Data set: Birdnest
This data set was collected in the spring of 1999 for a class project by Amy Moore, a Grinnell College student. Each record in the data set represents data for a species of North American passerine bird. Passerines are “ perching birds” and include many families of familiar small birds (e. g., spar-rows and warblers) as well as some larger species like crows and ravens, but do not include hawks, owls, water fowl, wading birds, and woodpeckers. Moore took all North American passerines for which complete evolutionary data were available, which comprised 99 of the 470 species of passerines in North America (part of her study used this evolutionary information). One hypothesis of interest was about the relationship of body size to type of nest. Body size was measured as average length of the species nest type was categorized as either closed or open. Although nests come in a variety of types (see the Nesttype variable), in this data set “closed” refers to nests with only a small opening to the outside, such as the tree- cavity nest of many nuthatches or the pendant- style nest of an oriole. “Open” nests include the cup- shaped nest of the American robin.
a. Moore suspected that closed nests tend to be built by larger birds, but here we will treat the alter-native as two-sided, since her suspicion was based on scanty evidence. Use comparative dotplots or boxplots and summary statistics to describe the relationship between average body length and nest type (the Closed variable). (Closed = 1 for closed nests; Closed = 0 for open nests.) Does it appear that Moore’s initial suspicion is borne out by the data?
b. Run a permutation test using a two-sided alternative to determine if type of nest varies by body length and interpret your results. Be sure to state your conclusions in the context of the problem and address how random allocation and random sampling (or lack of either) impact your conclusions.