On Jul 25, 10:56 pm, John Kane <jrkrid...@gmail.com> wrote: > This sounds like an excellent and very ambitious project. Since it is > a high school project I'd suggest that you don't get too complicated. > As I thing Richard pointed out it can be very complicated but treat > the project as a pilot - keep it simple and just remember to point out > the limitations when you write the report :) Then like any good > researcher you can ask for more funding for future reseach :) > > > iii)Is there some easily available (preferably free) software that > > will let me do all this analysis (brownie points for fitting > > probability distributions and graphing)? It would be a nightmare to do > > this by hand since we usually work with less than 50 data points > > instead of several hundred. > > Perhaps the Royal Royce of free statistical packages is Rhttp://www.r-project.org/. > However it is not all that user-friendly and it might take longer to > learn than is worth it for one project. > > There is plenty of good documentation on the site that can help and > there is a very good R-help mailing list where you can ask questions > > Otherwise here are some sites for some dedictated statistiscs/ > graphing packages (Free or shareware in most or all cases). I have > not checked the links in a few months though. > > http://gsociology.icaap.org/methods/soft.htmlhttp://freestatistics.altervista.org/stat.phphttp://www.psychnet-uk.com/experimental_design/software_packages.htmhttp://data.fas.harvard.edu/micah_altman/socsci.shtml
Thanks a lot. A teacher said that it's effectively impossible to separate a heterogenous population into constituent stratas based only on the data and not external variables. If that is true, we can not approximate marks distributions to be normal as they are certainly far from normal. They often have multiple local maxima and are severely skewed.