It is important to know your scales because most statistics have specific scale requirements. For example, ordinal data (ranks or ratings) can not be used to find the mean, standard deviation, or variance. The same is true for nominal data (frequencies). Most of the standard statistical procedures require at least INTERVAL SCALE. Many errors have been made by authors who have not remembered their scales. This results in incorrect data analysis and the interpretation of this information is questionable.
To help you remember your scales, use the following acronym:
NOIR = No One Is Ready. Which represents Nominal, Ordinal, Interval, and Ratio scales.
The most important thing you can do when you have a set of data in front of you is to evaluate the scale you have. Know the characteristics these 4 different scales. If you fail to do this, you may make an analysis error which invalidates your findings or outcome data. It is always tempting to use, for example, ANOVA when the data are ordinal scale because this statistic is more powerful than non-parametric methods such as the Kruskal-Wallis or Friedman procedures. Remember no one is ready!
Ken Salzwedel