▷ RE: Homestuck Title Test
I see the test has been circulating again recently, so I’d like to reiterate that it is going to change very very soon! Maybe if I stick this on the front page of my blog, people will actually pay attention…?
Items that are slated for the revision, in order of likeliness-to-actually-happen, and why:
A new system for calculating your Aspect. There are three reasons for this: firstly, as the test is, there’s quite a bias towards Life/Doom and Mind/Heart, and that’s clearly not good. The test should assign Aspects about equally. Secondly, the math is complicated and not very well explained on my part. This is an important reason! If people can’t actually complete the test, then clearly I’ve done something wrong and it needs to change. And thirdly, the first time around, my qualities for determining which Aspect to choose were not as they should have been. I’ve been working with BYB to improve the way I select the Aspect (so no more Heart for all strong introverts, Life for anyone who is even remotely curious by nature, etc.)Already addressed; see below.
- A revamp of the Triarchic Intelligence Test. As of the time of posting, 28% of everyone who’s taken the test and responded to my Google form is a Sylph. There are 12 classes, so clearly this can’t be right! It seems that in general, the test skews creative and nearly nobody ends up practical. This is the only test that I have full control over, so I can attempt to balance it some.
- The selection of a different cognitive function test. This might not occur, as these tests aren’t exactly plentiful on the internet, and I would really rather not direct people to a traditional Myers-Briggs assessment, as that kind of defeats the purpose of looking specifically at cognitive functions. However, this test gives a lot of ties judging by the amount of questions I’ve seen regarding it, and while the number of people diagnosed as Ne and Si are about as expected, there are far too many people diagnosed as Ni and not nearly enough Se.
Aspect calculation: The Aspect portion is almost done, actually! I’ve a system set up that balances scores based on which tend to skew higher, which is determined using the ratio of the average score of a given subscore to the ratio of the average of the subscores in every category. Each new score calculated is another that’s added to the database, so the tool will essentially autobalance as more people take the test. Right now, I’ve nearly removed Aspect bias, and I just need to set up a prettier automated version that people can use without mucking about in my spreadsheets. In addition, the traits that determine Aspects have been augmented as per BYB’s ideas. The determination of an Aspect is now much more complex, but simultaneously much easier, because all you’ll have to do is input your scores and let it spit your title out!This is finished and implemented! Check out this post if you want details.
- Triarchic Intelligence test: As for the Triarchic test, this may take a while. I plan to add more questions, firstly, but I’ll need to do some more research before that happens. Secondly, I may adjust so that not every question is worth the same amount, especially for the ones that most people seem to answer ‘yes’ to. This means that things like the Creative intelligence question about listening to music would be worth less than the one about drawing, because many people listen to music but not as many draw, and drawing would be a greater indication of creativity than listening to music. (just an example, though, no guarantees. I’m still thinking this one through.)
- Jungian Cognitive Function test: The thing about Jungian functions and Homestuck is this: to me, it seems pretty likely that the type of person who enjoys Homestuck is fairly likely to also be the type of person who scores Ni/Ne on this sort of test! I don’t know how much ‘balancing’ can be done without outright changing people’s answers, and this isn’t the sort of change I’d feel comfortable instituting. So if I can find a new test that proves that the Ni/Se problem is due to the construction of the test rather than a bias in the sample set, this will change, but otherwise I probably won’t touch it.
- Miscellaneous: Also, I am probably going to include within the test itself clearer explanations of Classes and Aspects, or at least more obvious links to BYB’s discussions of the same. A little thing, but something that many people have requested.