Clarity - I thought the context made it clear that I was using the word "feel" in a medical sense, not an emotional, cognitive, or tactile one.
Somewhere in the last few days (so many forums... fora? forae?) I expanded on the nature of bias in testing. Basically, knowing that you are in a test situation automatically biases your responses. How depends on the nature of the test, etc., but people just think differently when being tested, or used as test subjects.
And don't heap too much praise on the double-blind protocol. It is merely a method to reduce the bias effects *of the test moderator/interview/whatever*. It does not and can not offset bad experiment design, let alone counter, or even discern, the biases of the subjects. If you ask two groups of people how they are feeling today, one group on a warm, bright, sunny day and one group on a cold, dark, rainy day, you've got bias; it won't matter if the experiment is single, double, or triple blind. Better yet, interview everyone on the same day, half in a room with flourescent lights and half in a room with incandescents. With a large enough population, the bias will show up in the results. This stuff was worked out in the 60's, yet it affects the majority of the "research" being done today.
I am part of a long-term knee study. Every 6 months I get a free MRI of my knees, plus the joy of a 20 page written questionaire and the thrill of a 30-minute interview, all to determine how my knees "feel". They DO record the outside temperature, humidity, and barometric pressure. I checked.