plotCIcorrelation.Rd
Function outputting a plot of confidence interval around a correlation for a range of sample sizes
plotCIcorrelation(
corr,
minN,
maxN,
step = 1,
conf = 0.95,
minY = NULL,
maxY = NULL
)
plotCIPearson(
corr,
minN,
maxN,
step = 1,
conf = 0.95,
minY = NULL,
maxY = NULL
)
numeric: observed correlation (assumes Pearson correlation, approximately OK for Spearman)
numeric: the smallest n for which you want the CI
numeric: the largest n for which you want the CI
numeric: the step by which you want n to increment (default is 1), with minN and maxN should give no more than 200 values of n
numeric: the confidence interval you want, defaults to .95
numeric: if you want to set the y axis limits input a number greater than -1 here
numeric: if you want to set the y axis limits input a number smaller than 1 here
a ggplot object so if you don't capture it it will plot to your default plot device
This is another function like plotBinconf
which plots the width of confidence
intervals (CIs) for a range of sample/dataset sizes.
Version 1: 5.iv.21
Other confidence interval functions:
plotBinconf()
Other demonstration functions:
plotBinconf()
if (FALSE) {
### simple example for observed 95% confidence interval (CI) around observed correlation of .7
### for sample sizes between 10 and 100
plotCIcorrelation(.7, 10, 100)
### similar but for dataset sizes between 10 and 500 by steps of 5
plotCIcorrelation(.7, 10, 500, 5)
### similar but asking for 99% CI
plotCIcorrelation(.7, 10, 500, 5, conf = .99)
### similar but setting lower y axis limit to -1
plotCIcorrelation(.7, 10, 500, 5, conf = .99, minY = -1)
### and you can set the upper y limit, observed R = 0 for a change
plotCIcorrelation(0, 10, 500, 5, conf = .99, minY = -1, maxY = 1)
### same but exporting to tmpPlot and then changing plot
plotCIcorrelation(0, 10, 500, 5, conf = .99, minY = -1, maxY = 1) -> tmpPlot
tmpPlot +
ggtitle("95% CI around observed Pearson correlation of zero for n from 10 to 500") +
theme_bw()
}