…and we’re back

After a bit of hard work over the weekend, and half a dozen interviews, things seem to be back on track.

The regular meeting today seemed to go well. Main points to come out of the meeting:

• I need to think about whether my three proposed characteristics of technologies for smoking cessation – persistent, interruptive and responsive, should be collapsed down to two dimensions. Steve suggested persistent-transient and responsive-unresponsive. I’m not so sure about this … needs a bit more thought
• It probably is worthwhile interviewing the long-time quitters in the jaiku quit group
• I don’t need to be so concerned about the research question at this stage. My earlier efforts have focused the research somewhat, which has been useful. At the moment, it’s best to focus on the data and re-build the research question around that
• I need to get my literature review re-written, and my methodology written before next week. Procrastination be gone!

Loss of focus

Dealing with the possible change in methodology seems to have brought with it some doubt as to the focus of the research. Last week, in an attempt to get a sense of what my contribution might be, I explored the possibilities of integrating the model of the Temporal Aspects of Usability (TAU), the Dynamic Model of Relapse (DMR) and possibly even the Transtheoretical Model of Behaviour Change (TTM). Looking at this in writing now drives home just how ambitious (and unwise) a pursuit this is. Needless to say, it proved problematic. Fabre’s TAU model provides four high-level dimensions to be considered when designing for a particular action: location, duration, frequency and contingency. The DMR describes a complex system and, while temporal considerations are important in the model, it is not a ‘temporal’ model of relapse prevention. A couple of simplified integrated models I attempted seemed so obvious as to almost be common sense, making me question value of their contribution.

Going through this process also made me wonder about the role of the data I was intending to collect. How would the data support the assumptions made in the models? I found myself at one of those points where it seems like you’re doing research in reverse. So, back to square one, and time to re-ask some fundamental questions – What is the research question? Am I trying to modify a model, integrate some models or just use some models to get at some information? The first of these questions really should have been more present in my mind, but seems to have been ignored. The second, about models, can only really be answered once I have my data. There’s only so much you can guess about how a model can be changed when you don’t have any data.

I’ve resolved, at least for the moment, to press on with the following research question, which I have a feeling is at a good level of abstraction and will let me get the most advantage out of the contextual data I hope to collect (more on my new methodology later).

What are the influences on the ability of quitting smokers to access online social support at times of need?

In answering this, I will explore the activity of accessing support. I will use the TAU dimensions to explore this activity, and will use aspects of the DMR to explore contextual issues and temporal issues related to smoking.

I’m going to define ‘access’ as the sending or receiving of social support ‘content’. A ‘time of need’ may be defined either by the individual (eg. a craving or experience of withdrawal symptoms, intrapersonal negative affect), their context (eg. a high risk situation, conflict with others) or by research into appropriate times for intervention. In including the last of these, I hope to be able to cover interruption or persistence as a part of the investigation. Research has hinted that this is ‘needed’, but it is not a need that quitting smokers would necessarily express (note the inability to interrupt in online social support). Online social support will be defined as support in online communities from other quitters as well as micro-blog users who publish their quit experiences to their friends online. Some interesting comparisons could be brought up between conventional online communities and new ‘lighter’ social networking in micro-blogs, where communication is not only with other quitters but with a regular extended social network. “Online” may include mobile access to online technologies, eg twitter.

If I were to pre-empt some of the findings, I would be thinking about the interplay of rhythms and interruptions. The rhythms of the physiological effects of quitting, the rhythms of device use, rhythms of social interaction and other life rhythms may all have an effect. Interruptions would feature in a couple of ways. Experiencing a craving in a high risk situation could be seen as an interruption to the activity the individual wishes to be engaged in. We could also think of ‘remedial’ interruptions such as reminder emails.

Surely this needs another iteration, but I’m going to sit with it for at least 24 hours. I need some focus at least for the next little while. And I would really, really like to be able to keep most of my literature review! A possible problem: I think I (and other researchers) already implicitly know what many of the influences are. Also, the rhythms of internet use are perhaps not as interesting and “rich” (ie tied to interesting contexts) as the rhythms of mobile use. Then again, perhaps nobody has explicitly pointed out these inadequacies of online social support. Maybe I need to push for the distinction between tonic and phasic social support, and underline that an online community supports the tonic but is inadequate at supporting the phasic. This leaves a window to talk about mobile technology in the discussion, also highlighting the capacities of the different media to convey different kinds of social support. Another idea is to look at access to support in general at times of need. This could take in a variety of technologies … but I suppose it would better to just focus on one.

Re-thinking methodology

Well, failing some kind of miracle, it looks like I won’t have time to run the trial I was planning to. Hours of work building the website for the project and planning the logistics gone to waste. I knew there was a chance it would all be in vain, but I had to be prepared. Enough negative talk – how can I make this work? What data to I have at my disposal? How would that data help me to answer the question?

Possible sources of data

Interviews with online support community members: In order to get at ‘the influences on their ability to access online social support’, I would use the TAU and DMR concepts to probe their use of the community. This should lead to some discussion on the benefits of mobile technologies From members that have already contacted me, I should hopefully be able to tee up four or five online chat interviews. I could ask the administrators of the community if I could message some other members about the interview directly. Hopefully, I would have a few more bites this time. It would be good to get a list of members that had quit in the previous month or fortnight.

Interviews with quitext users: Quitext is a new service launched by an Australian company that sends SMS messages of advice for quitting smokers on a schedule defined by the individual. They’re quite a new service, though, and I’m not sure how many subscribers they have. I don’t think it’s that likely that this option will come through.

Channel on Jaiku of quitting smokers: These guys formed the group I was hoping to in my study. I could join the group and ask them to post more often, also ask if they’re posting from mobile or web. Could then do email pre-interview and web-chat post interview. Could be a bit risky. This is the last chance to get the micro-blogging trial idea off the ground. Otherwise, just interviewing some/all of them would be interesting. They seem to have stopped posting about their quit so much recently, but they have been posting to each other on other topics. I don’t know how many of these people knew each other beforehand. There is surely something interesting in here, I’m just not sure how relevant the TAU is to this any more. Need a rethink on the question if I go down this path.

Twitter feeds of people quitting: Across both micro-blogging platforms, there is one user on Twitter that posted the most about her quit attempt. About 20 times over the course of three weeks. She was already a regular twitterer, and she slotted her quit-related posts in between all her regular posts. She kind of ‘hijacked’ the communication channel in her quit attempt. The feeds themselves would be interesting to look at and the intervening posts would give some idea of the context. Interviewing her would be great, and her micro-blog would serve as a perfect aid to her memory, to get her back into the moments when she was posting and how she was thinking/feeling. If I could land an interview, then there could be a valuable methodological contribution regarding the use of micro-blog feeds to aid in interview recall.

I better stop crapping on now. I’ve read, re-read and re-written this post about 20 times. I’m feeling that focusing on the temporality of Quitnet use is a bit dull. I would ideally like to look at the micro-blogging quitters, but I need to define a research question and a contribution angle. I don’t know if TAU is going to cut it any more if I make so many changes to the methodology.

Hmmmmmm

Change of plan

A number of factors, time being the most prominent one, are forcing me to abandon my plan to run a trial as part of my thesis. The sending of an internal message to members of the online community I was hoping to recruit from has taken longer than anticipated. There have also been some unexpected complications, such as the fact that text messaging is not included by default in many North American mobile phone plans. Short of a miracle occurring in the next couple of days, I think I have to lay the plan for a trial to rest, and come up with something else quickly and smartly.