4 ms·

Good sociology papers are just as scientific as physics or engineering papers, they are just as a percentage in their field more rare.

tgv 5y ago
Super rare then, because I don't know of any quantitative sociological theory that holds under all (reasonable) conditions. Nor of a qualitative one, by the way. And this paper stems from psychiatry.
- Retric 5y ago
  What would you say where the best papers in the field then and where did they fail?
njarboe 5y ago
Possibly, but in physics the standard for stating something is "true" is 5 or 6 sigma while in sociology, maybe 2 sigma. In the first case, you can take the finding as a fact and it is very unlikely to change in your lifetime, in the second case, one needs to constantly keep in mind the conclusion might not be true. Not really that useful, actually.
- mattkrause 5y ago
  Snobbery about the conventional levels of significance (etc) is almost entirely beside the point. Doing sociology is hard and physics is, in some ways, absurdly easy. As a physicist, you can generate an endless supply of particles. They are all identical, and so behave in the same way. They do whatever they do, regardless of the experimenters' expectations, social pressures, or the like. The results can be slotted into a formal mathematical framework, which also helps you design better experiments. The work of these experiments is often difficult (e.g., holding a detector just about 0K), but the concepts are often relatively clear. None of this is true in 'softer' sciences. Every person (or animal) is different and you often can't even test the same subject twice because the very act of doing so changes them. People often don't know what they want or why they do something. When they do know, they sometimes lie about it, for myriad reasons. This makes designing, analyzing, and interpreting experiments conceptually hard, even if the actual work is pretty easy.
  - MichaelGroves 5y ago
    >>> Good sociology papers are just as scientific as physics or engineering papers, > Snobbery about the conventional levels of significance (etc) is almost entirely beside the point. It seems like entirely the point. Also, the difficulties you describe are reasons sociology can't be as scientific as physics, not reasons to pretend sociology is just as scientific.
    - s1artibartfast 5y ago
      I think the problem is conflating being scientific with numerical. This paper is not very numerical, but is scientific. It lays out observations and a hypothesis that might explain those observations.
      - MichaelGroves 5y ago
        > It lays out observations and a hypothesis that might explain those observations. The part you're missing here is where those hypotheses are tested. A field that consistently makes accurate predictions is more scientific than one which rarely does, regardless of how much effort is expended. A field being more scientific than another isn't a matter of researchers trying harder. The quality of their results, independent of the magnitude of their effort, is what really matters. If researchers dutifully follow the scientific method by the textbook and, due to the subject of their research, can only get slipshod results, their field is less scientific than one in which researchers easily get high quality results.
        s1artibartfast 5y ago
        >The part you're missing here is where those hypotheses are tested. There is no rule that every science paper needs to do testing and cover the whole scientific method from start to finish. Some can just be building blocks; data, hypothesis, ect. >A field that consistently makes accurate predictions is more scientific than one which rarely does I couldn't disagree more. That just means the field is more predictable
        5y ago

tgv 5y ago

Super rare then, because I don't know of any quantitative sociological theory that holds under all (reasonable) conditions. Nor of a qualitative one, by the way. And this paper stems from psychiatry.

Retric 5y ago
What would you say where the best papers in the field then and where did they fail?

njarboe 5y ago

Possibly, but in physics the standard for stating something is "true" is 5 or 6 sigma while in sociology, maybe 2 sigma. In the first case, you can take the finding as a fact and it is very unlikely to change in your lifetime, in the second case, one needs to constantly keep in mind the conclusion might not be true. Not really that useful, actually.

mattkrause 5y ago
Snobbery about the conventional levels of significance (etc) is almost entirely beside the point. Doing sociology is hard and physics is, in some ways, absurdly easy. As a physicist, you can generate an endless supply of particles. They are all identical, and so behave in the same way. They do whatever they do, regardless of the experimenters' expectations, social pressures, or the like. The results can be slotted into a formal mathematical framework, which also helps you design better experiments. The work of these experiments is often difficult (e.g., holding a detector just about 0K), but the concepts are often relatively clear. None of this is true in 'softer' sciences. Every person (or animal) is different and you often can't even test the same subject twice because the very act of doing so changes them. People often don't know what they want or why they do something. When they do know, they sometimes lie about it, for myriad reasons. This makes designing, analyzing, and interpreting experiments conceptually hard, even if the actual work is pretty easy.
- MichaelGroves 5y ago
  >>> Good sociology papers are just as scientific as physics or engineering papers, > Snobbery about the conventional levels of significance (etc) is almost entirely beside the point. It seems like entirely the point. Also, the difficulties you describe are reasons sociology can't be as scientific as physics, not reasons to pretend sociology is just as scientific.
  - s1artibartfast 5y ago
    I think the problem is conflating being scientific with numerical. This paper is not very numerical, but is scientific. It lays out observations and a hypothesis that might explain those observations.
    - MichaelGroves 5y ago
      > It lays out observations and a hypothesis that might explain those observations. The part you're missing here is where those hypotheses are tested. A field that consistently makes accurate predictions is more scientific than one which rarely does, regardless of how much effort is expended. A field being more scientific than another isn't a matter of researchers trying harder. The quality of their results, independent of the magnitude of their effort, is what really matters. If researchers dutifully follow the scientific method by the textbook and, due to the subject of their research, can only get slipshod results, their field is less scientific than one in which researchers easily get high quality results.
      - s1artibartfast 5y ago
        >The part you're missing here is where those hypotheses are tested. There is no rule that every science paper needs to do testing and cover the whole scientific method from start to finish. Some can just be building blocks; data, hypothesis, ect. >A field that consistently makes accurate predictions is more scientific than one which rarely does I couldn't disagree more. That just means the field is more predictable
      - 5y ago