4 ms·

I think they used the well-established and well-researched personality methodology called OCEAN. This has been well documented as being able to identify the big

I think they used the well-established and well-researched personality methodology called OCEAN. This has been well documented as being able to identify the big five traits (OPENNESS, CONSCIENTIOUSNESS, EXTRAVERSION, AGREEABLENESS, NEUROTICISM) and its sub traits. This framework was then used by Aleksandr Kogan to profile and harvest facebook users - who then sold it to Cambridge Analytica..etc. By relying somewhat on self-selection Aleksandr's firm was able to profile facebook users. Since there was no second-guessing by the subject of the test as what a "question might say about them" I think it was pretty robust methodology. And then when you look at the impact of his work, you can see it was effective. Not sure there is any snake oil to be found here.

anotherman554 5y ago
I've taken one of those big 5 tests, it's based on self reported answers to survey questions. While I would not call it snake oil, if self reports are not a reliable method of understanding people, it's fancy nonsense. (Like most psychology, it seems it may or may not be B.S.) I googled the Cambridge thing you mentioned and the first article that came up said it wasn't particularly accurate and was mostly just profiling based on demographics. (Shrug).
- dougSF70 5y ago
  Yes that article, sponsored by Cambridge Analytica, does play it down :)
  - anotherman554 5y ago
    There's no such article that came up. There's a long tradition of psychologists making grandiose claims that are short on facts. B.F. skinner claimed "Give me a child and I'll shape him into anything." Not sure why you would assume the Cambridge thing did what someone selling it said it could do.
disgruntledphd2 5y ago
> By relying somewhat on self-selection Aleksandr's firm was able to profile facebook users. Since there was no second-guessing by the subject of the test as what a "question might say about them" I think it was pretty robust methodology. Having worked in adtech, and as a psychologist who's attempted to use OCEAN for said adtech, I never found it to work better than standard approaches. Additionally, Costa and McCrae created the Big 5, found it didn't replicate in confirmatory factor analysis, so they created a new method of CFA which did replicate it. Like, E O and N are probably real, but the overall structure does not replicate and should not be taken very seriously.
ekianjo 5y ago
> This has been well documented as being able to identify the big five traits (OPENNESS, CONSCIENTIOUSNESS, EXTRAVERSION, AGREEABLENESS, NEUROTICISM) and its sub traits. It's just a simplified model to put personalities in a certain bucket, typical PCA-style. This has limitations just like everything else that uses self-reporting to make a judgment/categorization. it's not like we are able to observe the structure of someone's brain and confirm that with actual facts. In other words, that's not even remotely what you call Science.

anotherman554 5y ago

I've taken one of those big 5 tests, it's based on self reported answers to survey questions. While I would not call it snake oil, if self reports are not a reliable method of understanding people, it's fancy nonsense. (Like most psychology, it seems it may or may not be B.S.) I googled the Cambridge thing you mentioned and the first article that came up said it wasn't particularly accurate and was mostly just profiling based on demographics. (Shrug).

dougSF70 5y ago
Yes that article, sponsored by Cambridge Analytica, does play it down :)
- anotherman554 5y ago
  There's no such article that came up. There's a long tradition of psychologists making grandiose claims that are short on facts. B.F. skinner claimed "Give me a child and I'll shape him into anything." Not sure why you would assume the Cambridge thing did what someone selling it said it could do.

disgruntledphd2 5y ago

> By relying somewhat on self-selection Aleksandr's firm was able to profile facebook users. Since there was no second-guessing by the subject of the test as what a "question might say about them" I think it was pretty robust methodology. Having worked in adtech, and as a psychologist who's attempted to use OCEAN for said adtech, I never found it to work better than standard approaches. Additionally, Costa and McCrae created the Big 5, found it didn't replicate in confirmatory factor analysis, so they created a new method of CFA which did replicate it. Like, E O and N are probably real, but the overall structure does not replicate and should not be taken very seriously.

ekianjo 5y ago

> This has been well documented as being able to identify the big five traits (OPENNESS, CONSCIENTIOUSNESS, EXTRAVERSION, AGREEABLENESS, NEUROTICISM) and its sub traits. It's just a simplified model to put personalities in a certain bucket, typical PCA-style. This has limitations just like everything else that uses self-reporting to make a judgment/categorization. it's not like we are able to observe the structure of someone's brain and confirm that with actual facts. In other words, that's not even remotely what you call Science.