research

So out of boredom and a sense that I need to restrain myself from spending money, I asked my advisor to show me his database last Monday. Then, I reasoned, I would have something exciting and new to play with and occupy me. It definitely worked. So my advisor has access to a database of 15 years of health checkup records for the people in this area. It has all sorts of info about them, though he and I are mainly interested in things pertaining to eating, metabolism, etc. So...it's a wealth of data. The kind that the Japanese government collects, just for fun, and lets the researchers loose on. So now I'm playing with a gigantic Microsoft Access database with millions of people in it. It's interesting. This whole week has basically been dedicated to overcoming the research language barrier. I've been trying to understand and translate for myself every field name in the database. I've also miraculously made it possible to switch the user interface for Office programs from Japanese to English on my lab computer. I have no idea how that happened, but it took a morning of fiddling and I officially love Office now. Meanwhile, I've downloaded a very helpful ebook on data analysis in Access 2007, and I've been learning the program by reconstructing and Englishifying the database. That has been my week in a boring nutshell.

I've come to the conclusion that the research put out by this group is unfortunately not very good. Perhaps I've been around Harvard for so long that I've gotten used to a certain standard?　Maybe I'm being snobby, and maybe I'm only flattering myself in thinking that I can be a judge of good and bad research--supposedly the whole point of my science education. But I like to think that while maybe I can't produce stellar research (yet! I can still hold out hope), I can at least recognize its presence or absence.

I was helping my advisor do an English proofread for a manuscript he is helping a student put out. My God, it was horrible. Well, the English was good, but I was really surprised the study was being published! It was a case report on delayed language skills or something like that. I guess the main problem was that it really tried to generalize too much based on one isolated case, and it couldn't even persuasively suggest cause and effect in that one isolated case! Let alone trying to generalize to the population. It was horrific.

Then, on Monday as my advisor showed me the database, he gave me two files. One was a Methods type of file, outlining where the data came from and how metabolic syndrome criteria were calculated. That was extremely helpful. The other file was labeled Discussion part 1 !! How, pray tell, does one discuss data that has not yet been analyzed? He told me he just predicted the outcome and wrote a discussion. Hmm...right...that so sounds suspect. He told me he does things backwards--start with some discussion, do some results, write the rest of the discussion, and then come up with the objectives of the study. Shouldn't some sense of what one is analyzing for guide the analysis? Gee...now I'm confused.

Random tidbit for the day: The cute little kids in the school are now singing the Beatles' "Hey Jude" in Japanese. I have no idea what they are saying, but they are so convincing that "Hey Jude" sounds like a traditional Japanese song.