Now we've constructed a very basic Stroop experiment with only words (red and blue) and two colors (again, red and blue). We arbitrarily determined that 'f' would be the response for blue-colored words, and that 'j' would be the response for red-colored words. So far this is a perfectly acceptable design.
Now imagine that we include only right-handed subjects in our experiment - a common practice for fMRI studies. Experience with the computer keyboard shows that 'f' naturally falls under the left index finger and 'j' under the right index finger. What if we found that responses to red-colored words - i.e., stimuli that required a response with the right index finger - were both faster and more accurate than responses to blue-colored words? Is it a coincidence, or do right-handed subjects respond better to Stroop stimuli mapped onto their right hand?
To control for this possibility, we use a method called counterbalancing to randomly assign different response mappings to different subjects. In our study, we will create a List object that will determine which button maps onto which color. In our case there are only two responses, so only two counterbalanced conditions need to be created. You can create these with the List object, clicking on the Property Pages button, and selecting the "Selection" tab. If we counterbalance by Subject, the subject number will determine which level of the List object is chosen.
For example, if our counterbalancing List object has two levels, and we enter a subject number of "1", then level 1 of the List will be chosen. If level 1 has mappings of 'f' for blue and 'j' for red, then these mappings will be used for that subject for the whole experiment. On the other hand, let's assume that level 2 of the List reverses the mappings: i.e., 'f' for red and 'j' for blue. A subject number of 2 will select those mappings and use them for the duration of the experiment. This is a simple yet effective way of eliminating many common confounds.
As you can see, we have made our experiment progressively more sophisticated without making it unwieldy. We have relatively few objects in our experiment, but by using Lists we can present several different combinations of trials and control for confounds. In the next tutorial we will maintain our economical design but add more variety through Slide States - different conditions presented through the same Slide object.