Current Events > Sports medicine overrun by a statistical methodology that has no basis

(+) Yeah! (+)

https://fivethirtyeight.com/features/how-shoddy-statistics-found-a-home-in-sports-research/

Very interesting article.

Excerpt:

At first blush, the studies look reasonable enough. Low-intensity stretching seems to reduce muscle soreness. Beta-alanine supplements may boost performance in water polo players. Isokinetic strength training could improve swing kinematics in golfers. Foam rollers can reduce muscle soreness after exercise.

The problem: All of these studies shared a statistical analysis method unique to sports science. And that method is severely flawed.

---

I quit stretching before workouts and I'm less sore afterward
---

PoopPotato posted...

I quit stretching before workouts and I'm less sore afterward

all thru high school and collegiate athletics coaches and trainers told us to stretch after workouts, not before

warming up > stretching
---

Check out Dr. Mike on YouTube.

I forget which video, but he talks a out stuff like that and work out stuff are basically useless.
---

Read the article, basically someone came up with an EZ-analysis machine, which is wholly unreliable compared to standard analytical methods but requires less data and generally less effort. It has come under fire by a considerable number of statisticians who all say it'll create misleading results but there's a passionate population of users that refuse to give it up.
---

Hopkins views this as a benefit because it means that more studies turn up positive findings worth publishing.

lol
---

Balrog0 posted...

Hopkins views this as a benefit because it means that more studies turn up positive findings worth publishing.

lol

Indeed, it's either pure arrogance or pure naivety, or both.
---

COVxy posted...

https://fivethirtyeight.com

FAKE because this is by the same WEBSITE that predicted HILLARY would WIN LOL
---

Konflikt posted...

COVxy posted...

https://fivethirtyeight.com

FAKE because this is by the same WEBSITE that predicted HILLARY would WIN LOL

To be fair, they are the experts on fake statistical analysis.
---

538 had the highest chances of a Trump win out of anyone...
---

All of which has nothing to do with this topic (and that criticism relies on not really understanding the statistical processes 538 was using...).
---

Up.
---

Its not about finding statistical significance its about the magnitude of the change and is the effect a meaningful result.

omg
---

Balrog0 posted...

Its not about finding statistical significance its about the magnitude of the change and is the effect a meaningful result.

omg

Yeah, I mean, the idea in-and-of-itself isn't really that bad, as explained in the Sainani article (https://journals.lww.com/acsm-msse/Abstract/publishahead/The_Problem_with__Mag nitude_Based_Inference_.96919.aspx, remove space between Mag and nitude).

Defining an effect size before hand that would be too trivial to matter and then seeing if this trivial effect size falls in or out of the 95% confidence intervals seems perfectly fine, in fact a bit more rigorous than simple NHST. It's just what they do to sort of cheat the system by misinterpreting 95% confidence intervals which is the issue.

It's funny, that in the days of statisticians telling people to simply drop NHST, this drops and people are like "well, you could at least do proper NHST.
---