T test between price and size-CodePudding

Im trying to do a t test between 2 variables, price of a building and the size of a building.

Size consists of either "large" or "small" and Price can be any number (54,2123,325...)

i did this:

data$size <- as.factor(data$size)
t.test(data$size ~ data$price, data = data)

But i get this error and I cant solve it:

Error in t.test.formula(data$size ~ data$price, : 
grouping factor must have exactly 2 levels

I am a beginner in R and would appreciate any help here you can see the data

CodePudding user response：

You must reverse the order of the variables that you put in t.test, first comes the numeric variable (price) and then the binary variable (size).

size <- c("large","large","small","small","small","small")
price <- c(3536,2239,1982,1769,1769,1769)
data <- data.frame(size, price)

data$size <- as.factor(data$size)
t.test(data$price ~ data$size, data = data)

CodePudding user response：

I think you need this:

with(data, t.test(price[size == "large"],  price[size=="small"]))

CodePudding user response：

First, subset the data into small and large:

small <- subset(data, size == "small",price, drop = T)
large <- subset(data, size != "small",price, drop = T)

Next ensure that your data pass a test of homoscedasticity--are the variances homogenous? We can use var.test() for that:

var.test(small, large)

Finally, depending on whether the p-value came out >0.05 (var.equal = TRUE) or <0.05 (var.equal = FALSE) we perform a t.test():

t.test(small, large, var.equal = TRUE)