How to Get a Research Study Started
Essential reading. Overview of steps and approvals required prior to real data collection.
Discuss experimental design with your supervisor, and as many other people with relevant expertise as you can find (e.g., committee members).
Apply for research ethics board (REB) approval.
Develop the stimuli for the experiment
Program the experiment.
Run through the entire stimulus program yourself.
Fix any errors (typically an iterative process) until it runs as it should.
Look at the log (output) files from your stimulus program. Check that timing is as you expect it to be. Ensure that all the data you need to analyze the study are in the log files. Great sorrow has befallen those who ran an entire study before checking their log files.
After fixing any bugs you identify when running through yourself, pilot test the program on a friend or lab volunteer. Ask for feedback on the experiment - anything that was hard to understand, hard to see, boring, seemed like an error, etc.. Sometimes people won't give you helpful feedback unless you specifically ask. NOTE: if your study involves neuroimaging (EEG or MRI), you should not include imaging in your pilto testing until you have run through the program with yourself and others, beacuse otherwise you're wasting resources.
Write a complete protocol for your study. This is the "how to" manual for your study. This is critically important, and no one is allowed to begin data collection (or even pilot test using EEG or MRI) until the protocol is complete and approved by Aaron.
Recruit a pilot participant, and go through the entire experimental protocol, being sure that you have the protocol open and are actually following it. This is where you realize things you missed, or that are incorrect, and allows you to fix/update the protocol as necessary. Aaron, Cindy, or another designated senior lab member should be present for this pilot so they can provide feedback.
Copy all data from the pilot run to the server, then open the files and write the basic script(s) you need to read and process the data. Again, this is critical to ensure that you're getting all the data you need, in the format you need, before you invest significant time and resources in collecting multiple data sets.
Last updated