3 ms·

I'm surprised greater attention is not focused on the fact the "science" is in the midst of a reproducibility crisis. This is a big issue in the hard sciences;

I'm surprised greater attention is not focused on the fact the "science" is in the midst of a reproducibility crisis. This is a big issue in the hard sciences; I can only imagine what it's like in the squishy science of gender behavior. Tying your reputation to such a soft foundation is just inviting trouble.

josephg 9y ago
The people-things orientation has an extremely large effect size for social science. The Su et al. meta-analysis from 2009 found an effect size of 0.86 with N > 500,000 participants. Lippa (N > 200,000) found an effect size of 1.40. http://sci-hub.cc/10.1111/j.1751-9004.2010.00320.x http://sci-hub.cc/10.1111/j.1751-9004.2010.00320.x If you don't believe the research is solid, please post some convincing counter studies.
- jpmattia 9y ago
  > If you don't believe the research is solid, please post some convincing counter studies. It sounds like you don't understand the reproducibility crisis: It's not up to me to produce counter studies, it's up to you to produce replications of the already published work, using the exact same methods and techniques. So far, the reproducibility in the psychological sciences has been pretty poor. Nature summarized the original test a couple of years ago: 61 out of 100 replications failed in the test. https://www.nature.com/news/over-half-of-psychology-studies-fail-reproducibility-test-1.18248 https://www.nature.com/news/over-half-of-psychology-studies-...
  - ahh 9y ago
    I think you might not know what a meta-analysis is: it's actually a summary of replications and duplicate studies on some topic. Generally (I have not read this particular one!) they are specifically designed to detect potential publication bias and other effects that would imply non-reproducibility.
    - jpmattia 9y ago
      > it's actually a summary of replications and duplicate studies on some topic. ... they are specifically designed to detect potential publication bias and other effects that would imply non-reproducibility. Not quite right: Meta-analyses combine several studies to improve the statistical power of a given topic. They do not address reproducibility, because the individual papers forming the basis for the meta-analyses make no attempt at reproducing prior results before introducing new ones. I'd go so far as to say part of the necessity for the original reproducibility study of 100 seminal psychology papers was because meta-analyses fail to address reproducibility. The repercussions of 61/100 papers failing to reproduce are still not as widely understood as they should be.
pat2man 9y ago
I'm sure for every article like this one there are solid arguments coming to the opposite conclusion: https://www.linkedin.com/pulse/differences-between-men-women-vastly-exaggerated-adam-grant https://www.linkedin.com/pulse/differences-between-men-women...

josephg 9y ago

The people-things orientation has an extremely large effect size for social science. The Su et al. meta-analysis from 2009 found an effect size of 0.86 with N > 500,000 participants. Lippa (N > 200,000) found an effect size of 1.40. http://sci-hub.cc/10.1111/j.1751-9004.2010.00320.x http://sci-hub.cc/10.1111/j.1751-9004.2010.00320.x If you don't believe the research is solid, please post some convincing counter studies.

jpmattia 9y ago
> If you don't believe the research is solid, please post some convincing counter studies. It sounds like you don't understand the reproducibility crisis: It's not up to me to produce counter studies, it's up to you to produce replications of the already published work, using the exact same methods and techniques. So far, the reproducibility in the psychological sciences has been pretty poor. Nature summarized the original test a couple of years ago: 61 out of 100 replications failed in the test. https://www.nature.com/news/over-half-of-psychology-studies-fail-reproducibility-test-1.18248 https://www.nature.com/news/over-half-of-psychology-studies-...
- ahh 9y ago
  I think you might not know what a meta-analysis is: it's actually a summary of replications and duplicate studies on some topic. Generally (I have not read this particular one!) they are specifically designed to detect potential publication bias and other effects that would imply non-reproducibility.
  - jpmattia 9y ago
    > it's actually a summary of replications and duplicate studies on some topic. ... they are specifically designed to detect potential publication bias and other effects that would imply non-reproducibility. Not quite right: Meta-analyses combine several studies to improve the statistical power of a given topic. They do not address reproducibility, because the individual papers forming the basis for the meta-analyses make no attempt at reproducing prior results before introducing new ones. I'd go so far as to say part of the necessity for the original reproducibility study of 100 seminal psychology papers was because meta-analyses fail to address reproducibility. The repercussions of 61/100 papers failing to reproduce are still not as widely understood as they should be.

pat2man 9y ago

I'm sure for every article like this one there are solid arguments coming to the opposite conclusion: https://www.linkedin.com/pulse/differences-between-men-women-vastly-exaggerated-adam-grant https://www.linkedin.com/pulse/differences-between-men-women...