5 ms·
The problem with these kinds of statistical snippets that papers like the FT produce, is that they are fishing for some kind of shocking/surprising easy to unde
by JR1427 3y ago
The problem with these kinds of statistical snippets that papers like the FT produce, is that they are fishing for some kind of shocking/surprising easy to understand finding.
That doesn't mean that such statistics and charts actually show anything useful in a scientific sense.
The problem is that the information is presented as a scientific "truth", and many people don't have the understanding to know that all statistics and analysis involve human input (e.g. which tests to use, how to visualise, which data to exclude etc etc), and that this input introduces biases, whether accidentally or intentionally.
- JumpCrisscross 3y ago> they are fishing for some kind of shocking/surprising easy to understand finding The FT has a hard paywall. They aren’t pandering for clickbait. (The author’s tweet doesn’t even bother linking to the article.) > many people don't have the understanding to know that all statistics and analysis involve human input The Financial Times’ readership?
- JR1427 3y agoI'm not saying it's clickbait in the sense of "this one amazing trick to shrink belly fat", but it still trying to show something surprising (people find graphs showing things they know boring). Don't over estimate the scientific/statistical understanding of FT readers. They are just normal people. (I know a few of them.)
- JumpCrisscross 3y ago> They are just normal people Like Hacker News readers, who presumably need a general disclaimer about statistics. Got it.
- bee_rider 3y agoThey are normal people, our people follow a uniform distribution instead.
- JR1427 3y agoOh, and JBM is absolutely fishing for something by posting on Twitter. Otherwise, why would he post it?
- brickteacup 3y agoYes we all know it's possible to lie with statistics. What are your criticisms of this article in particular? If you don't have any then you're just baselessly casting aspersions.
- JR1427 3y agoI don't have access to the article - my comment was about the figures specifically, as they were posted on Twitter (where I can see them. I don't have an FT subscription). I'm critical/suspicious whenever a complex topic is presented as if it is apparently very simple, e.g. when people's political opinions apparently clearly follow a general rule across different countries and cultures (as in this case). Without any evidence to back it up, I can imagine a different set if questions leading to different outcomes. Ideally, there would be a different newspaper who would set out to show that Women are more conservative (or some other different finding to the FT's), and the we'd be able to decide which method we find more compelling. Having different opinions published by outlets that have the same readership is how scientific findings are disseminated and judged (e.g. two different theories both published by Nature, and read by Nature readers). In the case of newspapers, people of one opinion tend to read one subset of newspapers, and people of another opinion the other, which means that the kind of weighing up of evidence doesn't happen in the same way. Anyway, my comments were made in good faith, in the spirit of someone just saying what they think, as is the way with HN. I didn't mean to offend anyone.
- tiahura 3y agoWithout anything specific to the article, your comment is fishing for a point.
- JR1427 3y agoI wasn't fishing for anything. I'm not sure what I could even be fishing for, by commenting on HN?