Wednesday, April 21, 2010

Kind of depressed today. The weather doesn't help - cloudy and apparently it rained during the day and it looked like it was constantly on the verge of rain while I was on my way home - and I've been short on sleep lately, blah blah blah, but part of the problem is work, of course.

The latest issue isn't a high-stress problem like the last one I talked about. It seems so minor and trivial that that is probably part of the problem itself. But the main issue is that doing this project correctly seems almost impossible. I need to compare the latest stats on something to what the stats were when it was last done a few years ago. I have the numbers from last time, but I have no idea what they are based on, because the "raw data" is very unhelpful. My best guesses for what they are based on yielded two different sums for the latest stats, one about 40 percent more than the total in the previous renewal and one about 25 percent less. As far as I know, there's no reason for variation that's anywhere near so large. (And the sum from last time is around 5,300, so the variations are like 1,000 in either direction, much greater than I think random noise could account for.)

So it seems to me that I have two options. I could work my ass off on this very minor assignment, and it's possible that no amount of work would be enough if the previous renewal's data was just plain flawed. Or that might just prove that one of my initial best guesses were right, which would be a relief but very frustrating. Alternately, I could do a half-assed job and just handwave away the huge difference in numbers and hope no one notices. And maybe they wouldn't. I don't know if anyone checks my work on this before it leaves the building, and I don't know how closely the office that we submit this to looks at it either. I'm leery of that, partly because recent comments from my supervisor make me think this will get more attention than it would have six months ago and partly, of course, because who wants to knowingly do a half-assed job? What would that say about me? And what does it say about my job that I can do a half-assed job on this and reasonably think no one would notice?

As I write this, a third option occurs to me. The incomplete, very unhelpful numbers from last time are actually less clear than from the time before that. And the numbers from the time before last are actually very close to one of my two best guesses, and the difference is in the direction one would "expect". So maybe I need to just throw out the stats from the last time this was done, pass the buck to whoever did it then, and use the time before that as my baseline.

Well, tomorrow I will put more time trying to figure out what the numbers from last time are based on, of course. But I'm glad the third option occurred to me.

No comments: