Meta MBTI thread: Things to Consider

[Xander]

You would think so... Ti is logic applied inside to understand and Te is logic applied outside to problem solve. Just like Fi is self values and Fe is shared social values. But that's just what they are on the surface. Ti is not the same as Te. INTP will examine an idea in depth for understanding, the Te dominant ENTJ ought to explore all external options to understand each one. They do that, but only to the degree necessary to making a decision. They're not equal mirrors. Or is that J/P getting in the way? But then again what is J/P? The link between Yx and J/P is inherent but their relationship isn't obvious.

J/P is what you do when you extravert is it not? Js organise, define and generally make sense of things where as Ps flex, problem solve and flake ( ).

Taking the Ti vs Te split.
If Ti is not problem solving then why are ENTPs and INTPs good problem solvers??
If Te is not based upon complete understanding then why is it that when you try and persuade an INTJ/ENTJs point of view or enlighten them to something you think they haven't seen do you always find that they've considered everything (usually to a limit of understanding gauged more on their individual intelligence and level of care taken)?

It always seems that the splits in the functions aren't well defined enough to really be of use. Maybe I've not found the pattern yet, I'm open to that.

 

[mlittrell]

What about the checking procedure where they basically ask the same question again later on to try and verify that you are answering honestly. I think it has something to do with your opinion may change as you progress through the testing procedure and it's a deliberate lead. If you start limiting the person's type too early then how would you incorporate such a checking feature or do you think it's a bit of a waste of time?

I would consider a checking procedure but really if done right, I wouldn't need one. But thanks for brining that up.

Kinda like a workflow / flowchart?
I like it! Do I count as "people"?
Do it! Let me know if I can help

Yes you count as people lol glad you like the idea. its going be a b*tch for me to do because of my schedule but ill try to spit it out in the next couple of months. (Christmas vacation maybe...?). and sure if you wanna help ill keep in contact.

 

[Xander]

I would consider a checking procedure but really if done right, I wouldn't need one. But thanks for brining that up.

That told me.

 

[mlittrell]

That told me.

DId I miss something lol?

Another mlittrell retard moment?

 

[Xander]

DId I miss something lol?

Another mlittrell retard moment?

I read it as snippy... it just made me laugh. Kinda like you were looking down your nose going "well I'm going to do it right so I don't need to but it's soo nice that you are thinking so hard about it" whilst patting me on the head.

Sounded funny to me

Seriously though if the subject is supposed to be the foremost authority on their own type then would it not be contrary to this to attempt to build a test which was prescriptive?

Don't get me wrong I'd love it if it was done and worked... I'm just a little doubtful of the premise.

 

[mlittrell]

I read it as snippy... it just made me laugh. Kinda like you were looking down your nose going "well I'm going to do it right so I don't need to but it's soo nice that you are thinking so hard about it" whilst patting me on the head.

Sounded funny to me

Seriously though if the subject is supposed to be the foremost authority on their own type then would it not be contrary to this to attempt to build a test which was prescriptive?

Don't get me wrong I'd love it if it was done and worked... I'm just a little doubtful of the premise.

yaaa i didn't mean to sound snippy lol my bad. and yes a test which is prescriptive would be nice. it would be nice if a test could be made with a top down type of design though. hey ill keep any ideas you have in mind and im not just saying that to shut you up, im seriously down for ideas...

i started outlining the program a little btw

 

[Xander]

yaaa i didn't mean to sound snippy lol my bad. and yes a test which is prescriptive would be nice. it would be nice if a test could be made with a top down type of design though. hey ill keep any ideas you have in mind and im not just saying that to shut you up, im seriously down for ideas...

i started outlining the program a little btw

I think I've got a good suggestion....

Reference this double check idea, how about if the program takes you through and test you using different methods? So one tries to establish your temperament and goes down that route, perhaps another checks for function usage and another tries a more direct approach. That way you could then perhaps average out the results or do a more complex comparison before giving the results of the analysis to the subject.

I'm figuring that if the subject is tested more than once within one test then an average will hopefully reduce the impact of any "outliers" in their personality.

 

[mlittrell]

I think I've got a good suggestion....

Reference this double check idea, how about if the program takes you through and test you using different methods? So one tries to establish your temperament and goes down that route, perhaps another checks for function usage and another tries a more direct approach. That way you could then perhaps average out the results or do a more complex comparison before giving the results of the analysis to the subject.

I'm figuring that if the subject is tested more than once within one test then an average will hopefully reduce the impact of any "outliers" in their personality.

that is a good idea. the only problem is i wont do a function test mostly because i push function orientation so much. i figure it would look like this (for an ENTP)(nasty pseudocode):

test to determine N/S

if intuitive
then
/*test to determine NT/NF*/
ifelse intuitive = sensing /*if test results are even*/
then
/*Ask a previous question and then a new question. if previous question remains the same then ignore and ask new question to fully determine N/S. if previous question changes then change it and still go onto the next question*/
else
blah blah blah

and do this for NT/NF, then NTP/NTJ, then E/I

i was considering writing it in ksh just to test it out. im quite a bit better in ksh then C++ or java for that matter. though i can do it. its kinda a top down design i spose... what do you think?

EDIT:
thats an extremely rough outline, i a chunk of it layed out already

 

[Xander]

^^It'll be interesting to see how that lot turns out.

You plan on testing it here I presume. That'd be kinda cool.

 

[mlittrell]

yes it will be tested here, problem is i dont think many people will like it being a .exe for security reasons so i wanna figure out a better way of doing it for that reason

 

[Jack Flak]

mlittrell, I've thought about test structure like that, and I think parallel redundancy has less chance of going wrong than linear selection, unless your test is perfect, which none have been yet. I guess it depends on your goal, whether it's ULTIMATE ACCURACY or OK accuracy and "not taking forever."

 

[mlittrell]

well the current tests dont make sense to me at all. why go for each individual part? that doesn't make sense considering they all effect each other. a person is either an N or an S. they are either an NT or an NF. they are either an NTP or an NTJ and either an extrovert or introvert. i thought of this when i first took the test before i really knew anything lol. it just didn't make sense when i was taking the test why they did it the way they did it, so ive been looking into doing this for quite a long time. its quite an easy thing to program too. do you have any ideas? do you like the idea? my goal would be ultimate accuracy btw. when it comes to type there is no "OK accuracy" it needs to be quick and accurate. i know three people who are INFJs that where convinced that they were ISFJs until they read keirseys profiles. that kinda stuff bothers me for reasons im not going to go into. i figure this is how i type in real life so why not apply it to a test? any ideas are welcome

 

[entropie]

If you need programming help, let me know, I do speak different languages fluently .

Can even try to make an ASP website and program it in c#. I've got some free space for that.

 

[mlittrell]

If you need programming help, let me know, I do speak different languages fluently .

Can even try to make an ASP website and program it in c#. I've got some free space for that.

really? hmmmm well then that might come in handy if we wanna post it on mbtic. i know a little asp but if you know it fluently...

just give me time to write the program, i have 3 weeks left in school so im insanely busy.

thanks for all the suggested help people it is much appreciated

 

[Jack Flak]

well the current tests dont make sense to me at all. why go for each individual part? that doesn't make sense considering they all effect each other.

But you know people screw test-taking up all the time. The redundancy of testing several categories separately allows you to have a null result like "You fucked up, try again."

 

[mlittrell]

But you know people screw test-taking up all the time. The redundancy of testing several categories separately allows you to have a null result like "You fucked up, try again."

well if people are too stupid to take a normal test and i send them a response like that imagine how confused they will be then lol. good point nonetheless. that is why i need to make it stupid proof. the only problem i have with separate parts is its too similar to the other tests. if you want i could do it both ways and see which one is better...

what do you guys think about jacks proposal?

 

[Jack Flak]

what do you guys think about jacks proposal?

It'll take longer to take it, that's for sure.

 

[INTJMom]

...
what do you guys think about jacks proposal?

Which part?
The parallel redundancy?
What does that mean exactly?
Is that where you choose from a comparison of words which one best suits you?

I think MBTI is all about "compared to who" and because they use words that mean something different in another world, it's important to place them next to words that show what they are not.

 

[Jack Flak]

Which part?
The parallel redundancy?
What does that mean exactly?
Is that where you choose from a comparison of words which one best suits you?

No. How could it mean that?

What I had in mind, and I think ml understood, is effectively more than one test side by side. For example: One for N/S, one for NT/NF/SP/SJ, and one for whatever else you want.

 

[entropie]

I see the goal lies in redundancy aswell. You can have a test that clearly asks for N or F, because the questions could be easily made. And then again you need more question asking concretly for NT or even a whole remarkable personality trait.

The only new problem that could bring is, if people take the test who are fairly biased in their answering. Maybe thinking that there answer will lead to a T outcome they choosed their answer. There is the need then to prevent that, by a lot of good an cryptic questions.

Maybe that is the hardest part in developing the test. Should probably make a new thread and search for candidates for the questions