Quality is a measure of relative value. Qualitative data deals with meanings.
"Is this movie interesting?"
["Enemy of the State", "Stepmom" movies used for class discussion]

Meanings are usually mediated through language and action. Text is by no means the only means of communicating, but most effective qualitative information in an electronic age; art and design have become powerful media tools. Action is a medium for communicating meaning. To interpret is to make action meaningful to others.

["A Pattern Language" from Christopher Alexander used as example]

The richness of qualitative data encompasses virtually any kind of data: sounds, pictures, videos, music, songs, prose, poetry, embracing an enormously rich spectrum of cultural and social artefacts.

When we categorise data, we make a distinction between this observation and others. The categories can be "fuzzy" and overlapping. A concept can convey very different connotations.
["Western" movies is a thematic classification overlapping with chronological classification "Movies of 60s"]

Categorising brings together a number of observations, which we consider similar in some respects by implied contrast with other observations.

In assigning something to one category, we do not automatically exclude it from others. The categories are inclusive rather than exclusive. At a more sophisticated level of classification, we can differentiate more explicitly between observations. This can be done by identifying some characteristics which observations have in common, the better to understand what distinguishes them.

With nominal values, the values we use must be mutually exclusive and exhaustive. "Mutually exclusive" means no bit of data fits into more than one category.
[ISIC-International Standard Industrial Classification discussed as class example]

If values are put into a rank order, nominal variables can be converted into "ordinal" variables. Ordinal variables give more numerical information about the data, since one bit of data can be indicated if it is higher or lower in the pecking order than another.

Comparable conventions fixing the distance between values (or categories) can be rarely established. Social scientists do not have standard units in terms of which to measure things like poverty, health or quality of life. The prime reason is that we cannot agree in the first place about the meaning of what we are trying to measure.
[Class discussion: What do you understand under "Quality of life"?]

To interpret data in social research, it may be more important to use meaningful categories than to obtain precise measures.

Back to home