On Thu, 18 Apr 1996, Robin Allenson wrote:
> A friend of mine is doing her final year project, and
> needs to do some stats on a set of data, which is
> comprised of two variables. The first takes on three
> values: -1, 0, +1. The second is always 1. We'd like
> to analyse the correlation between the two. Currently,
> we're using Spearman's co-efficient.
First issue... Your friend has a real problem! The second variable is a
constant and thus, it can not correlate with any other variable. That
is, without variance, there can be no covariance.
> Does anyone have any idea of other tests to apply,
> and how what they measure differs from Spearman's?
> If anyone has references, or text, on what Spearman's
> Coeff. measures, or some background, this would also
> be extremely useful. If you can email directly I can
> summarise to the net if anyone else is interested.
Spearman's Rank Order Correlation is an algebraic conversion of the
Pearson formula for r when one is working with ranked data. With the
widespread availability of computer programs, most people just use Pearson.
______________________________________________________________________________
William B. Ware, Professor and Chair Educational Psychology
CB# 3500 EMAIL: wbware@unc.edu
University of North Carolina PHONE: (919)-966-5266
Chapel Hill, NC 27599-3500 FAX: (919)-962-1533
______________________________________________________________________________