## Statistics, map, and the fair

Posted by: | July 29, 2012 | No Comment |

My main project last week involved analyzing the data from the previous week’s cockle surveys and comparing them to the surveys conducted in February. Thus, much of last week was spent in the lab running stats and assembling maps on the computer. By now my GIS skills have sharpened somewhat, and I was able to assemble a map that compares the spatial distribution and densities of cockles in the July survey to those in the February survey. Although there does appear to be a difference, it seems random, and I cannot yet ascertain any significant conclusion from it. My mentor Scott was out of the office all week, so perhaps he will be able to provide some insights on Monday.

I’ve often told people that statistics is the math class I’ve been waiting my whole life for. I have yet to refine my abilities in it, but I find statistics a fascinating way to interpret data and answer questions of interest. The challenge in statistics for me is determining which method is most appropriate to use for a certain situation. Scott left me some suggestions, but it was still up to me to determine which method to use and how to interpret the results. I may have gone a little overboard: I ran two regression analyses of cockle weight vs shell length (one using the “raw” data, which turned out to be nonlinear, and one using transformed data to establish a linear relationship and constant variation); I also ran summary stats for both the February and July surveys. I ran ANOVA tests comparing cockle weights in July vs February and shell lengths in July vs February. I wasn’t sure if ANOVA was appropriate to use, so I additionally ran four nonparametric tests (two with one-sided hypotheses, and two with two-sided hypotheses) for cockle weights and shell lengths in July vs February. I generated boxplots, histograms, and scatterplots comparing cockle sizes and densities. On Monday I anticipate spending time whittling down the stats and analyzing the data. Scott will help me confirm my findings, but so far it looks like there is no significant difference in weights and shell lengths of cockles in July vs February. Maybe they just haven’t grown that much, or maybe I’m out in left field. Who knows?

Last Tuesday I helped staff the ODFW booth at the Coos County fair. We had a 1200-gallon aquarium with rainbow trout, baby cutthroat trout, baby smallmouth bass, and some perch. Despite the awesome showpiece, we weren’t very busy and didn’t have many visitors. It was probably due to it being a Tuesday morning, which constituted a slow day at the fair.

This week we’re scheduled to do a new cockle survey in a new location. We’re going to try and take down more data in the field this time, which should decrease the amount of time we need to spend in the lab. I also look forward to discussing the statistics with Scott and seeing if I’m on the right track with my data interpretation.

under: Eric Post