Table 1

Key differences between an online forum approach and an interview approach to data collection

AttributeOnline forum
Recognised feature identified in the literature
versus features identified in the present study
Semistructured interviews
Recognised feature identified in the literature
versus 
features identified in the present study
Context
 LocationRespondents from across a wide geographical area can participate at own convenience.54 Interviewees geographically restricted.
Patients’ own home: UK. Patient’s own home: East of England.
 SamplingVoluntary participation/self-selection. Recruitment does not require collaboration between clinical sites or support of professional staff.55 Purposive recruitment in healthcare settings guided by sampling techniques based on population demographics including age, gender and disability.
Voluntary self-selection by participant: no sampling criteria, no restriction on age. Verification of stroke/TIA diagnosis not possible.
Most participants taking part within the first 5 years since the cerebrovascular event.
Purposive sampling: GP screened, predefined criteria to achieve maximum spread of gender and disability.
Age: 55 years and over. Confirmed stroke/TIA diagnosis.
Mechanisms

 Participation
Multiple participants per conversation thread: stroke survivors or caregivers. Conversation possible between survivors, survivors and caregivers or caregivers with other caregivers. Maximum of two or three participants per single interview conversation: researcher, stroke survivor and caregiver.
 Dynamic of interactionDiscussion conducted remotely. Relative anonymity can encourage users to feel uninhibited.5 56
Likelihood of expressing honest opinions about sensitive issues.
Engagement can be actively encouraged.
Face-to-face approach enables development of rapport between interviewer and participant.57
Less knowledge of participants and participants remain anonymous. No influence of researcher on participation. Researchers gain knowledge of interviewees, development of researcher–participant rapport, active encouragement of participation.
 Response contributionLonger conversations allow for a broader understanding of the subject matter and potential for greater depth.58 Discussion threads generate reflection and greater description among users.33 Probing questions from researcher seeking clarification or to pursue a more detailed response.40
Potential for significant individual contribution.
Responses shaped by other peer contributors.
More opportunity for self-reflection.
More frequent comments.
Maximum of two individual contributions (patients and caregiver).
Responses shaped by researcher interaction.
Less opportunity for self-reflection.
In-depth comments encouraged by researcher.
 Timing of eventUsers can post repeatedly and frequently on many topics over a long period of time.Interview is a single event occurring at one point in time. Maximum of two contributors to the interview discussion.
Single or multiple participation over time.
Ability to contribute to discussion on multiple occasions/topics.
Single participatory event.
Contribution fixed to a single time period.
 GuidanceUser freedom to choose what to discuss and how frequently to contribute to free-flowing discussion threads.5
Posts created through peer-to-peer communication, without professionals’ involvement and influence.59
Response shaped by contribution of other survivors or caregivers.
Follows a predefined line of questioning. Several key questions define the area to be explored.40
Researcher oversees the direction of conversation.
Free or peer-guided discussions. Guided conversation: responses to predefined questions in topic guide.
 CommunicationPermits broad accessibility and asynchronicity with online communication.5 60
Restricted to those with internet access.34
Direct face-to-face, synchronous communication.
Indirect communication, via computer, no physical proximity, asynchronous. Direct communication, face to face, synchronous.
Outcomes
 ActivitiesNo physical transcription is required; user contributions printed automatically, improving credibility of data.7
Potential for inaccurate interpretation through misunderstanding nuances in the data may still exist.7
Transcription is key to representing the individual and dependability of data. Transcription opens data to misinterpretation or misunderstanding.61
Rigour and accuracy in transcribing is integral to the analysis process, influencing the degree of dependability of data.61
No audio recording.
Automatic transcription printed directly from forum
No field notes.
Interviews are audio recorded and interviews are transcribed. Potential for ambiguity through inaccurate transcribing
Field notes taken during interviews
 ReportingForum posts are moderated before appearing online, effect on the data collected is relatively unknown. Moderation processes can influence engagement in online communities.62
Third-party moderation leading to possible exclusion of data. No exclusion of data prior to analysis.
  • Location: geographical area of the research. Sampling: sampling method used to recruit participants. Participation: individuals participating in conversations. Dynamic of interaction: knowledge of participant determined by level of engagement. Response contribution: level of contribution to the conversation by individuals. Timing of event: frequency of participation over time. Guidance: level of conversation guidance and level of freedom to discuss. Communication: face to face versus distance communication. Activities: need for audio recording and transcription activities. Reporting: moderation of data before analysis. 

  • GP, General Practitioner; TIA, Transient ischaemic attack.