4 ms·

When we do it: A/B Testing When FB does it: Psychological Experiments

cronix 8y ago
I think most people would categorize this as Psychological Experiments vs A/B testing. Most companies don't A/B test whether they can intentionally manipulate people into generally feeling negative. > To test that, Facebook data scientists tweaked the newsfeed algorithms of roughly 0.04 percent of Facebook users, or 698,003 people, for one week in January 2012. During the experiment, half of those subjects saw fewer positive posts than usual, while half of them saw fewer negative ones. To evaluate how that change affected mood, researchers also tracked the number of positive and negative words in subjects’ status updates during that week-long period. Put it all together, and you have a pretty clear causal relationship. The results were published last month in the Proceedings of the National Academy Sciences, which, for what it’s worth, is a pretty prestigious journal. https://www.washingtonpost.com/news/the-intersect/wp/2014/07/01/9-answers-about-facebooks-creepy-emotional-manipulation-experiment https://www.washingtonpost.com/news/the-intersect/wp/2014/07...
- travisoneill1 8y ago
  OK so they published the results of an A/B test. My point is that those two phrases mean the same thing, not that it wasn't a psychological experiment.
  - bobthepanda 8y ago
    One describes the how and the other describes the what and the why. They're not the same thing at all.
    - sidr 8y ago
      Yup, but "do people click on a blue button than a green button?" could also be accurately described as a psychological experiment. Every single A/B test that involves user interactions could be described as a psychological experiment. Yet we don't hear the mass media outrage directed at literally every other website in existence.
      - pm90 8y ago
        > Yup, but "do people click on a blue button than a green button?" could also be accurately described as a psychological experiment No it wouldn't. Only to a person who knows nothing about psychology or about experiments would that phrase accurately describe a psychological experiment.
      - cronix 8y ago
        > Yet we don't hear the mass media outrage directed at literally every other website in existence. Sorry, let's rephrase. "Harmful psychological testing" Sure, A/B tests have a psychological component, although the tests don't specifically test for that or they'd be after the reason WHY they clicked on blue vs green. They don't care why, they only care about which got more clicks in order to choose which one to go with. They usually also tend to actually test something, like a feature. What feature was FB A/B testing? They weren't. It was a pure psychological test, with harmful results. You honestly don't see a distinction?
    - scarejunba 8y ago
      Which is which? Facebook’s study was “Does emotional contagion occur?” Right? So that’s “what”, not “why” or “how”.
      - bobthepanda 8y ago
        The "how" is the psychological experiment. The hypothesis is the "what", as well as the "why" we are doing this. There were a lot of ethical failures on the part of Facebook, such as the failure to get informed consent, the failure to review the experiment with an independent review board until after it had occurred, misleading said review board to say that the dataset was previously approved, etc., and all for results that the authors think were not very valuable: https://www.theatlantic.com/technology/archive/2014/06/everything-we-know-about-facebooks-secret-mood-manipulation-experiment/373648/ https://www.theatlantic.com/technology/archive/2014/06/every...

I think most people would categorize this as Psychological Experiments vs A/B testing. Most companies don't A/B test whether they can intentionally manipulate people into generally feeling negative. > To test that, Facebook data scientists tweaked the newsfeed algorithms of roughly 0.04 percent of Facebook users, or 698,003 people, for one week in January 2012. During the experiment, half of those subjects saw fewer positive posts than usual, while half of them saw fewer negative ones. To evaluate how that change affected mood, researchers also tracked the number of positive and negative words in subjects’ status updates during that week-long period. Put it all together, and you have a pretty clear causal relationship. The results were published last month in the Proceedings of the National Academy Sciences, which, for what it’s worth, is a pretty prestigious journal. https://www.washingtonpost.com/news/the-intersect/wp/2014/07/01/9-answers-about-facebooks-creepy-emotional-manipulation-experiment https://www.washingtonpost.com/news/the-intersect/wp/2014/07...

travisoneill1 8y ago
OK so they published the results of an A/B test. My point is that those two phrases mean the same thing, not that it wasn't a psychological experiment.
- bobthepanda 8y ago
  One describes the how and the other describes the what and the why. They're not the same thing at all.
  - sidr 8y ago
    Yup, but "do people click on a blue button than a green button?" could also be accurately described as a psychological experiment. Every single A/B test that involves user interactions could be described as a psychological experiment. Yet we don't hear the mass media outrage directed at literally every other website in existence.
    - pm90 8y ago
      > Yup, but "do people click on a blue button than a green button?" could also be accurately described as a psychological experiment No it wouldn't. Only to a person who knows nothing about psychology or about experiments would that phrase accurately describe a psychological experiment.
    - cronix 8y ago
      > Yet we don't hear the mass media outrage directed at literally every other website in existence. Sorry, let's rephrase. "Harmful psychological testing" Sure, A/B tests have a psychological component, although the tests don't specifically test for that or they'd be after the reason WHY they clicked on blue vs green. They don't care why, they only care about which got more clicks in order to choose which one to go with. They usually also tend to actually test something, like a feature. What feature was FB A/B testing? They weren't. It was a pure psychological test, with harmful results. You honestly don't see a distinction?
  - scarejunba 8y ago
    Which is which? Facebook’s study was “Does emotional contagion occur?” Right? So that’s “what”, not “why” or “how”.
    - bobthepanda 8y ago
      The "how" is the psychological experiment. The hypothesis is the "what", as well as the "why" we are doing this. There were a lot of ethical failures on the part of Facebook, such as the failure to get informed consent, the failure to review the experiment with an independent review board until after it had occurred, misleading said review board to say that the dataset was previously approved, etc., and all for results that the authors think were not very valuable: https://www.theatlantic.com/technology/archive/2014/06/everything-we-know-about-facebooks-secret-mood-manipulation-experiment/373648/ https://www.theatlantic.com/technology/archive/2014/06/every...