My Photo


  • Michael Moorcock
    "a genuine philosophy for the 21st century"
  • Mary Midgley
    "this matters - read it!"
  • Kendall Walton
    "wonderfully refreshing and inventive"


Game Design

Blog powered by Typepad

« In Control | Main | Introduction to Temperament Theory »


Feed You can follow this conversation by subscribing to the comment feed for this post.

I guess one major question concerns your use of words like "type", "character" or "behaviour" and "the way people react".

It seems to me that even quite simple circumstances not considered in your presentation may play a critical role in the picture you seem to present here, e.g. whether a person tries to actively/passively *achieve* something or actively/passively *prevent* something from happening.

Can you supply more details on the context you find suitable for application of this model?

What is your general idea on the connection between "temperament type" and action? What would "as a rule..." mean in this regard?

I see the need to justify, or prove, concepts like Temperament Theory to be futile and rather beside the point.

The goal of most "personality" testing is not to place you within a box. They should not be used to determine action, compatibility, direction. In fact, they shouldn't be used to *determine* anything.

Personality tests are useful as a tool. They can be used for self-analysis – to become aware of how you act, and present suggestions as to why you act. They can be used to predict likely actions of individuals – as much as logical analysis or game theory can. They are a way to profile and understand, in general terms, the actions of someone (or a group of someones). They can also be used to classify groups or actions, so long as they are taken to be broad and incomplete classifications.

No system is going to tell someone (even you) how you will act in all situations, under all conditions. Personalities, while often firm and unchanging, are prone to alteration under the right circumstances.

Temperament Theory seems, to me, to be just a way of looking for recognizable behaviours patterns within human actions. Good enough.

Duncan - I couldn't have said it better myself! I believe "as part of a process of self-discovery" is the phrase Berens uses to describe how this model can best be employed.

translucy - by 'application', do you mean how might it be scientifically employed? In this regard, only ever at a statistical level. This is a model that can be used to help one think about behaviour - but it is not a system for predicting action (or inaction) at an individual level.

That said, as an example of using this model to make a general prediction, I can hypothesise that plumbers (who statistically tend towards Artisan in Temperament Theory according to the data I have) will file their tax returns later on average than teachers (statistically tending towards Guardian). But I could never make an accurate prediction about when an individual plumber or teacher would file their tax return.

Hope this is clarification and not a tangent! :)

(Note: This response was written on the train after a 2-day workshop, so does not take note of other discussion on the thread and may be even more disjointed than usual)

A friend [note to Chris: Malcolm Harwood] pointed me to the DOS version of "Please Understand Me" in 1992. I took the test then and have re-taken it and its variations several times over the intervening years. As far as I know, I still have the detailed logs of most of these if anyone's interested. Typical scores (from memory, there could be outliers beyond these values):

90% to 70% Introverted, slowly decreasing over the years as I learn to get along with people better;
80% to 60% iNtuitive, no obvious correlation with time or mood;
15% Thinking to 15% Feeling, no obvious correlation with age, some bias towards T when I also show symptoms of depression*;
10% Judgemental to 20% Perceptive, no obvious correlation with age, Judgemental bias only appears when I also show symptoms of depression.

So in one sense I can agree with the criticism: over the years, I could have been pigeonholed as INTJ, INTP or INFP (with one dip into INFJ territory with both the F and J barely present). One person, four different temperaments over time - clear refutation of the theory, m'lud.

In another*** sense, however, this provides some evidence for the theory, or at least for a close relative. Firstly, the scores on several different tests over a decade and a half show considerable consistency - more so when you consider that I have an episodic mood disorder and that the variations in two of the axes correlate reasonably well with variations in the strength of that disorder. Secondly, if one considers statements about a personality, one would expect each of those statements to apply to personalities in *some subset of* the 4-dimensional space. There is no particular reason why they should apply only to the extremes. Why should a statement not apply to (say) moderate introverts with no particular bias towards Thinking versus Feeling? I elect to interpret the statements in the results I typically obtain in that light, and statements common to the IN*P or INF* profiles do, indeed, seem to correspond better with my observations of my own personality. Forer effect? Perhaps... let's see if we can construct an experiment that is at least somewhat resistant to that effect, and examine its possible outcomes.

The Myers-Briggs and Keirsey work was undertaken before the present explosion in automated information processing and communication, and may have been limited because of the difficulty of gathering and analysing large data sets. Given the Internet, we can (and arguably should) try to gather the data for a finer-grained approach. I suspect that a modified theory which represented scores as points within a four-dimensional space and associated statements about people with (say) an ellipsoid within the same space would provide a reasonably good match to a personality. That hypothesis is testable fairly easily. One could construct an experiment where all subjects took a test; all then received statements matching some random point in the space. Subjects could then be asked to mark each statement on an agree / neutral / disagree, and the number of agreements could be tested against the distance between the point identified by the test and the point where the statements were presented to the subject. While not bombproof - some care would have to be taken to avoid statements being presented that matched questions in the test, for example - this could provide a more sensitive experimental base than the present typology.

Setting up the initial site to ask questions and gather raw scores would not be hard; neither would presenting a random selection of responses and asking people to
rate them. Detecting systematic fraud (noise) due to deliberate mis-filling of the questions would be harder, I suspect. For maximum amusement, such a site could
be used to *infer* the positioning of statements relative to M-B scores once a certain number of ratings had been received, simply by examining the two sets of results
and correlating appropriately. That would in itself be an interesting use of such a site, as the set of inferred areas corresponding to statements could then be treated as a set of hypotheses, to be reinforced (but never confirmed) or falsified by subsequent observation and tests. If the site continued to emit a proportion of random results, a figure**** for the probability of the Forer effect coming into play could be calculated.

Hmmm. I feel a project coming on. If I could only construct a set of questions that was not considered derivative of the Myers-Briggs or Keirsey questions... That said,
this is an obvious experiment and has almost certainly already been done. Can anyone point me to the results?

* I can recognise** at least three episodes of major depression during the time I've been aware of Myers-Briggs typology and occasionally taking the tests, of which the most recent two episodes have been independently diagnosed. There is one other possible episode that may have been sub-clinical.

** My GP accuses me (with a smile on his face) of being one of the most self-aware patients he's ever treated.

*** And to my mind more valid, but feel free to disagree.

**** I'm not stating that the figure would be *accurate*, owing to possible systematic biases introduced by those taking the test - an ever-present risk when you have a self-selecting population of subjects.

Peter: I also tend to see the Myers-Briggs axes as defining a four dimensional "personality space", although I'm not certain how helpful this model is in general terms.

Psychologists who are pro-Temperament theory generally lament paper tests as so wildly inaccurate as to not be worth using any more (although to my knowledge no-one has attempted a long term correlation of data from paper tests). An additional problem is that the commercial exploitation of the Myers-Briggs model renders this untenable as a research concern - expect lawyers. :(

This is one of two reasons I now favour Temperament Theory, which still lies in the public domain, and Berens addition of Interaction Styles provides a complete correlogue for the Myers-Briggs scales. Unfortunately, learning that she has trademarked the Interaction Style terminology is a frustration. That said, I can render the same system in a public domain form and, quite frankly, may have to.

All this being said, if you are happy to wait for me to finish going through the Temperament and Interaction Style work on the blog as a preliminary step, I would be happy to work with you to build a long term online personality study - especially if we can also (voluntarily) collect data on what people are playing at the times they are tested (data regarding games of all kind, including sports, puzzles etc) so that I can use this in my professional research.

I already have some experience constructing and calibrating tests, and don't see this aspect as a problem. The calibration is tricky, though, as one must choose to correlate with something - either MBTI itself, or an independent evaluation. In other words, we might need another Temperament theory-literate psychologist to help with calibration unless we can come up with a reasonable compromise. There may be an argument that any personality instrument is its own point of reference, in which case there may be other legitimate approaches.

Happy to discuss this with you further once I've finished chewing through all the Temperament Theory work on the blog (which is a preface for the next round of research I'm planning anyway). It should only take a few months, I think, and I would like to have everything straight in my head before proceeding.

We should get together and discuss this at some point. I did email you about this weekend, but I guess you didn't get the email in time. :( No matter - some other time, no doubt.

Best wishes, and look forward to talking to you more about this at some future point!

Addendum: When I said 'four dimensional personality space', I actually meant 'eight dimensional personality space'. Each axis is best understood as two independent factors, not as a single measurement, since people may express the contrary elements strongly or weakly in an independent fashion.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Your Information

(Name is required. Email address will not be displayed with the comment.)