3 ms·

Many social scientists lean left and tend to "find" results which coincidentally align with their political viewpoint. If the results are truly reproducible, th

Many social scientists lean left and tend to "find" results which coincidentally align with their political viewpoint. If the results are truly reproducible, then someone on the other side of the spectrum should also be able to reproduce it. On the other hand, if the work was politically motivated then it won't be reproducible.

aaplok 4y ago
> if the work was politically motivated then it won't be reproducible. Why not? If indeed there is a bias ib social science then we could expect some conclusions to remain unpublished if they contradict the bias. But I don't see why the results that do get published wouldn't be reproducible.
- ceeplusplus 4y ago
  > But I don't see why the results that do get published wouldn't be reproducible The very article we are reading illustrates why published results can be irreproducible and indeed completely made up. The Losada research cited was invalid [1] yet nobody questioned it because it fit their views of how humans should interact (positivity ratio implies some form of constructive criticism, which is often taught in Western schools as the ideal way to give feedback to someone). Any attempt by a sceptic to reproduce it would have realized that the "math" used to model the positivity ratio was wrong. The authors themselves state in the article they still believe the positivity ratio idea even though the Losada research was shown to be egregiously wrong. [1] https://retractionwatch.com/2014/07/29/positivity-ratio-research-now-subject-to-an-expression-of-concern/ https://retractionwatch.com/2014/07/29/positivity-ratio-rese...
  - ModernMech 4y ago
    Most research isn’t reproduced because it’s not economically lucrative for researchers to reproduce research. It’s that simple. If you want scientists to reproduce research, make funding available for that purpose. It has nothing to do with political biases. I mean, if you’re talking about completely made up research then that’s a matter of fraud, not political biases.
ModernMech 4y ago
Do you have a basis for this assertion that has been studied and reproduced by someone of the opposite political spectrum? I’m sorry but as a scientist, your assertion that we must hear from someone with the opposite political view before we can determine the veracity of some work is just so foreign to me. Must we also hear from creationists on evolution research? Why can’t that stand alone? If some experiment is reproducible it shouldn’t matter the political bias of the person running the study. Moreover how would this even be implemented? The idea of a “political spectrum” is just a rule of thumb at best, and a dangerous misleading heuristic at worst. According to you, how would we identify an author’s political bias? Force them to self identify or take some sort of certified test? Will we divide all of our departments between liberals and conservatives? Right now I don’t know the political affiliation of any of my colleagues, and they don’t know mine. Are you suggesting we should bring our politics explicitly into our research? I mean… have you even thought any of this through?

aaplok 4y ago

> if the work was politically motivated then it won't be reproducible. Why not? If indeed there is a bias ib social science then we could expect some conclusions to remain unpublished if they contradict the bias. But I don't see why the results that do get published wouldn't be reproducible.

ceeplusplus 4y ago
> But I don't see why the results that do get published wouldn't be reproducible The very article we are reading illustrates why published results can be irreproducible and indeed completely made up. The Losada research cited was invalid [1] yet nobody questioned it because it fit their views of how humans should interact (positivity ratio implies some form of constructive criticism, which is often taught in Western schools as the ideal way to give feedback to someone). Any attempt by a sceptic to reproduce it would have realized that the "math" used to model the positivity ratio was wrong. The authors themselves state in the article they still believe the positivity ratio idea even though the Losada research was shown to be egregiously wrong. [1] https://retractionwatch.com/2014/07/29/positivity-ratio-research-now-subject-to-an-expression-of-concern/ https://retractionwatch.com/2014/07/29/positivity-ratio-rese...
- ModernMech 4y ago
  Most research isn’t reproduced because it’s not economically lucrative for researchers to reproduce research. It’s that simple. If you want scientists to reproduce research, make funding available for that purpose. It has nothing to do with political biases. I mean, if you’re talking about completely made up research then that’s a matter of fraud, not political biases.

ModernMech 4y ago

Do you have a basis for this assertion that has been studied and reproduced by someone of the opposite political spectrum? I’m sorry but as a scientist, your assertion that we must hear from someone with the opposite political view before we can determine the veracity of some work is just so foreign to me. Must we also hear from creationists on evolution research? Why can’t that stand alone? If some experiment is reproducible it shouldn’t matter the political bias of the person running the study. Moreover how would this even be implemented? The idea of a “political spectrum” is just a rule of thumb at best, and a dangerous misleading heuristic at worst. According to you, how would we identify an author’s political bias? Force them to self identify or take some sort of certified test? Will we divide all of our departments between liberals and conservatives? Right now I don’t know the political affiliation of any of my colleagues, and they don’t know mine. Are you suggesting we should bring our politics explicitly into our research? I mean… have you even thought any of this through?