confidence: Plot confidence intervals for variables from your linear...

In loganstundal/POL3085: Automates linear regression model plotting

Description

This function takes the linear regression model you estimate with lm() for your final projects and returns a coefficient plot appropriate for you to use in your poster or final paper.

Usage

confidence(model, vars = NULL, var.labels = NULL, level = 0.95,
  grid = T, grid.int = 5, mai = NULL, ps = 12, pt.type = 16,
  pt.size = 1, pt.color = "black", lwd = 1, lty = 2, legend = F)

Arguments

model	Model takes the object name of your linear regression model. Example - if you run the following command: m1 <- lm(y~x), then enter m1 for this argument.
var.labels	New variable names for the plot to be entered as characters here. Note, this must be the same length as the variables in your model. If you have four variables, you must enter four names. See example code below.
grid	Include vertical grid lines in the plot. Default is True.
grid.int	Spacing interval for grid lines. Default value is 5.
mai	Change plot margins (inches).
ps	Change text size in plot. This will be useful for your posters (30 would be good for a poster, 12 for a paper).
pt.type	Point style of regression parameter point estimate. Default is 16, circle. See ?points for information on different styles.
pt.size	Side of point estimate, default is 1.
pt.color	Color of point estimate, default is 'black', Any color argument will work here.
lwd	Line width for confidence intervals, default is 1.
lty	Line type for confidence intervals, default is 2 (dashed line). See the lty argument in ?par for additional details on line types.

Examples

# Estimate a regression using 'mtcars' data:
m <- lm(mpg~disp+hp+wt,data=mtcars)
confidence(m)

# Plot without intercept:
confidence(m,vars=c('disp','hp','wt'))

# Plot with intercept and one covariate:
confidence(m,vars=c('(Intercept)','wt'))

# Plot with two model variables AND new variable names:
confidence(m,vars=c('disp','wt'),var.labels=c('Displacement','Weight'))

# Plot with higher confidence interval and legend:
confidence(m,level=0.99,legend=TRUE)

# Dynamic margins for longer variable names:
confidence(m, vars=c('hp','wt'),var.labels=c('A really long name for variable: Horse Power','Weight'))

loganstundal/POL3085 documentation built on May 21, 2019, 10:32 a.m.