Monday, March 10, 2008

Spearman correlation

Spearman correlation coefficient determines the degree of relationship for two sets of ranked data. Spearman correlation is also called the rank-order correlation coefficient. although spearman correlation is far less common than pearson, occasionally variables are ordered according to rank, or variables may be subsequently ranked on the basis of a continuous variable. The formula for spearman is the following:
see page 181.
Spearman might be most typically used in situations where there are a number of variables and they are all ranked by two independent judges. For example, a psychologist wishes to determine how alike husbands and wives are their vegetable preferences. The member of the couple were independently asked to rate their preference for seven vegetables from most preferred (#1 rank) to least preferred (#7).
Significance test for spearman r
Spearman cannot be tested for significance in the same manner as Pearson r.

No comments: