I am not a Harry Potter fan but I do like the idea of sorting people by personality. If it’s not MBTI, then it’s Hogwarts houses and sorting hats. I looked up Hogwarts houses on Wikipedia and found the following descriptions.
- Gryffindor values courage, chivalry and boldness.
- Hufflepuff values hard work, loyalty, determination, patience, friendship and fair play.
- Ravenclaw values intelligence, knowledge and wit.
- Slytherin values ambition, cunning and resourcefulness.
On reading this, I thought, ‘Hmm. Hufflepuff, I guess. Maybe Ravenclaw.’ So I searched for ‘what harry potter house quiz‘, did the first test and got these results.
Hufflepuff – 14
Ravenclaw – 12
Gryffindor – 11
Slytherin – 8
Sheesh… this sorting hat quiz is intense! There’s a consent form.
I just did that test. It was very long but at least the questions weren’t leading. I got:
Ravenclaw – 88
Hufflepuff – 85
Gryffindor – 81
Slytherin – 41
Hufflepuff – 16
Ravenclaw – 12
Gryffindor – 8
Slytherin – 6
i guess mine almost similar to yours!
and the next quiz, i had
hufflepuff 77
ravenclaw 62
gryffindor 69
slytherin 60
pretty much matches the previous quiz results. haha
I know it’s just for fun… but what good are tests that give results for categories where people score so similarly on the different categories??!
It got me all wrong. Apparently, it thinks that I am “Intelligent, witty and capable!”
Ravenclaw – 12
Hufflepuff – 11
Slytherin – 10
Gryffindor – 7
Who is this Harry Potter guy anyway?
Good question vera.
A valid test in many situations can assign similar scores to multiple categories. In fact, valid tests should be able to make such assignments, which represent cases that are borderline or indeterminate.
Of course, this is all dependent on context. You might imagine a series of ordered categories which represent degrees of magnitude, say the categories “happy”, “neutral”, “angry”, “very angry”, “ballistic”. In that case, if similar scores are given to both “happy” and “ballistic” you would probably be suspicious.
In this case, suppose a highly courageous person that is also very knowledgable (and where these are the most prominent characteristics) takes the test. You would expect high and similar scores for both Gryffindor and Hufflepuff.
The problem with interpreting these test results is not in the fact that scores are similar in some categories. Rather, it is due to the fact taht we have no sense of scale or separation. What does ‘Ravenclaw – 12’ mean? Is that a high score? Suppose it is the highest, is it higher enough than the other scores to be meaningful? Is a difference of 5 in the max and min scores significant? How would persons X and Y score on these tests?
Sure, this is all fun and probably doesn’t warrant deeper analysis, but I think it is worth reflecting on the nature of the weaknesses of these tests because they teach you about the interpretation and use of statistics in more realistic situations.
Apparently, anything but Slytherin…
Gryffindor 85
Ravenclaw 83
Hufflepuff 83
Slytherin 37
Hi Beldar,
Yes, I understand that it’s reasonable to have borderline cases. But I think when a large number of people (not that we have a big sample size here) score similar scores for different categories, you have to re-examine the test.
The thing that concerns me most is — why group the traits the way they are grouped? I think these kinds of groupings can be all too easily misinterpreted. Say you are highly scored for Hufflepuff — what does it actually say about you? Does mean you value hard work, loyalty, determination, patience, friendship and fair play? Or maybe you just value hard work, loyalty, determination, friendship and fair play, but don’t give a damn about patience?
Bring other categories in, and it complicates things even more simply by throwing in a few more variables.
And you’re right, it’s all fun and games until someone actually tries to write a real test, or use the results from a real test.
(I’m a big skeptic when it comes to psychometrics.)
I agree, the categories in this test are so broad and inclusive that I would expect most people to score similarly for multiple categories.
If you want to do this properly, you should have scores for each individual characteristic of interest. I.e. one score for loyalty, one for patience, one for friendship, etc.
As for real tests, the MBTI seems to be quite popular around the world and is apparently quite useful. It has 16 categories, which is quite a lot!
Hahaha, I didn’t even think about the sorting algorithm. Like the engineer that I am, I went directly to the application. Let me say that the application (having the similar people live and play together) concerns me.
According to Wikipedia, ‘There is no rule against students from different houses mingling, but in practice, a good majority of social interactions occur within the same house.’
Now, this is a worry. If you’re in the Ravenclaw house, you’re likely to grow up with smart, wise and witty people. You could pretty screwed up when you go out into the real world. Similarly, a Hufflepuffian will never learn to deal with the sneaky and underhanded. Gryffindorians will be egged on to boldness, without seeing the example of the meek. And you can imagine what a hothouse of intrigue and backstabbing the Slytherin house would be!
If Hogwarts wants well adjusted, tolerant wizards and witches who have learned to work together, the house sorting system would be better to mix up the personalities!
By the way, the MBTI differs from most psychometric tests because it sorts on preferences (dichotomies), rather than by trait. You can’t have ‘too much’ or ‘too little’ of something. Rather, it says, ‘You prefer to act in this way.’
I did this test a long time ago so I can’t remember the scores for Hufflepuff or Ravenclaw. I confidently expected to get a score placing me squarely in the Gryffindor camp.
I’m a Slytherin.
Is there a distinction between preferences and traits? If a preference is consistent, surely you can describe it as a trait?
Most personality metrics use Trait theory. You end up with a score to say how ‘orderly’, ‘open’, ‘trusting’ and ‘altruistic’ you are. I believe that the scores are distributed normally. You can have ‘average’ self discipline, and so on. Therefore, there is some implication of having ‘too much’ or ‘too little’ aggression, modesty, compliance…
Preferences are distributed bimodally. You can have one or the other. You’re always on one side or the other. When Myers and Briggs came up with this, they were mocked by the psychology community for having variables with bimodal distributions, not understanding that MBTI doesn’t use traits.
There is a score associated with preferences. The score is a measure of consistency of preference, which is different to strength of preference (which MBTI does not measure) or skill.
The difference between consistency and strength of preference:
If you scored 21-0 on the extrovert-introvert scale, that means you pretty much consistently are extroverted in a range of situations. This means you are energised by most social interactions.
It doesn’t say how much you’re energised, or how loud or open or confident you are.
If you score 11-11 on the judging-perceiving scale, it means that from situation to situation, you can switch between judging (making decisions) to perceiving (being spontaneous, avoiding decisions) quite easily. You could be an very structured pedantic organiser/decision-maker, even if you have no particular preference for it. An even score just means that you don’t consistently pull out the same tool from the toolbox.
This is in fact an advantageous personality type. Jung was all for people developing over their lifetimes so that they could exercise their non-preferred modes. Makes you more well rounded.
Correct me if I’m wrong, but I think what Beldar means is — what is the difference between a preference to being orderly, and having the trait of orderliness.
I think (there’s no evidence behind this, apart from the vague recollections of the many personality lectures I’ve sat through over the last few years haha) that preferences are seen to be “symptoms” of traits. What trait theory is trying to do is to try and find a sort of “DNA” for personality, trying to find the fundamental building blocks. I think preferences are kind of like the “phenotype”.
I think whenever you’re measuring anything, you’re really measuring a preference of sorts. A trait really is just a construct, not something observable. So, it’s like, you could measure some behaviour (and if it’s a frequent behaviour, a preference), and extrapolate what kind of traits would lead to that behaviour.
What’s interesting though is that a lot of the major personality theories today don’t actually have much… well, theory. As you can imagine, you could easily identify thousands of “traits” to choose from to explain any behaviour, but for a model to be really effective, you’d try to filter it down to a few basic ones. How do you do this? One of the popular ways these days of doing that is just to use factor analysis!
So… in a lot of personality theories, there’s not much reasoning — just a bunch of statistical correlates haha… 😐
(There have been interesting studies on different traits identified in cultures with different languages, by the way.)
By the way: I don’t believe that trait theory implies you can have “too much or too little” of something, or at least not in the judgemental sense (which I think you’re hinting at?). I think that’s kind of like saying a (continuous) height measurement says that someone is too tall or too short, rather than taller than average or shorter than average…
I only have experience in MBTI and the theory behind seems pretty comprehensive and self-contained. It’s based on Jung’s theory of psychological types.
I don’t believe Jung theorists would agree that preferences are the result of traits. If anything, I would call his ‘functions’ and ‘attitudes’ the DNA of personality and traits the emergent properties.
Jung proposed only three variables, each with two preferences, for a total of eight personality types. Myers and Briggs added a fourth variable, taking it up to sixteen types.
The combination of four functions and attitudes interact to produce a whole personality (or the possiblities of a personality).
The two attitudes are: how one is energised and one’s preferred mode of interacting with the external world (by making decisions or collecting information).
The two functions are (surprise, surprise) how one makes decisions and how one collects information.
I haven’t studied trait theory except in contrast to MBTI but it seems quite fundamentally different to the MBTI, which is based on preferences for attitude and function.
The theory was developed and MBTI refined based on statistics. There is strong statistical backing but the statistical evidence came after the theory.
I think I see what you’re getting at, actually. The question you ask might be ‘Why do you like making decisions, rather than collecting information?’ or ‘Why are you energised by social interactions, rather than drained by them?’
It has never occurred to me to ask ‘why’. People just are. People are different. Erm. Genetics?
By the way, I don’t regard MBTI as ‘truth’ or having a monopoly on validity. I’m not bothered about if it’s true true or not. To me, it’s useful, a structured way to understand yourself, your strengths and weaknesses, how to assemble a team, how to communicate with people and defuse (or promote!) conflict. Systematic analysis in whatever framework leads to insights.
Based on this ‘application’ worldview, all personality theories are valid if they have something that helps you function effectively. If there is something more effective than MBTI, I’ll use that too (not ‘instead’).
There are criticisms of MBTI. Ah, there they go again, fixated on the unconventional bimodal distribution again. I’ve already explained that one. The criticism on ‘utility’ has been made by someone who doesn’t understand the purpose of MBTI. You’re not meant to use MBTI to select people for jobs, it’s unethical first and ineffective second.
The question of reliability… I’ve heard conflicting statistics on this.
Well, Jungian theory more belongs to the “old” theories, I think — it’s the newer theories that are more statistically based. I think it’s a backlash from the old theories being NOT statistical — most of the old theories were all theory and no empirical study (I would say Jung with his psychoanalytic background falls into that category).
Both extremes are bad, but it’ll probably even out sooner or later — psychology as a science (if you can call it that at the moment… hahaha) is a pretty new field — in my opinion, it’s only just beginning to find its feet… and still has a long way to go. I think one of the problems is psychology (in its non-empirical form) has such a long history that it’s really hard for people to break away from that. So… even though now researchers realise the need to have empirical evidence, they still build on all the non-empirical theory that’s come before… and it’s all just a mess.
And I know what you’re saying about MBTI just being a tool for you, but I mean, personality psychology as a whole is NOT just a tool — as a science it SHOULD be concerned with the “truth”.
Yes vera, that was pretty much what I was trying to ask. It wasn’t clear to me what was meant by ‘trait’ and ‘preference’ and the distinction between the two.
I like your analogy with genotype and phenotype. It’s not quite that clear, thought, since it seems to me that even traits are described in a behavioural sense, which makes them more like phenotypes. Maybe some phenotypes are thought to be more fundamental than others? Maybe there are multiple levels of types, each serving as a base for the levels above them?
Regarding the use of ‘bimodal’: in probability and statistics this refers to a distribution with two maxima (in contrast to, say, a normal distribution which is unimodal). If you describe a classification as bimodal, you might be interepreted as saying that most things lie in one of two places, which doesn’t seem to be what you meant. Maybe ‘bimodal’ has a different meaning in psychology? For example, you don’t seem to claim that most people are either mostly introverted or mostly extroverted. (Although you say that lots of statistics has been collected for the MBTI, maybe it is true that their four measures are statistically bimodal?)
I think it’s a bit airy-fairy, which is why it isn’t really clear. I think the idea is sort of like, you can’t measure “orderliness” as a concept, you can only measure the behaviours which reflect this concept. Yeeeeah…
But yes, I think in most personality models there are base traits, and secondary traits, etc. I think in the five factor models there are actually something like 60 traits altogether, which is a bit ridiculous, really… but hey, if that’s what factor analysis of dictionary adjectives spat out, well it must be true! hahaha
Yes, what’s the deal with this bimodal distribution? Why’s that expected?? Although, I had a quick glance at the criticism, and it appears to say that introversion-extraversion is a SUB-scale, and that this is normally distributed (which make sense). I’m not sure which score is bimodally distributed, and why that would be expected…
(But yes beldar, bimodal in psychology is definitely referring to the usual statistical meaning.)
Okay, now you’re wandering into territory that I have not studied.
“f you describe a classification as bimodal, you might be interepreted as saying that most things lie in one of two places, which doesn’t seem to be what you meant… For example, you don’t seem to claim that most people are either mostly introverted or mostly extroverted.”
I believe this is exactly the claim. Most people have a preferred mode of operation. Some people record ‘slight’ preferences but people are generally identifiable as extrovert/introvert, sensing/intuitive, and so on.
I have read studies disputing the statistical base for this. I have read stuides that support it as well. I don’t know where the current debate is.
Why is a bimodal distribution expected? You’ll probably have to go back to the theory base. I understand that people have preferences for one or the other, that it’s not typical for people to sit exactly in the middle.
Remember, the distributions are of consistency of preference, not strength of preference. A bimodal distribution just says that people who prefer extroversion are mostly consistent in their preference from situation to situation. If the distribution was normal, you would be saying that many people are comfortable switching from extroversion to introversion in a given situation.
“…I had a quick glance at the criticism, and it appears to say that introversion-extraversion is a SUB-scale, and that this is normally distributed (which make sense).”
The extrovert-introvert scale has, I think, five subscales. This is the next level of accreditation, which I don’t have. I have only worked with the primary scales, which supposedly have a bimodal distribution. The Wikipedia article is saying that the subscales elements are normally distributed, which makes it unclear as to why the primary scales are bimodally distributed. As I say, I don’t know the theory.
I do have a lot of material on it but it wasn’t on the exam so I didn’t study it 🙂
What?! I just discovered that the response format of an MBTI test is forced-choice dichotomous response! That might explain a lot about the bimodality. Moo, to satisfy your statistics-ness, I have some references for you.
I’m going to continue this over email.