4 ms·

> Your own source would be a good example You say this then go on to shit on it and then only agree with parts that support your argument? > they are unlikel

> Your own source would be a good example You say this then go on to shit on it and then only agree with parts that support your argument? > they are unlikely to ever have children This reports the same thing: https://www.pewresearch.org/short-reads/2021/11/19/growing-share-of-childless-adults-in-u-s-dont-expect-to-ever-have-children/ https://www.pewresearch.org/short-reads/2021/11/19/growing-s... > That study is 'p-hacked' Not typically how p-hacking is used? > Having the age range be 18-49 is illogical It looks like they selected the buckets on the criteria of being adults and physically able to have children. > 18-49 is not reasonable for fertility issues I guess? It probably made the survey easier to conduct though. > Even with their heavily biased sample Why is their sample biased? It was a survey of that population.

somenameforme 2y ago
No, you're not understanding. Amongst all people that are not having children there are two subsets - those that plan to never have children, and those that are unsure or plan to have children - just not right now. Those numbers Pew is publishing are only from the former. This is what I meant by saying that their headline is most aptly stated as, "[a very slight majority of] people who don't plan to ever have kids say they don't want to have kids." This is essentially what p-hacking is. By organizing variables in a certain way, you can create statistically significant results to say basically whatever you want them to say. A huge tell-tale sign of p-hacking is constraining one's sample in ways that don't really make sense and/or are likely to be misleading. The thing we were discussing is 'why people aren't having kids.' This is very different than why 'people who don't plan to ever have kids aren't having kids.'
- someothherguyy 2y ago
  > This is essentially what p-hacking is. By organizing variables in a certain way, you can create statistically significant results to say basically whatever you want them to say. Do you know what statistically significant results means?
  - somenameforme 2y ago
    Yes, and it doesn't mean "true" as you might think. There's a great demo of p-hacking here. [1] It's based on real data too, which makes it extra fun. Your goal is to prove that one party or another being in power leads to a stronger economy. That statement is almost certainly false as you can quickly see by how awkward it is to "prove" it. But you'll find when you arrange your variables in certain ways you can indeed find statistically significant evidence to "prove" literally whatever you want. But again the one tell-tale sign somebody like me would notice from your study is 'why in the world would you limit it to that sample'? Well because that sample gives you the answer you want! It's science, baby! [1] - https://projects.fivethirtyeight.com/p-hacking/ https://projects.fivethirtyeight.com/p-hacking/
    - someothherguyy 2y ago
      This is a survey. There is nothing under test. Its like saying you are p-hacking an election. I don't understand what you mean.
      - somenameforme 2y ago
        This is the entire point of our dialogue. You claimed my statement that most people [who don't have children] claim they aren't having children because of economic reasons was false. Your evidence was a study that showed a negligible majority of people [who actively don't plan to ever have kids, age 18-49] are simply doings so out of personal choice, and then seemed confused about why this was an inappropriate sampling for what we are actually testing, so to speak. It'd be akin to if I claimed (as I do) that the party in office has no meaningful effect on the economy, and you responded 'Uh, yeah they do! Democrats have a majorly positive effect on the economy. Look at this study!' [President only, employment only, excluding recessions]. That survey you linked to is not just presenting the survey. It's presenting a rather specific and odd subsampling.
        someothherguyy 2y ago
        I didn't claim anything. I was just asking where you heard what you heard. I provided a survey that showed a different response. You weren't happy with the population it sampled from. You thought the buckets should be sized differently. You called this p-hacking. I provided another survey from a larger population with a similar response to the first one I shared. You didn't acknowledge it. You still haven't provided a source.

somenameforme 2y ago

No, you're not understanding. Amongst all people that are not having children there are two subsets - those that plan to never have children, and those that are unsure or plan to have children - just not right now. Those numbers Pew is publishing are only from the former. This is what I meant by saying that their headline is most aptly stated as, "[a very slight majority of] people who don't plan to ever have kids say they don't want to have kids." This is essentially what p-hacking is. By organizing variables in a certain way, you can create statistically significant results to say basically whatever you want them to say. A huge tell-tale sign of p-hacking is constraining one's sample in ways that don't really make sense and/or are likely to be misleading. The thing we were discussing is 'why people aren't having kids.' This is very different than why 'people who don't plan to ever have kids aren't having kids.'

someothherguyy 2y ago
> This is essentially what p-hacking is. By organizing variables in a certain way, you can create statistically significant results to say basically whatever you want them to say. Do you know what statistically significant results means?
- somenameforme 2y ago
  Yes, and it doesn't mean "true" as you might think. There's a great demo of p-hacking here. [1] It's based on real data too, which makes it extra fun. Your goal is to prove that one party or another being in power leads to a stronger economy. That statement is almost certainly false as you can quickly see by how awkward it is to "prove" it. But you'll find when you arrange your variables in certain ways you can indeed find statistically significant evidence to "prove" literally whatever you want. But again the one tell-tale sign somebody like me would notice from your study is 'why in the world would you limit it to that sample'? Well because that sample gives you the answer you want! It's science, baby! [1] - https://projects.fivethirtyeight.com/p-hacking/ https://projects.fivethirtyeight.com/p-hacking/
  - someothherguyy 2y ago
    This is a survey. There is nothing under test. Its like saying you are p-hacking an election. I don't understand what you mean.
    - somenameforme 2y ago
      This is the entire point of our dialogue. You claimed my statement that most people [who don't have children] claim they aren't having children because of economic reasons was false. Your evidence was a study that showed a negligible majority of people [who actively don't plan to ever have kids, age 18-49] are simply doings so out of personal choice, and then seemed confused about why this was an inappropriate sampling for what we are actually testing, so to speak. It'd be akin to if I claimed (as I do) that the party in office has no meaningful effect on the economy, and you responded 'Uh, yeah they do! Democrats have a majorly positive effect on the economy. Look at this study!' [President only, employment only, excluding recessions]. That survey you linked to is not just presenting the survey. It's presenting a rather specific and odd subsampling.
      - someothherguyy 2y ago
        I didn't claim anything. I was just asking where you heard what you heard. I provided a survey that showed a different response. You weren't happy with the population it sampled from. You thought the buckets should be sized differently. You called this p-hacking. I provided another survey from a larger population with a similar response to the first one I shared. You didn't acknowledge it. You still haven't provided a source.