var.test and t.test in r-CodePudding

I am a newbie in R and currently I'm crash into this problem when trying to do a t-test.

I am working on the data set Cars93 in MASS, filter out type t1 (Compact) and type t2 (Large) from Type in Cars93 to get 2 separate dataframe called df.t1 and df.t2

Now I want to compare the means of MPG.City of type t1 and type t1

What I have done are

var.test(df.t1$MPG.city ~ df.t2$MPG.city)
t.test(Cars93$MPG.city[Cars93$Type ==`type t1`]~ Cars93$MPG.city[Cars93$Type ==`type t2`])

However R shows:

Error in model.frame.default(formula = df.t1$MPG.city ~ df.t2$MPG.city): 
variable lengths differ (found for 'df.t2$MPG.city')

Please if anyone could help me look into this. I would be truly appreciated

CodePudding user response：

You dont need to subset the data in the way you have described (creating two separate data.frames, which result in different lengths and throw an error) - try this:

data("Cars93", package = "MASS")

df <- Cars93[Cars93$Type %in% c("Compact", "Large"), ]

var.test(df$MPG.city ~ df$Type)
t.test(df$MPG.city ~ df$Type)