Recent observations from CHEST 2017 by the inestimable Pat Rich!

Days of Past Futures

Poster

Last week, I spent valuable time jealously guarding the only power outlet in a conference hall of about 3500 people so I could live tweet the presentations without fearing suddenly losing power in my laptop. At least at that meeting I had a chair pilfered from the rows of interlocking seating rather than having to sit on the floor next to the outlet which has often been the case.

Imagine my awe to read that the American College of Chest Physicians annual meeting being held in Toronto this week was actually holding designated seating for live tweeters at its most important sessions. Wait, there’s more: Delegates were able to add an “I tweet” ribbon to their name badge at the conference as well as find designated selfie areas throughout the conference to take and share photographs.

In addition, many of the sessions were live-streamed via Facebook and YouTube, an educational…

View original post 296 more words

I tweet 10-20 times per week on average. But I’m far more active on Twitter when I attend conferences. At conferences, I write as many as 300 tweets over two days. I use conference hashtags to share my thoughts, generate discussion, and disseminate others’ work. Conference tweeting is a way to electronically meet other attendees and have an online conversation with them. I try to use live-tweeting etiquette as much as possible.

Along with tweeting, I normally take notes on my laptop during conference presentations. Lately, along with my usual conference tweets, I have been writing what I call tweetnotes. These are tweet threads I write by tweeting my notes and threading together replies. The tweets become my notes, hence the term tweetnotes.

Before creating tweetnotes, I used to write a dozen or more discrete tweets during presentations. But I can thread tweetnotes together, meaning they can be saved as a narrative instead of a basketful of discrete tweets. In the past, I would scan through old tweets to find tweets from a specific presentation. Now, I just save or embed the tweetnote links and I can easily access the entire thread. Here are a few examples from the annual Qualitative Health Research conference I just attended.

Tweetnotes are useful for me and for others. For example, when I finish a tweetnote I will often retweet it with a summary of the presentation I attended. The presenter, conference organizers, and other audience members can now interact with either individual tweets or to the entire tweetnote. For example, CADTH retweeted my tweetnote on some presenters form their organization.

So far, I think the uptake of using tweetnotes has been well received by my network. What do you think of tweetnotes or other ways of using Twitter threads? Share your thoughts as a comment.

Sally Thorne’s Interpretive Description: Qualitative Research for Applied Practice is now in its second edition. Among many other qualities, it’s a thoughtful and pragmatic invitation for researchers in applied disciplines to rethink coding. I know this because something interesting happened when I recently attended Thorne’s workshop at the annual Qualitative Health Research (QHR)  2017 conference.

A bit of background: QHR is an excellent conference hosted by the University of Alberta’s International Institute of Qualitative Methods (IIQM). For full disclosure, Sally and I volunteer on the advisory board for IIQM. You might think I am writing this post because Sally asked…But it was actually another respected qualitative researcher who requested a summary of Sally’s workshop.

(It was at this point in the workshop that @trishagreenhalgh asked for a blog, so here we are. When Oxford’s inestimable Trisha Greenhalgh lurks the conference hashtag #QHR2017YQB, you’re doing something right.)

One of Sally’s most relatable points was that qualitative research often is taught as a set of highly procedural, theoretically dense steps. She’s not against using qualitative methodologies, but she reasons that educators, clinicians, and other applied researchers can be distracted by unnecessary methodological procedures.

Sally endorses an approach to qualitative data analysis that centres on how applied researchers understand their data, not how closely these researchers followed a procedural series of methodological steps. A breath of fresh air blew through the room when Sally said she believes coding can become the motivation and driver for research rather than research questions.

Sally defended her position on coding with two key points. First, she reminded us that it is easy to code but hard to uncode. Once we reify an idea into a code it becomes a knowledge object. From this position, then, it becomes difficult for us to ‘uncode’ our work because this involves destroying a knowledge object.

Second, Sally reminded us to consider how coding involves signification, the act of assigning a concept or object to a symbol. While signifying our ideas occurs as an essential part of qualitative analysis, Sally noted a common trend where researchers assign codes too early. Coding too early can negatively impact how researchers think about their data later in their projects. “Coding,” says Sally, “is not the same as deep, real exploration of qualitative data.” When we establish codes too early, they demand we feed them our data instead of supporting the development of new ideas.

Sally invited us to consider some alternatives to our impulse to name and label codes too early in the analytic process. Instead of labelling a code with a word, like Frustration, Sally recommends using labels like Category A or I don’t know why this is important yet. These techniques resist the urge to early labelling that can impact later and potentially deeper thinking.

Sally explained that she sees these coding problems frequently in manuscripts submitted to the journal Nursing Inquiry, of which she is the editor-in-chief. She describes under-theorized qualitative manuscripts that foreground codes as if they are analytic results in and of themselves. “I have 14 codes and 2 categories is not a finding,” quipped Sally to gales of nervous laughter. (Come on, we’ve all done this!)  Writing results involves not only summarizing the analysis, but also creating a rich narrative that can be applied in practice.

Oh. And hell no,  don’t even bring up saturation with Sally! She calls this concept a relic of the early days when qualitative researchers had to measure up to their quantitative colleagues. It’s “absolutely antithetical to applied health disciplines” because after all, “we would never trust a health professional who told us there was no more information to be found” on a topic.

I’ll let Sally’s last slide speak for itself. Needless to say, she’s amazing. Thanks for reading. I hope this was helpful. If you’re interested in this topic and you’d like to know more, please consider following Sally on Twitter as well as the IIQM. 

DMSLGglW0AA2ReI.jpg

Word

I created a simple word cloud based on the text of the recent anti-diversity manifesto published by a now-disgraced former Google engineer. The visual speaks for itself, but I wanted to share some thoughts on the sentence where I simply stopped reading:

“Being emotionally unengaged helps us better reason about the facts.”

There are many statements I disagree with throughout this diatribe–and if you haven’t read it you’re not missing out on much. But this sentence rubs me raw.

Where do facts come from? The author should note well that emotional engagement is central to what propels science, research, and the people who choose scientific careers.  We gain nothing by separating so-called “matters of fact” from the emotional human contexts in which those facts exist, profligate, and influence human action.

Not only do we observe this artificial separation of reason and emotion throughout this manifesto, we frequently observe this separation used to justify the mythos of innate male and female skills and virtues.  Here is the link to the Gizmodo article containing both the original anti-diversity memo as well as the corporate reply from Google’s new Vice President of Diversity, Integrity & Governance, Danielle Brown.

What do you think? Let me know!

Dear Lorelei,

Nearly 5 years ago, Stella Ng and I decided to organize the process of nominating you for the Schulich School of Medicine & Dentistry’s Distinguished Leader award. We brought the idea to Mark Goldszmidt and he was quick to get the ball rolling, whatever we needed he would make it happen.

As it goes with these types of awards, it was a layered process and we knew it would be a big job. Many letters of endorsement would need to be written, documents signed, etc. We were, of course, glad– but unsurprised–that everyone we approached enthusiastically agreed to write a letter. In fact, it came to a point where we had too many offers and needed a backup list. Physicians, research scientists, and even a former Dean wrote on your behalf.

As it also goes with these types of awards, you would not have read my letter, but with the news you have won yet another mentorship award–the 2017 Meredith Marks award!–I thought I would post it online for you in a de-identified format.

I’ve had the pleasure of working with Dr. Lorelei Lingard for 4 years, first as a part-time research assistant through the University of Toronto’s Wilson Centre, next as a full-time research associate at the Centre for Education Research & Innovation, and today as a PhD student in the Faculty of Health Science Department of Health & Rehabilitation Sciences. In the medical education research universe, Lorelei is the sun: faculty gravitate toward her ability to support their research, residents are drawn to her ability to connect them with ongoing medical education and surgical education research projects, and graduate students gravitate toward her brilliance, positivity, and poise. She is the best possible ambassador for graduate and post-graduate education at the Schulich School of Medicine & Dentistry and worthy for consideration for the Distinguished Leader Award.  

When I met Lorelei, I was in the second year of my Masters degree. Although she was a researcher with a major international profile, she welcomed me onto one of her research projects with kindness and understanding. Being so novice, I was intimidated by her confident, expert approach to a research project involving residents’ comfort-levels with end-of-life care communication. While I could tell Lorelei was extremely busy, I left each meeting with her feeling empowered that I was capable of contributing to her research program and inspired that I was collaborating with someone so generous and knowledgeable.  

As luck would have it, Lorelei was hiring a full-time research associate to begin working with her at Western University at the time I completed my Masters degree. As the successful candidate, I have seen first-hand the impact she is having on the medical education community at Western. I have had the privilege of seeing her work with residents and faculty who come to her with problems and leave with research questions. She never declines to consult with any faculty or student, regardless of their level of understanding of research methods or the stage of development of their research project. Further, no matter how busy she is, she never cancels meetings with her colleagues. Lorelei enables the residents, students and faculty she works with to transform their half-baked ideas into feasible, rigorous, research programs–often projects that end up being funded! She is an inspiration to me, her colleagues, and the students and faculty who work with her.  

I am routinely reminded by others of how fortunate I am to have Lorelei as my PhD supervisor. One colleague has confided in me that they have received more mentorship working with Lorelei for two months than they received from supervisors during their entire doctorate. A recent visiting researcher from [outside Canada] told me that she received more face-time with Lorelei in three weeks than she had received from her doctoral supervisor in six months. These are two simple anecdotes from dozens I’ve observed during my years working with Lorelei. 

Lorelei is world-renowned for her research, but her graduate mentorship, post-graduate student mentorship, and faculty development deserve recognition. I sincerely hope you will select her as the top candidate for this award.

5 years past but the words ring true today. Congratulations (again), Lorelei!

Sincerely,

Allan

McMaster UniveGeoff Normanrsity’s Professor Geoff Norman is a prominent figure in the Medical Education/Health Professions Education community. Along with co-authoring one of the most popular handbooks in Biostatistics, Norman is also a long-time editor for the Advance in Health Sciences Education journal.

In his journal’s most recent issue Norman offers up some tips on how to not get your paper published. I’ve highlighted my favourites–sound advice for academics and graduate students alike:

  1. Support your problem statement with the best available evidence (“Nice ideas finish last”)
    • Your paper must provide sufficient theoretical and empirical justification for why its problem statement is an important one.
  2. Conduct a thorough literature review early (“The world doesn’t need another mousetrap”)
    • Before beginning your study, ensure the same work has not been conducted before. If it has, ensure your problem statement and study design sufficiently expand past findings.
  3. Ensure data are consistent with the journal’s/disciplines expectations (“We don’t really care if your students loved the course”)
    • Norman reminds us that studies involving self-assessment as data are consistently uncorrelated with performance measures. (Eva and Regehr 2005)  While another journal may publish self-assessment data, Norman begs authors to consider whether their data is appropriate for journal/disciplinary conventions before submitting their manuscripts.
  4. Ask whether your findings are applicable in another context (“We are scientists, not inventors”)
    • In studies where a costly simulation device or complex intervention is designed, Norman warns authors not to presume that their findings should be of interest to a wider audience–taking special care when the research is aligned with an commercial interests (e.g., a device manufacturer)
  5. Ask whether your findings are interesting in another context (“The only people who care about education of pediatric gerontologists in Lower Volga are pediatric gerontologists in Lower Volga”)
    • Norman reminds us to think about an international audience for our work. It is incumbent on authors to figure out why researchers in different countries will care about their work.
  6. Disclose other venues in which your study has been published (“Salami slicing belongs in butcher shops”)
    • Salami slicing is the colloquial term for publishing the same study in multiple journals. While Norman admits that this is sometimes an appropriate practice, he reminds researchers to avoid self-plagiarism by writing the second study de novo rather than cutting and pasting swathes of text.
  7. Represent quantitative value with appropriate detail (“P values tell you whether it’s zero but little else”)
    • While a research team’s findings may be statistically significant, simply displaying an appropriate P value is insufficient. Norman reminds authors to include standard deviations, confidence intervals and effect sizes, best represented by good figures and graphs.
  8. Leave room for future research (“More research is always required”)
    • Avoid overstating the universal applicability of your results. Be specific about other research questions that remain to be answered in your work.

Leave your thoughts

What do you think? Am I missing anything obvious? Do you find the advice pedantic? I’d love to read your commends below.

 

A recent post on the LinkedIn NVivo users page inspired me to write this post on collaborating with NVivo. This is a topic that has been an important component of my work for the past 4 years. I’d surmise that pervasive communication technologies and increased interdisciplinary research mean that NVivo is being used on more research teams now than ever before. This post captures 6 key ideas when collaborating with NVivo. An early caveat is that this post does not discuss the implications for teamwork offered by NVivo Server. With the understanding that I’m excited to see how NVivo Server will develop, I’m not convinced it’s widely available enough to warrant an expanded discussion here.

1. NVivo 10’s collaborative functions

As it stands, only one team member at a time can edit an NVivo project file. However, team members’ identities are recorded when working in a project file at different times. Alternatively, each team member can work on a copy a project file that can later be imported back into a “master” file using the import project dialog box. The “master” file is the main copy of the NVivo 10 project file that contains the most up-to-date source data and data analysis. Crucially, the “master” file should be stored on a computer that is frequently backed up. Anyone who has ever suffered the loss of a project file would likely encourage multiple back-ups on your NVivo project. I use Dropbox as a solution for backing up my NVivo projects, but this solution isn’t perfect as I will expand on below.

Assuming that the work has come to a stage where different members have submitted contributions to the NVivo project file, make sure that the team uses easily identifiable standardized user profiles when they work with their respective parts. User profiles, represented by each individual user’s initials, are an important function when the time comes to compare coding.

Coding stripes can be a useful features for NVivo projects where multiple team members are coding data. Coding stripes allow a team to visually identify how data has been organized into Nodes. NVivo also allows coding stripes to be filtered by team member in order to visually identify how an individual coded the data.

When reporting data analysis findings, a more comprehensive summary of the variance in team members’ coding may be needed. In this case, Coding Comparison Queries allow multiple team members’ coding to be quantified in metrics of inter-rate reliability (e.g., Kappa coefficient).

2. Team roles & responsibilities

Beyond the technical features available for NVivo 10 users, team roles and responsibilities should include some attention to data analysis. On any research team on which I work, I advocate the team appoints an NVivo coordinator for the research project. The NVivo coordinator is the team member responsible for maintaining the team’s “master” project file. The coordinator role entails 4 primary duties:
◆ Ensuring that each team member’s independent data analysis is routinely imported into the “master” project;
◆ Backing up the”master” project;
◆ Importing new data into the “master” project;
◆ and Distributing up-to-date copies of the “master” project.

3. Team workflow plan

Create a team workflow plan that includes how to manage your team’s data and analytic findings. The team workflow plan is a document that capture the team’s shared understanding of filenames, read-only and read-write file access, storage and backup locations, and rules for file distribution and archiving. For example, when it comes to audio and video files, what file formats will the team use? Will those files be embedded items or linked as external items?

The workflow plan should include the team’s approach to creating and maintaining nodes. I always recommend that team members write ‘instructions’ in every Node’s Description field (max 512 characters). Unless using In Vivo Coding, creating a new node results in a new node dialog box that includes a blank description field. What better place for a researcher to capture their thinking at a given moment? A team member’s definitions or reflections on coding and nodes can also be written as a linked Memo, which is easier to write, read, print and code.

4. NVivo and cloud computing

I use cloud-based file sharing services like DropBox, SkyDrive and Google Drive as a working solution for backing up and sharing collaborative NVivo Project files. In theory, these services also allow changes to a NVivo project file to be made across several computers using the ‘cloud’.

I recommend you turn off the live syncing features of these programs while you are running the NVivo 10 software client. Most of these services allow you to toggle live-syncing on a folder by folder basis so that you can sync all other folders but the folder containing your NVivo project. I learned this the hard way by losing hours of coding time due to an error caused by simultaneously using NVivo and syncing its attendant .nvp file in a cloud-based utility. Other colleagues of mine have had similar experiences. Cloud-based utilities can be useful for team collaboration, but taking the proper precautions can avoid costly loss of analysis time due to software crashes.

5. Team meetings

Finally, add the NVivo project file as an agenda item team research meetings. While the meeting agenda will no doubt be packed with discussions of the research process, briefly talking about the tools of your is a good idea.

These insights and many more are contained my technical manual on QSR’s software, “NVivo 10 Essentials” (co-authored with Bengt Edhlund).