| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Family Medicine, University of Colorado at Denver and Health Sciences Center, Aurora, Colo
2 JSI Research and Training Institute, Denver, Colo
CORRESPONDING AUTHOR: Douglas Fernald, MA, PO Box 6508 F496, Aurora, Colorado, 80045-0508, doug.fernald{at}uchsc.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS We drew on our experiences in working on numerous projects of varying, size, duration, and purpose. Through trials of different tools and techniques, expert consultation, and review of the literature, we learned to improve how we build teams, manage information, and disseminate results.
RESULTS Attention given to team members and team processes is as important as choosing appropriate analytical tools and techniques. Attentive team leadership, commitment to early and regular team meetings, and discussion of roles, responsibilities, and expectations all help build more effective teams and establish clear norms. As data are collected and analyzed, it is important to anticipate potential problems from differing skills and styles, and how information and files are managed. Discuss analytical preferences and biases and set clear guidelines and practices for how data will be analyzed and handled. As emerging ideas and findings disperse across team members, common tools (such as summary forms and data grids), coding conventions, intermediate goals or products, and regular documentation help capture essential ideas and insights.
CONCLUSIONS In a team setting, little should be left to chance. This article identifies ways to improve team-based qualitative research with more a considered and systematic approach. Qualitative researchers will benefit from further examination and discussion of effective, field-tested, team-based strategies.
Key Words: Qualitative research • evaluation studies • research methods • team work
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Despite published accounts of team-based qualitative research,1–11 there are few concise, concrete, practical suggestions for how a team goes about improving the team and its work. Furthermore, investigators must be confident that processes and findings will stand up to peer scrutiny,12–15 which can be more difficult as data, analysis, and results are spread across team members, computers, departments, or institutions. Teams must find ways to manage and analyze large amounts of data, deal with subtle change that occurs with time, and reduce the competing demands and crises that may disrupt the effort. We do not propose a specific qualitative method; instead, we discuss issues we have encountered and synthesize the tools and techniques we used to improve our team-based qualitative research.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Data Management and Analysis As projects grow in size and length, data and thinking are spread across time, space, and people, resulting in potential threats of losing track of data, good ideas, and assignments. As the volume of data builds, finding relevant information and managing the different items, such as transcripts, notes, and early ideas, becomes more difficult. Also, there are potential technical challenges of working across institutions in terms of computer hardware and software.
Analytically, details like definitions and uses of codes can drift or change altogether. Analysts may begin work with incomplete or outdated data, duplicate efforts, or lose keen insights and connections if they do not have clear guidelines for managing files and notes. Biases or differences in analytical styles can impede analytical progress when output does not match expectations. In our own experience, for example, some analysts prefer to work with paper copies, while others prefer electronic copies. We have to determine how to accommodate these 2 approaches to have a single, perpetual record of analysis that can be easily accessed for future iterations of the analysis. Finally, lack of early planning for writing up results can lead to delays and hard feelings about contributions and authorship.
Strategies to Improve Teamwork
Team Formation
Team leadership.
Early in a project a team leader—often the principal investigator (PI)—sets the tone for how the team can work together. These messages should be clear and can start with a discussion about the project’s purpose and any unwritten hopes or expectations for the project. One technique is to craft a 1- or 2-sentence sound bite that captures the essence of the project; repeat it often until the team adopts it as a common description of the project. The PI is responsible for making sure that the team has a common understanding of the project throughout the life of the project. To ensure a common focus, the PI should be involved from the beginning and be able to hear all voices, which he or she can meld into one. A team leader must also watch for signs that the team is not working well (silence, outbursts, missing deadlines, etc) and restore its function. Essentially, a team leader needs excellent management skills or assistance from others who can help manage people, time, and information.
Team timing.
Convene the team as soon as possible. Even members who may not seem central to the project will benefit from early discussions. For instance, if the project entails a participatory style of research, it is important to include representatives of the communities, participants, agencies, or others at this early stage. Early discussions help build an overall clearer sense of project purpose, possible roles, and details of the project that will inform and guide work to come. Importantly, discussion starts orienting team members to the vocabulary of project—technical language, medical terms, context, names, and acronyms.
Team Building
Regular team meetings.
Regular team meetings have been among our most effective methods for keeping members engaged and on track. Begin these meetings early in a project and schedule them in advance so they are on all the team members’ calendars. When it seems there is little to justify a meeting, the team should still meet, if only to avoid getting out of the habit of meeting. Consider changing the format and focus of meetings to reenergize the team: focus a meeting on emerging ideas or engage in an open forum on ideas for publications. The team leader sets the tone of meetings, which should be comfortable. Establish norms for conducting open, respectful meetings. Have the team members agree to a few ground rules for listening, speaking, and handling disagreements.
Roles, responsibilities, and expectations.
Early in a project ask each team member (do not assume) what specifically interests them about the project and any particular parts they feel strongly about. Ask again later in the project as members acquire more skill or better understand the goals and the remaining work, which may have changed.
Data Management and Analysis Strategies
Analytical Biases and Work Styles
Spend a sufficient amount of time talking about these differences and troubleshoot how to accommodate them. The following are some major considerations to discuss: theoretical approaches to analysis (eg, phenomenology, hermeneutics, grounded theory), analytical style, (eg, lots of codes vs few codes, structured vs unstructured), and use of computers (eg, e-mail, electronic data vs paper). An analyst who is accustomed to unstructured coding, for example, may find a highly structured coding scheme difficult to work with, especially if coming into a project at an intermediate phase. These discussions are helpful for understanding how to structure data collection and analysis and where talents and methods can be put to their best use.
Data Management At the start of a project, think through how and where notes and files will be kept, who will be oversee them, and how they will be safeguarded. One of our most effective tactics is to appoint one person as a data manager. Almost all files, whether computer or paper, pass through him or her. The data manager cleans, compiles, and merges files for analyses and updates all team members on file status and accessibility. Early in a project, it is also worth a quick inventory of computers and software to determine members’ preferences, training needs, software version compatibility, and data security.
Common forms.
Regularly use simple, familiar, and common forms at all phases of the analysis. Miles and Huberman8 describe a number of ideas, but we have found a simple 3-question summary form invaluable for all types of qualitative data collection: (1) What were the main issues or themes? (2) What are your impressions, concerns, ideas, or reflections? (3) What new or remaining questions do you have? Without consistent use of such forms, retrieving certain information can become a time-consuming search for data or ideas embedded in detailed notes or codes.
Coding conventions.
If coding data is a major part of analysis, remind the team to routinely write down code definitions, annotate them, and revisit them. This recommendation sounds obvious, but it is especially important to update codes as the analysis progresses and discuss them periodically to identify inconsistencies, changes in meaning or use, and emerging sets of new codes. We found it helpful to agree on a few common coding conventions (using punctuation or specific naming conventions) to find codes or data of a certain type quickly. We often use a code called "Quotables," alone or in combination with other codes, to quickly retrieve segments of text that are particularly illustrative or potentially useful in a report.
Trim as you go.
As the data grow and ideas emerge, the team needs to document these ideas and continue documenting as the process flows and decisions are made as to what to keep, discard, or reserve for later review. To keep members focused on the right tasks, revisit the original research question or project goals, and trim the ideas that might be important but are not within the scope of the project. Data grids have helped us identify major ideas and themes or major gaps in our analysis across data sources or team members (Figure 1
). Sometimes we assign scores or symbols to themes to identify essential ones easily; or we use 0s and 1s for quick calculations to count the presence or absence of themes across cases or within cases.
|
Regular documentation.
At team meetings, keep notes about project administration, decisions, task assignments, and the analytical process. For example, document why specific codes were used instead of others, how merging specific codes came about and what they include, or who is responsible for the next data collection or analytical steps. Keep track of the smaller tasks that might get assigned and forgotten because they seem less important. During more-intensive analytical meetings, consider tape recording and transcribing conversations. Send a meeting summary to members and keep summaries in an easily accessible place.
Writing up results.
Books have been written as guides to writing up qualitative research16–21; however, these guides do not address the process of collaborative writing. Writing collaboratively mentors inexperienced writers, alleviates blank-page anxiety,17 and provides a good check on the accuracy of the reporting. It saves time in the long run to start brainstorming and outlining potential manuscripts and writing tasks early in the process of qualitative analyses. Establish author guidelines, roles, and expectations early in the process to avoid misunderstandings later. We have had the unfortunate experience of routing a manuscript to coauthors to find that we had inadvertently left a substantive contributor out of the loop. As ideas for manuscripts start, ask of the whole team which members would like to contribute to the written report and which members would like to lead; then route early drafts of an abstract to those team members, and ask whether there are others who should contribute. Simultaneously, be extremely clear about what contributions meet the guidelines for authorship.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication May 4, 2004. Revision received January 21, 2005. Accepted for publication January 30, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  K. C. Stange In This Issue: Patient Outcomes, the Process of Care, and the Capacity for Innovation Ann. Fam. Med, July 1, 2005; 3(4): 290 - 291. [Full Text] [PDF]
|
|
Read all TRACK Comments
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
