Thread by @JCSkewesDK

As someone who's taught with both, the discussion of R vs SPSS for teaching stats neglects two facts.
1) Learning SPSS is no picnic
2) SPSS obscures concepts and makes stats HARDER to learn
Trust me, from a purely didactic perspective, R is miles ahead.
It's not just about coding

You don't need to do fancy coding, or use markdown or tidyverse.
You can tweet a whole psych undergrad friendly exercise on t-tests.

muA<-50
muB<-55

n1 <-100
n2 <-100

sigma1<-20
sigma2<-20

A<-rnorm(100,muA,20)
B<-rnorm(100,muB,20)
hist(A)
hist(B)
t.test(A,B)

Change the last 3 lines and now it's an exercise on linear models!

muA<-50
muB<-55

n1 <- 100
n2 <- n1

sigma1 <- 20
sigma2 <- 20

A<-rnorm(n1,muA,sigma1)
B<-A+rnorm(n2,muB,sigma2)
plot(A,B)
summary(lm(A~B))

Four more lines and you can show how t.test are really just linear models anyway.

scores<-c(A,B)
groups<-c(rep(1,n1),(rep(2,n2)))

summary(lm(scores~groups))
t.test(A,B)

Mess around with variables, and talk about interpretation.

Discuss content from texts, lectures, etc.

Have students write their own simulations.

Bring real data in.

Introduce some ggplot.

But really you can code the stats part from intro to psych methods in 3 tweets.

Forget psychology undergraduates. We should be doing this in highschools.