InteractionAnalysis

The following steps outline a process of analyzing collaborative interactions for use in the VMT Project:

Design the Collaborative Situation

It is important to plan the interactions to make it likely that the kinds of collaboration that you are interested in will occur, will be observable by you and will be recorded by your equipment. This may mean simply selecting likely settings or it may mean structuring the events that will take place. For instance, your subjects may not be used to collaborating and may attempt to carry out tasks individually whenever possible, so that you must structure the situation to minimize that and to encourage or force people to engage in the kind of collaboration that you want to observe.

Record the Interactions

It is important to capture both visual and audio with as much detail as possible

Use directional mikes and PMZ mikes on tables or personal mikes worn by individual subjects (Alan Zemel has some more advanced MicTips

Use wide-angle on video recorder to capture as much of the environment as possible

Use tripods to stabilize cameras and make them less intrusive

Use high tripods to get viewing angle down on work area

check that batteries are fresh and volume controls are turned up

Position subjects and camera(s) to get clear view of each subject's face and bodily gestures (e.g., to identify speakers)

Position cameras to capture any work done on table or paper -- and any gestures pointing to or manipulating the papers

Position cameras to capture any work done on computers -- and any gestures pointing to the computer screen or use of mouse or keyboard

Collect any documents created; photograph any whiteboard diagrams; archive any computer logs

See (Jordan & Henderson, 1995) for more suggestions

Preserve the Recordings

Make backup copies immediately, before anything can get lost

Store data in a secure place, in accordance with IRB agreements

Add timecode where appropriate, to allow for synchronization and coordination of multiple sources

Remove identifying information about the subjects, in accordance with IRB agreements

Log the Recording

Create a log that describes in a sentence or two what takes place every minute of the recording

Index the log to the timecode, etc. for the recording(s)

Identify Interesting Moments

Identify natural episodes in the entire recorded period and note them as headings on the log

Generally, the episodes should be divisions that the participants would recognize as distinct

Select a small and manageable number of moments that seem to have collaborative interactions of the sort you are interested in analyzing and mark natural boundaries for these moments; these should be relatively self-contained interactions lasting on the order of a minute or two

Transcribe Interesting Moments

Prepare a Jeffersonian transcription of each interesting moment

See (Psathas, 1995) for a detailed guide to Jeffersonian transcription

Start with a simple transcription of the words said

Work from a digital recording on the computer so you can easily loop and repeat and slow-motion play to make sure you are correct on who is saying what

Add pauses, with number of seconds in parentheses (.3)

Add false starts and overlaps to represent the basic flow

Add emphases and most important gestures

More detail can be added later as necessary for the analysis

Software is available to support the transcription process and sortable storing of transcribed clips, e.g., the open source, multi-user system at http://www2.wcer.wisc.edu/Transana/Transana

Hold a Data Session

Gather a number of people -- perhaps a half dozen -- to observe one or two episodes

Hand out the log for that section of the recording to provide context

Describe the setting, etc.

Hand out transcript for the episode(s) to be observed

Play the digital video/sound several times, repeating sections upon request

Perhaps suggest a general interpretation

Have the group interpret the interaction until they are satisfied that it accounts for all the data

The collaborative construction of an interpretation by the group will bring several perspectives to interact with the data, with suggested lines of interpreation and with each other, producing a robust questioning of the data under an evolving interpretation

Interpret the Transcript

Write up a fairly comprehensive interpretation of the data, tied to the transcript

Build on the insights of the group data session(s)

Identify patterns in the data and developments of changing discourse

See (Duranti, 1998) for discussion of some typical discourse patterns

Provide arguments for all interpretive claims, based on evidence in the transcript

Refine the transcript to show relevant detailed features, such as gaze, bodily positioning, gesture, intonation

Hold Another Data Session

Hold another data session -- or perhaps several of them iteratively

Make sure that the interpretation of the data holds up to the group's review, meets the standard of general plausability

The collaborative data sessions are essential to the methodology; they are a form of group construction of knowledge

They ensure that the interpretation is intersubjectively acceptable; they are the analogue of inter-rater reliability in quantitative analysis, avoiding effects of an idiosyncratic researcher perspective in addition to bringing the benefits of group cognition

Review the Interpretation with Participants

Include some of the participants, such as the teacher, in a data session

Use this to confirm the reasonableness of the interpretation

Use this to solicit other possible interpretations

There is no ultimate judge of the correct interpretation; intersubjectivity and evidence in the data make an interpretation more plausible; plausibility is the only judge

Evidence of how the participants took something in the original interaction is particularly important; statements by participants after the fact (in interviews or data sessions) may be helpful, but must be subjected to interpretation and analysis themselves and not taken at face value

Analyze the Collaboration

Write up a formal analysis of the collaboration

See (Stahl, 2002) for an example

Show this to several people familiar with the data and revise the analysis in accordance with their comments

Iterate

The whole discourse analysis process is an iterative one, with multiple refinement to the transcript and even to the identification of the interesting moments and their boundaries

Go back and see how the analysis of moments fits into the understanding of the entire recorded event

Perhaps even change the definition of the research questions that you are pursuing

Design new events to observe and analyze that might shed light on questions or puzzles raised by this analysis

Publish Findings

Present your findings in professional conferences to get feedback from a larger audience that shares your research interests

See (Garfinkel, 1965) for examples of publications based on the ethnomethodology approach to interaction analysis

See (Sacks, 1992) for examples of publications based on the conversation analysis approach to interaction analysis

See (Streek, 1983) for examples of publications based on a microethnographic approach to interaction analysis stressing bodily gesture among children in a classroom

References

Duranti, A. (1998). Linguistic anthropology. Cambridge, UK: Cambridge University Press.

Garfinkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall.

Jordan, B., & Henderson, A. (1995). Interaction analysis: Foundations and practice. Journal of the Learning Sciences, 4(1), 39-103. Retrieved from http://lrs.ed.uiuc.edu/students/c-merkel/document4.HTM.

Psathas, G. (1995). Conversation analysis: The study of talk-in-interaction. Thousand Oaks, CA: Sage.

Sacks, H. (1992). Lectures on conversation. Oxford, UK: Blackwell.

Stahl, G. (2002). Understanding educational computational artifacts across community boundaries. Paper presented at the International Society for Cultural Research and Activity Theory (ISCRAT '02), Amsterdam, NL. Retrieved from http://www.cis.drexel.edu/faculty/gerry/cscl/papers/ch03.pdf.

Streeck, J. (1983). Social order in child communication: A study in microethnography. Amsterdam, NL: Benjamins.

Related Research Methodologies

Ethnomethodology

Conversation Analysis

Microethnography

Ethnography

Discourse Analysis

Grounded Theory

Participatory Design

Action Research

Participant Observation

Design Research

Social Network Analysis

other

Acknowledgments

Alena Sanusi and others around the CU-Boulder Communication Department in 2000/2001 introduced me to most of these steps