I almost spat out my cereal this morning when reading this conference proceedings by Anderson et al. (2015) presenting “*the technique of joint likelihoods*“. What’s the technique of joint likelihoods, you ask? Just the most obvious thing that you’ve been doing since forever: when you have N separate datasets with individual likelihood functions, , for common parameters of interest, , and nuisance parameters, , the technique of joint likelihoods suggests that to perform inference on you *multiply the likelihood functions together *(as opposed to blowing them out your arse, I presume).

### Like this:

Like Loading...

*Related*

Statisticians and your fancy-pants “multiplication”!

Fantastic advance! Now we can analyse data sets that are more than a single number.

You still eat cereal? You should read a little from this blog for a change from astro and stats; it will surely do a lot more for your health and your ability to continue reading astro and stats for years to come: healthfully.wordpress.com

Ha! You’re right: my wife & i are trying to kick our cereal habit thanks to a new blender!

There was a biology paper a while back which introduced the “new” concept of Riemann integration.

True, that paper came up a number of times in my conversations with bioinformaticians last week at the Wellcome Trust … though in each case they themselves brought up this embarrassing example!