Monday, April 7, 2008

Data, interpretations, and field work

You know what I find really interesting about all the various geopuzzles that have been popping up in the geoblogosphere lately? Reading the answers from people with various backgrounds.

When I read other scientists talking about data and interpretation, I occasionally feel like I'm from another planet. My problem usually isn't pushing students to come up with interesting interpretations of their data. My students love to tell a good story. But in field work, it can be difficult to tell what's data and what's interpretation - what's repeatable, and what is one of a large set of reasonable (and testable) explanations.

There are things in geology that are obviously data. Measurement of discharge of a river. A chemical analysis of an igneous rock. GPS measurements showing the displacement of a site after an earthquake.

And then there's field geology. I walk up to a rock:

I look at it.

I measure the orientation of the surface, and the orientation of the lines, and write the data in my field notebook, with some kind of cryptic annotation that says:

Shear fracture = orientation
Slickenlines = orientation
Slip direction = direction

And I move on and collect more data.

But... wait a minute. Didn't I just assume the most interesting part, there? How do I know what that surface was? It's just a crack with shiny grooves on it.

And that was a pretty minor interpretation. During my first semester of grad school, I went on a field trip to an ophiolite in the Diablo Range in California. We hiked up the canyon, and at every stop, we had a quick overview of what this rock used to be, back in the day when it lay below the ocean floor. And, well... I, ummm, wasn't convinced. I was lousy at looking at ultramafic rocks (all those green and black minerals were hard to tell apart for someone with little experience), and we moved quickly, talking about what the rocks meant rather than what we saw in the rocks.

When I do fieldwork, I need to think about interpretations. It's possible that I will never be able to go back to a site: maybe I was dropped off by a helicopter, maybe I backpacked twenty miles into a wilderness area, maybe the property owner will deny me permission to come back. I need to simultaneously observe, document my observations with photos, take measurements, collect samples, and think about what it all means. And I need to construct my mental model and think about how to test it, all while slapping mosquitoes or putting on suncreen.

We don't talk about the importance of that skill set very often: that field geologists have to simultaneously conduct one "experiment" and design experiments to test competing hypothesis and develop models. It takes a lot of mental flexibility, and a combination of wildly creative egotism (to imagine an entire mountain belt while looking at a tiny outcrop) and humility (to recognize that the interpretation could turn out to be, you know, totally wrong).

Perhaps it shouldn't be a surprise that students find it difficult to figure out what's data and interpretation. Field trips are like professional field work, only worse, in some ways. Part of their goal is to get students excited about geology - and telling the story of the big picture is one of the ways to get the students engaged. Part of their goal is to help students learn to observe (at progressively more sophisticated levels as they go from freshman to senior). And part of the goal is for the students to understand how the observations relate to the big picture, so they can go to a new outcrop on their own and make up their own stories.

I've tried a number of ways to make students be explicit about what they've observed versus how it's interpreted. I make my structure students write up their work in two sections: data (or observations), and interpretation. The data, I grade based on whether its correct; their interpretation, I grade based on whether it's reasonable. (Perhaps I should also grade based on whether it's interesting, but geology students seem quite happy to extrapolate from the scale of outcrop to plate.) In my intro class, I play devil's advocate at the Great Unconformity - I try to convince the students that the granite is younger than the quartz-cemented sandstone. (It's pretty easy to convince them that the sandstone might really be a quartzite, and that maybe the granite intruded it.) Then I make them give the evidence that supports or disproves each possible interpretation.

It doesn't always work, so I'm always looking for new things to try. I'm thinking of pointing my students (especially the intro students) towards some of the geopuzzles. (Maybe I'll do something similar this fall on my class web site, and let students ask questions about things that can't be observed in a picture. I wonder what kind of carrot or stick would get them to participate? Hmmmm.)


Christie Rowe said...

This post inspired this post

Anonymous said...


Here Here. Especially when trying to map tectonites (which in itself is an interpretations), the line between observation and interpretation does get quite hazy.... am I looking at a sheared granite at this outcrop, or is it a volcanic? What evidence is in nearby outcrops to support or refute this? In the meantime, how do I denote it on my map? Etc... Etc... The funny thing is that the users of my maps often refer to them as "data", when they are very much interpretation. At NEGSA this past year, I was in a discussion with other mappers where it was said "No two geologists would produce identical maps of a given area. The same geologist may not even produce the same map if they came back later and mapped it again..."

-Joe K.