NJTI-4 feedback

Thread posts

See all posts in this thread below.

  • User
  • Message
  • Actions
    • Published: 03-04-2016 05:01 am
      Updated: 03-04-2016 05:13 am
    • Ok, so almost 500 have done the new test now - which is great.

      The test confirms the statistics from the Swedish test that the dichotomies are real and valid.

      But there have been two general complaints:

      1. Questions are too similar
      2. The use of a Likert scale (normative) instead of a ipsative test (bimodal)

      So to answer these complaints:

      1. Yes, some questions are too similar. We are working on this. We agree that some question, not all, are too similar.

      2. We want to use a Likert scale because it is much more reliable as a tool of producing valid science. Ipsative tests can enforce false dichotomies and for that reason it's better to use a likert scale. See this.

      It is intuitive that if you independently rate that you like A higher than B - it would equal you would prefer A over B.

      But if you only prefer A over B, not rating them indepedently, it doesn't mean that A and B are dichotomous in reality and it excludes that you could like them equally, it could just be a feature of the test to present A and B as dichotomous even though they are not.

      Since MBTI has received criticism for propagating false personality dichotomies we consider it important to not fall into the trap of false assumed dichotomies. Therefore a Likert scale is a better option. Using the Likert scale we can answer the question whether our dichotomies are real or artificial.
    • ErikThor likes this post.
    • Published: 06-19-2016 04:38 am
    • A major criticism to the test at the moment is the redudancy of asking the same question multiple times. How are we doing with that?
    • Published: 06-19-2016 04:56 am
    • I'm not sure it's an real issue other than it's boring.
      The questions about the same trait are not identical but still asks about the same latent factor. Since people respond incoherently it improves validity and is standard procedure.

      It is recommended to use at least 4 questions about the same latent factor to get a confident mesaure.

      If you compare established tests it's pretty much the same for them so I don't think this is an issue any more.
      http://ipip.ori.org/newNEOKey.htm#Neuroticism
    • Published: 06-19-2016 12:44 pm
    • Yeah, if we need it for precision and accuracy, you might be right that it's best to keep them. I usually give that explanation to people but alot of people seem to be annoyed by it.

      At some point we might wanna consider removing the control questions because alot of people mistake them for being stupid questions. We only need the control questions for ensuring statistics are healthy.

      Here's a suggestion for how to change the note above the personality test: Basically shorten it alot and add these three symbols.

      https://docs.google.com/document/d/16S3ytZ9-9TvSkTcHAMPzq2RHJkQ5D0KsPQr3kkeij3o/edit
    • Published: 06-20-2016 01:51 am
    • Alright I will fix this
    • Published: 06-20-2016 01:57 am
    • I think this document makes some copyright infringement by referring to the test as a MBTI-test. CAPT has been after me already.

      We can change it to which I think is acceptable

      • The first ever Myers Briggs-inspired test based on science
      • Most accurate and scientific MBTI-inspired test on the internet.

    • ErikThor likes this post.
    • Published: 06-20-2016 04:08 am
    • That works, I was actually wondering about that. But what about the images? I think icons8 is a great service, for finding and using good images and icons when necessary. The three icons in the page are from icons8. But if we use them we should credit Icons8 at the end of the page. They use creative commons and all their icons are free to use.

      https://icons8.com/license/
    • Published: 06-20-2016 07:13 am
    • That change should be implemented now. I added a license link to icons8 as well.
  • User
  • Message
  • Actions

Related

You might be interested in.