The “Observer Effect” in Self-Tracking
Chris Volinsky discusses his findings of how fluctuations in his daily weight measurements are correlated with whether or not he is logging his food intake: “Obviously the most important link to weight is food intake. I’ve struggled with being consistent with my food logging- it requires a dedication and diligence I cannot seem to keep up. I was hoping to see how the foods that I eat related to other aspects of my life, and that is hard to do without consistent data. But I did learn something very interesting thanks to my missing data! It is not necessarily *which foods* I log, but rather *whether I log my food at all* which is a good indicator of whether I am losing weight or gaining weight.”
Chris concludes that this is a psychological effect: the process of measuring affects what is being measured.
Recording your diet reminds you why you are doing this, and as a result, you continue eating healthy. This phenomenon is known in psychology as a Hawthorne effect or “obsrver effect”..(not to confuse with “placebo” effect), according to which people tend to change their performance when they are under observation.
In social sciences, changes in subject’s behavior due to certain observer effects definitely suggests flaws in experimental design. But what about self-tracking, when observer himself is a subject? In my opinion, that depends on the purpose of the project. If self-tracking is used for accountability, that is, to simply track the progress, then you should be fine. In fact, continuous feedback from self-tracking will work as an additional incentive in your endeavor, helping you stay the course. If, however, you are self-experimenting with certain treatments, or observing yourself under different conditions to make certain comparisons and conclusions, you need to be cautious about how you measure the impact. This especially applies to measurement of performance metrics (concentration, productivity at work, performance at gym, and other things that require conscious effort) under different conditions. You need to make sure that the instruments that you are using are not influenced by the treatment or observation process. I will try to elaborate more on that in the future posts.