ABSTRACT
Ranked (ordinal) data must be analyzed using different statistics than continuous (interval or ratio) data. This is true for validity, reliability, and objectivity, and also true for research applications. Although there are non-parametric tests of the difference in ordinal means, for our purposes, we will only consider the Spearman’s rho to quantify the relationship between an ordinal variable and another variable. The Bland-Altman error analysis allows simple evaluation of continuous measures by evaluating the error between measurements. It also provides information on how the new measurement might be improved. It is a bit simpler than the t-test and correlation and gives us more information, so is a good way to assess validity.
