Results 1 to 10 of 10

Thread: interrater reliability and socionics

  1. #1
    implied's Avatar
    Join Date
    Aug 2005
    Posts
    7,750
    Mentioned
    7 Post(s)
    Tagged
    0 Thread(s)

    Default interrater reliability and socionics

    let's discuss this phenomenon.
    6w5 sx
    model Φ: -+0
    sloan - rcuei

  2. #2
    expired Lotus's Avatar
    Join Date
    Jun 2007
    TIM
    Se/Ni sx/sp
    Posts
    4,494
    Mentioned
    99 Post(s)
    Tagged
    0 Thread(s)

    Default

    ahem. what does this mean
    maybe a saint is just a dead prick with a good publicist
    maybe tommorow's statues are insecure without their foes
    go ask the frog what the scorpion knows

  3. #3
    ILE - ENTp 1981slater's Avatar
    Join Date
    Aug 2005
    Location
    Spain
    TIM
    ILE (ENTp)
    Posts
    4,866
    Mentioned
    15 Post(s)
    Tagged
    0 Thread(s)

    Default

    Wikipedia:

    "...for example by determining if a particular scale is appropriate for measuring a particular variable. If various raters do not agree, either the scale is defective or the raters need to be re-trained."

    Since x experienced user and y experienced user do not agree on their typings:

    a) one of them, at least, is an idiot
    b) Socionics is crap
    ILE "Searcher"
    Socionics: ENTp
    DCNH: Dominant --> perhaps Normalizing
    Enneagram: 7w6 "Enthusiast"
    MBTI: ENTJ "Field Marshall" or ENTP "Inventor"
    Astrological sign: Aquarius

    To learn, read. To know, write. To master, teach.

  4. #4
    Hello...? somavision's Avatar
    Join Date
    Sep 2008
    Location
    London
    Posts
    1,474
    Mentioned
    8 Post(s)
    Tagged
    0 Thread(s)

    Default

    You really need the raters to initially do the rating independently or in seperate groups to avoid groupthink.

    When I've done Cohen's Kappa, we'd start with a discussion to try to clarify what it is we're looking for, followed by independent coding and a discussion of the results to find out why we differed,

    I'd imagine that across the board reliability on this forum would be pretty iffy, mainly due to the omplexities of inferring individual types from, complex human behaviour expressed in written text.

    However it may perhaps be interesting to post a high queality video of someone, who is unaware of socionics but is at the same time giving rich and socionically valuable information. The types could then submit privately their opinions on the individuals type/ and or the varius dichotomies/groups. This information could be analysed statistically to measure reliability and repeated with different individuals.


    It woud also perhaps be interesting to give the participant a test after the intervoew or allow them to select a type which they think suits them the most, to measure the correlation, between this and the results of the coders.

  5. #5
    Punk
    Join Date
    Feb 2010
    TIM
    ESE
    Posts
    1,659
    Mentioned
    2 Post(s)
    Tagged
    0 Thread(s)

    Default

    Quote Originally Posted by 1981slater View Post
    Since x experienced user and y experienced user do not agree on their typings:

    a) one of them, at least, is an idiot
    b) Socionics is crap
    you know what's funny...

    if b happens to be true, then we are all 100% idiots for choosing to play.
    if a happens to be true, then we are only an idiot if we are wrong.

    The relative probability that we are idiots is 3/4 or 75% for choosing to play socionics. Now start to add in facets of meta-cognition that leads to philosophical inquiry and the potential to supersede our propensities in completely unpredictable and new ways; you will find that the relative probability that we are an idiot, for playing socionics, slowly increases from 75% on its convergence path to 100%. Then the relative probability leaves no room for bias error and becomes the absolute probability of 100%. FUN

  6. #6
    ragnar's Avatar
    Join Date
    Aug 2005
    TIM
    ILI
    Posts
    635
    Mentioned
    12 Post(s)
    Tagged
    0 Thread(s)

    Default

    Quote Originally Posted by 1981slater View Post
    Wikipedia:
    "... If various raters do not agree, either the scale is defective or the raters need to be re-trained."
    As long as the condition
    Quote Originally Posted by Kingsley View Post
    You really need the raters to initially do the rating independently ...
    is fulfilled, then the condition that all typers type correctly more than 1/16th of the time is quite obviously sufficient to establish the existence of a 16-way partition that is likely to reflect socionics theory (or something difficult to distinguish from it), provided the sample of typings is large enough and that the errors are random.

    Having my math degree from a public uni and being a sloppy student on top of that, I'm at a loss to formulate the necessary as opposed to some sufficient condition, anyone who can in a concise and simple to understand way is welcome to do so.
    Last edited by ragnar; 11-25-2011 at 12:48 AM. Reason: typo 16 vs 6
    Greetings, ragnar
    ILI knowledge-seeker

  7. #7
    Glorious Member mu4's Avatar
    Join Date
    Aug 2007
    Location
    Mind
    Posts
    7,966
    Mentioned
    568 Post(s)
    Tagged
    3 Thread(s)

    Default

    Quote Originally Posted by 1981slater View Post
    Wikipedia:

    "...for example by determining if a particular scale is appropriate for measuring a particular variable. If various raters do not agree, either the scale is defective or the raters need to be re-trained."

    Since x experienced user and y experienced user do not agree on their typings:

    a) one of them, at least, is an idiot
    b) Socionics is crap
    Alternative explainations.

    c. The subject is hard to analyze due to erratic behavior and good acting ability
    d. lack of experience on the part of rater with certain types leads to a inability to type some people

    I think most people will not have close personal experiences with all 16 types of people and have a indepth knowledge or understanding of these types.

    Aushra herself was not able to write all 16 descriptions and she was someone who had interviewed and worked with hundreds and thousands of people.

    Knowing socionics really well is only one component of typing people accurately, you have to really know people as well, which requires a good deal of socializing.

  8. #8
    Grand Inquisitor Bardia's Avatar
    Join Date
    Aug 2007
    TIM
    ESI
    Posts
    1,258
    Mentioned
    7 Post(s)
    Tagged
    0 Thread(s)

    Default

    I think this thread should be renamed interrater unreliability and socionics.
    “No psychologist should pretend to understand what he does not understand... Only fools and charlatans know everything and understand nothing.” -Anton Chekhov

    http://kevan.org/johari?name=Bardia0
    http://kevan.org/nohari?name=Bardia0

  9. #9
    InkStrider's Avatar
    Join Date
    Aug 2011
    Posts
    419
    Mentioned
    1 Post(s)
    Tagged
    0 Thread(s)

    Default

    Quote Originally Posted by hkkmr View Post
    Alternative explanations.

    c. The subject is hard to analyze due to erratic behavior and good acting ability
    d. lack of experience on the part of rater with certain types leads to a inability to type some people

    I think most people will not have close personal experiences with all 16 types of people and have a indepth knowledge or understanding of these types.

    Aushra herself was not able to write all 16 descriptions and she was someone who had interviewed and worked with hundreds and thousands of people.

    Knowing socionics really well is only one component of typing people accurately, you have to really know people as well, which requires a good deal of socializing.
    I couldn't agree with you more. Having close interpersonal relationships with all 16 types (or 32 subtypes, since many have noticed good differences between the two subtypes), is I think, essential in order to be able to discern and pick out the patterns/indicators that matter in a particular type from traits that do not.

    Socionics theory isn't difficult to grasp, nor is it difficult to type *in theory*. But to know people as who they truly are rather than who we simply think them to be requires an entirely different skillset altogether, which oft times has a direct relation to the degree and depth of socializing we engage in. Many a time, it is not lack of understanding of theory that is the cause of mistypings (for the more seasoned typers), but a lack of understanding of the subject himself, coupled with lack of close experience with the type/subtype in question (which means that the risk of attributing the wrong type indicators to observed behavior is high).

  10. #10

    Join Date
    Jun 2011
    TIM
    LIE-Se 8w9
    Posts
    15
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)

    Default

    Typology is sound, Socionics is sound as a framework, Socionics does not have sound content.

    Anyone who tries to base their typings directly on this content is thus going to get it wrong, and if they are not even looking at a real phenomenon in the first place, there is no basis for reliability. As a framework it's not bad, but it needs a lot of work (which I currently have in the works).

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •