Logging in western North America impacts populations of western trillium, a long-lived perennial that inhabits conifer forests (Trillium ovatum; see the photo at the beginning of the chapter). Jules and Rathcke (1999) measured attributes of eight local populations of western trillium, confined to forest patches of varying size created by logging in southwestern Oregon. Their data, presented in the following table, compare estimates of recruitment (the density of new plants produced in each population per year) at each site with the distance from the site to the edge of the forest fragment.

a. Display these data in an appropriate graph. Examine the graph and describe the shape of the distribution. What departures from the assumption of correlation analysis do you detect?

b. Choose a transformation and transform one or both of the two variables.

Plot the results. Did the transformation solve the problem? If not, try a different transformation.

c. Using the transformed data, estimate the correlation coefficient between the two variables. Provide a standard error with your estimate.

d. Calculate an approximate 95% confidence interval for the correlation coefficient.