Martin John Brown

I do formal statistical analyses, scientific literature reviews, and research designs.   If you feel I might be helpful to your work, please contact me by phone (503-736-9588) or email.

My subject expertise is forest ecology but at the end of the day, data analysis is data analysis, and the same principles apply. I’ve analyzed data about birds, fungi, and people.

My clients are typically scientists in federal agencies (past clients include the Smithsonian Environmental Research Center, National Park Service, and Forest Service PNW Research Lab), but occasionally include social service agencies and marketing firms.

My computer and “number crunching” skills are respectable: 10 years experience with the statistical program SPSS, 1 year experience with R, understanding of parametric and nonparametric statistical tests, Monte Carlo techniques, power analysis, working knowledge of SAS, PHP, etc.

However, the most important assets I bring to scientist clients are an ability to ask clear questions, a schedule that allows me to concentrate entirely on the task at hand (no agency meetings to attend), and, from my work as a popular writer, the ability to express results clearly in writing.

The experience of dozens of analysis projects has made me wily, too.  Data analyses can be morasses hazardous to time and focus, but I know the pitfalls of the process. I am not an undergraduate fishing around for a p value (I understand sample size effects, for one thing). My goal is to make a clear, honest summary of the data at hand, and note substantial differences (or samenesses) not just “significant” ones.

Some of my favorite projects are listed below, for the most part with reprints of the ultimate report or paper.   You’ll see that some of my work involves heavy computing or intellect, while others are more data management jobs.  I’d like to think that I’ve approached both types with the same goal of clarity and accuracy.  If you have any questions about my experience please send them along.

—

Question: how does sampling effort (i.e., the number of plots and years in which studies were conducted) affect field evaluations of fungi diversity and abundance?

• Answer: Sampling effort has a huge influence on surveys for rare or hard-to-detect species.  In three separate reports for Forest Service fungi researchers (here’s one), I used subsampling and curve fitting to describe the effect of sampling effort on detected fungi diversity and abundance.  Many species simply won’t be detected without tremendous effort.

Question: would scientists really be interested in purchasing green products for the lab, or are they just greenwashing themselves?

• Answer: Yes, they truly want green options.  I created an unusual survey and statistical strategy for this piece of  market research.  An extensive survey asked parallel questions about purchasing behavior in home and lab life, which could then be interpreted using a simple but powerful paired-samples test.  The gap between home and lab scenarios clearly demonstrated the existence of unmet demand in the lab.

Question: How much potential marbled murrelet habitat is there on federal land in the Pacific Northwest?

• Answer:  Several hundred thousand acres, as described in Forest Service General Technical Report 650, where I contributed to Chapter 4, “Estimating the amount of marbled murrelet nesting habitat on federal land by using a systematic grid sampling strategy.”  Though work on the marbled murrelet was famously troubled by political interference, the basic logic of this analysis (creating a logistical regression predicting murrelet occupancy from habitat descriptors, and then applying that regression across an entire landscape) was very sound. My job on this analysis was to check the other statistician’s calculations.  I caught a few errors. :)

Question: How does the forest canopy filter ultraviolet radiation?

• Answer: it filters it a lot, but in a way that is different than visible light.  For my paper in Journal of Ecology, “A Survey of Ultraviolet-B Radiation in Forests” I designed numerous simple surveys, then combined them to construct the first working description of UV radiation within the forest canopy space.  I also made predictions about the effects of declining stratospheric ozone on future forest composition.

Question: How has the forest at Warner Creek responded to the fire that took place there in 1991?

• Answer: with an exquisite pattern of decay and regrowth.  For more than 10 years I have been the primary data analyst for a long-term study of tree mortality and forest regeneration at the Warner site.  The job hasn’t involved higher math, but I’ve successfully tracked thousands of individually labelled trees and snags observed over a 15-year span, spanning multiple field crews and computers OS’s — a data management triumph, I’d say.  The results produced from this nice clean dataset will be published in an upcoming Forest Service General Technical Report.

Question: Is the concept of forest canopy stratification, so often used in forest ecology, scientific?

• Answer: Not really — which shows how far you can get just by doing a good review of the literature.   In my American Naturalist contribution, “Forest Canopy Stratification: Is it Useful?” I collected the many meanings of the term “stratification.” Then I applied/calculated them all for a single well-studied forest.  They didn’t agree, and rarely could be used to make scientific predictions. A basic review of words ended up being a meaningful contribution to knowledge.