1

I am conducting the Levene's test as well as Shapiro-Wilk test on a large set of data. I am wondering how do I access to particular parts of a result and store them in a CSV file to simplify reporting results.

For example, a Levene's test results look like this in R:

> leveneTest(Var~Label, data = df)

Df F value Pr(>F)
group   3  0.7595 0.5183
      166

How can I get only F-value and p-value stored in separate variables? Same as W and p-value in the following:

> shapiro.test(df[,i])

Shapiro-Wilk normality test

data:  dfc[, i]
W = 0.70952, p-value < 2.2e-16
mfaieghi
  • 570
  • 2
  • 9
  • 24

1 Answers1

2

Like this below? I don't know what is i in your code though:

library(car)
df = data.frame(Var=runif(100),Label=rep(c("A","B"),50))
test = leveneTest(Var~Label, data = df)

To know to access the data, do:

class(test) [1] "anova" "data.frame"

So test is a data.frame and you can either use $ or access it via rows and column names. Since the column names contains spaces in this case, using the matrix way is a bit better (and you know the column you called):

Fvalue = test[1,"F value"]
pvalue = test[1,"Pr(>F)"]

Fvalue
> [1] 0.3516209
> pvalue
[1] 0.554563

For shapiro, unfortunately they created a new class, so you can check the vignette:

Value:

 A list with class ‘"htest"’ containing the following components:

statistic: the value of the Shapiro-Wilk statistic.

p.value: an approximate p-value for the test. This is said in Royston (1995) to be adequate for ‘p.value < 0.1’.

So we treat it as a list:

names(test)
[1] "statistic" "p.value"   "method"    "data.name"

test$p.value
[1] 0.0002292068
StupidWolf
  • 45,075
  • 17
  • 40
  • 72
  • Perfect -- i is just an iterator to iterate over several vectors I want to test. Can you also tell me how to do this with shapiro.test results? The output for that is a bit different. Also, is there any general way to figure out which approach to take in dealing with these situations? – mfaieghi May 19 '20 at 16:38