Experimental Design Assignment

Assignment: Design an experiment using any of the methods of human cognitive neuroscience that we have discussed in the class to address the brain basis of intuitive physics. The biggest part of this task is to come up with a good idea for an experiment that asks an important, interesting, and theoretically motivated question about intuitive physical inference in the human brain. Because a single experiment rarely definitively answers a question, you should include two linked experiments, the second one addressing questions left unanswered by the first. Think hard about and provide important details about your task and stimuli. The difference between a beautiful experiment and a “meh” experiment is often in the specifics of the stimuli. (Work to come up with good ones and show examples!) Extra points for creativity, but note that usually the best experiments are simple and elegant. (Crazy-complicated experiments almost never work.)

Beyond coming up with two successive nicely designed experiments that ask important questions, it is also important to write up your ideas clearly. So, edit your prose multiple times to remove extra words and promote clarity. This assignment counts for 15% of your course grade. Expect to spend many hours on it. You may not propose an experiment that has already been published.

You have already read the very relevant Fischer et al. (2016) paper. I recommend you brainstorm for a while on your own, taking notes, and only after that read Ullman et al. (2017) and/or Kubrich et al. (2017), then revisit and perhaps revise your ideas.

Your writeup may be only a few pages (short is good), but must do the following (for each of your two experiments):

  1. Clearly state your hypothesis and motivate on theoretical and empirical grounds why it is important. This should require 1–3 clearly written paragraphs referring to what we already know from common sense / everyday experience and/or prior studies (including citations, with full references at the end).
  2. Describe your experimental design, listing all factors manipulated, all conditions within each factor, and measures collected.
  3. Describe what subjects will actually do (the task) and what happens in each trial including the precise sequence of events and their timing.
  4. How will the data be analyzed? I don’t need the level of detail of a methods section in a published paper, but I do need enough to understand the logic of the experiment (at the level used in lectures in class).
  5. State or draw the precise predictions of your hypothesis and note which (if any) are main effects and which are interactions.
  6. Discuss what you can infer from each of the main possible data outcomes. Consider alternative accounts of each of these possible outcomes. Especially if the prediction of your hypothesis is borne out by the data, what would you still need to worry about, and what other experiments might address those remaining concerns? This is how you will motivate your second experiment, and even after the second experiment you may still be left with alternative accounts and unanswered questions, which you should describe.

Course Info

Learning Resource Types

theaters Lecture Videos
assignment Written Assignments
notes Lecture Notes
co_present Instructor Insights