By Bill Wallace, Science Teacher at Georgetown Day School in Washington, DC

My Physiology class of high school juniors and seniors is alive with various sounds, smells and lights all directed at pill bugs in choice chambers on the lab benches. One pair of students is intently playing the same By the Seaside ringtone on their iPhone over and over again. Another pair is spraying Axe deodorant leading the lab to smell like a middle school boy’s locker room. A third pair has borrowed klieg light gels from the theater in order to produce red and yellow lights on their bench top. All of the groups are absorbed in counting pill bugs and recording their results in a spreadsheet on their laptops. Occasionally, I hear “No! How can you pill bugs be so stupid?” or “Yesss” and an exchange of high fives by the two students with cooperative pill bugs. They are doing something very different from the typical 20th century classroom.

Why is this important?

Scientific thinking is as much an attitude as it is a particular way of thinking. At the high school level, the attitude of the student towards laboratory activities should change from asking “Are my results right?” to asking “How do my results support or refute my hypothesis?’ In other words, the student’s laboratory activity is more than merely reinforcing a concept that was presented in the classroom. The student is carrying out an experiment that asks an original scientific question, proposes a useful hypothesis based upon her existing knowledge of the topic, designs a valid experiment, and analyzes her experimental results in terms of her hypothesis. The asking, proposing, designing, and analyzing are the fundamental intellectual skills used by research scientists.

I do not expect high school students to become expert scientists (in the way Daniel Willingham defines professional expertise as creators of new knowledge, Why Don’t Students Like School, 2009) but I do expect them to understand the thinking processes that scientists use when they create new knowledge as experts in their field. This understanding of how scientists think is crucial in recruiting our next generation of scientists! Equally importantly, it will enable our students who don’t want to become scientists to make scientifically-informed decisions on policy issues such as climate change.  “Educating global citizens is one of the most important charges to universities, and the best way we can transcend ideology is to teach our students, regardless of their majors, to think like scientists.” (Peter Salovey, Scientific American, 2018)

This experiential approach to science education engages the students in open-ended problem-based learning rather than a cookie-cutter lab. A cookie-cutter lab has a predetermined outcome, known to both the teacher and the student, that needs to be obtained in order for the activity to be successful. These types of labs are useful to reinforce a concept taught in the classroom. The student faithfully follows a set of technical instructions. Ok. But such labs have very little utility in teaching the higher level scientific thinking that challenges the student to think in terms of asking questions, proposing hypotheses, designing experiments, and analyzing data.  Thus, in the age of Google, the scientific thinking process becomes the teaching objective.

Each student is therefore actively engaged in conducting an authentic research investigation. They are learning how to think like a scientist. As a former research neuroscientist, I used scientific thinking professionally throughout my career. While I learned the scientific method from a textbook in school, I did not truly learn to think like a scientist until I was actually asked to practice it on an authentic research problem.

Here’s an example. You can do this in your classroom or do something similar. And all levels of students can do this. It’s better than AP!

How can you challenge your students to apply scientific thinking to an original problem?  I ask my students to ask the question: can very simple animals learn? The students use pill bugs which are a commonly used high school experimental animal that is readily available and easy to maintain. Based upon research on other simple animals such as bees, worms, and fruit flies, it is a reasonable hypothesis that pill bugs can be taught.

In this investigation, the students (1) ask questions, (2) propose hypotheses, (3) design experiments, and (4) analyze the results in terms of the hypotheses.

As investigators, students first ask what appropriate stimuli are to which pill bugs are instinctively attracted (the unconditioned stimulus) and what stimuli can they detect but show no preference for (the conditioning stimulus).

Second, they propose hypotheses as to what environments (damp, dark, noisy, smelly) to use for both types of stimuli.  In order to propose these hypotheses for appropriate stimuli to teach the pill bugs, the students research/google the literature to understand pill bug sensory capabilities and behavior.

In the third step, the students design experiments to test their hypotheses with a choice chamber in which the pill bugs show their preference by occupying one chamber over the other. Designing an experiment begins by defining the independent and dependent variables. In this case, the independent variable is the presence (in one chamber) of a hypothesized stimulus to which the pill bugs are instinctively attracted (for unconditioned stimulus) or show no preference (for conditioning stimulus). Examples of unconditioned stimuli include dark and/or damp environments, and the lack or presence of food. Examples of conditioning stimuli my students use include music, deodorant, pencil erasers, suntan lotion, light of various colors, and sandpaper. The dependent variable will be the preference of the pill bugs as measured by their presence in one of the two chambers at any given time. The experiment must be validated by a negative control in which neither chamber has a stimulus in it. The pill bugs are expected to equally occupy both identical chambers.

See the video here!

In the final step, the students critically analyze their experimental results to determine whether the pill bugs were attracted to their unconditioned stimulus, did not prefer their conditioning stimulus, or ultimately, learned by associative conditioning.

Results that support the students’ hypothesis for a hypothesized unconditioned stimulus would show that more pill bugs were present in the chamber with the stimulus than the control chamber. This result can be validated by a statistical analysis (for example, using a t test).

Typical results (taken here for single trial from one student’s experiment)

Time Interval (seconds)                      (Number of pill bugs in stimulus chamber)

Unconditioned Stimulus         Conditioning Stimulus


30                                                        5                                              5

60                                                        5                                              4

90                                                        6                                              4

120                                                      7                                              5

150                                                      8                                              4

180                                                      8                                              4

210                                                      9                                              4

240                                                      8                                              3

270                                                      9                                              5

300                                                      8                                              4


Unconditioned stimulus using carrot slices

Conditioning stimulus using presence of pink eraser

Once both unconditioned and conditioning stimuli have been discovered, the pill bugs are ready for training. Can the students train the pill bugs to change their behavior to prefer a stimulus that prior to training they did not prefer and thus support the hypothesis that pill bugs (a simple animal) can learn?

At this stage, several trials are used in which both stimuli are present in the same chamber. For these training trials, the pill bugs should show preference for the chamber with the unconditioned stimuli but at the same time, detect the conditioning stimuli.

In order to determine if the pill bugs learned, they are tested. After the multiple training trials (students typically use five consecutive), the pill bugs are once again subjected to only the conditioning stimulus. Remember that the first time the pill bugs were exposed to the conditioning stimulus, they showed no preference for that chamber. The preference of the pill bugs for the conditioned stimulus chamber after training is evidence that they learned a new behavior.

The entire investigation including a discussion of the results can be completed in six class periods. Thus, it can be easily embedded in various types of science courses.

For example, a typical schedule would be: Day One: Introduction to scientific question, students research pill bug behavior (for homework); Day Two: students conduct negative control experiment (three trials); Day Three: students conduct unconditioned stimulus experiment (three trials); Day Four: students conduct conditioning stimulus experiment (three trials); Day Five: students train pill bugs with both stimuli present (five trials) then immediately test pill bugs with only conditioning stimulus present (three trials); and Day Six: Analysis of data using the t test and assignment of investigation report.   For the results obtained in days two through four, the homework should task the students to determine standard deviations for each time interval of their three trials and to make individual graphs of their results for negative control, unconditioned stimulus, and conditioning stimulus. The day five results should be used to make a graph comparing the pill bug choices in response to the conditioning stimulus before and after the training. Any differences between these two lines can be analyzed by a ‘t’ test for each time interval using p< 0.05 as the measure of significance. Differences between the two lines are evidence that the pill bugs changed their behavior to prefer a conditioning stimulus due to their training. In other words, they learned by association.

Pill bugs that have learned will present a graph that looks like this one (taken from a student’s investigation report). Note that the upper line represents the behavior of the trained pill bugs and the lower line represents the behavior of the pill bugs before training.

I have been conducting this investigation for about ten years and my student groups have made some interesting discoveries. Within the same boisterous classroom, groups have found that pill bugs can be taught (for example, using the dark as the unconditioned stimulus and suntan lotion as the conditioning stimulus) while other groups showed no association with their various stimuli. They have not yet been able to prove the hypothesis! As a science teacher, the thrill for me is that all of the students experience an authentic research investigation that help them think like a scientist. The students learn to design an original experiment, conduct the experiments, and analyze their results using statistics.

Through this investigation, the students have: (1) learned to ask soluble scientific questions, (2) learned to propose testable hypotheses for the different stimuli used to train the pill bugs, (3) learned to design and test the validity of experiments to include an independent variable (the stimuli), dependent variable (counting the number of pill bugs at various time intervals) and a control (a negative one in which no stimuli are present), and (4) learned to critically analyze experimental results (using graphs and statistics) in terms of the hypothesis (did the pill bug behavior change with training?). Through this process of scientific thinking, they are experiencing science as a way of exploring the world. Their attitudes have matured. They no longer ask “Is this the right result?” Instead, they have learned to think of their results in terms of their own hypotheses. They are thinking like scientists.

Note:  Eric Kandel won the Nobel Prize using the simple sea slug, Aplysia to characterize the molecular biology of short-term memory. This investigation invites students to explore whether pill bugs can learn by associative conditioning using unconditioned and conditioning stimuli.

Bill Wallace

Bill Wallace received a PhD from Case Western Reserve University and was a post doctoral fellow in the laboratory of Nobel Laureate Paul Greengard at Yale and The Rockefeller Universities. He led laboratories as a neuroscientist researching Alzheimer’s disease at Mount Sinai School of Medicine and National Institute on Aging. He has been a science teacher at Georgetown Day School in Washington DC for the past 20 years.