Thanks for your comment. I actually thought about the withdrawal point after posting this but before seeing your comment, but didn’t have the (good) idea of using blocks to mitigate it. I’m now pretty uncertain about whether blocks would be better or not. The rest of this comment should be read from the perspective of me thinking out loud not an authoritative response.
From a practical perspective, I don’t want to go through withdrawal more than once in two weeks (because from my prior experience, it will be horrible).
From an experiment perspective though, I actually am more interested in the question of “conditional on being addicted to caffeinated coffee, does my body detect the difference between regular and decaf?” I’m also interested in whether, conditional on not being addicted to coffee, drinking caffeinated coffee enhances my performance, but I question whether trying to answer that as part of the same study makes sense. Given that, doing blocks would be good because it would isolate the withdrawal period but bad insofar as it would reduce my samples of “conditional on addiction, do I notice” under different conditions.
Thanks for your comment. I actually thought about the withdrawal point after posting this but before seeing your comment, but didn’t have the (good) idea of using blocks to mitigate it. I’m now pretty uncertain about whether blocks would be better or not. The rest of this comment should be read from the perspective of me thinking out loud not an authoritative response.
From a practical perspective, I don’t want to go through withdrawal more than once in two weeks (because from my prior experience, it will be horrible).
From an experiment perspective though, I actually am more interested in the question of “conditional on being addicted to caffeinated coffee, does my body detect the difference between regular and decaf?” I’m also interested in whether, conditional on not being addicted to coffee, drinking caffeinated coffee enhances my performance, but I question whether trying to answer that as part of the same study makes sense. Given that, doing blocks would be good because it would isolate the withdrawal period but bad insofar as it would reduce my samples of “conditional on addiction, do I notice” under different conditions.