Re: Effect sizes for Friedman's ANOVA

Thank you both for your anwers.

mOOes: you're right: it's Phi of course, not rho.
It was pretty late when I wrote my question and it seems I wasn't that focused anymore...

I also think Ray Koopman answers your question and mine.

I think I found also an alternative to this method by simple rank transform m data according to Friedman and then perform a parametric repeated measures ANOVA on ranks. I than simple estimate the effect size by computing a partial eta². The normality assumption is stressed and the homogenity of variances assumption does not stand in my design. Since I am not interested in investigating interactions between independent variables, according to Beasley I do not need to align the Friedman ranks... Do you think it's a valable alternative?
.

Relevant Pages

  • Re: Integrals In Spherical Coordinates
    ... If phi and theta are fixed, what are the allowed values of rho ... These rho values can depend on both phi and theta. ... on theta *but not rho* and give the limits of the middle integral. ...
    (sci.math)
  • Re: Integrals In Spherical Coordinates
    ... You can think of the rho values as sweeping out ... These rho values can depend on both phi and theta. ... > line sweeps out a planar cross-section of your volume. ...
    (sci.math)
  • Re: Integration in polar coordinates?
    ... No, assuming you are talking about spherical coordinates (rho, theta, ... phi) here, there is a sinfactor in addition to the r^2, which you ...
    (sci.math)