If the NFP scores are at the top, and the STJ scores are near the bottom, and then there's types with the opposite function also near the top... it seems strange, given that this test was based on cognitive functions. I get that an Ne test item would add more points to ENFP than ISTJ, but still I would expect some kind of correlation.
I guess though that noname scored highly in Ne, Fi, Ti and Ni?
Actually, its because Im not purely using cognitive functions anymore. Im mostly using them. Perhaps I need to put a disclamer on that and count how many questions are not directly related to cognitive functions.
Just pointing what I can remember:
- I originally used the enneagram only to E/I questions but then I expanded the enneagram and I had used an enneagram table to determinate the points. For example, I had put enneagram 5 stuff, and these stuff points a lot for INTJ/INTP.
- There are the pair questions of words where you have 8 options. These questions were taken from truity descriptions at first, as truity has these words from the manual. However, looking at [MENTION=35566]Luminous[/MENTION] results, I had decided to do some little changes and add or remove some words so they are closer to the type, but that question points directly to the types.
- I have imported and adapted some of the questions I considered most convenient and reliable from Sakinorva cognitive function test. This website has stats of the average marking on the questions on agree/disagree from one to five and I did the poiting system to mirror that statistics directly from type to type. I did that to make the test better and because I knew that some of my "crazy" questions were in risk to derail the whole test. Actually, a whole version using this "method" can give good results.
- Also note that Im NOT using Grant Cognitive Function stack, although I am using pairs and pairs play an important role to the result. The system I use is more smooth, so there is not that big switch to INFJ for INFP for example where INFP and INFJ dont share any functions.
- The quiz website forced me to have points for every type on every question, so I added as low weight as possible but on these cases I used dichotomy. But dichotomy on these questions weigheted 12.5-25% on these questions and is less than half of the questions had that problem.
- Its important to note that I didnt have added any dichotomy (except E/I) questions, except in this update where I, due to the high concentration of mistypings in J/P dimension, added 1 question (and it has 4-6 different options) as a last resort and desperation. But I extracted that question from indirect stats, it used Big 5 stuff, not regular wording in J/P.
I said way much that it was needed, but just for the sake of the curiosity.
Perhaps I should consider renaming the test and doing one purely with cognitive functions. Or doing a more regular one with only cognitive functions and no indirect questions from other systems.