###R code for Lecture 19 - relationship between T-tests and Regression


### import the data
datum=read.csv(file.choose())
head(datum)

### Note that x-variable can be dummy coded, but doesn't have to be. 
### If not dummy coded, R will dummy code for you when it analyzes data

### Also note that with newer versions of R (Version 4.0 or greater), 
### R doesn't automatically treat words at categorical x variables
### To get R to treat words as a categorical variable, you must convert
### the variable to a 'factor'

datum$X=as.factor(datum$X) # Replace 'X' with the variable name


### Plot the data
plot(Y~X,data=datum)
### Replace 'Y' with y-variable, and 'X' with X variable



### Run a regression with categorical x (technically a t-test)
results=lm(Y~X,data=datum)
### Replace 'Y' with y-variable, and 'X' with X variable
summary(results)
### Note that if X is a text variable you might have to figure out what your 'reference' is

### Get confidence intervals
confint(results)



### Predictions when x is categorical are the same as when x is continuous, 
### but x must be set to a categorical group

NewX=data.frame(X="group1") #where 'X' is the name of your x variable and "group1" is the group you want to make
### predictions for. 
### Note that "group1" must be in quotes
predictions=predict(results,NewX,interval="prediction")

### If you want to make predictions for all groups:
NewX=data.frame(X=c("group1","group2","group3"))