A raspberry of an idea: How to do inspired science as a group

Note from Jeremy: this is a guest post from Rachel Germain.

****************

Let me begin by making one thing clear: I love working groups. I have participated in ~10 meetings so far, three of which I organized. I have one more coming up. They have all been net positive experiences — if not, I wouldn’t keep doing them. They are also an incredible effort to organize (see here for a past post on this topic).

What’s so great about working groups? For one, when they work well, they can have a transformative effect on science. They can open up new research fields while shifting energy away from older ones. They can show connections among subdisciplines and establish high-level consensus among many individual studies. Working groups can also have lasting effects on participants, creating long-term collaborative networks and providing researchers a window into how others approach science. The average working group paper accumulates citations 2-5x faster than non-working group papers (Wei et al. 2020). The catering is also usually pretty solid 👌.

So what’s the downside? A recent analysis of NSERC-funded researchers showed that, in terms of career progression, the benefits of working groups are not equally distributed among all participants in ways you can probably predict (Figs. 10 & 11 in Wei et al. 2020; also reported as a CIEE report). This doesn’t surprise me. In recent years, a lot of discussion has gone towards the importance of maintaining a balanced demographic composition in working groups — ensuring that participants are invited equitably to include a diverse representation of backgrounds and perspectives, ultimately leading to better science (Campbell et al. 2013; Hofstra et al. 2020). Funding committees evaluate working group applications with this in mind. This is great, but we need more discussion of what best practices for supporting Equity, Diversity, and Inclusion (EDI) during meetings look like. This is the purpose of this post — to initiate such a discussion and provide concrete actionable recommendations for working group leaders and participants. Note that our discussions here are not reserved to working groups only. They apply to any activity involving more than two people, for example, a paper discussion, lab meeting, trivia night.

Let’s begin with an hyper-exaggerated illustrative example designed to point out common working group interactions and how they align with inclusive discourse and equitable benefits. Imagine two groups of 10 people. Imagine each group of 10 people has been assembled to maximize diversity of perspectives, both based on demographics, career stage, and research expertise, and each is tasked with answering the same research question: “which fruit is objectively the best?”* Both groups meet for 5 days. Below are snippets of discussion from each group.

Group 1

“I like bananas” 🍌

“I like bananas too but they bruise easily and I’m pretty active so my bananas get knocked around. Apples have the portability of a banana but with less bruising”

“I’m not the biggest fan of the texture of apples (so grainy, also the inconvenience of the core) – grapes seem to combine the benefits of apples while also having a juicy texture. What I’m hearing though is that we all value different things in a fruit – why don’t we make a list of our favourite fruits and why we like them, to find commonalities?”

“Great idea – person X is great at multivariate analyses and person Y makes amazing figures, why don’t we summarize fruits based on where they fall on each “desirability” axis?”

“Amazing! That sounds like a ….fruitful…. avenue. Is everyone on board with exploring this plan? We can always pivot later. *all nod* – Which component would each of you like to work on?”

“Person Z also had an interesting point in our smaller breakout group about exploring the savoury fruits, e.g., wheat, avocados.”

“True – botanically speaking, grains are fruits. Let’s define the scope of our question so we can decide if that fits in or not”

End result: The group identifies key desirability axes and ranks popular fruits along each, weighting axes based on importance to overall desirability (e.g., ‘portability’ ranks higher than ‘shelf life’). Such an organizing schema allows researchers to identify the best fruit for a given application (Fig. 1) while also identifying the single best fruit in multivariate desirability space: raspberries. The group verifies their findings using consumer spending habits. The working group not only meets its goal but builds new supporting ideas around it. All participants feel like they contributed and arrived at an answer that none would have arrived at without the full group. They are excited to see what the group can do together next and have ideas for spin-off projects led by different people.

Group 2 

“Oranges are the best”

“I’m not sure, oranges have a lot of faults – for example, the handling time” (note: person spent 5 years as a professional orange taster)

“I disagree – you’ve clearly never had a good orange” *subset of group chortles

*turns red, stops speaking*

“….personally, I prefer apples”

“Actually apples are the best” 🍎 (aka person Y)

“You’re right, person Y, apples are the best. Let’s pivot our discussions towards what’s good about apples. You should lead this project.”

*half the group hates apples but wants to avoid confrontation and remains silent* (side note: I hate apples)

“I want to point out that most of you have yet to contribute.”

End result: An article comparing apples to oranges (Fig. 1), mostly reflecting the views of a few participants. Note that, in this uber contrived scenario, the group ended up with an answer to a question narrower in scope than intended (i.e., ‘are apples or oranges better?’ vs. ‘which fruit is better than all the rest?’). Additionally, the group arrived at a suboptimal answer because, as we’ve already established, apples aren’t very good. The project leader finds it hard to motivate the group after the meeting ends and feels unhappy that half the group did not pull their weight. People stop buying citrus, scurvy spreads to the masses.

Now, let’s analyze what happened in the two groups. Both groups included a mix of folks that differed in communication style – some direct or confrontational, others more democratic, some less assertive or deferential. The problem is, different personalities and communication styles are a product of background and life experience, and different communication styles are not always compatible or conducive to EDI (see related post on ‘devalued introverts’). In the two fruit groups, all 10 participants had something valuable to contribute even though they communicated differently, but not all were provided the space to do so. Let’s contrast a few interactions observed in each group.

Knee-jerk nay-saying™: Working groups benefit from outside-of-the-box thinking. Often discussions meander in all sorts of weird directions before arriving at the thing the group wants to write about. That’s absolutely fine and an essential part of the discovery process. One of my favourite digressions was in a macroevolution group where we explored the “rise of crabs” on Earth 🦀 as an example of clade replacement. You can observe ‘knee-jerk nay-saying’ in the actions of group 2 with the simple statement, “I disagree”, before any real discussion had ensued, with no solutions or alternatives offered (not very helpful). This halts any potential meandering which may lead somewhere unexpected and interesting. Now, my point here is not that we should withhold criticism — critical thinking is crucial. Rather, my point is simply to think about how we voice our criticisms and whether an idea truly is or isn’t worth exploring, beyond knee-jerk reactions. This is where a group perspective becomes so valuable.

Attribution error: This is perhaps the most upsetting of the interactions a participant can experience in a working group, and it tends to happen most to a non-random subset of people. In group 2, one participant (using passive phrasing) proposed that they personally like apples. Another participant, perhaps half listening, confidently declared that apples are the best. The group responds to the confidence of the second participant and attributes the idea to them. The first participant realizes, ‘it’s not my ideas that are not valued, it is me as an individual’. That’s not an inspiring way to feel. That person may withhold their best ideas in the future or withdraw from discussion. These kinds of interactions affect authorship orders and who we value as future collaborators, leading to real inequities in career benefits of working groups. Once again, group 1 got this right. All ideas are properly attributed and acknowledged for their intellectual contribution.

Unnecessarily definitive statements: The phrasing, “Oranges are the best”, conveys right and wrong thinking. It establishes that one person is objectively right and anyone with a different viewpoint is wrong. Are definitive statements wrong to use? No – but they could be the difference between a productive discussion and an unproductive one. It sets up others to either agree, take a defensive stance, or doubt their own knowledge due to the confidence of the speaker. What seems like a subtle phrasing issue has now set a weird tone for the meeting. Group 1, by contrast, got it right! “I like bananas” conveys an opinion without a definitive statement of correctness, allowing others to contribute opinions. Other forms of overly definitive statements that pop up in working groups: “This is not novel”, “I did this in my 1980 paper”, “This is simply incorrect”. More often than not, after much expended energy arguing and presenting counter-evidence, that person’s mind can change, meaning that a whole lot of time is spent not moving forward. Now, what if someone *is* objectively right? Easy – if so, their comment will hold up to group discussion and as the speaker expands on their argument. In science, few things are truly definitive. True confidence is a willingness to (respectfully) be proven wrong.

Discounting expertise of others: Working group participants are invited because everyone is viewed as an expert in something. When, in group 2, the professional orange taster voiced their expert opinion, two things happened: their expertise was not acknowledged and their opinion was discounted. Yes – the professional orange taster used more passive phrasing, perhaps deliberately to welcome alternative viewpoints. Their phrasing may erroneously be perceived as a lack of confidence, and thus, a lack of knowledge. Even worse, certain experts (like the orange taster) end up having basic knowledge about their own research field explained to them by others 😬. In general, we humans must fight against our tendency to conflate confidence with capability. Otherwise, we risk hearing only from those who speak the loudest.

Now, in the fruit working group scenario above, let me throw in 🗲a twist🗲: group 1 and group 2 are composed of the exact same 10 people in two alternate realities ✨. The difference in realities is that group 2 had an 11th, super domineering participant, whereas group 1 did not. Notice that in group 2 many participants were negative, whereas in group 1, no one really was despite these groups literally being the same people. My point here is that one ….bad apple…. can spoil the bunch. On the flip side, one single positive person can inspire and motivate. They can amplify others, curb negative behaviour, push ideas further, and create an environment that is more conducive to respectful, engaged discourse. This is what we should strive for – this is true leadership.

What can working group leaders and participants do to better support EDI?

General advice

1. Approach every interaction with others from a place of mutual respect, free of assumption about a person’s value. Everyone has something to teach you if you give them the space to do so and if you listen. I’ve had fascinating conversations with complete non-experts on topics I know well as they try bridging ideas to their own knowledge. A person who is silent in one situation may be a chatter box in a different environment. The onus should not be on individuals to change themselves in order to be heard – hear them as they are.
2. Approach new ideas from the perspective of exploration and discovery, rather than defaulting to knee-jerk nay-saying. It is much harder and more impressive to see and develop the potential of a new idea than it is to shut one down. Ask yourself why your first instinct was to respond negatively – was it based on the idea itself or your subconscious feelings about the person that proposed it? If the latter, ask yourself why those feelings exist in the first place and what you can do to combat those feelings.

Specific, actionable recommendations

1. Training: The first step to change is realizing this is a pervasive issue and being able to identify the many ways it can play out. Once you can identify what poor EDI practices look like, you’ll notice them when they happen and can respond appropriately. Reading this post is a good start 😉 but beyond the here and now, try thinking of what an EDI support plan might look like for your meeting. Share your plan with your group before you meet so everyone knows what the deal is.
2. Establish a rotating ‘facilitator’ role: Nominate different participants to keep track of ideas with proper attribution and provide support or counter arguments if someone is being dismissed (“wait, let’s revisit person X’s point, we haven’t fully explored that yet”). Doing so can change the vibe of the meeting and prevent negative behaviours from spreading to other group members.
3. A simple phrasing pivot: Instead of saying, “this is not novel”, say, “how do you see this proposal building on what’s already known on XYZ field or famous-person Z’s research?” – aka, give the person an opportunity to expand on their reasoning in a non-confrontational way. If you took a hard stance on an issue and were proven wrong, simply say so: “you’re right – I see that now” rather than finding loopholes for why the right answer was actually consistent with yours given some underlying set of implicit assumptions you had in mind but never stated. We convey much about how we view ourselves and each other (e.g., level of respect) with the specific words we say and how we say them.
4. Acknowledge and amplify the contributions of others: When building off an idea initiated by someone else, simply acknowledge them. Say, “Like person X was saying earlier…”, or, “I think this is what person X was getting at – to build off of it….”. If someone makes a good comment in a breakout group, repeat it to the larger group with proper attribution (like group 1 did re: the savoury fruits).
5. A bit of damage control: As the group leader, speak privately to participants who may be creating a negative dynamic. Reach out to participants who haven’t spoken up much to see if there’s a reason why and what they need. Hot tip: start with, “What do you think about the discussions so far?” and not “I noticed you haven’t said anything yet” 👀.
6. Know the expertise of all group members: At the beginning of the meeting, have the group leader present the various expertise of invitees and how their knowledge aligns with the working group topic. Alternatively, have participants present short talks themselves. It is absolutely worth the time. Acknowledge this expertise. Say, “Person X is an orange tasting expert so please feel free to weigh in, but here’s how I’m approaching this idea.”
7. Provide (optional) materials for participants to use to prepare in advance of the meeting: Some people feel more comfortable contributing with preparation, especially if they have less working group experience or are new-ish to a topic. Set up a google doc with some leading questions people can fill in with thoughts (which also makes it easy to see who proposed what) and suggest a reading list of key papers.
8. Have small breakout groups: Large groups can be intimidating, but breakout groups can be a chance for folks to feel comfortable contributing ideas. Properly attribute ideas to individuals when summarizing the breakout group discussion to the full group.
9. Diversify tasks: Group leaders can ask participants in advance about what they hope to contribute or what skill they can share. Everyone, regardless of career stage, has some unique skill not shared by others, so best to find ways to play to those strengths. Someone may really like writing text, others building models, someone else figure construction, or another synthesizing the literature.
10. Diversify projects: Say the direction of a project dips outside the expertise of a participant or is trending in a direction that is of less interest to them. One way to keep that person engaged and involved is to encourage them to lead a spin-off project that aligns their interests with the general theme of the meeting. People in the group can then sort into different projects based on interests, thus increasing the number of participants who end up in notable authorship positions across all working group papers.
11. Equitable task allocation: There are many tasks associated with producing a working group product. Be mindful of who is allocated a lower level task (e.g., note taking) vs. higher level tasks (e.g., idea generation). Even better, assign lower level tasks to participants at random, with duties rotated throughout the meeting. If someone wants to take notes, then great! Note taking is important – but what’s also important is not assigning certain tasks to certain people who haven’t voiced a desire to do them.

I’ll end by reiterating that I love working groups, and that in my experience, everyone is well meaning. I’ll also say that I’m not always perfect at all the things discussed above – I’ve certainly caught myself getting riled up in the odd paper discussion. What’s important is the desire to learn what it takes to better support others, to self-reflect aways, to put in the work, and as a result, to have the best dang working group you can possibly have.

Have suggestions I haven’t covered? Add them to the comments section! 

Acknowledgements: Thank you to feedback and discussions with Germain lab members, as well as feedback from Diane Srivastava, Jeremy Fox, and Tess Grainger. *hypothetical working group topic was inspired by weeks of fruit-related debates initiated by J. Sakarchi of the Germain lab. Many people have contributed ideas to our fruit debates, not always by choice – notably, those stuck in the car with us en route to the ASN meeting.