In Block 3, our sixth and seventh graders looked at the scientific method as applied to psychology.
We started by reading a scientific journal article reporting on four experiments on ego depletion—the idea that self-control is like a muscle and can be temporarily exhausted through intense use. The paper consisted of four mini-lab reports, including introduction, methods, results, and discussion, on each experiment. I broke the paper up into these four parts and gave each student one part. They started reading these during class. It became clear that they needed some background in statistics to interpret the results, which were reported with p-values.
We spent the next few weeks looking at statistics. First, students did activities simulating data collection and analyzing the results using R (a statistical programming language that also graphs results). They got the idea of a normal distribution, average, standard deviation, and that the more data you collect, the lower the noise and the better you know the true average of the distribution you are drawing from. We calculated some simple t-tests and p-values.
During this activity, students started to wonder about the probability that the hypothesis is true. I found an interesting paper that surveyed psychologists, instructors of psychology methodology classes, and students who had completed a class in statistics, and found that almost all respondents in every category had some misconceptions about the meaning of p-values. I gave the survey to our class, and indeed, all of them had many misconceptions. Following the paper’s recommendation, we then discussed Bayes' Law and conditional probability, and the students retook the survey. Their understanding improved, with many of them now answering like the respondents in the survey—still not perfect!
Returning to analyze the original ego depletion paper, students presented the four experiments in groups to the class. I then asked them to look at a newer paper challenging the ego depletion results, and finding that the effect appears to disappear with larger sample sizes, and controlling for publication bias. To my surprise, many students remained convinced that ego depletion exists: it just makes intuitive sense, they said. Finally, for a real-life example, we looked at the well-known paper showing that parole judges make more favorable judgements (from the prisoner’s point of view) first thing in the morning, or after a snack break.
The last two weeks were spent working on individual research projects. Students picked three papers on a particular topic and wrote a paper summarizing and synthesizing the results. They also gave short presentations to the class. The students discovered this block that while it isn’t always easy, they can actually read the original scientific literature in certain fields. I hope this experience will help put the common student task of lab-report writing in perspective, so they get a better idea of what it really looks like at the professional level, and what qualities distinguish good and bad reports.
-- Emily Eames