Height

[Guest]

I bet you have a hard time finding button up shirts that fit right.

[Guest RayG]

Yup. I don't know how many times I've gone to get a dress shirt, the guy whips out a tape measure to wrap around my neck, and I have to tell him to ignore my neck and measure my arms/shoulders instead. I typically take a size or a size and a half larger than my neck size, and either end up with something HUGE in the chest/waist, or just select a short-sleeved shirt instead. It's just a joy shopping for a new jacket too. To get one to fit nicely in the sleeves, I can nearly fit another person inside the body of it. (a really skinny person of course)

My wife always said I had the widest shoulders she'd ever seen on anyone. When I remarked to a friend earlier this year that I was packing on too much weight and getting fat, she commented that I wasn't fat, I was just wide like a fridge.

RayG

[Incorrigible1]

Ray, I never considered you an "******." (Oops, your avatar.) Pencil-necked geek, on the other hand..........

Edited October 30, 2010 by Incorrigible1

[Guest midnightwalker1]

Just tossing this out there because I thought it was very interesting. I found a word document online from "Anthropometric Guidelines - Department of Defense Ergonomics Working Group, A table of measurments and guidelines developed and published by the DoD Ergonomics Working Group."

The direct link is found here, on a page filled with links to info on Anthropometry.

Unfortunately there's no data included to indicate how many subjects were measured, but here's a pic with the info I found most interesting:

According to their measurements, the average height for a man (line 9) is 68.7 inches, with only 5% of the population greater than 72 inches tall. If I stand really straight, I top out at 72 inches tall.

Line 24 shows that the average shoulder breadth on a man is 17.9 inches, with only 5% of the population having a shoulder breadth greater than 18.8 inches. My bare shoulders are more than 2 inches wider at a whopping 21.5 inches, and I have no idea what percentage of the population would fit into that category.

The last category that I found interesting was foot length, with the average length being 10.5 inches, with only 5% having feet greater than 11.2 inches in length. My feet are 11.5 inches.

So, while my height is right at the cutoff for the 95th percentile, and my feet are slightly longer, my shoulders are a great deal larger, and demonstrates the problem with trying to take average measurements and applying them to the general population.

RayG

Ray,

You'll find an exception to any rule. That's why you lean on the mean average and the law of large numbers to validate your findings in a statistical model. You don't get hung up on the minutia of exceptions. I'd rather be right 80-90 % of the time (with minor % error ratio) than throw out the model in its entirety. Bottomline, we don't have a statistical sampling of sas and we're guessing but the statistical premises seem valid. Valid comment though and worth discussing.

Edited October 30, 2010 by midnightwalker1

[Guest]

Makes sense and seems solid statistically (law of large numbers).

midnightwalker1,

I wouldn't call my analysis "solid statistically." I merely make a comparison to anthropometric data from two populations. Odds are that the members measured from the populations were not randomly selected. Nonrandom selection may impose biases on the data. Both populations represent but a portion of the North American human population, which may bias the averages and certainly biases the variances, such that reported standard deviationis underestimate relative to the values that would characterize the broader population of interest.

There are certainly better, more applicable anthropometric databases (one is available for a mere $10,000, which would likely be representative of the broad North American population). All models are wrong; some models are useful. I believe the comparisons I offered, while wrong, are sufficiently useful to help answer the question posed in the original post. The adequacy of the foot length-to-height conversion I presented seems supported by similar results some others have offered herein. Everyone can test the conversion or others by measuring his own foot and seeing how close the conversion gets to his own height.

The biggest uncertainty in the original post revolves around the size of the footprint in question. While a human foot length up to 18.5in or so has been documented, there are very few humans with 16.5in long feet. Unfortunately, no measurements of print breadth, at both the ball of the foot and the heel, are provided. I understand that purported bigfoot prints are substantially wider than human feet, especially at the heel. If the print in question is from a bigfoot, I estimate that a human conversion may overestimate the bigfoot's height (for a given height, a bigger critter will require bigger feet to carry the bigger mass).

Pteronarcyd

[Guest]

According to their measurements, the average height for a man (line 9) is 68.7 inches, with only 5% of the population greater than 72 inches tall. If I stand really straight, I top out at 72 inches tall.

Line 24 shows that the average shoulder breadth on a man is 17.9 inches, with only 5% of the population having a shoulder breadth greater than 18.8 inches. My bare shoulders are more than 2 inches wider at a whopping 21.5 inches, and I have no idea what percentage of the population would fit into that category.

The last category that I found interesting was foot length, with the average length being 10.5 inches, with only 5% having feet greater than 11.2 inches in length. My feet are 11.5 inches.

So, while my height is right at the cutoff for the 95th percentile, and my feet are slightly longer, my shoulders are a great deal larger, and demonstrates the problem with trying to take average measurements and applying them to the general population.

RayG

RayG,

First, allow me to thank you for your military service, Master Corporal.Because you couldn’t find the specific source of the anthropometry in the table you posted, let’s look at the NASA compendium to see if we can’t find a source that might apply to you. NASA numbers each study, and each measure.

In 1974 a survey of 565 Canadian military personnel was conducted (study 62). It included 32 anthropometric measurements, including the three measures you focus on:

shoulder breadth (aka bideltoid breadth, measure 122)

• mean of 18.81in
• standard deviation of 1.08in

foot length (measure 362)

• mean of 10.41in
• standard deviation of 0.47in

stature (measure 805)

• mean of 68.92in
• standard deviation of 2.49in.

Assume one found your footprint in the Canadian wilderness. Using the Canadian military anthropometrics, let’s see how well one would predict your stature and shoulder breadth.

First, simple deterministic predictions based on using average values only:

• predicted stature of 76.1in (5.7% too high)
• predicted shoulder breadth of 20.8in (3.4% too low).

Predictions accurate within +/-6% relative error, under the circumstances, seem pretty good.

Now more involved predictions based on propagating the natural variability using probability distributions. Two additional assumptions are required. Mine are:

• the data for the three measures are lognormally distributed, and
• each of the three measures is positively and strongly correlated with one another (i.e., Spearman’s rho = 0.87).

The latter assumption accounts for the observation that guys, such as you, with large feet tend to be tall and broad.

Let’s define the range of interest being that encompassing 99.9% of the participants in the study, the probabilistic predictions are then:

• predicted stature of between 69.6in to 83.8in (your 72in stature is at the 1.2
  ^nd
  percentile of the distribution of predictions), and
• predicted shoulder breadth of between 18.5in and 23.9in (your 21.5-in stature is at the 88.5
  ^th
  percentile of the distribution of predictions).

The prediction of your height is getting close to the lower fringe, but both predictions appear to be successful, if one regards +/-6% relative error as successful. The NASA summary of study 62 indicates the original report provides a correlation matrix for the various measures. Thus, if one obtained the original report, one would not need to assume a degree of correlation for each pair of measures.

Pteronarcyd

[gigantor]

What is Spearman’s rho?

Thanks.

p.s. Thanks RayG for your service.

Edited November 1, 2010 by gigantor

[Guest RayG]

Are Canadians in the military wider than average?

It's interesting how the measurements stack up against each other.

While the standard deviations for all three measurements are close, and the mean measurements for height and foot lengths are very close to one another, the mean numbers for shoulder breadth are not.

mean height: 68.7 (chart) 68.9 (military) -- very small difference

mean foot length: 10.5 (chart) 10.4 (military) -- even smaller difference

mean shoulder breadth: 17.9 (chart) 18.8 (military) -- a huge difference compared to the other measurements, nearly one full inch

The 18.8" mean from the Canadian military survey equates to the 95th percentile in the chart I provided. So the average width was 18.8" for the Canadians, while on the chart I sourced, only 5% of the men measured had shoulders 18.8" or wider. This image shows the mean, and the 95th percentile.

Was there any indication what the 95th percentile was for that Canadian military survey? And just how big do shoulders get anyway?

RayG

[Guest]

What is Spearman’s rho?

gigantor,

Spearman's rho is the non-parametric equivalent to the parametric correlation coefficient, Pearson's r. Parametric methods require an assumption of distributional form for the data; non-parametric methods do not. To calculate Spearman's rho one rank-transforms both the x and y data (i.e., convert them from measured values, such as foot length and height in inches to ranks, with 1 being the smallest measure, and n being the largest of the n data points). Tied ranks are averaged (Microsoft Excel's RANK function does not average tied ranks, so one needs to adjust them). Once the data are ranked, calculate Pearson's r on the ranks (e.g., using Excel's CORREL function) and you have obtained Spearman's rho. My simulation software uses Spearman's rho to define correlations among input variables.

Of course, I did not calculate the correlation (insufficient data was available). I merely subjectively assumed a correlation.

Pteronarcyd

[Guest Angie]

Great responses from everyone! Thanks for the thread. This kind of conversation makes me feel like there's a little bit of progress going on in this subject.

With all due respect to everyone, Im perfectly comfortable using Chris Bs calculations.

Thanks again!

[Guest]

RayG,

Are Canadians in the military wider than average?

The vast majority of the large number of anthropometric surveys summarized by NASA are for various military populations around the world. Compared to the rest of the world, Canadian servicemen do appear, on average, to be rather wide. Some US military subpopulations are, on average, a bit wider, but this could well be an artifact of the US surveys being dissected into multiple surveys -- certain parts of a whole must be larger than the average.

It's interesting how the measurements stack up against each other.

While the standard deviations for all three measurements are close, and the mean measurements for height and foot lengths are very close to one another, the mean numbers for shoulder breadth are not.

mean height: 68.7 (chart) 68.9 (military) -- very small difference

mean foot length: 10.5 (chart) 10.4 (military) -- even smaller difference

For most practical purposes, I'd consider such differences to be zero.

mean shoulder breadth: 17.9 (chart) 18.8 (military) -- a huge difference compared to the other measurements, nearly one full inch.

The 18.8" mean from the Canadian military survey equates to the 95th percentile in the chart I provided. So the average width was 18.8" for the Canadians, while on the chart I sourced, only 5% of the men measured had shoulders 18.8" or wider. Was there any indication what the 95th percentile was for that Canadian military survey? And just how big do shoulders get anyway?

RayG

I looked for an alternative dataset because I interpreted your initial post to indicate you didn't know what population was represented by the tabulated data and could not find how many people were included in the survey. Also, I wasn't sure what "shoulder breadth" meant. In the NASA publication I initially found no reference to a shoulder breadth measurement (it's in a pdf format that I can't search electronically). Possible shoulder breadth measurements appeared to be biacromial breadth (measure 103), bideltoid breadth (measure 122), and interscye (measure 506). NASA provides a diagram showing how each measure is made. Biacromial breadth is measured from the top of one upper arm across to the top of the other upper arm. Bideltoid breadth is measured from the outside of one upper arm (where the deltoid muscle is) across to the other. Interscye is measured from one armpit to the other.

I recall you posting a photo of you measuring your back with what appeared to be a checkerboard of known size. I couldn't tell what exactly you were measuring, and the checkerboard was in front of you, which will exaggerate any measurement from the photo (the checkerboard should lie across your back, and you should have someone else holding it).

I settled on bideltoid breadth when I stumbled upon a part of the NASA publication defining that to mean shoulder breadth.

The NASA data summary lists the 95th percentile of bideltoidal breadth to be 20.7in for study 62. However, the summary of study 62 says the original report lists only mean, standard deviation, and 25 percentiles, which I assume means the 25th and 75th percentiles. The NASA report states that 95th and other percentiles were sometimes calculated from the study mean and standard deviation assuming a normal distribution -- perhaps not a terrible assumption, but likely not as good as a lognormal assumption. However, with the standard deviations being small in relation to their corresponding averages, it likely doesn't make much difference whether one assumes lognormality or normality over most ranges.

Using the published average and standard deviation for bideltoidal breadth I calculate a 95th percentile of 20.6in for either lognormal or normal assumptions. The upper end of the range I reported represents the 99.95th percentile, 22.5in for lognormality, 22.4in for normality.

How wide can shoulders get? Let's go nuts and invoke the 6-sigma philosophy and calculate the value at the 99.999983rd percentile. For a normal assumption that value is 24.3in; for a lognormal assumption, it's 25.2in. I doubt either assumption holds that far out on the distribution tail, so take these numbers with a grain of salt.

Pteronarcyd

[Guest RayG]

Compared to the rest of the world, Canadian servicemen do appear, on average, to be rather wide.

Interesting. Any thoughts as to why this might be?

For most practical purposes, I'd consider such differences to be zero.

I agree.

I recall you posting a photo of you measuring your back with what appeared to be a checkerboard of known size. I couldn't tell what exactly you were measuring, and the checkerboard was in front of you, which will exaggerate any measurement from the photo (the checkerboard should lie across your back, and you should have someone else holding it).

Yeah, this is the one:

I wasn't trying to measure anything, it was originally posted to show how something can appear to be larger than it really is. My actual shoulder width is 21.5", yet they can appear to be nearly 27" across even from a short distance. As there were only two of us, me holding the two chessboards against my chest, and the other person taking the photo, there was no other option for the placement of the chessboards. There were two chessboards, so it would have taken four people to take the same photo with the chessboards across my back. As the chessboards were flat against my chest, I felt there was minimal exaggeration between their width and mine.

RayG

[Guest]

RayG,

I don't know why Canadians might be broader than normal. Americans tend to be broad, too, so it may just be that North Americans are better nourished, on average, than the rest of the world. Keep in mind that Canadian military personnel (in general, all North Americans) are likely somewhat broader today than they were 3.5 decades ago.

That's the photo I referred to. Part of the reason the wider pair of boards appears to be narrower than they are (which makes your back seem broader than it is) has to be the boards are at a different distance from the camera lens. Think about it -- if you put a 12in ruler far enough out in front of you, with the right lens you could "demonstrate" your shoulders to be 5ft wide if you wanted to.

If you did want to measure your shoulder breadth, you'll want to take a peek at the NASA compendium to see the rules for how to do each specific measure. For example, I believe interscye is measured by a tape along the curve of the back, while bideltoid and biacromial breadths are measured with an anthropometer (a big caliper; http://www.nexgenergo.com/ergonomics/lafayettea.html#Sec1). In addition, a specific posture is assumed for each measurement. The photo accompanying the description of the large anthropometer looks to be showing someone's bideltoid breadth being measured, except he should have his shirt off.

Pteronarcyd

[Guest midnightwalker1]

Interesting. Any thoughts as to why this might be?

I agree.

Yeah, this is the one:

I wasn't trying to measure anything, it was originally posted to show how something can appear to be larger than it really is. My actual shoulder width is 21.5", yet they can appear to be nearly 27" across even from a short distance. As there were only two of us, me holding the two chessboards against my chest, and the other person taking the photo, there was no other option for the placement of the chessboards. There were two chessboards, so it would have taken four people to take the same photo with the chessboards across my back. As the chessboards were flat against my chest, I felt there was minimal exaggeration between their width and mine.

RayG

Ray you're a friggin haaauuusss.

[Guest RayG]

That's the photo I referred to. Part of the reason the wider pair of boards appears to be narrower than they are (which makes your back seem broader than it is) has to be the boards are at a different distance from the camera lens. Think about it -- if you put a 12in ruler far enough out in front of you, with the right lens you could "demonstrate" your shoulders to be 5ft wide if you wanted to.

Except I didn't hold the chessboards at arms length, I had them pressed against my chest, so the extra distance from the camera shouldn't be great enough to cause more than minimal distortion, if any. The best way to measure would have been to have the chessboards disect my body, but that wasn't desirable.

It's the equivalent of painting a wide strip on the wall and then having me stand in front of it, with my back pressed against the wall.

If you did want to measure your shoulder breadth, you'll want to take a peek at the NASA compendium to see the rules for how to do each specific measure.

Yes, the method I settled on was to stand with my back flat against the wall and have someone put a pencil mark at the top of each arm at the edge of my shoulder. I then used a tape measure to measure the distance between each mark, and it was 21.5".

Ray you're a friggin haaauuusss.

Ha, I got the width, not the height. I'm a runt compared to my three brothers, and three of my sons, they're all 3-4 inches taller.

Hell, I've even got a daughter who's 5'11". It's like being surrounded by basketball players.

RayG