making a better test

[Society]

based on your understanding of the MBTI and experience with past tests:

what is your criticism of current tests?
in what ways can they be improved?
how would you design them differently?

 

[burymecloser]

what is your criticism of current tests?

1. Self-reporting tests are deeply unreliable. You're telling the test about yourself rather than the other way around.
2. Most tests make some answers sound better than others (especially N > S), so people may -- consciously or otherwise -- answer inaccurately.
3. Most tests are suited to Keirsey's sliding scales rather than to Jungian functions. The tests measure behaviour -- something that can change from day to day -- rather than static preferences which give us more meaningful information about personality.

in what ways can they be improved?

1. ... the key here is probably just more reliable interpretation of data. Ideally, test-takers would understand their results as a guide or a starting point rather than as the final word.
2. Neutrality. The language used to describe Sensors should rely more on input from Sensors, and ideally, some sort of detailed data would be used to insure that results tend to be accurate and lay people could consistently expect such results.
(Not always, just consistently. And not perfect, just accurate.)
3. Questions aimed at discovering people's functions and preferences rather than just their behaviour.

how would you design them differently?

See above. I'd also like to see specific examples and clarity in word choice so vocabulary issues and/or unclear questions and answer choices don't distort individual results.

 

[KDude]

The main problem is that most are based on dichotomies. I don't have to point out how simplistic that is.

I don't have much of a problem with Nardi's cognitive test (although I liked his first version that isn't available now.. it followed Beren's definitions more closely).

If you were to make a dichotomy based test, I think it'd be better served with "narrative" questions.. situational problems. Illustrate the functions in specific ways, rather than in the abstract (which many people will read into too often).

 

[WALMART]

I have some thoughts.


Remind me in Vent to share them.

 

[KDude]

Oh, another thing worth testing is "functions as blocks". Too often functions are isolated, as if we're in "Ti mode", then "Se mode", when it's a whole block of how one experiences things. There's a reason why ENFJ doesn't resemble ESFJ exactly or ISTP isn't merely some "plain and dry" version of INTP. Functions are expressed differently, even among two types that share it as a dominant. Lenore Thomson's book is an excellent source of this typing theory (as is Myers herself). ISTPs don't rely on abstracted "logical" approaches like INTPs. Their logic is used emergently, what she calls tactile or "body logic".

 

[A window to the soul]

in what ways can they be improved?
how would you design them differently?

For the free tests, I would improve the education that's available to the test takers and put in a wizard format to guide them through the training and test taking process. I don't remember the website's address, but I took the official (paid) MBTI test once and was required to go through a thorough training prior to the test. That was different and helpful.