Hacker News Reader

Comment by bqmjjx0kac | Hacker News Reader

[flagged]

To put some numbers here: Upper body strength and grip strength are around 3 standard deviations apart for men vs women, meaning that an average man is stronger than 99/100 women.

Those numbers might be true, but talking about standard deviations seems difficult to visualize or reason about, because knowing the width of the distribution is required, and because the width of the distribution of one sex doesn’t reflect on the other sex, so the standard deviation framing is potentially misleading. Almost all sources talk about the ratio of mean strength of males vs females, and depending on the study and the specific exercise in question, the conclusion is typically somewhere in the neighborhood of 2/3, plus or minus 10%-15%. Saying women have 65% of men’s strength sounds very different and sends a different impression from saying average man is stronger than 99% of women, even though both statements can be true at the same time.

Importantly, those stats you brought up are limited to upper body. When it comes to lower body, the ratio of female strength to male is more like 80% and your stddev number is smaller, so at the very least your conclusion should preserve that qualifier: average man’s grip or upper body strength is higher than 99% of women’s grip or upper body strength. It’s not quite true to generalize to an across-the board statement that avg man is stronger than 99% of women. It appears most studies widely agree the discrepancy is due primarily and simply to body weight / muscle mass; this is all mostly a simple byproduct of women being smaller than men on average.

Googling I see some results on grip strength (GS) saying mean male GS is higher than 97% of females. The graph/study in this article happens to have an even smaller ratio than what you mentioned with male GS at higher than 99.9% of females. But look at the rather large overlap of the distributions (first image). Saying the average male is strong than either 97% or 99% or 99.9% of women one might not realize that each distribution overlaps with something like 30% of the area. (@bqmjjx0kac is correct - the distributions do overlap) https://briefedbydata.substack.com/p/female-vs-male-grip-str...

>It appears most studies widely agree the discrepancy is due primarily and simply to body weight / muscle mass; this is all mostly a simple byproduct of women being smaller than men on average.

But muscle is closely tied to testosterone, something men have much more of. It makes me think of the iconic study [1] where men were compared with/without testosterone injections, and with/without exercise. The men given testosterone that did nothing grew more muscle mass than men strength training in the gym. So it isn’t women being smaller, it is that most men build more muscle laying on the couch than women going to the gym.

[1] https://www.nejm.org/doi/full/10.1056/NEJM199607043350101

Yes, testosterone and other parts of male physiology are what lead to men having more muscle mass on average. Men and women have different average sizes, so we don’t need to debate whether men are larger or women are smaller, both are true. The main takeaway from the muscle mass point for me is just that there’s very little difference in strength between men and women per kilo, when normalizing against lean body mass. It’s just a good thing to keep in mind when taking about these sex differences, since there’s a lot of misleading framing and wrong impressions, right?

You are absolutely correct, and framing makes a huge difference on how such statements are perceived ("women are 65% the strength of men" vs "men are 50% stronger" vs "1 in hundred women stronger than average man").

But pointing out that there is a 30% area overlap glosses over the main point in this context in my opinion: If you are going to have any benchmark ("need to be this strong to safely pilot"), then that benchmark is either going to be trivial for most men or it's going to be almost impossible to meet for almost every woman.

I put this number here because it was quite surprising to me when I learned about this; naively, I would have expected something like 1/10 women to be stronger than an average man, but that is off by a lot.

> If you are going to have any benchmark

This is actually why IMO saying women are 65% as strong as the average man is the better framing. It gives you an immediate, practical, and tangible goal. If the force required to do a given operation is acceptable for men and not acceptable for women, then we don’t need to reduce the force by 100x, we need to reduce it by 1.5x, which is not very much.

Talking about the standard deviation doesn’t give you a usable result, while talking about the ratio of means does.

BTW this argument should be moot. It’s very, very poor safety design to suggest that controlling a vehicle or doing anything else safety critical should require more than 50% of my maximum strength. My own strength varies considerably based on how much I have to use it. I can’t maintain my max strength if I have to do it repeatedly, and it will affect my ability to think if I have to use my max strength in a crisis situation. There simply should not be controls that require more force than most women can apply, not even close.

And women need to be in the top 10% of female grip strength to have stronger grip strength than the bottom 5% of men.

> I challenge you to consider that men and women's strength have overlapping distributions.

This is true. However, most pilots in the early days came from the military, which likely selects towards the stronger end of the men's distributions.

There is overlap, but not a whole lot with the same amount of physical training.

In the military, they have standards for fitness.

The military selected early pilots based on how close they were to average dimensions. They measured dozens of attributes and designed airplanes around the median of each attribute, assuming it would be easy to find people that fit into that design. It turned out that it's really easy to find people that fit into one or two median measurements, but effectively impossible to find someone that fits into even half a dozen median measurements.

Also, averages change over time, as the environment we are raised in changes, so what was designed around an average height in the 1920's required shorter than average pilots by the 1940's.

Here's an article on the topic: https://www.thestar.com/news/insight/when-u-s-air-force-disc...

Here's a book with several chapters going even more in depth: https://www.goodreads.com/book/show/24186666-the-end-of-aver...

IIRC the distribution doesn't actually overlap all that much.

That’s a subjective statement. Here’s a picture, and I’d say there’s a decent amount of overlap. https://briefedbydata.substack.com/p/female-vs-male-grip-str...

>In fact, just 17 of the 2,515 females, or less than 0.1%, have a summed grip strength greater than the median summed grip strength of males.

You’re right, it’s subjective, but that’s not what I would call a decent overlap.

You’re not quoting the overlap, you’re quoting the (IMO) misleading framing that compares extremes. If you look at the picture, roughly one third of the distribution has overlap. It’s not a majority, but it’s certainly not negligible either.

BTW, the 0.1% number may be sampling noise, as similar studies have come up with numbers like 1% or 3%. While they’re all small, 3% is 30x larger than 0.1%. Don’t put a lot of stock in that number; it’s trying to draw conclusions from the noisiest lowest-data part of the distribution, and therefore is highly prone to error.

Specifically you need to be in the top 10% for female grip strength to be stronger than the bottom 5% of male grip strength.

I challenge you to consider that the overlap is small enough to be ignored

Earhart was a dude.

Reply on news.ycombinator.com