Over the past few weeks, I've been more actively reviewing some of the data I've collected since 2FNs got started. Last night, I sat down and decided it might be worth a little bit of work to review my weight-loss trend compared to my food-log activity. To describe my question, I wanted to see if the frequency of the entries in my food log could predict or, at the very least, suggest a good or bad weigh-in week based on how many days of the week leading up to Wednesday I recorded my food intake.
Hypothesis: If I keep track of my food intake more frequently during the week, then I will lose more weight come Wednesday morning.
To do this, I looked back into my food logs on my smartphone app and recorded either the presence of an entry for a given day or the absence of an entry for a given day. This has a small caveat, which I may revisit later, in that not all of my entries were for an entire day, some including only breakfast and lunch. If the day had only a breakfast logged, I decided to throw that day out of the count.
The results were pretty interesting. My correlation values seemed lower than I had anticipated, only getting to about 0.48, which clearly suggests that there are other factors going on. But when I graphed the data and assigned one of three categories to each week—either a positive expected outcome, a neutral expected outcome, or a negative unexpected outcome—the results looked a bit more in tune with what I had a firm suspicion of.