We’ve been talking about “learning skills” as if it was clear what that means. It isn’t.
Consider an auto mechanic’s skills. They include:
- Lifting a car on a hoist.
- Removing a bolt with a torque wrench.
- Fixing a tire.
- Getting a car to start normally.
The first two are low-level, or specific, skills. The goal is detailed (remove a bolt). The steps to meet the goal are clear. Inspect the bolt, find a torque wrench…
The third is less specific. “Fix” could mean replace the tire, or maybe repair it. Which one? Depends on the damage, the customer’s budget, etc. Some cars have special procedures for removing the tire. For example, some sports cars have screw-on hub caps that are threaded in reverse on one side of the car. Instead of “lefty loosey, righty tighty,” it’s “lefty tighty, righty loosey,” but only on one side.
The actions needed to fix a tire can’t completely be specified in advanced. The mechanic needs to ask questions, inspect the tire, etc., and make decisions about what to do. That’s problem solving.
The fourth goal (making a car to start normally) requires even more problem solving.
Good problem solvers have many patterns in their heads. A pattern, or a schema, is a rule or procedure. For example:
If nothing happens when you hit the ignition switch, it’s probably an electrical problem.
Part of learning to be a good mechanic is learning patterns. An expert will have thousands of patterns in his/her head meat.
Abstract patterns apply across many different problems, like:
Break a big goal into smaller goals. Work on the smaller goals independently.
However, most patterns are context-specific. For example, the pattern about ignition and electrical problems won’t help someone analyze a balance sheet. An expert in one area won’t automatically be an expert in another.
How do people learn patterns? They can learn them from experience, inferring the patterns. They also can learn them directly, if patterns are made explicit. Explicit patterns should describe the situations in which they apply. That is, when they can be triggered.
Patterns are triggered by cues in the environment. For example, “I’m at work, here’s a car the doesn’t start, the boss says I should fix it.” In essence, a cue is another type of pattern.
Sometimes students use cues that don’t make a lot of sense. For example, a finance textbook might talk about internal rate of return (IRR) in chapter 5. Students create the pattern:
If the question is about chapter 5, the IRR procedure might work.
The pattern works for the chapter 5 quiz, so it gets reinforced. The pattern works for the midterm as well, which was over chapter 4 to 7. The pattern is reinforced again. In fact, the pattern is so good that some students might not have any other patterns for when they should use IRR.
Time for the cumulative final, with five problems. Now there are no cues about which chapter to use for each problem. Argh!
Students take the second finance course the next semester. The professor complains:
It’s like they never learned about IRR. But I know they did! I taught them IRR last semester! Argh!
The “chapter 5 could mean IRR” pattern only works for the first course. It won’t work for the second.
Students don’t automatically know when to use which patterns. If you use them in teaching, you should make sure that students know when each patterns applies, and that the cues are not course-specific.
Researchers differentiate between shallow learning and deep learning. Shallow learners memorize definitions, and learn low-level skills. They don’t learn problem solving, however. Deep learners know facts and low-level skills, but they learn problem solving as well.
One person might use deep learning in one situation, and shallow learning in another. Motivation makes a difference. However, the structure of a course makes a difference, too.
Deep learning takes more time than shallow learning. In particular, it takes practice and feedback. Lots of it.
Professors underestimate the time deep learning takes. They were most likely the best students in their discipline, but what was easy for them is not easy for everyone. Further, researchers have shown that experts often forget how they solve problems. It sounds strange, but it’s true.
Consider driving. Drivers make small steering adjustments to keep their cars on course. Experienced drivers have made thousands of these adjustments. Their subconscious takes over the task, and they do it without even thinking about it. It takes no attention at all. They do the task well, but might forget how they do it. They don’t need to remember.
Why is this important? When students don’t have time for deep learning, they switch to shallow learning. A student has, say, 120 hours to put into a course over a semester. Deep learning might take 200 hours for that particular student. That’s an extra two work weeks for that one course. Will s/he find the extra time? Maybe not, depending on life circumstances, and motivation.
The structure of a course affects learning at least as much as the course’s execution. Many courses make deep learning impossible for all but the handful of students who are as good as their professors were. We’ll talk more about this later.
Cyco has some features for deep learning. One is patterns. Authors can create patterns. Here’s part of one for a programming course:
Each pattern has a situation (when to use it), action (what to to), and explanation. It can also have attachments.
Students use patterns in two ways:
- Learning patterns
- Using patterns
First, students learn about the patterns themselves. Authors embed patterns in the content, and explain them.
This is what students see:
The second way students encounter patterns is after they have learned them, when they need to use the patterns in exercises, projects, and even on the job. For this, there’s a separate list of patterns students can use:
Students can look through the list, and choose patterns that suit their needs.
Authors can design pattern in sets. For example, there can be high-level patterns (“Make the car start”), low-level patterns (“Hoist the car”), and mid-level patterns (“Check the electrical system”). Patterns can refer to each other.
Danger, Will Robinson, danger!
Patterns can be a two-edged sword. Students wanting shortcuts will look for simple rules that apply all of the time, without having to think very much. They’ll doom themselves to shallow learning. Deep learners know why different patterns make sense in different situations.
Authors can help students build mental models of processes. They can explain how patterns emerge from the model. (In fact, patterns offer insights into how processes work.) When no patterns fit a situation, good problem solvers fall back on their mental models.
Humans use patterns. That’s simply the way our brains work. Patterns can help people learn skills. However, patterns should be used with care.
Jill is a pseudo student, or pseudent (the p is not silent).
Pseudents don’t always succeed. They get stuck, and need help getting unstuck. They misunderstand tasks, and use the wrong patterns. This helps the real students understand problem solving, what can go wrong, and what to do when that happens.
Deep learning means learning problem solving, as well as low-level skills. Deep learning takes practice and time. Professors often underestimate the time deep learning takes. When students don’t have enough time, they switch to shallow learning.
Patterns are rules showing what to do in various situations. CyberCourse helps authors define patterns, and embed then in content. There’s also a pattern list that students can use when doing exercises.