Tuesday, June 3, 2014

The Interplay of Affect and Cognition

I had dinner with my friend Alec again Friday night, and we had an interesting discussion about the roles that affect and cognition play in learning. We agreed that affect and cognition are both essential. To learn something, you have to want to learn it and you need to have the tools to understand what you are learning. But Alec felt that, in general, the cognitive tools are there and it’s a lack of affect that impedes learning in schools, while I felt the opposite.

We went back-and-forth on this topic for most of the evening without settling anything. But when I got home, I suddenly realized that I had been misstating my position. When I was a classroom teacher, I spent a significant chunk of my time thinking about and working on the affect of my students. So I had clearly identified a lack of affect; it’s just that the lever I used to shift their affect was providing them with the most powerful cognitive tools I could.

I watch a number of reality shows (Kitchen Nightmares, Restaurant Takeover, Restaurant Impossible) where expert chefs try to help owners turn around a failing restaurant. Most of the restaurants are failing because the owners lack the tools to run a restaurant, are afraid of change, and/or are too proud to admit they are wrong. Some of the tools they are lacking are fairly basic, such as keeping the restaurant clean and giving your staff clear direction, so that has to come down to affect (fear and ego).

The expert chefs work directly on affect, but they also work on helping the owners develop restaurant management skills at the same time. Most owners won’t change or admit to being wrong until they can envision a realistic chance that they can pivot and be successful. These tools range in scope and impact: from teaching the owner to cook a few new dishes, to focusing on food quality and always cooking from fresh, or being mindful to give your guests something special so that you stand out from the competition.

The more powerful the tool, the more empowered the owners and the greater the shift in affect. If the tools are powerful enough, some owners even report on follow up visits that their passion for the restaurant has only grown since the initial intervention and that they are experiencing a level of passion that they never expected or experienced before. For me, as a classroom teacher, I am always looking for cognitive tools that are easy enough for my students to use within their current level of affect, but powerful enough to shift their affect so that they can begin developing and applying a series of more powerful tools.

Apples and Bananas

So what does this look like in an academic setting? In 7th-grade, students learn to simplify expressions by combining like terms and applying the distributive property. However, instead of presenting it in that way, I give them an alternative cognitive tool that is both easier to understand and more powerful.

3 apples + 5 bananas + 2 bananas + 6 apples + 4 bananas

We start by adding apples and bananas. How many apples and bananas are there in total? Students sit up and take notice at this point because they know that I’m up to something. This is way too easy.

4 ( 2( 7a + b ) + 3( 5b + 6a + 2b ) + 10a )

Then I introduce the concept of baskets and nested baskets. Here we have 4 large baskets. Inside each of the large baskets there are 5 smaller baskets and 10 apples. Inside two of the smaller baskets, there are 7 apples and 1 banana. Inside the other three baskets, there are 5 bananas, 6 apples, and 2 bananas. How many apples and bananas are there in total?

At this point, all I am doing is making sure that students can translate the apple, banana, and basket notation into the real world. After that, it is completely up to them to unpack the baskets, and combine apples and bananas. (The answer is 168 apples and 92 bananas.)

This cognitive tool for thinking about and simplifying expressions is easy because it builds on top of existing mental models. The students can predict what they’ll have when they unpack a basket; they can reason through problems without trying to remember a set of procedures. It’s also powerful because they can simplify complex expressions with nested parentheses to any depth… and they can do that in the first 30 minutes of class.

It’s also more powerful because I can ask students to unpack the large baskets first (some students do that on their own anyways).

4 ( 2( 7a + b ) + 3( 5b + 6a + 2b ) + 10a ) = 8 (7a + b) + 12 (5b + 6a + 2b) + 40a

Most teachers would never do that because they would be afraid of confusing their students with too many rules to remember. The standard cognitive tools for simplifying expressions are more fragile and less flexible, and students experience that.

Now, if this is all we did, students would think it was a fun interlude, but that’s about it. If I taught a few more topics this way, it might shift some of their affects a little bit, but they would still attribute their success more to me than to themselves. In other words, they would still see themselves the same way.

To truly shift their affects, I try to get them to learn A, and then B. Then I try to get them to use A and B to learn C, where learning C doesn’t involve any creative tricks on my part. I want them to see that if they really understand A and B (to a greater extent than the standard cognitive tools allow), then they can apply their understanding to learn something really challenging, such as C… and that’s all them. Take that up another level and students start to internalize that they really are good learners, which shifts their affect and can ignite a virtuous cycle. They become passionate about learning and deeply understanding A and B because they know it will enable them to understand C, and they can see themselves doing that in any field they choose.

Teaching a Robot How to Dance

For apples and bananas, you are going to have to take my word for the impact it can have on students. But you can see for yourself how powerful cognitive tools can shift affect by watching two high school juniors, Amalia and Ariana, figure out how to program a robot with follow-the-leader behavior. The 90-minute session is documented in a series of video clips on the Computing Explorations website.

A group of us designed a five-week introductory computer science course for high school students with no background or prior interest in programming or robotics. The course is project-based: students learn how to program and how robots work in order to “teach” a robot to improvise a lead-follow dance with an unknown robot partner in time with unknown music. One of the intermediate challenges is to figure out how to get your robot to track and follow another robot.

If you watch the video clips, you will see that Amalia and Ariana are captivated by the robot and it’s life-like follow-the-leader behavior. It behaves like a cute pet. But that initial interest won’t take them anywhere unless we can give them the cognitive tools to replicate and understand the behavior themselves. After all, they will have to program much more complex behaviors in order to complete the project.

We start by giving them basic cognitive tools for understanding differential steering and the IR sensors. Then we show them how to link them using behavior-based programming (condition-action rules and behavior arbitration). They explore these concepts until they feel like they have a concrete understanding of how the robot works. Then they try to apply what they know to program the robot to orient itself to and follow the beacon. When unexpected behaviors emerge, they need to debug the robot.

At this point, Daniel (the instructor) models how to use the real-time instrumentation to figure out what the robot is “seeing” and to identify the behaviors being triggered. Amalia and Ariana quickly see the utility of this approach, and begin eagerly and capably debugging the robot themselves. They go through six iterations of the engineering-design process to get a robot with working follow-the-leader behavior.

By the end of the session, Amalia and Ariana are noticeably more engaged by the robot and the task than they were at the start of the session. This happened because of what they were able to do: they understood differential steering, IR sensors, and behavior-based programming (level 1); they applied those three elements to program the robot to do something (level 2); and they used real-time instrumentation and the engineering-design process to debug the robot (level 3), resulting in a rewarding accomplishment. And they did it in 90 minutes.

If we stopped there, Amalia and Ariana would have happily gone on to program more cute life-like behaviors into the robot. To create the behaviors they envisioned, they may even have gone out and acquired more powerful cognitive tools for understanding how the robot works. But in case that doesn’t happen (their affect is such that they stay in their comfort zone and plateau), we designed the final project so that they would be compelled to deepen and expand their understanding in order to complete it. Our goal was to get them to a place by the end of the course where they would be eager and ready to take on new challenges, and ideally transfer their new tools and affects to other domains.

Wrap Up

  • Affect and cognition are both essential for learning.
  • Helping students develop powerful cognitive tools is an effective and necessary way to shift affect.
  • Cognitive tools should be easy enough for students to develop given their current level of affect, but powerful enough to shift their affect, hopefully triggering a virtuous cycle.
  • In order for learning to have a significant impact on affect, students must see themselves differently because of what they learned.
  • Students will see themselves as more capable learners if they go vertical (rapidly building on top of what they have just learned in order to reach a place that they never thought they could reach).