# Why you should plot data before doing statistical tests 
# CJ Brown 2020-11-06
#More at www.conservationhackers.org

library(ggplot2)

#
# Make some data 
#
n <- 50 #Sample size per group
x <- 1:n
sd <- 1 #SD for errors

y1 <- x - 0.02*x^2-8 + rnorm(n, sd = sd)
y2 <- -1*x + 0.02*x^2 +8 + rnorm(n, sd = sd)

dat <- data.frame(x = c(x, x), y = c(y1, y2),
                  grp = rep(c("y1", "y2"), each = n))

#Do a t-test first - the wrong way
t.test(y1, y2)
#shows no 'significant' difference of mean y1 to mean y2

head(dat)

#Now plot the data
ggplot(dat) + 
  aes(x = x, y = y, color = grp) +
  geom_point() + 
  stat_smooth()
  
#Clearly y1 and y2 are different