ESM 206 Problem Set 4
Mini-Exam will be due Tuesday, May 30 at noon.
Comparing streamflow
The files SCflow.csv and SCflow2.xls contain data on annual streamflow in each of
five streams in Santa Barbara County (the data are the same in the two files, they are
just structured differently). Your task is to answer the question: “On average, does
streamflow differ between these five creeks?”
a. What statistical procedure should you use?
b. State your null and alternate hypotheses.
c. Answer the question, reporting the relevant statistical quantities.
Impact assessment
In 1990 the park service renovated a jetty on Santa Rosa Island. The park service has
been conducting annual surveys of the giant keyhole limpet, Megathura crenulata, at a
variety of sites around the Channel Islands since 1983; one site was close to the project
and another was around the point. The project took place between the 1990 and 1991
surveys; your job is to determine whether it had an impact on limpet densities, and if so,
how large of an impact.
The data are in the file limpet.csv. “Limpet” is the mean number of limpets per square
meter at a site, based on 12 quadrats. The “North” site is directly adjacent to the jetty,
and should be considered the impact site; the “South” site is around a point from the jetty,
and should be considered a control.
Gas prices
In the solution to part B, question 4 of problem set 2, I presented a model that included a
coefficient of -85.3 for the effect of price on gasoline consumption. This parameter
estimate was significantly different from zero (P < 0.01). Fit this model, and test the null
hypothesis that this coefficient is really 100 (that is, a doubling of gas price from 1995
levels – at which gas price is indexed – will lead to a consumption reduction of 100
gallons per year).