Définition Qualitative data

Qualitative data are attributes, the values of which (e.g. replies to a survey) can clearly be categorized. They cannot, however, be a mathematical value (=number). Typical examples of qualitative data are gender, religion or political affiliation. For these kinds of attributes, we can only reproduce percentage values as results of a survey (x in 100%).

Strictly speaking, ordinal data such as education level or personal satisfaction levels are qualitative data as well. We can establish a hierarchy of ordinal data; an exact numerical scaling is not possible though. One level of education is not "50% better" than another, it can only be specified as "superior to". Ordinal scales are not interval-scaled.

Nevertheless, ordinal features are described as 'seemingly quantitative', because approximately, they allow for mathematical operations. Thus, when comparing two groups, we can state that group A 'on average' has a higher level of education than group B. In order to achieve this, we would translate the ordinal levels of education into numerical values (1=lowest education level, 5=highest education level). From these numerical values, we can calculate a mean for each group and then compare both values. Technically speaking, this calculation is incorrect though, because the statement "group A has a higher education average, scoring 3.3 over group B's 3.0" actually implies that all educational levels have the same distances or intervals from each other (=are interval-scaled).