4 ms·

How is a criticism for small sample size an arm chair dismissal? Lack of existing data doesn’t mean that the data that does exist is predictive. It mostly mean

How is a criticism for small sample size an arm chair dismissal? Lack of existing data doesn’t mean that the data that does exist is predictive. It mostly means we can’t say anything meaningful. Science is largely an exercise in bad epistemology.

ckrapu 4y ago
A small sample size is, by itself, not an issue. It only means that you need larger effect sizes to come to any conclusion. Fortunately for the researcher referenced, the changes are indeed rather pronounced.
- leetcrew 4y ago
  the objection is that large-scale weather patterns can be cyclical over a period of years or even decades (eg, north atlantic oscillation). sample vs effect size in the statistical sense is a red herring here. an oscillation will look like an increasing/decreasing trend until you have data for one or more full cycles.
  - ckrapu 4y ago
    I understand your point about oscillations as an alternative cause, but we have enough data for at least a single full cycle of NAO. While skimming some of the related work, I found a line in this paper (https://journals.ametsoc.org/view/journals/bams/100/10/bams-d-18-0189.1.xml https://journals.ametsoc.org/view/journals/bams/100/10/bams-...) which roughly estimated six decades of data required to assess a modest (10%) increase in frequency of severe hurricanes in the Atlantic, taking into account natural variability from a range of sources. I don’t know how comprehensive that is, though. The big picture view I have on this is that the skeptic perspective on this article has a big hurdle to overcome, namely that physical theories do support the empirical finding here. I think it is reasonable to keep alternate causes in mind here, but I don’t see too many that seem viable globally.
random314 4y ago
Small sample size? It is clear that you are disputing science using whatever jargon you hear in science articles and you have no understanding of statistics. You would dismiss Newton's theory of gravitation when he proposed it in the 1600s because it can only explain the motion of 5 planets, clearly a SMALL SAMPLE SIZE. > Science is largely an exercise in bad epistemology. You have used a very big word here. I am very impressed, you sound very smart and impressive. My recommendation to you is to avoid science and go back to Javascript.

ckrapu 4y ago

A small sample size is, by itself, not an issue. It only means that you need larger effect sizes to come to any conclusion. Fortunately for the researcher referenced, the changes are indeed rather pronounced.

leetcrew 4y ago
the objection is that large-scale weather patterns can be cyclical over a period of years or even decades (eg, north atlantic oscillation). sample vs effect size in the statistical sense is a red herring here. an oscillation will look like an increasing/decreasing trend until you have data for one or more full cycles.
- ckrapu 4y ago
  I understand your point about oscillations as an alternative cause, but we have enough data for at least a single full cycle of NAO. While skimming some of the related work, I found a line in this paper (https://journals.ametsoc.org/view/journals/bams/100/10/bams-d-18-0189.1.xml https://journals.ametsoc.org/view/journals/bams/100/10/bams-...) which roughly estimated six decades of data required to assess a modest (10%) increase in frequency of severe hurricanes in the Atlantic, taking into account natural variability from a range of sources. I don’t know how comprehensive that is, though. The big picture view I have on this is that the skeptic perspective on this article has a big hurdle to overcome, namely that physical theories do support the empirical finding here. I think it is reasonable to keep alternate causes in mind here, but I don’t see too many that seem viable globally.

random314 4y ago

Small sample size? It is clear that you are disputing science using whatever jargon you hear in science articles and you have no understanding of statistics. You would dismiss Newton's theory of gravitation when he proposed it in the 1600s because it can only explain the motion of 5 planets, clearly a SMALL SAMPLE SIZE. > Science is largely an exercise in bad epistemology. You have used a very big word here. I am very impressed, you sound very smart and impressive. My recommendation to you is to avoid science and go back to Javascript.