The Scientific Method

They teach you from an early age about the hypothetico-deductive scientific method. Let me just say right now that that's a load of crap. The true scientific method is dominated by slow, steady research with no hope of an end in sight, punctuated by bursts of frantic work precipitated by conference or grant deadlines.

As a case in point, let's examine my past 2 weeks.

Two weeks ago my contact from CDFA suggested that I present our project at the State of the Estuaries Conference in Oakland this fall. And oh yeah, abstracts are due this Friday. I forwarded this request to my advisor. She and I agreed that it would be a good idea, especially if we could present some subset of the greater project since the general project has already been presented in several venues. Naturally, this all points to me presenting my research. So it's very unfortunate that I have absolutely no satisfying results from last year's data. And since we won't get this year's imagery for at least another month, I spontaneously generated a project that could be completed entirely in 2 weeks. And if the results were still unsatisfying, we'd have the general project as a fall-back.

The idea is that it will be a nice little spectral uniqueness study which will serve as a first step towards mapping my species in the imagery. I bounced this idea off of my advisor, who agreed that it would be worthwhile. I was a little flabbergasted, however, when she told me that she'd had a student do that very thing a few years before I came to Davis. It seems like that would be something you'd expect to come up over the course of the last year since she's known that I've been planning on working on this species... Even worse, apparently this phantom student was not able to discriminate my species in spectral data. Doesn't really bode well for my thesis, now does it? Though Susan wasn't so confident in his results. Apparently he was vocally upset with doing that project for her, and she thinks he just said it wasn't possible without actually trying, so she recommended that I find his dataset and analyze it. Not surprisingly, the dataset has resisted our efforts to find it, so I prepared to go collect my own data. Yay. That's much more fun anyway. Plus it puts me in control of the entire process. Most of the times I'm unsuccessful at mapping a species it's because someone else collected the field data improperly. So I'm generally much happier using my own data.

Through some act of god amazingly working in my favor, the spectrometer was actually free last week. So I spent the week cruising around Central California recording the reflectance spectra of various species, mostly weeds. Sunday I spent in lab preprocessing the data and calculating physiological indices. And this week is devoted to analyses and writing and submitting the abstract.

Not too suprisingly, univariate statistics weren't getting me very far. So today I tried out a classification tree. Unqualified success!! This is amazing! It's pretty rare for things to work! Especially given the extremely short amount of time I've had to work on this. I'm very excited about all this. Sure, it'll just be a poster at a regional conference, but I've never presented my own work before. I've had to give talks about the bigger projects I've worked on at a few meetings, but this is my research! I'm all starry-eyed! Right now I'm feeling very optimistic. I bet I'll even be able to get a publication out of this. Of course, there are more analyses I want to perform beyond the quick-and-dirty-get-something-to-say-in-the-abstract stuff that I did this week, but they should be pretty straight-forward.

This could even be the first chapter of my thesis! I might actually graduate!