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

1-cdf value vs chi-squared statistic mislabeled on site?

  1. Mar 1, 2013 #1
    I am confused by the labeling of the (only) graph in
    http://en.wikipedia.org/wiki/Pearson's_chi-squared_test
    The values on the vertical axis appear to be the values of 1-(the chi-squared cumulative probability distribution values*)
    as I would expect in finding the p-values.
    However, this graph is labeled the "chi-squared distribution" -- which is ambiguous enough, but when one clicks on this term under the graph, one is directed to a site where the term means the chi-squared probability density function. But the graph cannot be the probability density function: the values are different*, and a probability density function has probability densities on the vertical axis, not probabilities.
    So, is this mislabeled?

    (*The pdf and cdf graphs were taken from http://en.wikipedia.org/wiki/Chi-squared_distribution#Probability_density_function, and individual values, for more precision, from http://www.danielsoper.com/statcalc3/calc.aspx?id=11)
     
  2. jcsd
  3. Mar 11, 2013 #2
    The graph isn't actually labeled as the chi-squared distribution. It's merely saying it's a variation of the chi-squared distribution. The density function given on the second page is correct. Chi-squared is just a gamma distribution so it's obvious the first page doesn't show a standard chi-squared. The graph on the first page is just the function of a statistic, not a random variable.
     
  4. Mar 11, 2013 #3
    ahhhhh. That solves that mystery. Thanks, jwatts.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: 1-cdf value vs chi-squared statistic mislabeled on site?
Loading...