Yesterday I wrote a post about Debunking the Hockey Stick. What I learned was that, as expected, many have already discussed this in some detail and, also, that there’s much about this that I don’t fully understand (no surprises there then).
To see if I could learn more about this I tried, today, to get McIntyre & McKitrick’s analysis code to work. I installed R on my macbook and copied the final version of their code. After a bit of fiddling around (which included finding and copying various data files into a local sub-directory) I managed to get the code to start. It was clearly generating the ACFs and started to produce the figures, but then crashed with an error that I couldn’t seem to fix.
Apart from confirming my own ignorance, I haven’t really learned anything new. In a sense, what was interesting (for me at least) was how much of this seems to rely on libraries and/or functions that are included in the R package, but that aren’t explicitly visible in the code. This might be normal for some, but is a little unusual for me. Typically my programmes have almost everything explicitly in the code. Even if I could get the McIntyre & MKitrick code to work, I’d still need to learn quite a lot about how R works, and what these various functions are actually doing.
Given that I’m off on holiday on Saturday and am trying to finish a draft of a paper before I leave, I’m going to give up for the moment. Was quite interesting and although I’m not really any more enlightened about this than I was when I started, I’ve certainly learned some things. I may have another go when I’m back at work, so if anyone has had any success getting this code to work or has any experience with this type of code, any advice would be gratefully accepted.