysis (SA): Statements that define how the group will go
about developing its solution in general terms, including criteria and
general directions
3b. Solution Design (SD): statements that propose solutions
3c. Solution Elaboration (SEB): statements that alter or amend solutions
3d. Solution Evaluation (SEV, +,-,/): statements that support (+),
criticize (-), or offer evaluation (/) of solutions. 3e. Solution
confirmation (SC, +, /): statements that ask for confirmations (+) or
votes (/) for final confirmation of decisions.
4. Non-related Activity
4a. Tangents (TA): moving to an unrelated subject
4b. Disorganization (DI): Disorganized or nonfocused discussion.
The GWRCS consists of several categories and associated codes as follows.
1. Work Focused Relationships
1a. Focused Work (FW): Periods when members are task focused and do not
disagree with one another
1b. Critical Work (CW): Periods when members disagree with each other, but
the disagreements are centered on ideas and no opposing sides have been
differentiated.
2. Conflict
2a. Opposition (OP): Periods in which disagreements are expressed through
the formation of opposing sides.
2b. Accommodation (AC): A mode of resolution of opposition in which one
side gives in.
2c. Tabling (TAB): A mode of resolution of opposition in which the subject
is tabled or dropped.
2d. Open Discussion (OD): A mode of resolution of opposition that utilizes
problem-solving discussions, negotiation, or compromise.
3. Integration
3a. Relational Integration (RI): periods when the group is searching for
task-focus, but may wind up on tangents, joking or distracted.
1. Descriptive site map (DM). Site locations and names.
2. Rankmap (RM) - displays ranks of the sites
3. Graduated Circle map (GM) - shows at tributes via graduated circle
4. Bar Map (BM) - shows attributes using bars as in a histogram
5. Choropleth map (CM) - a gray scale shaded map
6. Orthoimage (OI) - shows area using photo image
7. Table Display (TD) - a table of attribute information
8. Graph Display (GD) - a graph dis play to show relation
9. Text help (TH) - text help explains a particular software capability or
data category
1. Aggregation method choice
1a. weighted summation (WS) aggregation method - the familiar rate and
rate approach
1b. aspiration level (AL) aggregation method - sets a level of attainment
and telss the user how close to this level has been achieved.
2. Weight method choice
2a. aspirations (AS) - sets a level of attainment for an attribute
2b. pairwise comparison (PC) - each attribute is compared against every
other for preference
2c. ranking (RK) - assign on a scale of 1 to 9 2d. rating (RA) -
proportion 100 points across all criteria
1. non-rank vote (NRV - simple majority of yea answer for vote
2. rank vote (RV) - adds the ranks (1 to 9) using Borda social preference
1. direct appropriation (DIR) - represents active use of a single decision
aid, i.e. software capability and/or organizational guideline for a
cognitive task;
2. substitution (SUB) - one decision aid replaces another to carry out the
cognitive task.
3. combination (CMB) - two decision aids are melded together to carry out
the cognitive task.
4. enlargement (ENL) - two decision aids are compared to eachother, but
only one may be used to carry out the cognitive task.
5. constraint (CST) - constraint attempts to interpret and understand a
single decision aid in light of the cognitive task.
6. contrast (CNT) - two decision aids are placed in opposition and one is
chosen to carry out the task.
7. affirmation (AFF) - represents the positive modes of response to
others' appropriations for carrying out a cognitive task
8. negation (NEG) - represents the negative modes of response to others'
appropriations for carrying out a cognitive task
9. ambiguity (AMB) - Ambiguity represents uncertainty and confusion for
what should be used to carry out a cognitive task.
Rather than develop two coding systems for decision aid use, one for maps, and one for decision models, a single, synthesized coding system for decision aid use has been devised. Literature for map use operations includes basic reading tasks (Board 1978, 1984, Morrison 1978) and analytical operations (Nyerges 1991). Literature for the MCD model perspective includes individual decision making (Payne 1982), individual decision support system use (Todd and Benbasat 1992), an overview of group-based MCD appro aches (Hwang and Lin 1984), and a particular MCD model in a GDSS (Dickson et al 1991). When DAUCS is combined with the DACS subsystems, different decision aids and their manners of use can be identified. The keywords and respective codes are as follows.
1. Acquire (AQ) information from a source and internalize it
1a. read (RD)
1b. retrieve (RE)
1c. search (SE)
1d. identify (ID)
1e. locate (LO)
1f. istribution (DT)
1g. abel (LB)
1h. define (DE)
2. Save (SV) information for later use
2a. save temporarily (ST)
2b. save permanently (SP)
3. Interpret (INP) meanings for a source of information to develop a
perspective
3a. distinguish (DT)
3b. categorize (CZ)
3c. simplify by location (SL)
3d. classify by location (CL)
3e. aggregate (locate, cluster, bind, describe) (AG)
3f. generalize by attribute (GA)
3g. classify by attribute (CLA)
3h. simplify by attribute (SIA)
3i. weight (WT)
4. Analyze (ANL) to derive or synthesize information, or change into a
different form
4a. associate (AO)
4b. cluster (CS)
4c. rank (RK)
4d. count (CT)
4e. correlate (CR)
4f. measure (MS)
4g. interpolate (IP)
4h. add (AD)
4I. subtract (SB)
4j. multiply (MU)
4k. divide (DV)
5. Evaluate (EVL) to explore information usefulness
5a. compare (CP)
5a1. within (CPW)
5a2. between (CPB)
5b. contrast (CN)
5b1. within (CNW)
5b2. between (CNB)
5c. verify (VE)
6. Judge (JU) to determine information usefulness
6a. prefer (PF)
6b. like (LK)
6c. choose (CH)
Each coding system has an emphasis, making it easier to encode data from a different perspective. The decision functions coding system describes phases in a decision process. The group working relations coding system describes person to person social inte raction as part of conflict management in a group. The decision aid coding system consisting of a map type coding subsystem and a multicriteria decision model subsystem is used to decision the kinds of aids being used. The aid appropriation coding system captures how the decision aids are brought into use by the group. Finally, the decision aid use coding system focuses on the social-cognitive act of putting the decision aids to use in various ways.
Both general and detailed hypotheses will make use of the data captured by way of the coding systems. As an example of general hypotheses, it is expected that both the MTCS and DMCS codes will be applied to idea differentiation and idea integration decisi on sequences. However, it is expected that MTCS will be applied more during coding of idea differentiation and DMCS will be applied more during coding of idea integration. In addition, data observations are used in some of the more than twenty variables t hat have been identified to describe the CSDM processes.
Whether development and use of coding systems is the appropriate way to perform behavioral research on decision making is still being debated. It is accepted as an approach to systematically characterize a decision process. However, the concern is whether it does so artificially. Do users of technology really make decisions in the way that they are coded? Or, are decision makers and users of software technology interacting in a way different than what is being summarized. One way of assisting with insurin g valid data is to: 1) video tape decision makers, 2) ask them to comment on what they did and tape record that, 3) code the video tapes, and 4) use the audio tapes as a quality control check on the coding.
The interaction coding process is the core of a social-behavioral science investigation of decision making. There are several approaches to coding and coding validation, but they are nonetheless part of a coding process with many commonalties. There are o ther levels of coding of data about small groups other than at the level of the group as a whole. For example, how do individuals in the group behave? Are there leaders and followers, and is this caused by expertise with information technologies. How do g roup decisions fit into organizational contexts as part of the decision process.
Through research experience we can come to gather better qualitative and quantitative data characterizing CSDM processes. Only with better data can we evaluate the input, process, and outcomes of the decision process in a systematic fashion. Everything el se will really be anecdotal speculation.
Board, C. 1984. Higher Order Map-using tasks: geographical lessons in danger of being forgotten, Cartographica, 21(1):85:97.
Davies, C. Medyckyj-Scott, D. 1995. Feet on the Ground: Studying User-GIS Interaction in the Workplace, in T. Nyerges, D. Mark, R. Laurini, M. Egenhofer (Eds.) Cognitive Apsects of Human-Computer Interaction for Geographic Information Systems, 123-142. Do rdrecht, Netherlands: Kluwer Academic.
DeSanctis, G. and M. S. Poole, 1991. Understanding the Differences in Collaborative System Use Through Appropriation Analysis, Proceedings, the Hawaii International Conference on System Sciences, Los Alamitos CA: IEEE Press. pp 547-553.
DeSanctis, G. and M. S. Poole, 1994. Capturing the Complexity in Advanced Technology Use: Adaptive Structuration Theory, Organization Science, 5(2):121-147.
Dickson, G. W., G. DeSanctis, M. S. Poole, M. Limayem, 1991. Multicriteria Modeling and "What If" Analysis as Conflict Management Tools for Group Decision Making, Proceedings, 11th International Decision Support Systems Conference - DSS '91, (ed.) Ilze Zi gurs, The Institute of Management Sciences.
Ericsson, K. and Simon, H. A. 1984. Protocol Analysis: Verbal Reports as Data. Cambridge: MIT Press.
Hwang, C. and M. Lin 1987. Group Decision Making Under Multiple Criteria, Lecture Notes in Economics and Mathematical Systems 281, New York: Springer-Verlag.
Krippendorf, K. 1980. Content Anlaysis: an Introduction to its Methodology, Beverly Hills, CA: Sage.
Lewis, C., R. Reitsma, and E. Zigurs, 1992. A Decision Support System for Group-based Hydrological Modeling, Research Proposal funded by the National Science Foundation, Program for Robotics, Information and Intelligent Systems. University of Colorado, CA DSwes Research Center.
McGrath, J. E. 1984. Groups: Interaction and Performance, Englewood cliffs, N J: Prentice-Hall Inc.
Morrison, J., 1978. Towards a Functional Definition of the Science of Cartography, The American Cartographer, 5(2):97-110.
National Oceanic and Atmospheric Administration, 1994. Concept Document, Elliott Bay / Duwamish Restoration Program, report prepared for the Elliott Bay / Duwamish Restoration Program Panel by the King County, Washington, Department of Metropolitan Servic es, Panel Publication 7, June 1994.
Nyerges, T. L. 1991. Analytical Map use, Cartography and Geographic Information Systems, 18(1): 11-22.
Nyerges, T. L. 1993. How Do People Use Geographical Information Systems?, in D. Medyckyj-Scott and H. Hearnshaw (Eds.), Human Factors in Geographical Information Systems,pp. 37-50, London: Belhaven Press, available now from John Wiley & Sons.
Nyerges, T. L. Jankowski, P. 1994. Collaborative Spatial Decision Making Using Geographic Information Technologies and Multicriteria Decision Models. research funded by the National Science Foundation, Directorate of Social and, Behavioral Research, Geogr aphy Program, SBR-9411021.
Payne, J. W. 1982. Contingent Decision Behavior, Psychological Bulletin, 92(2):382-402.
Poole, M. S. and J. Roth 1989a. Decision Development in Small Groups IV: A Typology of Group Decision Paths, Human Communication Research, 15(3):323-356.
Poole, M. S. and J. Roth 1989b. Decision Development in Small Groups V: Test of A Contingency Model, Human Communication Research, 15(4):549-589.
Poole, M. S., Holmes, M. and DeSanctis, G. 1991. Conflict management in a computer-supported meeting environment, Management Science, 37(8):926-953.
Roschelle, J, and Goldman, S., 1991. VideoNoter: a productivity tool for video data analysis. Behavior Research Methods, Instruments, and Computers, 23(2):219-224.
Sambamurthy V. and M. S. Poole 1992. The effects of variations in capabilities of GDSS Designs on Management of Cognitive Conflict in Groups, Information Systems Research, 3:3, Sept, 1992, 225-251
Sanderson, P. M., James, J. M., and Seidler, K. S. 1989. SHAPA: an interactive software environment for protocol analysis, Ergonomics 32(11), 1271-1302.
Todd, P. and I. Benbasat 1992. The Use of Information in Decision Making: An Experimental Investigation of the Impact of Computer- Based Decision Aids, MIS Quarterly, September 1992, p. 373-393.
Walton, R. 1969. Interpersonal Peacemaking: Confrontations and Third Party Consultation. Reading, MA : Addison-Wesley.
Tim Nyerges received his Ph.D. from the Ohio State University in 1980. From 1980- 85 we was a software project lead and consultant. From 1985-present he has been with the Univ of Washington. His teaching and research responsibilities include topics for de sign and application of GIS. Current research focus is studying spatial problem solving and decision making through human-computer interaction in the context of group decision support systems. He specializes in applications in transportation and environme ntal management. Presently, he has two National Science Foundation funded projects. One is for studying the inluences of use of computer-assisted decision aids on collaborative spatial decision making . The second is for curriculum development emphasizing teamwork with geographic information technology. He is currently a member of ACSM, ASPRS, URISA, ACM and AAG.