Home » Cover Story

Navigating Tough Conversations in Statistical Collaboration

1 September 2021 2,253 views No Comment

Emily Griffith is an associate research professor of statistics at North Carolina State University. She is also a fellow in the Office of Research Innovation, working on development and strategy to further innovation in the university’s data sciences initiatives. In her free time, Griffith enjoys running (even in the North Carolina summer), cooking, and hanging out with her family.

 

 

Megan Higgs is a statistician who has worked in academia and private industry and is now working independently as Critical Inference LLC and writing posts for a blog of the same name when she can. She volunteers as editor of the International Statistical Institute’s Statisticians React to the News blog. Higgs loves living in Montana—for the winters more than the summers.

 

 

Julia Sharp is an associate professor of statistics and the director of the Graybill Statistics and Data Science Laboratory at Colorado State University. She is a widely recognized expert in statistical collaboration and was recently awarded the Outstanding Mentor Award from the ASA’s Section on Statistical Consulting. When she is not working, Sharp enjoys baking, hiking, and enjoying the company of family and friends.

 

 

Zach Weller is an assistant professor of statistics at Colorado State University, where he works as a consultant in the Graybill Statistics and Data Science Laboratory. He also teaches statistics courses and advises undergraduate and graduate students on applied statistics projects. When he is not working, Weller enjoys hiking, camping, and fishing in the great outdoors of Colorado.

 

 

Uncomfortable conversations are inevitable in many research and professional relationships, including statistical collaborations. Topics can vary widely due to the different backgrounds and expectations of researchers. Common discussions include those about coauthorship, the role of the statistician, and even technical concepts. Reasons for discomfort can include misaligned expectations, the tone of either participant in the conversation, and particular topics or words (e.g., bias) that might escalate the conversation from congenial to difficult.

Although it feels difficult, it is important to navigate our way through uncomfortable conversations with the goal of ending in mutual understanding and a clear path forward. In The Champion Forum podcast, Jeff Hancher argues that avoiding difficult conversations is a disservice to everyone and gives the following five reasons we should have them:

  • To build mutual trust – As collaborative statisticians, we build trust and respect by being honest with our collaborators, even around difficult topics.
  • To communicate value – We show ­collaborators we value them enough to have tough conversations.
  • To clarify the future – Without having a ­conversation, each party is left to their own interpretation and conclusions, which can be misunderstood and result in misaligned ­expectations.
  • To increase self-awareness – It is important to come to an understanding of where their (or our) understanding may be falling short.
  • To stretch our leadership – With each difficult conversation we have, our ability to have productive conversations improves and the range of topics we are confident in tackling expands.

We have found tough conversations are easier if we are prepared for them. When preparing for a conversation we expect to become uncomfortable, we strive to consider both our own point of view and the point of view of our colleague(s). We try to identify goals of the conversation beforehand, including our expectations for actions during and after the meeting. We might consider the mode of our conversation, as well.

For example, email communication is sometimes necessary, but tone can be easily misinterpreted. To prepare, we might compose email drafts and let them mature for a few days before sending or have a neutral third party read the draft email, while respecting privacy. For in-person conversations, we suggest drafting talking points for particular topics ahead of the meeting. We also try to anticipate potential detours and roadblocks that may arise and plan paths around them.

Finding common ground is difficult, and success is far more likely if potential conflicts are approached from a calm, understanding, and patient mindset. While it may seem cliché, we have found using “I” statements to be more helpful than using “you” statements (e.g., “I feel the data could be organized better.” vs. “The way you have organized the data could be improved.”).

Regardless of the conversation modality, being willing to slow down and gather our thoughts allows for a more meaningful response. Slowing down deserves particular emphasis here, as most of us are working in fast-paced, extremely busy environments. It’s important to realize our collaborator in the conversation is likely under similar stress and might not come into the conversation from a calm space. This recognition provides an opportunity for empathy, which can lead the conversation in a productive direction.

Being a good listener and having a positive demeanor can improve the likelihood of positive outcomes in uncomfortable conversations. Being an active, empathetic listener while understanding nonverbal cues and emoting friendliness and confidence can help move a conversation forward. On the other hand, being judgmental, using unwarranted technical language, and not paying close attention to reactions—both verbal and nonverbal—can reduce the chances of a conversation going well.

We can also play to our strengths and use helpful resources when preparing for and navigating tough conversations. For example, we may ask a colleague for advice or solicit suggestions (if appropriate) from an online forum like the ASA Section on Statistical Consulting community. We may advocate for having the meeting in-person instead of over the phone to facilitate clear communication or organize our ideas and goals ahead of the conversation.

We mentioned the importance of preparing for uncomfortable meetings, but the preparation itself can be challenging. It can be difficult to predict how conversations will go, especially for those with little practical experience. Facilitating effective and active training for early-career collaborators, such as graduate students, is essential. In our own work, we have developed resources to engage students in potential discussions with rich enough context to provide them a chance to reflect and think through the communication of others, as well as their own, in a safe environment.

In our series of educational videos on statistical collaboration, we provide rich, realistic scenarios to engage students in exploring and reflecting on strategies to navigate tough conversations. Video topic three, for example—turning down requests for new work from a current collaborator—highlights challenges a collaborating statistician may face when being asked for additional work, as well as strategies for handling the request.

In the video, the statistician is prepared for the conversation and has strategies in place to enforce their boundaries. First, the statistician gathers information from the client about the scope of work and desired timeline. After the statistician explains they are at capacity for the next couple of months, the client offers the statistician additional financial support and coauthorship, both of which can be hard to turn down. The statistician thanks the client for the offer, but makes it clear their answer is still no. The statistician then offers to connect the client with another consultant or revisit the request in the future. The client graciously acquiesces as the conversation ends.

Many collaborative statisticians are helpful by nature and saying “no” can be challenging. However, having a strategy for turning down new requests is a way to say “yes” to maintaining high-quality work and good work-life balance. The statistician’s goal for the meeting was to refuse future work with their collaborator, and they were successful.

On the other hand, in video topic 10—pseudo replication and refusal to use a statistician’s advice—both the statistician and collaborator express different needs or expectations from the collaboration. As statisticians watching the scenario, we may initially only be sympathetic to the statistician whose advice is ignored, but the client is also in a difficult situation. Because the statistician explained their reasoning in a kind way, there is still potential for the collaborator to ask for advice and run a different analysis in the future when the collaborator no longer has to worry about their adviser’s unhappiness.

It is important to recognize a bad interaction with a statistician can have a lasting impact on a researcher’s willingness to seek statistical advice or collaboration. We should strive to provide positive experiences for collaborators that will encourage them to continue to work with statisticians.

As video 10 illustrates, even with the best intentions and preparation, conversations occasionally escalate into discomfort. Some strategies for de-escalating intense discussions include the following:

  • Check your own assumptions or stories about why the other person is reacting the way they are
  • Consider taking a break
  • Remain calm and professional without scolding or shaming the other person
  • Stop talking and start listening—really listening, without preparing—while the other person is talking

As Motomi Mori and Rongwei Fu state in their chapter—Competencies Needed for Statistics Leadership from an International Perspective—in Leadership and Women in Statistics, “Uncomfortable and difficult situations are frequently only the result of misperception. You can make an effort to view them differently.” It can be hard to confront conflict but communicating expectations and desires from both parties is important to having a productive and pleasant collaborative relationship.

1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)
Loading...

Comments are closed.