Statistics of physics (separate from stat mech)

[Simfish]

So, when people speak of "statistical physics", they're usually referring to "statistical mechanics". But statistical mechanics is really just a small subset of all of statistical physics (unfortunately, that makes googling for statistical physics difficult, because you'll mostly get stat mech sites if you want some more general methods about statistical physics).

There is definitely statistics involved in measurement, and you could just google things about physical measurement. But this doesn't count all the Bayesian methods people use to predict data these days.

So what exactly would you call the branch of physics that correspond to all the statistical methods people use for nuclear and particle physics these days? (especially modern statistical methods like statistical inference and Bayesian statistics?) Certainly, they do form a huge core of nuclear/particle physics research (as you can see at http://gordonwatts.wordpress.com/). But nuclear/particle physics contain theoretical/analytical/observational components too. Statistical methods are often at the intersection of these components. And are there even any scientists who specialize in this sort of stuff? There are certainly particle/nuclear physicists who specialize in these statistical techniques. But what about people who specialize about statistical techniques over many different branches of physics? Just as you have computational astrophysicists who specialize about computational techniques over many different branches of astro.

Here's a good quote:

In particle physics this is what we do. We have petabytes (1000 terrabytes!) datasets consisting of billions of physics interactions. For the particularly rare ones we need to pick out several 100 or 1000 and study them in detail. As you might expect, we are drowning in data and have developed many tools to help us. Computers are central – without them we would not be able to do the science we currently do!

Edit: Wow, so I looked further into the gordonwatts blog. Could you become an experimental physicist without even knowing how to work with laboratory apparatuses these days (just as astrophysicists don't even need to know how to use telescopes anymore)? A lot of the high-throughput data now seems to be done in giant laboratories elsewhere - and experimental physicists have to just analyze the data, right?

Yeah, I know that it's helpful to learn laboratory apparatus, but some of us simply have poor fine-motor skills. We can still learn to do things, but we probably shouldn't be trusted with these apparatuses in a real research setting.

 

[R.Mesri]

That's definitely true. With the development of high-throughput and data-driven experiments, experimental physicists now have access to huge datasets that they never had before. As a result, there is a greater emphasis on data analysis and statistics. So while it still helps to be familiar with laboratory apparatuses, there are plenty of opportunities for those who don't have strong fine motor skills. In fact, many of the most successful experimental physicists today are those who combine their knowledge of laboratory apparatus with strong statistical and data analysis skills.