Changing Mental Models

In my last few posts, I’ve been exploring how building shared mental models is key to integrating data teams into biotech organizations, particularly “tech biotechs” that have a significant AI/ML component. I’m working towards a series of posts about specific things you can do to help build these SMMs, but before we get there I want to discuss how individuals change their mental models, and what can get in the way of this happening.

In this post, I’ll cover two frameworks: First, I’ll summarize a model from the study of adult cognitive development that describes the different relationships people can have with their mental models. Then I’ll go through a list of conditions that, in my experience, individuals in a biotech setting need before they’ll shift their mental models.

If these ideas resonate with your own experience, please let me know in the comments below. Also consider signing up for my weekly newsletter where I send out short ideas to get you thinking about this in your day-to-day, and announce upcoming blog posts that will go deeper into this topic.

The Four Stages

In her book Changing on the Job, Jennifer Garvey Berger describes four stages in most adults’ cognitive development. Not everyone makes it through all four stages, and many people spend a lot of time in in-between stages, but it’s useful for framing how we think about individual’s relationships to their mental models. She describes the stages in terms of perspectives, which are technically different from mental models, but the stages effectively the same either way. So I’m going to describe them in terms of mental models because that’s my MO.

As you’re reading through this, you may start to wonder what stage different members of your friends and family are at. But RESIST THE TEMPTATION. Berger stresses that it’s impossible to determine someone’s stage from only observing them. Instead of trying to diagnose individuals, you should think about approaches to shared mental models that will be effective for team members at all the different stages.

For the first stage, you’re born with no mental models. You observe things happening around you, but you struggle to understand why and to figure out how to interact with others. While you quickly develop mental models for the physical world, mental models for how other people think and feel and react take longer to develop. Most people get past this stage while they’re still a child, and to be a functioning adult you pretty much have to be in a later stage.

In the second stage, you adopt a mental model from your family, peers, teachers, etc. Because this model is externally imposed, you don’t feel a sense of agency to change it. So you treat it as a narrow, rigid set of rules and rely heavily on external validation to reinforce the model. Think of a high school or college student suddenly discovering a philosophy or an academic subject that seems to explain everything they ever wondered about. They can become obsessed because this new mental model defines how they think about nearly everything.

The third phase starts when you adopt your own, internally constructed mental model. It may be based on the model from Phase 2, but the key is that you now control it. You choose when to update it and how flexible it can be. The external world continues to be an input for you to update your model but you no longer expect others to have the same model. They’re no longer a source of validation. This autonomy is the key difference between Phases 2 and 3.

In the final stage, you learn how to switch between mental models, applying different ones in different situations, and leveraging the different mental models that other people apply. Recognizing that there are different equally valid mental models is an important step in getting here, but the hard part is learning when and how to employ them. So reading these blog posts about shared mental models won’t instantly get you to Phase 4, but it’s a start.

According to Berger, it is possible to be professionally successful, even as a leader, in Phase 2 but this mostly happens with people who spend a long time in a single, narrowly scoped career. Because their externally imposed mental model is highly compatible with their job requirements, it can serve them very well. The issue is when the job changes or they move to a new context, and all that goes out the window.

So most of the people you encounter in a biotech organization will be in Phase 3 or 4. It probably goes without saying that it’s easiest to build a shared mental model with the folks in Phase 4. But in practice you’re going to have a mix of people from at least Stages 2-4.


Now that we understand the four phases that most adults go through with respect to mental models, the question is what allows someone to shift their mental model in a particular context, particularly in a biotech organization. Below is a list of things I’ve noticed, based on a mix of observation and educated speculation. In other words, this has not been as carefully researched as the four phases above but I think it’s still helpful.

Psychological Safety

Shifting the way that you approach your work is risky, particularly when it requires changing your mental models. When something like AI/ML is added to a workflow, it might eliminate something else that you relied on, maybe even all or part of your job. What if you can’t keep up with the changes, or just don’t like where things end up? What if the changes cause your work and timelines to slip? What if someone uses this as an excuse to criticize you, whether or not you actually screwed up?

For members of an organization to be willing to change, they need to know that their colleagues and leaders have their back. They need to know that the organization will support and protect them. They need to feel like they have leeway to experiment and fail without damaging their reputation, their position or their ability to work effectively.

A lot has already been written about psychological safety in the workplace, how to recognize when it’s missing and what to do about it. Without it, you can forget about creating a shared mental model. If your organization does not provide psychological safety, it will be treading water until you or someone else fixes it.

Recognizing the Problem

The next thing someone needs before they’ll shift their mental model is to recognize that the current situation isn’t the best you can do. Maybe the data scientist thinks that spending half their time cleaning up messy data from the lab is just the way things always are. Maybe the bench scientist doesn’t think that their most vexing question could be answered with a new machine learning or data science technique. Maybe everyone thinks there’s just no way to automate the current slow, manual processes.

You won’t recognize a pain unless you know what it feels like without it. You won’t recognize something that needs to be fixed unless you think there’s a better way. If you notice that members of your organization aren’t interested in trying to fix a problem, it may be because they don’t think it’s a problem.

Seeking Change

Just because you recognize there’s a problem doesn’t mean it’s worth fixing. Maybe cleaning all that messy data seems easier than trying to convince the bench scientists to do things differently. Maybe explaining the biology to a data expert feels more painful than getting a good-enough solution on your own. Maybe figuring out how to automate all those manual processes seems like more work than the manual steps you’re doing today.

Shifting mental models is a slow, exhausting and sometimes painful process. If you expect that pain of creating change to be worse than the status quo, then the logical choice is to stick with the status quo. If you recognize someone who seems unhappy about the situation but won’t admit it or doesn’t want to fix it, it may be because they don’t think it’s worth the effort.


Just because you want to change something doesn’t mean you know how, or feel like you have the power. The data scientist may not think the bench scientists will listen to their suggestions about improving data collection processes. The bench scientist may think that the data team is too busy with more interesting and important problems. With big, complex processes, any one person may not feel empowered to recommend changes outside what they’re directly responsible for.

Organizational change is a complex and subtle process, and most members of your organization won’t have experience with it. Most won’t even know where to start. They may not recognize that others will have to update their mental models, not just their actions. If you recognize someone getting increasingly frustrated about a situation, it may be that they want to change it, but don’t think they have the power or ability to get others to change in ways that will address the problem.

Willingness to Change

If you’re unwilling to be flexible in how you solve a problem, you’ll quickly become the last thing that needs to change. The data scientist may need to adopt vocabulary and conventions that are closer to realities in the lab. The bench scientist may need to structure their experiments differently to support a new type of analysis. The person responsible for the one efficient step in that long complex process may have to adjust to the changes that make the rest of the process efficient too.

It’s easy to recognize what other people need to do differently, but most of these problems require everyone to change. Even if you’re in Stage 3 or 4 of the framework above, you may be unwilling to shift your mental model in this one particular instance. When someone has successfully gotten the organization to shift in a better direction but they’re still unhappy, it’s possible they’re being too inflexible to accept that they’re the final piece of the problem.


The good news is that for each type of obstruction, there are ways to push things towards a solution, depending on your particular leadership style. You may need to communicate a vision of a better way. You may need to persuade someone that the solution is less painful than the problem. You may need to give them the formal authority that they lack. Or you may need to convince them to change their expectations for what the solution looks like. None of these are necessarily easy, but they’re all possible, once you recognize the importance of shared mental models.

4 thoughts on “Changing Mental Models

  1. While I agree about the whole idea around mental models. Often the issue is that incentives aren’t aligned within the company. It’s easy to overthink/over engineer the problem, maybe it’s not that complicated 🙂


    1. That’s a great point – Incentives will trump mental models every time, so misaligned incentives will tend to defeat even the best shared mental models. But in my experience, even when incentives are aligned, different mental models can impact how different teams respond to those incentives.


  2. Reading this reminded me of George Lakoff’s work. If you haven’t already, you should check him out. A lot of it has a heavy skew towards the political realm, but the concepts he discusses are general.

    Lakoff’s central thesis is we each have a set of fundamental conceptual metaphors we use when processing information. This impacts our frames of reference, and thus how we react to a given input. Being metaphor based, one can look at wording someone uses to understand the metaphor in play. One can use that knowledge to cross the bridge in one way or another, e.g. most of his books focus on coaching the American Left on how to outmaneuver the Right. But in most people’s day to it is important to understand how to shift to another metaphorical context. Or even if it’s worth the effort to try to shift the other person out of theirs.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: