More from productive discussions at Aarhus University

Standard

Thinking about the big picture

I cannot underestimate how valuable it is to talk to people in other disciplines, in other schools, with other perspectives. I was challenged yesterday to think about the big picture of my work and philosophy on labs, and to really step back to see the forest from the trees. What I have come to is this:

The way we teach in labs is not new. There is lots of fundamental research that tells us that the components that we used are all helpful and productive for learning. The key features are as follows:

  • Students make comparisons
  • They reflect on them and make sense of them
  • Then they decide what to do about them (how to act), ultimately leading to new decisions
Cycle

Figure from Holmes, et al. (2015) in PNAS with “Decisions” added to the center to reinforce the student autonomy involved.

The importance of making comparisons has been well documented (e.g. Gick & Holyoak, 1980; Bransford et al., 1989; Bransford & Schwartz, 1999). The importance of revising and iterating with feedback has been well documented (e.g. Ericsson, et al. 1993; Schwartz, et al. 1999 — I should have more references than this… hmm…). Clearly, combining the two is going to be helpful.

And while we focused very specifically on the sorts of comparisons that introductory physics students make with data and models, these comparisons could be a number of things. For example, coming up with multiple experimental designs, comparing them, determining which is best, and then trying one of them. Of course, then one has to evaluate the outcome of acting on it to try to determine whether it worked as well as expected (compare to expectations or predictions). This may then lead to trying the other design or modifying it in a small way. The initial comparison, however, can happen before students take any data in a lab, which can be beneficial when iterating and repeating measurements is costly (e.g. non-reusable materials in chemistry or biology labs). Another example would be to compare to another group, each try the designs, then compare how well they worked to determine the optimal design.

Comparing to predictions is also useful, but the problem with most courses is that students stop at the comparison. They reflect, but often draw funny conclusions, such as “they disagree because of human error.” [What does that mean?!] Encouraging students to then test that interpretation (or act on it in some way) is going to make those comparisons much more meaningful.

I’m pretty sure I wrote about a lot of this in our papers, but I think I needed to pushed on it to really come to believe or understand it.

Advertisements

Ideas and perspectives on inquiry labs from discussions at Aarhus University

Standard

My first day (morning) visiting the Centre for Science Education at Aarhus University has already elicited some very productive discussions and new perspectives.

The first is the notion of a continuum of inquiry.

One member of the group described how people often see labs as either being cookbook (where the recipe for how to conduct the experiment is given with no space for critical thinking or decision making by the students) or fully open inquiry (where students come up with their own goals, design, etc.). They then described how the framework for labs I presented shows that there is a continuum to that inquiry.

More recently, I’ve been thinking about lab course design based on the “cognitive task analysis” for experiments in physics, recently published by Carl Wieman (here). The cognitive task analysis presents a list of all the decisions and cognitive tasks that one must do to conduct an experiment in physics. This list, merged with the notion of a continuum of inquiry, can present a very clear perspective for developing labs.

Each lab can focus on which decisions you want to leave open to the students. No single lab needs to address all of the items, rather each course (or experiment) can focus on developing the skills necessary to complete one of the tasks. In our SQILabs framework, we’ve been focusing on the experimental design, data analysis, and evaluating results. In something like the ISLE labs, the focus also extends to determining the goals and criteria (which hypothesis are you going to test, for example, and what outcome will help you decide whether the hypothesis is correct).

This can also inform curricular design, such that by the end of the program, students can complete all of the tasks in, for example, a senior thesis project.

The second is the use of invention activities for experimental design.

Several members of the group have been working on evaluating and improving chemistry lab courses. In many of these, the fact that materials are not re-usable (e.g. chemicals get used and then you need more chemicals) makes the iterative, repetitive cycling of experiments undesirable (and costly). We came up with the idea to iterate on the experimental design before they actually conduct the experiment. This can be done using the invention as preparation for future learning framework.

That is, students can be given the research goal and equipment and then be tasked with coming up with an experimental design to meet that goal. They would then work in groups to define all the different elements that need to be considered. Then, a class discussion can compare the various designs, ultimately leading to a best design, presumably that intended for the experiment by the instructor. Students can then be handed the regular cookbook protocol to follow, but now with a more fundamental understanding of why we perform each of those steps. This maps right on to the invention activities we’ve been using to teach data analysis. Students are given a problem and have to invent a way to solve it. Then we discuss the different approaches, ultimately leading to the expert solution (or equation). This type of activity has been shown to help students apply the equation, but also to understand the various components of the solution, and to better recognize faults in variations on the equation and what components they fail to address. It, essentially, breaks apart the black box into various components and pieces.

Another option (or to do in addition) would be for students to explain next to the protocol why we perform each step, what it should accomplish, and to group them by different phases of the process. This again forces them to make sense of each step, rather than following the instructions blindly. One could also get students to predict what they should see at each step. That way they have comparisons in hand while conducting the experiment to allow them to make sense of the outcomes as they get them and think on their feet, while doing the experiment.

One of the group members had used this approach (getting them to explain why we perform each step) and found that a 10-15 minute activity at the beginning of lab saved the students 1-2 hours doing the lab, because they knew what was going on. The instructors also got students to assign who was going to do what at each stage to better balance the division of tasks. This was in response to some observations (both in my work and theirs) that students distribute the work unevenly during the lab (often based on gender or other demographics).