Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

How useful is probability?

  1. Jun 23, 2011 #1
    Hello PF

    No doubt that techniques in probability and statistics are useful to an aspiring physicist but I was wondering how useful exactly (or approximately). One of my upper division classes is called "thermodynamics and stat mechanics". I have no idea what level of probability or stats is needed, but is it sufficient if I learn the math as I go along? I'm currently an undergrad student looking to get a masters and then go to industry, preferably in CM.

  2. jcsd
  3. Jun 24, 2011 #2


    User Avatar
    Science Advisor
    Homework Helper

    It's about 50-50 that you will need it.
  4. Jun 24, 2011 #3
    Basically none for the thermo part, at least in the theory. For undergrad stat mech: basic probability (expectation values, variance, mean, gaussian/poisson distribution, etc), basic combinatorics (e.g. binomial theorem). I took both without a strong pobability/statistic background and did pretty well though I would advice to do a little self-teaching before the stat mech part.
  5. Jun 24, 2011 #4
    Hahahahaha, well played.
  6. Jun 25, 2011 #5
    In my experience, I learned statistic and probability measures as I went along.... in both coursework and labwork/research (with no detrimental effects). I don't think it's necessary to take a special course on statistics... and as I heard from others who did take such a course, the course probably quite dry. (of course I'm just one data point...)

    The statistical mechanics text should include the measures important to the field... and in a more interesting way. Basically temperature is a measure of the average energy per particle... but there's a spread of energies (and different "types" of particles with different ways of inhabiting available states)... and all this relates to other bulk measures and behaviors of the system. Since you are taking an undergraduate course (albiet an upper-level), your professor will probably "hold your hand" a bit and help you through parts that the text may have unclear.
  7. Jun 25, 2011 #6
    Not to hijack, but how much statistics do you need to do experiments/interpret the data? Can you get away with the simpler stuff: mean, standard deviation, confidence intervals, etc?
  8. Jun 26, 2011 #7
    thanks everyone. I have a habit of chopping the vegetables with a butcher's knife, as they (or just me) say but I'll try and self-study a little bit of stats can't hurt. I think stats and probability are really powerful mathematical tools but it doesn't interest me as much as its applications in physics.
  9. Jun 26, 2011 #8
    Interesting thread. Not really related to statistical mechanics, but there are a lot of physics/computer science/biology professors doing research with large data sets, and I was thinking that a good way to get undergrad research was to learn data mining/statistical analysis and try to wiggle myself in as a tool.

    If you learn graduate level probability and statistics (proof based classes), will that be helpful for analyzing these data sets? Or do professors just look for people who know how to program, and don't really care if you know the theory behind the analysis (obviously this is not research under the pure math department)?
  10. Jun 27, 2011 #9


    User Avatar
    Science Advisor

    I can't comment on specific areas of physics that you have mentioned above, but I can say that in the general context of experiments, it's pretty damn important.

    The whole notion of measuring, analyzing, and controlling experimental bias lies on a solid foundation of statistical analysis that has been developed (and is still developing) of a long period of time.

    Detecting bias is crucial in any experimental context, so that it can at least be identified and possibly corrected, or identified and analyzed further.

    I'm guessing that physicists like any other scientist, have to consider this especially when drawing conclusions and supporting evidence for their argument in whatever context that is made in.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook