Welcome to cocor! This is a website allowing to conduct statistical comparisons between correlations. Click "Start analysis" to begin!

The calculations rely on the tests implemented in the package cocor for the R programming language. An article describing cocor and the cocor R package documentation are available. Here you find an overview of all implemented tests.

You can integrate the R code generated by this web interface in your own R script. To do so, you can download the cocor package from here or r-project.org. Alternatively, you can type one of the following commands in the R console to install:

# cran repository
install.packages("cocor", lib="/my/own/R-packages/")

# alternative repository
install.packages("cocor", lib="/my/own/R-packages/", repo="http://comparingcorrelations.org/repo")

The cocor package also includes a GUI extension for the R front-end RKWard, which you can use instead of this web interface.

Are the two correlations based on two independent or on two dependent groups? (If the data were taken from measurements of the same individuals, the groups are dependent.)

The two correlations are based on

Are the two correlations overlapping, i.e., do they have one variable in common?

The two correlations are

Examples
Overlapping: Nonoverlapping:

Correlation 1: age ~ intelligence
Correlation 2: age ~ shoe size

These are overlapping correlations because the same variable (age) is part of both correlations.

Correlation 1: age ~ intelligence
Correlation 2: body mass index ~ shoe size

These are nonoverlapping correlations because no variable is part of both correlations.
Please provide the correlations you want to compare:
r1.jk:

r2.hm:
Please indicate the size of your samples:
n1:

n2:
Please provide the correlations you want to compare:
r.jk:

r.jh:
To assess the significance of the difference between two dependent correlations, you need to provide the correlation between k and h:
r.kh:
Please indicate the size of your sample:
n:
Please provide the correlations you want to compare:
r.jk:

r.hm:
To assess the significance of the difference between two dependent correlations, you need to provide additional related correlations:
r.jh:

r.jm:

r.kh:

r.km:
Please indicate the size of your sample:
n:
Please choose an alpha level:
Alpha level:
Please choose a confidence level:
Confidence level:
Please choose a null value. The null value is the hypothesized difference between the two correlations used for testing the null hypothesis. If the null value is other than 0, only the test by Zou (2007) is available.
Null value:
Do you want to conduct a one- or two-tailed test?