The purpose of the second stage of the study (and the extra CoPs) was to expand on and inform the initial insights. Glaser and Strauss (1967) observe that the fuller coverage comes at the beginning of the research as categories and themes emerge. Subsequent studies do not need such full coverage as the aim is to gather more data pertinent to the themes and categories that have emerged, generally looking for confirmation or elaboration. To this end, Glaser and Strauss point out that earlier stages will take a fuller approach, such as reading documents, interviewing, and observing at the same time. Later stages can be more focused with shorter interviews and more direct questions. This was the approach that I took with Stage Two of the study and with the extra CoPs. The Contextual Design tools were not as applicable for the second stage, as the focus was tighter than in Stage One. In Stage Two and with the extra CoPs, the focus was not on the whole work but on the use of artefacts and the importance of relationships and face-to-face communication. Additionally, Contextual Design focuses on regular work with the overall aim of designing a system to support the work. During Stage Two of the study, the regular work had been at most suspended, or at least interrupted for the period where the two cores were together. Therefore, the tools used for data gathering and analysis tended to be a mix of Contextual Design techniques and interviews.

Data Collection

The greater part of the study was spent observing the members of WWITMan in meetings and interacting together. It was not possible to follow a Contextual Interview Model with the researcher playing the role of apprentice. Rather, the observation was simply observation. However, there was also the opportunity to speak with the three established American WWITMan members, the Japanese member, and with Carol, a member of the Client Server team. In the interviews with these people, an open semi-structured interview seemed to be the most suitable approach. This took the highlighted issues from Stage One and tackled them directly but still allowed for flexibility to pursue other relevant avenues that might arise. The interview schedule (Appendix 9) was intended purely as a guide for the interview and was not intended as a script. To finish the interview, a subset of the propositions that had been presented to the UK core was presented and discussed. The interview also offered the opportunity to refer the respondents back to the meetings for clarifications and explanations. Detailed notes were made during the meeting observations, and all interviews were recorded.

Following the period spent in California, I spent a further period with the UK core to ascertain what effect the American visit had had. This period also consisted of observation and interviews that were annotated and recorded.

Analysis of Stage Two

The first stage of the analysis, as with Stage One, was to transcribe the interview transcripts. The coding of the transcripts followed the pattern in Stage One; that is, notes were made as the transcripts were read, as if it was being done in a Contextual Design interpretation session. These notes were then added to the Affinity, for example, creating a new category or being added to an existing category.

The Contextual Design Models were then explored to see if they could be of use. Physical Models had not been drawn for the American core; however, Physical Models had been drawn for the meetings.

The Artefact Model in Stage One had not proved to be as useful as had been expected. Additionally, only one new artefact had been retrieved from Stage Two—Stan had made notes and a PowerPoint presentation to present to his Informatics Team. Therefore, the Artefact Model was not used in Stage Two.

A Cultural Model was not created for Stage Two. The Cultural Model that was created in Stage One seemed to be sufficiently comprehensive, and the few points that arose in Stage Two fitted better in the Affinity.

The Flow Model proved to not be relevant in Stage Two for a number of reasons:

  • The focus was much tighter than in Stage One, which took a broader view of the work in general.

  • The everyday work was suspended in view of the visit of the UK—and, like the UK core, there is no such thing as a "representative week."

  • The consolidated Flow Model of Stage One had highlighted the importance of collaboration in the role of the CoP members. This became one of the propositions that was presented to the UK core. A subset of the propositions was also presented in the interviews in Stage Two, and this seemed to cover the Flow Model aspects satisfactorily.

In Stage One, meetings were seen where:

  • the UK and the US cores were distributed, and

  • the UK core was co-located.

In Stage Two, the UK and US cores were co-located; therefore, Sequence Models were created for the meetings to see if they showed anything different.

As one of the focus points of Stage Two was the use of shared artefacts and the Artefact Model had not proved useful, a different approach was needed. A tracking approach seemed to offer a way forward. The main artefact in evidence was still the planning document and so an artefact timeline was drawn to show the planning document's development up to the visit to America and then its use in America.