There is another big problem (related to replication): the career-related pressure to find something publishable. That will lead people to try to get a publishable result even if the underlying question is not an important one, I.e. one that really adds to human knowledge as opposed to being good for getting the investigator a job. We might be better off allowing people to fail to find answers to important questions rather than making the find answers to unimportant ones.
Sorry, Rob - the replication crisis is the result of a wrong turn in the academe that began with John Venn attacking Laplace, and was perpetuated through Fisher and Popper and has come to dominate across most of academic primary research. Venn et al denied the meaning of the "probability of a hypothesis" and instead looked only at the "probability of the data" against a "null" - an anti-hypothesis - which they *assume* is true as a starting point and then tell themselves that unlikely data (p <.05) means (by convention) that it's "extremely unlikely" that they're "incorrect." They compare hypotheses to infinite trials that never occurred.
That is all bullshit; it is absolutely NOT science. The reason none of it replicates is because it was never intended to replicate. See E.T. Jaynes "Probability Theory: the Logic of Science."
This is before you even get to the problem that most of the social "sciences" don't measure anything at all. They record memories and feelings with no scale or well-defined error rate. Non-science from start to finish. Academics polluted (and inverted) what science is and its sole metric for validation: a hypothesis' predictive strength is both its objectivity and the measure of its validity.
All true. The easiest way to develop your own believability, though, is this.
Read the paper. Especially the 'Methods' section. Ask yourself if that makes sense as a way to observe a phenomenon. You'd be amazed at what researchers will declare is a good methodology.
I love that the original WSJ title of your piece implied exactly the opposite conclusion from that you reached - proving your point about moaners at the "replication crisis" perfectly.
There is another big problem (related to replication): the career-related pressure to find something publishable. That will lead people to try to get a publishable result even if the underlying question is not an important one, I.e. one that really adds to human knowledge as opposed to being good for getting the investigator a job. We might be better off allowing people to fail to find answers to important questions rather than making the find answers to unimportant ones.
Good post I think we all have to keep this in the back of our minds. It’s an easy trapped to fall into.
This quote isn’t really the point of our goodly author’s piece but it’s something worth mentioning:
“Mathematician Stanislaw Ulam challenged economist Paul Samuelson to name a single social science proposition that was both true and non-trivial. “
For me, there is a taint on the social sciences, they are a midwit playground.
Sorry, Rob - the replication crisis is the result of a wrong turn in the academe that began with John Venn attacking Laplace, and was perpetuated through Fisher and Popper and has come to dominate across most of academic primary research. Venn et al denied the meaning of the "probability of a hypothesis" and instead looked only at the "probability of the data" against a "null" - an anti-hypothesis - which they *assume* is true as a starting point and then tell themselves that unlikely data (p <.05) means (by convention) that it's "extremely unlikely" that they're "incorrect." They compare hypotheses to infinite trials that never occurred.
That is all bullshit; it is absolutely NOT science. The reason none of it replicates is because it was never intended to replicate. See E.T. Jaynes "Probability Theory: the Logic of Science."
This is before you even get to the problem that most of the social "sciences" don't measure anything at all. They record memories and feelings with no scale or well-defined error rate. Non-science from start to finish. Academics polluted (and inverted) what science is and its sole metric for validation: a hypothesis' predictive strength is both its objectivity and the measure of its validity.
All true. The easiest way to develop your own believability, though, is this.
Read the paper. Especially the 'Methods' section. Ask yourself if that makes sense as a way to observe a phenomenon. You'd be amazed at what researchers will declare is a good methodology.
I love that the original WSJ title of your piece implied exactly the opposite conclusion from that you reached - proving your point about moaners at the "replication crisis" perfectly.