77
More than half of adults in U.S. say they lack basic statistical understanding (www.psu.edu)
submitted 5 days ago by to c/science@lemmy.world
43 comments fedilink hide all child comments
 

cross-posted from: https://piefed.world/c/scientificarticles/p/1359566/more-than-half-of-adults-in-u-s-say-they-lack-basic-statistical-understanding

you are viewing a single comment's thread
view the rest of the comments
[–] 5 points 5 days ago* (last edited 5 days ago) (1 child)

How many of you could fully explain how a p value is calculated (this may be the wrong sub to put this in, but I'll let my point stand)?

How much do you understand about statistics and p-values

I could tell you how it's used. I'd have to look up how to calculate it these days. I could tell you that's a shitty way to get to 67% of something using an ambiguous question that a high schooler could pick apart.

Shitty questions in, shitty conclusions out 🤷‍♂️

  • permalink
  • fedilink
  • source
  • hideshow 2 child comments
    •
  • [–] 4 points 5 days ago (1 child)

    You don’t need to look up how a p value is calculated, the calculation depends entirely on your model and hypothesis assumptions.

    The best way I’ve heard a p value described was “the degree to which the data are embarrassed by the null hypothesis”. It’s the most succinct, correct, and easy to remember description.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 2 points 5 days ago (2 children)

    P value just tells you how likely the difference between two data sets is just random chance. Having said that, there are many ways to p hack to achieve significant values.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 4 child comments
    •
  • [–] 1 point 5 days ago

    I’m probably being too pedantic, but that’s exactly correct, though it also captures the spirit and practically most common cases.

    Just change “two datasets” to “a statistic”, because hypothesis tests can be pretty general.

    I don’t love P values, they’re not quite an intuitive probability, they bake in a lot of assumptions, they can be deceptively over or under sensitive, and they’re hard to explain. As a tool they’re useful but they always need context and planning.

  • permalink
  • fedilink
  • source
  • parent