Pitfalls

Qualitative data in number codes is not the same as quantitative data

For example, to illustrate the ordering of the impact of training on staff performance from most negative to most positive, the three responses (negative impact, no impact and positive impact) may be coded as:

Code    Response
1           Negative impact
2           No impact
3           Positive impact

To take another example, while industry is a nominal variable, we may assign numerical labels to various industries for coding purposes or to illustrate its hierarchical structure. For example:

Code    Industry
46         Wholesale Trade
56         Food & Beverage Services
85         Education

When the values that a qualitative variable can take are expressed in number codes, we may mistake them to be quantitative variables, and think that we can perform further arithmetic operations on them such as computing their mean or subtracting one value from another. However, the numerical values taken by qualitative variables are merely labels rather than actual quantitative values, and are to be treated no differently from qualitative variables expressed in words or letters of the alphabet system.

How do we determine if a variable is quantitative, ordinal or nominal?

The following flowchart can help us to determine whether a variable is quantitative, ordinal (qualitative), or nominal (qualitative).