Chapter 1 Variables and Their Measurement
1.3.1 Nominal Variables
As the name of this level of measurement implies, the information contained in the categories of a nominal-scale variable is solely their… well, name. Think about the religious affiliation variable from the Do it! 1.2. exercise. You have already probably imagined people’s possible affiliations in terms of religion (i.e., what religion they subscribe to, if any) as something like Muslim, Jewish, Christian, Sikh, Hindu, Buddhist, not religious — though likely (and depending on your own religious affiliation) not in this particular order.[1]
Of course, I could have just as easily listed the possible categories (or “questionnaire answers”) as Christian, Muslim, Jewish, Buddhist, Hindu, Sikh, not religious. Or, as Sikh, not religious, Buddhist, Hindu, Jewish, Muslim, Christian. Or, as… virtually any possible variation in the ordering of the list.
In other words, the information we have about religious affiliation is simply in identifying the different categories, and that is all. We cannot do much more than count the different answers and specify what they are. We cannot even use some inherent order to them, as they are only that, names.[2]
Do it! 1.3. Nominal Variables
Try to come up with at least three different nominal variables. Can you explain why they are nominal? Try to defend your choice in identifying the scale for these variables as nominal.
- It's also likely that these general categories might have been disaggregated to list variations/denominations, e.g. Catholic and Protestant instead of simply Christian, or Shia and Sunni instead of simply Muslim, etc. For simplicity's sake, I choose to use the most general religious categories in the example. ↵
- Of course, we could order the categories alphabetically -- just like you can order pretty much anything alphabetically. That would be an arbitrary decision, however, not an inherent order contained in the names (like that in small to big, left to right, slow to fast, less to more, etc.). When researchers study religious affiliation in real life, they usually list the groups' names by the size of the religious group/popularity of a religion in their area. For example, in the Americas and Europe the listings usually start with Christian. In India, one can arguably assume they start with Hindu, etc. This type of ordering by size is still purposefully imposed, not an inherent one. ↵