Identification of the most deviant case — most

Identification of the most deviant case (= worst predicted case), based on regression estimates.

Usage

most_deviant(lmobject)

Arguments

lmobject: Object generated with lm

Value

The most deviant case with the largest absolute residual of all cases.

Details

Proposed by Seawright, Jason and John Gerring (2008): Case Selection Techniques in Case Study Research: A Menu of Qualitative and Quantitative Options. Political Research Quarterly 61 (2): 294-308. (doi:10.1177/1065912907313077 )

Examples

df <- lm(mpg ~ disp + wt, data = mtcars)
most_deviant(df)
#> Toyota Corolla 
#>        6.34844