This has been an exciting week. My own work may be soon coming to fruition, and I’m starting to the “pre-unblinding jitters”. I even took a walk from my office to the BaBar detector today, even in the oppressive heat, to relax. But what are “pre-unblnding jitters”? This is a little hard to explain. Let me give a medical analogy.
When you want to test the potency of a new medication, you want to do it as honestly as possible. The way to do this is a “double-blind” test. You have two groups of test patients. One half gets a placebo, and the other gets the drug. Before the medication is distributed, each bottle of pills is given a random serial number and the relationship between drug type (placebo, medicine) and serial number is hidden in a computer and never told to the researcher. This is called “blinding”, and it’s double-blind because neither the patients nor the researchers know who gets the real thing and who gets the placebo. The patients are then given their bottles, and take the medicine for the allotted time. At the end of the study, the patients’ records are compared to the codes in the computer, and the relationship between patient and pill-type is revealed – “unblinded” – to see whether the new drug performed better than placebo.
In physics, we do something similar. We define a region in which the signal we are looking for will live, if it’s there, in the data. We never look at the data in that region – we “blind” it – and we do all of our studies on control samples, simulations, etc. After we’ve settled on our analysis strategy and committed to a method to interpret the results, we “unblind”.
My own work is nearing this stage. I’m getting nervous about it. Will I find something? Will the results be consistent with a null signal hypothesis?
Only nature knows. Maybe, soon, so will I.
