More preliminary data on the blood glucose/mental performance ‘study’
The R-squared and the estimated effect size in a simple linear model both look almost identical at this point in time as they did 55 observations ago – I’ve posted both the old scatterplot (first) and an updated version (second) below – click to view the full size versions:
I have however been a little suspicious about a few data-points which were collected around the time of the London Chess Classics tournament last year – I spent a significant amount of time on chess during that week and my playing strength when playing blitz games went up a lot those days too (I gained ~150 elo points over 4-5 days, which is a lot – I’ve lost that rating again at this point). Here’s what the image looks like without those observations:
I am not convinced that ‘blood glucose has no effect on tactics trainer performance’ is the conclusion to draw from this data-set, so I’m still collecting data at this point. The true data generating process of course includes many variables not included above – you may want to reread the first article if you want to know more about the ‘true’ DTG.
I wrote in my first post that: “I know myself well enough to know that I don’t want to bother with non-linear models when I look at this stuff later; it’s a poor and underspecified model to begin with.” If I actually have to work with methods which prove useful when analysing this type of dataset during my statistics course this semester (do remember that I have not included all the data I’ve gathered in the above plots), I may change my mind about how much work I’ll do on this dataset. Maybe I’ll be reminded of useful ways to handle stuff like this during the course; stuff that I’ve forgotten about at this point. We’ll see how it goes.
If anyone else would like to have a look at the data, just leave a comment below – I’d be happy to send you a copy of the data.
No comments yet.