We re-analysed the data from Loyn (1987; see Box 8.2 ) by fitting a
simpler model that just included grazing and log patch area (based on
our model selection criteria – see Box 9.1). First we treated grazing as
a continuous variable like we did in the original analysis by fitting
the following model:
(bird abundance)i = β0 +
β1(log10 area)i +
β2(grazing)i +
εi
Preliminaries
First, load the required packages (car, lm.beta)
Import loyncat data file
This file is the loyn data, with grazing classified into five
categories (“grazecat”). We could have simply turned the numerical
grazing level into a factor, but this way the grazing levels have
meaningful names
loyncat <- read.csv("../data/loyncat.csv")
head(loyncat,10)
loyncat$grazecat<-factor(loyncat$grazecat)
Fit linear model with grazing as continuous variable
loyncat1.lm <- lm(abund~log10(area)+graze, data=loyncat)
tidy(loyncat1.lm, conf.int=TRUE)
Get standardized coefficients
lm.beta(loyncat1.lm)
Call:
lm(formula = abund ~ log10(area) + graze, data = loyncat)
Standardized Coefficients::
(Intercept) log10(area) graze
NA 0.521 -0.391
Fit linear model with grazing as a categorical predictor
Use zero grazing as the reference category
loyncat2.lm <- lm(abund~log10(area)+relevel(grazecat,ref="zero"), data=loyncat)
tidy(loyncat2.lm, conf.int=TRUE)
Check residuals - look fine
plot(loyncat2.lm)
Get standardised coefficients
lm.beta(loyncat2.lm)
Call:
lm(formula = abund ~ log10(area) + relevel(grazecat, ref = "zero"),
data = loyncat)
Standardized Coefficients::
(Intercept) log10(area) relevel(grazecat, ref = "zero")high relevel(grazecat, ref = "zero")intense relevel(grazecat, ref = "zero")low
NA 0.54847 -0.04948 -0.47202 0.01259
relevel(grazecat, ref = "zero")medium
-0.00788
Get added variable plots
avPlots(loyncat2.lm, ask=F)
LS0tCnRpdGxlOiAiUUsgQm94IDguMTEiCgpvdXRwdXQ6IAogIGh0bWxfbm90ZWJvb2s6CiAgICB0aGVtZTogZmxhdGx5Ci0tLQoKYGBge3Igc2V0dXAsIGluY2x1ZGU9RkFMU0V9CmtuaXRyOjpvcHRzX2NodW5rJHNldChlY2hvID0gVFJVRSkKYGBgCgpXZSByZS1hbmFseXNlZCB0aGUgZGF0YSBmcm9tIExveW4gKDE5ODc7IHNlZSBCb3ggOC4yICkgYnkgZml0dGluZyBhIHNpbXBsZXIgbW9kZWwgdGhhdCBqdXN0IGluY2x1ZGVkIGdyYXppbmcgYW5kIGxvZyBwYXRjaCBhcmVhIChiYXNlZCBvbiBvdXIgbW9kZWwgc2VsZWN0aW9uIGNyaXRlcmlhIC0tIHNlZSBCb3ggOS4xKS4gRmlyc3Qgd2UgdHJlYXRlZCBncmF6aW5nIGFzIGEgY29udGludW91cyB2YXJpYWJsZSBsaWtlIHdlIGRpZCBpbiB0aGUgb3JpZ2luYWwgYW5hbHlzaXMgYnkgZml0dGluZyB0aGUgZm9sbG93aW5nIG1vZGVsOgoKKGJpcmQgYWJ1bmRhbmNlKX4qaSp+ID0gKs6yfjB+KiArICrOsn4xfioobG9nfjEwfiBhcmVhKX4qaSp+ICsgKs6yKn4yfihncmF6aW5nKX4qaSp+ICorIM61fml+KgoKIyMjIFByZWxpbWluYXJpZXMKCkZpcnN0LCBsb2FkIHRoZSByZXF1aXJlZCBwYWNrYWdlcyAoY2FyLCBsbS5iZXRhKQoKYGBge3IgaW5jbHVkZT1GQUxTRSwgcmVzdWx0cz0naGlkZScsIGVycm9yPVRSVUV9CnNvdXJjZSgiLi4vUi9saWJyYXJpZXMuUiIpICAgI1RoaXMgaXMgdGhlIGNvbW1vbiBsaWJyYXJ5CmxpYnJhcnkobG0uYmV0YSkKYGBgCgpJbXBvcnQgW2xveW5jYXRdKGRhdGEvbG95bmNhdC5jc3YpIGRhdGEgZmlsZQoKVGhpcyBmaWxlIGlzIHRoZSBsb3luIGRhdGEsIHdpdGggZ3JhemluZyBjbGFzc2lmaWVkIGludG8gZml2ZSBjYXRlZ29yaWVzICgiZ3JhemVjYXQiKS4gV2UgY291bGQgaGF2ZSBzaW1wbHkgdHVybmVkIHRoZSBudW1lcmljYWwgZ3JhemluZyBsZXZlbCBpbnRvIGEgZmFjdG9yLCBidXQgdGhpcyB3YXkgdGhlIGdyYXppbmcgbGV2ZWxzIGhhdmUgbWVhbmluZ2Z1bCBuYW1lcwoKYGBge3J9CmxveW5jYXQgPC0gcmVhZC5jc3YoIi4uL2RhdGEvbG95bmNhdC5jc3YiKQpoZWFkKGxveW5jYXQsMTApCmxveW5jYXQkZ3JhemVjYXQ8LWZhY3Rvcihsb3luY2F0JGdyYXplY2F0KQpgYGAKCiMjIyBGaXQgbGluZWFyIG1vZGVsIHdpdGggZ3JhemluZyBhcyBjb250aW51b3VzIHZhcmlhYmxlCgpgYGB7ciB9CmxveW5jYXQxLmxtIDwtIGxtKGFidW5kfmxvZzEwKGFyZWEpK2dyYXplLCBkYXRhPWxveW5jYXQpCnRpZHkobG95bmNhdDEubG0sIGNvbmYuaW50PVRSVUUpCmBgYAoKR2V0IHN0YW5kYXJkaXplZCBjb2VmZmljaWVudHMKCmBgYHtyIH0KbG0uYmV0YShsb3luY2F0MS5sbSkKYGBgCgojIyMgRml0IGxpbmVhciBtb2RlbCB3aXRoIGdyYXppbmcgYXMgYSBjYXRlZ29yaWNhbCBwcmVkaWN0b3IKClVzZSB6ZXJvIGdyYXppbmcgYXMgdGhlIHJlZmVyZW5jZSBjYXRlZ29yeQoKYGBge3IgfQpsb3luY2F0Mi5sbSA8LSBsbShhYnVuZH5sb2cxMChhcmVhKStyZWxldmVsKGdyYXplY2F0LHJlZj0iemVybyIpLCBkYXRhPWxveW5jYXQpCnRpZHkobG95bmNhdDIubG0sIGNvbmYuaW50PVRSVUUpCmBgYAoKQ2hlY2sgcmVzaWR1YWxzIC0gbG9vayBmaW5lCgpgYGB7ciB9CnBsb3QobG95bmNhdDIubG0pCmBgYAoKR2V0IHN0YW5kYXJkaXNlZCBjb2VmZmljaWVudHMKCmBgYHtyIH0KbG0uYmV0YShsb3luY2F0Mi5sbSkKYGBgCgojIyMgR2V0IGFkZGVkIHZhcmlhYmxlIHBsb3RzCgpgYGB7ciB9CmF2UGxvdHMobG95bmNhdDIubG0sIGFzaz1GKQpgYGAK