The article here has been making the QS rounds today. Check it out.
I should start by saying that I don't really care for the tone of this article and so I could be overreacting. I think I have something smart to say either way. The author's use of 'obsessive' and 'virtually every' makes me think he might not be a big fan of the QS peeps. Again, maybe I am overreacting and the rest of my thoughts should still be valid even if I am overinterpreting.
First point - even if imprecise feedback was better than precise feedback, how do either of them compare to zero feedback (which is what most people have)? If any kind of feedback is better than no feedback...that is a argument for QS, not against.
Second point - This assumes that the person getting the feedback is positively motivated (so the can positively "distort the information") The discussion in the QS community about gamification and behavior change shows that new habits often require more than just data to actual get a change in results. I am not sure that any imprecision will get used for good.
That all being said the study showed what it showed and the results may seem a bit paradoxical to some QS'ers. If taken literally it means we might even consider fuzzying our data on purpose. That couldn't be right could it?
I have at least a stab at an answer. I think when things are directly connected it would be hard to have data that is too precise. If you are diabetic and you have a donut, I don't really think anyone would suggest that "your blood sugar is higher than it was" is precise enough to take action on. You need to know what your blood sugar is so you can determine how much insulin you need to take.
Contrast that situation with something like weight loss. There are a whole lot of factors that can get confounded with each other and if you are weighing yourself to the gram every hour you might very well find some connections that aren't there. I think a weekly update on your weight (perhaps with some sort of body composition check up) would be the way to go in that situation.
However, if you were a pro bodybuilder trying to get from 8% bodyfat to 7% bodyfat...maybe you would need to hourly weigh ins down to the gram.
Point is....don't read too much into this article. Keep on quantifying. To bastardize an Einstein quote - your data needs to be as precise as it needs to be, but not one bit more.