Why Face Masks Don’t Work, According To Science - Ben Swann

submitted by fred_red_beans from youtube.com

sorted by:
top
newinsightfulfun
you are viewing a single comment's thread.

view the rest of the comments →

[–]whereswhat 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (3 children)

I appreciate your attention to detail but I disagree with your conclusion:

I think that Canini et al. which showed no significant difference, which McIntyre say may be underpowered, is more compelling than a sketchy post-hoc analysis on a slightly higher number of subjects.

McIntyre already had shown a statistically significant difference between the groups. Therefore, a post-hoc analysis is required to optimally constrain the empistemic error. Failure to do so would leave unchecked error because some of the control group members actually did have masks on for a small part of the time and some in the masks group did not actually wear a mask. It doesn't mean that everyone in the masks groups only wore a mask for 1 hour. It was intended to ensure that the different groups were properly defined and that the binning did not skew the results in a significant way.

Canini et al. was clearly underpowered. Here is the conclusion summary from the article itself:

This study should be interpreted with caution since the lack of statistical power prevents us to draw formal conclusion regarding effectiveness of facemasks in the context of a seasonal epidemic.

Statistical power is closed form quanitative metric. Here we have one study (McIntyre) which has very high power and has been validated with a post-hoc analysis and another (Canini) which has low power and therefore does not warrant post-hoc analysis. I am going to trust the former over the latter, just like the authors suggest.

[–]noice 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (2 children)

Are we looking at the same paper?

The Kaplan-Meier curves showed no significant differences in the outcomes between two arms (p>0.050; figure 2).

There was no association between mask use by the index cases and rates of infectious outcomes in household members (table 3). Although the risks of CRI, ... ILI..., and laboratory-confirmed viral infections were lower in the mask arm, the difference was not statistically significant.

It's not until they conduct the post hoc analysis that they see a statistical difference.

I'm not suggesting that everyone in the mask group only wore the mask for 1 hr. I'm saying that their study as designed found no statistical difference, and then they found a way to achieve statistical significance by afterward moving more (based on the numbers, non-event) subjects into the mask group.

[–]whereswhat 1 insightful - 1 fun1 insightful - 0 fun2 insightful - 1 fun -  (1 child)

I believe we are.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5223715/

The part you quoted only pertains to household members who were likely spending time in thoroughly contaminated environments. Even with that being the case, the data did show a benefit from the masks but not a strong enough correlation to accept without post hoc analysis.

Conclusions

The study indicates a potential benefit of medical masks for source control, but is limited by small sample size and low secondary attack rates. Larger trials are needed to confirm efficacy of medical masks as source control.

This is why we must rely on more recent reviews that collate data from many sources. Like the original paper I cited.

[–]noice 2 insightful - 1 fun2 insightful - 0 fun3 insightful - 1 fun -  (0 children)

Yeah we're talking about the same paper. I stand by my statements, and I think that means I fundamentally disagree with your interpretation. McIntyre et al.'s review touting McIntyre et al's study is recent, yeah, but it's not as conclusive as you imply.