Redesigning our lab practices
posted on December 20, 2019


by Ari Benjamin.

Every lab is a set of practices, traditions, recurring events, and expectations. These practices determine the daily rhythms of its members. They affect personal growth, research quality, the relationships between its members, and how enjoyable it is to be a scientist. Unfortunately, for reasons relating to complacency and authority, I’ve found that a lot of labs just organize themselves in standard ways and don’t regularly switch things up.

Just picture that “toxic” lab down the hall. (You know the one). Communication is stifled, trainees don’t feel nurtured, science is competitive, and everyone is scared of being wrong. These are some common big failures, but even the best labs have room to improve.

Now, group leaders have the best of intentions. It’s just hard to be a perfect manager – too much is expected of advisors to be perfect at everything. For us, it was important to realize that we only had to try to improve things and things would improve just by virtue of our trying. It was the process of lab-improvement that matters, not just the new solutions. A good lab isn’t good from the outset, but rather has a process of reflection and regularly asks, “How is this, what we’re doing now? Does it support our values? How could it be better?”

This fall we challenged ourselves to make the Kording lab better suited for us. Our key realization was that we could design our lab towards the values we think make a good lab. Behavior follows structure!

We changed a lot, and learned a lot, and we hope our lessons can be of some use to others.

Overall approach

We took a look at the way we run our lab in the same way that a human-centered design consultant would. Essentially this just means we separated the overall process into three separate steps. These occurred in three separate all-lab meetings.

We came up with over 100 ideas for good things our lab could do. We settled on implementing just one new thing each week. We’ve put in place some awesome things! (Scroll to the bottom if you just want to see what our new practices are.)

At all times it was democratic: everyone had equal input. I can’t understate the importance of this. Since we were looking for a system that works for everyone, it was really helpful to hear from everyone what matters most to each of us. Everyone equally stated their priorities, came up with solutions, and decided what would be best for everyone. At our lab meeting each Wednesday, we now vote on one new thing to implement in our lab structure. What we ended up with is an environment where (I think!) everyone in the lab really feels empowered to contribute in their own way.

Where we got our process from

Very creative and smart folks in human-centered design have come up with ways of being creative about creativity. Many of the same folks have also thought a lot about what makes a great team and a great working environment. If you’re interested in learning more, I recommend checking out the courses put out by IDEO. I can vouch that these courses are entirely worth it. (I personally took the course entitled ‘Cultivating Creative Collaboration’).

How Konrad got involved

We were really lucky to have an advisor with an open mind to new possibilities. The entire process was led by labmembers, and Konrad even stepped out at various points in the process when he felt his presence would stifle suggestions, conversation, or disagreement. It ended up being really useful that everyone in the lab felt so empowered, even new members and rotating students. Our process won’t work for all labs, I’m sure, but usually the reason trainees won’t speak up or act to improve their situation is that they feel intimidated by their advisor.

Here’s what the process looked like at each stage.

Stage 1. What makes a good lab environment?

Our lab practices were established long before anyone (besides Konrad) had arrived. In our first session, we spent some time just outlining what those things were. We don’t have weekly lab meetings, unlike a lot of labs. Instead, we largely met spontaneously and individually about our research. We also had weekly “lab teachings” in which one lab member would teach everyone about some concept they found interesting. We ended up keeping most of these practices (free coffee!) but it was helpful to recommit.

We then outlined some of our goals in an open discussion. What should a lab be? How should people feel while a part of it? One of us stood at the whiteboard and put down what people said. You can see some of our priorities above.

Out of the values that came out of that discussion, I want to single out the concept of “psychological safety”. (You can Google that one and jump right into the management consulting world.) Any manager (or PI) should strive for this. When an environment promotes psychological safety, it promotes:

  1. A learning mindset (in which it is OK to make mistakes),
  2. A curious mindset (in which people are curious of everyone’s thoughts.
  3. A vulnerable mindset (in which people feel welcomed to express their shortcomings).

Once everyone had their say, the next step was to actually design towards what we just established as our collective goals. The central idea here is that behavior doesn’t just come about spontaneously. Recurring practices (like lab meetings) are essential. But what should they be?

Stage 2: How might we design our lab environment?

Diverging & Brainstorming

“Coming up with ideas” sounds organic, but it’s helpful to be systematic about it. We had two phases: 1) Diverging and 2) Converging. In the “diverging” stage, the goal was to come up with as many and as diverse ideas as possible. We saved any critical thought (“that won’t work because…”) until the “converging” phase. By brainstorming ideas without critiquing them at the same time, you can encourage creativity and diversity. People feel encouraged to try things out without risk of being shut down. This process also avoids the failure mode of running with the first good idea; the best ideas usually come later.

We spent 5 minutes in silence writing ideas onto post-its. I highly recommend the post-it approach as opposed to hand-raising or open discussion. Our best ideas came out from that silence. It’s efficient and gets ideas equally from everyone. IDEO folks will also tell you that post-its allow the idea’s originator to detach themself from their own ideas.

After a few rounds of post-its and idea presentations, we had over 100 ideas of awesome-sounding lab practices.

Converging

We wanted to ensure that the process was democratic throughout. (This is all about creating a lab environment that everyone feels they own, after all.) Since voting on 100 things is a lot, we took another trick from IDEO’s book and voted in parallel. Everyone had several votes in each of several categories, and in a mad dash we voted on everything and then dropped the half of ideas with the least votes.

Our shared google sheet for voting on ideas.

Stage 3: A continued process of weekly voting.

Instead of deciding right away what the full list of new practices will be, we took it slow and chose things one at a time, once per week. We did this weekly all through the fall. In the first 10 minutes of lab meetings (or rather, lab teachings), we nominate a few possibilities from the list. Everyone votes for their favorite, and the option with the most votes takes it. Then – ta da! – we have a new lab practice.

Every new practice has a czar

When we decide on a new practice, we also nominate a person to be in charge of bringing it into being and ensuring it continues to happen. We found this to be enormously helpful. The czars rotate once a semester so people don’t get burned out.

5. Overall takeaways

More than anything else, I’d say the democratic aspect of this process helped the most. A lot of standard practices – and things we already did – actually work quite well. The problem is that when they don’t have 100% buy-in, these practices tend to grow less helpful. Now everyone feels like they’ve decided to be doing them, we’ve really amped up our collective effort.

It’s been fun to be creative with this. I’m proud of some of the changes we made. We’re more social as a lab, more supportive, and I think the high morale is going to be self-sustaining.


Our current list of practices

Our practices generally fall into the categories of physical environment, scheduling, orientation, social, research communication, mentorship, documentation, and resources.

  1. Lab teachings: A member of the lab leads a ~1 hour ‘teaching’ on a subject of their interest. It cannot be updates on their current research. This is a longstanding Kording lab practice. http://kordinglab.com/2018/08/29/upcoming-lab-teaching.html
  2. 1-on-1’s: Every member of the lab is required to sit down with every other member of the lab and discuss their work for an hour. Refreshes once a semester.
  3. Physical environment: We now have
    • little magnetic pandas that are used to indicate if you would like not to be bothered (see pic)
    • yoga mats
    • a hammock
    • a hangboard
    • plants
    • a tea kettle
    • an espresso machine and free coffee
  4. Tea talks: The tea czar makes tea, and someone gives a 10 minute talk about anything. Usually someone talks about an idea they’re working on in their research. This happens weekly.
  5. Monday morning donuts: Our lab is not an early-morning lab. Some of us need a push out of bed. For that reason we have donuts in the lab served at 9:00am every Monday.
  6. A potluck lunch: Once a semester, we have a potluck.
  7. DJ Friday: We just play music in the lab on Friday afternoons! We have a shared lab YouTube playlist, and anyone can add to it.
  8. Failure Friday: Academia is full of times of disappointment, big and small. To bring these events into the open, we have a Failure Friday every now and then in which people are encouraged to discussed things that did not go as hoped, what they learned, and (if asked for) for advice going forward.
  9. ‘Extreme’ (pair) programming: All about learning to code! Everyone is matched with someone else, and the two of them code together. One codes, the other watches, and then they switch.
  10. Lab calendar: A simple Google Calendar that everyone can edit. For these lab practices but also for seminars and so on.
  11. Web redesign efforts: These are more one-off things. If we vote on a specific web improvement, then we appoint a czar and they are in charge of coordinating the lab to get this done. Often these are resources we want, like how-to’s.

That’s the list! There are still things to be voted on, too – it’s an ongoing process. An orientation buddy system and more social events are likely upcoming additions.