Visually comparing shrinkage and LS estimates

Source: R/group_level_residual_functions.R

This function creates a plot (using qplot()) where the shrinkage estimate appears on the horizontal axis and the LS estimate appears on the vertical axis.

compare_eb_ls(eb, ols, identify = FALSE, silent = TRUE, ...)

Arguments

eb

a matrix of random effects
ols

a matrix of the OLS estimates found using random_ls_coef
identify

the percentage of points to identify as unusual, FALSE if you do not want the points identified.
silent

logical: should the list of data frames used to make the plots be suppressed.
...

other arguments to be passed to qplot()

Author

Adam Loy loyad01@gmail.com

Examples


wages.fm1 <- lme4::lmer(lnw ~ exper + (exper | id), data = wages)

#> Warning: Model failed to converge with max|grad| = 0.00609024 (tol = 0.002, component 1)
wages.sepLM <- adjust_lmList(lnw ~ exper | id, data = wages)
rancoef.eb <- coef(wages.fm1)$id
rancoef.ols <- coef(wages.sepLM)
compare_eb_ls(eb = rancoef.eb, ols = rancoef.ols, identify = 0.01)

#> `geom_smooth()` using formula 'y ~ x'
#> `geom_smooth()` using formula 'y ~ x'
#> Warning: Removed 38 rows containing non-finite values (stat_smooth).
#> Warning: Removed 38 rows containing missing values (geom_point).