I am testing the reaction type of female birds (which can be + or -) and the length of reaction of females (in seconds) to male birds depending on the color of their feathers (highly colorful or not colorful) and their loudness (loud or quiet).

Independent Variables --> Male Bird Color (binary), Male Bird Loudness (binary)

Dependent Variables --> Female Reaction Types (binary), Female Reaction Length (continuous)

Which statistical test should I use to find the correlation strength of my data for these 4 variables?

  • 2
    $\begingroup$ It depends on your experimental design. Could you provide some information about how you went about selecting experimental subjects and measuring or observing their behaviors? $\endgroup$
    – whuber
    Apr 6, 2018 at 18:04
  • $\begingroup$ "Colorful" males had a high feather color index while "not colorful" males were albino. Sound frequency was used to differentiate between "loud" and "quiet" male birds. Female bird reaction length was measured with a timer if the reaction type was + (if a reaction occurred). If there was no reaction (-), the time was recorded as 0 seconds. $\endgroup$
    – orangebull
    Apr 6, 2018 at 19:29
  • $\begingroup$ Please add extra information to your question rather than replying in comments. Coding a lack of reaction as an instantaneous reaction seems bizarre on the face of it, by the way. $\endgroup$ Apr 6, 2018 at 20:11

1 Answer 1


The standard analysis would be some kind of linear regression. In this case, since your response variables are related but have different types, you'll get strongest results from a flexible multivariate generalized linear model. At this time, I only know of R packages for such analysis: mcglm for quasi-likelihood models and mcmcglm for a Bayesian approach. There might be others out there too. Note that the learning curve on both can be steep, especially if you don't already have a strong statistical background.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.