3 ms·

Like clockwork, every decade or so, someone says "Hey, we haven't looked at attachment theory in a while! We can get a few pubs and a book out of it.". A bunc

Like clockwork, every decade or so, someone says "Hey, we haven't looked at attachment theory in a while! We can get a few pubs and a book out of it.". A bunch of studies later, we see the repeat of all the previous findings: inconclusive, some interesting patterns, nothing super repeatable with strong effects. It's fine to revisit and revise past theory work. It's just surprising that after 70 years, this theory that should explain so much, after scrutiny, just... doesn't. Looking forward to the latest revisions to see how to apply attachment patterns to our modern world. I have popcorn in hand.

jschwartzi 6y ago
I would think it's very difficult to study human attachment using effect sizes. The sheer diversity of attachments a human makes and the rate at which each attachment changes means you're almost certainly seeing some confounding effects. You could bin attachment into styles and levels but it's more like a continuous function and you can never be certain that your bin sizes are adequate to describe the whole variability of the system.
sjg007 6y ago
Well if you want to make the argument that attachment theory is junk then you should provide some evidence.
- mwexler 6y ago
  That's not really the way science works. The theory drives testable hypotheses. We look at data to feel some confidence that the hypotheses are correct. If we don't find that data, we can continue to believe in the theory, but with less and less confidence. If a theory cannot be supported, then I can make the claim that its not correct in its current form. I don't think it's junk, but more that it's causal impact may be less than the theory purports. And sure, we can have an alt theory that goes against Attachment theory, and generate hypos from that, and put them to the test. Answering a different question, but same problem: can't just say "won't work", need to generate specific testable hypotheses. I'm just pointing out that the theory is nice and should have more provable (and supported) predictions, but it keeps having limited support from research, fades from view, comes back around, history repeats. After a while, our confidence goes very low. But it's not zero, and the fact that it keeps showing up again provides new hope. The theory is attractive so I hope we find its proper alignment through this latest iteration. But it has a tough road to hoe.
  - sjg007 6y ago
    Right now the scientific consensus is in for attachment theory.. I asked you to provide evidence or an alternative that it is not. What is the experiment that proves the null hypothesis? You claim it does not replicate? Ok, show me some papers. "That's not really the way science works. The theory drives testable hypotheses." Aside from the ethical considerations of using human subjects to test hypotheses, you also won't like genomic/bioinformatics/physics, if you faithfully adhere to the Popper doctrine.
    - mwexler 6y ago
      Perhaps we follow different sources. Bowlby and Ainsworth suggest that att drives many, many behaviors and thought processes. Clinical psych has found situations where the theory applies and some therapies have had success. Yet social psych, for example, has not found similar success using this basis to predict behaviors or mental processes, for example, on relationship behaviors as adults. Since so many other theories do drive predictions across branches of science, then it's hard to believe that this theory, even after all the evolution its had, is as general as it should or could be. Not to say it won't get there, or that it's wrong, but that it gets at something without quite getting there (to put it very vaguely). I'm not sure I get your final point: while Popper was never a fan of social sciences, psychologists have spent years running testable hypotheses and experimenting with them. Att is just a preexisting condition, and therefore should drive a repeatably different response to certain stimuli; I don't think anyone was suggesting we try raising babies in diff att styles. But yeah, if you take Popper at his word, lots of work goes out the window (I avoided saying "pops"). I'm gonna choose a middle ground here, and see where att theory goes. I have high hopes!

jschwartzi 6y ago

I would think it's very difficult to study human attachment using effect sizes. The sheer diversity of attachments a human makes and the rate at which each attachment changes means you're almost certainly seeing some confounding effects. You could bin attachment into styles and levels but it's more like a continuous function and you can never be certain that your bin sizes are adequate to describe the whole variability of the system.

sjg007 6y ago

Well if you want to make the argument that attachment theory is junk then you should provide some evidence.

mwexler 6y ago
That's not really the way science works. The theory drives testable hypotheses. We look at data to feel some confidence that the hypotheses are correct. If we don't find that data, we can continue to believe in the theory, but with less and less confidence. If a theory cannot be supported, then I can make the claim that its not correct in its current form. I don't think it's junk, but more that it's causal impact may be less than the theory purports. And sure, we can have an alt theory that goes against Attachment theory, and generate hypos from that, and put them to the test. Answering a different question, but same problem: can't just say "won't work", need to generate specific testable hypotheses. I'm just pointing out that the theory is nice and should have more provable (and supported) predictions, but it keeps having limited support from research, fades from view, comes back around, history repeats. After a while, our confidence goes very low. But it's not zero, and the fact that it keeps showing up again provides new hope. The theory is attractive so I hope we find its proper alignment through this latest iteration. But it has a tough road to hoe.
- sjg007 6y ago
  Right now the scientific consensus is in for attachment theory.. I asked you to provide evidence or an alternative that it is not. What is the experiment that proves the null hypothesis? You claim it does not replicate? Ok, show me some papers. "That's not really the way science works. The theory drives testable hypotheses." Aside from the ethical considerations of using human subjects to test hypotheses, you also won't like genomic/bioinformatics/physics, if you faithfully adhere to the Popper doctrine.
  - mwexler 6y ago
    Perhaps we follow different sources. Bowlby and Ainsworth suggest that att drives many, many behaviors and thought processes. Clinical psych has found situations where the theory applies and some therapies have had success. Yet social psych, for example, has not found similar success using this basis to predict behaviors or mental processes, for example, on relationship behaviors as adults. Since so many other theories do drive predictions across branches of science, then it's hard to believe that this theory, even after all the evolution its had, is as general as it should or could be. Not to say it won't get there, or that it's wrong, but that it gets at something without quite getting there (to put it very vaguely). I'm not sure I get your final point: while Popper was never a fan of social sciences, psychologists have spent years running testable hypotheses and experimenting with them. Att is just a preexisting condition, and therefore should drive a repeatably different response to certain stimuli; I don't think anyone was suggesting we try raising babies in diff att styles. But yeah, if you take Popper at his word, lots of work goes out the window (I avoided saying "pops"). I'm gonna choose a middle ground here, and see where att theory goes. I have high hopes!