Sometimes nothing is more amusing than taking a look at all the quizzes in magazines to determine what “type” you are. Fortunately, they serve to entertain the general public rather than strive for empirical confirmation. “Shopping type”, “Seasons Type”, “Single type”, “Flirt type”, etc … It seems that everything can be categorised into “types”. The reason is simple: type models are intuitive because everyone knows someone who is somehow at least a bit similar. It is no wonder that the use of ‘types’ is also popular in HR. However, here the types, are not “winter type” versus “summer type”, but rather “doer”, “instigator”, or “innovator”.

Historical origins

Even the ancient Greeks were concerned with dividing people into types. Following a theory that probably goes back to Hippocrates, there are four types of temperament, by the respective domination of one of the four fluids of the body (blood, phlegm, black bile and yellow bile). These temperaments are: melancholic, phlegmatic, sanguine and choleric.

Psychoanalysis paved the way to an entirely new way to ‘type’ things. The most prominent example here is certainly the psychosexual classification developments of Freud. Those who are stuck in an earlier phase of development are rather compulsive or hysterical compared to their contemporaries.

Today’s Type Models

There is a variety of competing paradigms for describing individual differences (an overview can be found in Asendorpf & Neyer, 2012). In aptitude testing and potential analysis property-based approaches are generally accepted, rather than type models (eg Big Five model). Here differences on various property variables are examined in terms of performance differences in the workplace. For example, an entertainer should not have a below average extraversion rate.

Unlike property based approaches, type models are people-oriented, which means that people who are similar, are divided into personality types. There are two basic approaches:

1. Firstly, types can be defined based on very high and very low occurrences of certain characteristics (extreme group approach). A basic problem with this approach, however, is that it gets complicated quickly when exceeding several feature dimensions. Alone among the Big Five (5 dimensions) there are, when only above and below average is used, 2 * 2 * 2 * 2 * 2 = 32 combinations.

2. With prototypes, fictitious persons are defined, whose prototypical features represent the prototypical ideal. Hardly anyone fits clearly into one prototype but is rather a combination of multiple types.

Type Models: Yes or No?

One advantage of type models is that they are more easily communicated than purely characteristic-centred approaches. However, it should be noted that as type models deal with generalisations, this can lead to over-generalisation. Add to this problem is the treatment of mixed types. If a type model is not replicable and valid then it is simply useless.

If type models are used, should at least be ensured that the postulated characteristics and behaviours that distinguish one type (or not) also (1) actually occur in clusters, (2) type classification is reliable and objective (ie, when a first test result is “doer”, the second should not be “Sociable Type”) and (3) allow an accurate statement to empirically confirm development potential or predict suitability. Otherwise, type models are not worth more than a test to determine which season you are.