COMM 2001: October 21 st
The level of measurement is critical to analysis
- Each question is measured at a different level of measurement
Levels of Measurement
Nominal: A set of distinct categories
Ordinal: Response categories with order to them (There is order but no mathematical meaning)
Interval: Response categories have order plus equal measureable distance between categories.
(There is mathematical meaning to the scale)
Example: On a scale from hot to cold…
High to low, so high will have the level of 1 and low will have a level of 4.
The advantages of higher level measurement are that there is precision and the ability to use better
statistical measures. The disadvantage is that it is difficult to look at distributions.
Tables: Describing distributions
Charts/Graphs: Visually presenting distributions
Describe the distribution
- Number of people or cases in each category
- Percentage of cases in each category
- % with non-response (missing values)
Changing Distributions: Recording Variables
- To collapse categories of an interval variable which has too many to display.
- To collapse categories of an ordinal variable which has two few people in them.
- To organize a large nominal variable into categories which make conceptual sense.
Non response categories of