Difference between [c"COLUMNNAME"] AND df$COLUMNNAME, giving 2 different test results

Question

I am running a Rayleigh Test of Uniformity in R. I have my data frame subsetted to 3 columns with slightly differing degrees. When it then run tests selecting one of the columns, my results vary depending on how I have selected the column to use in the test: [c"Direction"] or df$Direction.

My df:

Direction Dir180 CorrDirection    Sex Location
1       146    146           128 Female      Low
2       129    129           111   Male      Low
3       337    157           319 Female      Low
4       130    130           112 Female      Low
5       216     36           198 Female      Low
6       351    171           333   Male      Low

Test:

r.test(x=df2[c("Direction")],degree=T)
r.test(x=df2$Direction,degree=T)

Test results:

> r.test(x=df2[c("Direction")],degree=T)
$r.bar
[1] 3.265616

$p.value
[1] 0.002646553

> r.test(x=df2$Direction,degree=T)
$r.bar
[1] 0.03887638

$p.value
[1] 0.880773

I would like to know why I am getting different results when to me it looks like I am selecting the same column for the tests. What is the difference between using the two different methods and how does the test or program respond to each? Thanks

Is this from `psych` package or from any other package – akrun Oct 03 '19 at 16:39 — akrun, Oct 03 '19 at 16:39
this is from the CircStats package – Kyle Finn Oct 03 '19 at 16:49 — Kyle Finn, Oct 03 '19 at 16:49
Please check my solution posted below – akrun Oct 03 '19 at 16:54 — akrun, Oct 03 '19 at 16:54

akrun · Answer 1 · 2019-10-03T16:52:24.200

The issue is that first is a data.frame with single column and second is a vector. If we extract the column as a vector either by including the , for the column index (if it is tibble, use [[). According to the ?r.test

r.test(x, degree=FALSE)

x - vector of angular measurements in radians.

So it expects a vector and not a data.frame

library(CircStats)
r.test(x=df2[,c("Direction")],degree=TRUE)
#$r.bar
#[1] 0.2321319

#$p.value
#[1] 0.740345

result would be the same with $ extraction as we are dealing with vector

r.test(x=df2$Direction,degree=T)
#$r.bar
#[1] 0.2321319

#$p.value
#[1] 0.740345

data

df2 <- structure(list(Direction = c(146L, 129L, 337L, 130L, 216L, 351L
), Dir180 = c(146L, 129L, 157L, 130L, 36L, 171L), CorrDirection = c(128L, 
111L, 319L, 112L, 198L, 333L), Sex = c("Female", "Male", "Female", 
"Female", "Female", "Male"), Location = c("Low", "Low", "Low", 
"Low", "Low", "Low")), class = "data.frame", row.names = c("1", 
"2", "3", "4", "5", "6"))

Yes I understand thank you. – Kyle Finn Oct 03 '19 at 18:53 — Kyle Finn, Oct 03 '19 at 18:53

Difference between [c"COLUMNNAME"] AND df$COLUMNNAME, giving 2 different test results

1 Answers1

data