Thread: Average age of straight shavers?

Originally Posted by Jimbo

Mean = average; Mode = most frequent; Median = rank the ages and take the middle age.

Std. Dev = measure of dispersion around the mean
= square_root(1/(n-1) * sum[(Age_i - Mean)^2)

Skew = measure of magnitude and direction of deviation from symmetry.

Here's the results of a quick analysis of the data so far (just ran the data in the posts through one of my stats packages):

Attachment 3045

Attachment 3046

Attachment 3047

The age distribution looks approximately normal to me. Both the number of posts and number of years spent using a straight are highly right skewed.

No obvious relationship between age and number of posts, or age and number of years spent using a straight.

Probably not surprisingly, there's a significant linear relationship between the natural logarithm of the number of posts and the natural logarithm of the number of years spent using a straight (p = 0.002)- diagnostics show the model is reasonably adequate, although the R-square is low... But in any event, the longer you've been shaving with a straight the more posts you make, on average. This could either be evidence in favour of the idea that those with more experience have more information to pass on, or that the older you get, the more you talk. [Edit: Actually, it's more likely an indirect effect of SRP - people who've been using a straight longer probably discovered SRP earlier than others, and so have more posts...]

Of course, this is a self-selected sample...

James.

Once again for those of us who only speak English?

X

Originally Posted by Jimbo

Attachment 3047

HEY! I can see my house from here.

X

Originally Posted by xman

HEY! I can see my house from here.

X

...

Originally Posted by xman

Once again for those of us who only speak English?

X

OK - sorry X. There's a statistically noticeable relationship between number of posts and number of years spent using a straight. ie the longer you've been using a straight, the more posts you've made, on average. Which I guess is basically a no-brainer.

The form of the relationship is interesting though. Because both the number of years spent using a straight and the number of posts are very right-skewed (eg think Ilija's post-count) the natural log of these variables tends to be a bit better behaved (statistically). The fact that there's a straight line relationship between the logged variables actually means a multiplicative model on the original variables of the form:

Years = A * posts^B

James. (Sorry - It's clear on re-reading that I do not have the capacity to speak in plain English )

That helps some though. What does the standard deviation and the skew mean again?

In the 'years' part of the three way graph, you've got it going from -2 through 0 to 2. How does one straight shave for -2 years?

X

Originally Posted by xman

That helps some though. What does the standard deviation and the skew mean again?

X

I know they used to mean a headache for me when trying curve fitting with data in labs but since uni I've done my best to almost forget this. Thank you for bringing it to the forefront of my mind again and, no doubt, disturbing my sleep for the next few nights.

Originally Posted by xman

That helps some though. What does the standard deviation and the skew mean again?

In the 'years' part of the three way graph, you've got it going from -2 through 0 to 2. How does one straight shave for -2 years?

X

Standard deviation is a measure of how dispersed the data are, with reference to the average. So you work out the average age. Then, you work out how far each individual age is from the mean (by subtracting the mean from each age), square those differences and add them up. Then, find the average of this number by dividing by the number of ages you collected (basically).

So the bigger the standard deviation, the more spread out the ages are.

Skew is a measure of how non-symmetric the distribution is. If you look at the age graph (2nd attachment) you'll notice the blue age histogram is reasonably bell-shaped and symmetric. So age would have a small skew - depending on how the skew is calculated, the skew of age would either be close to 0 or close to 3. On the other hand, both the number of posts and years spent using a straight are highly right skewed - the distribution of these variables tails off noticably to the right indicating a few very large measurements (eg your and Ilija's posts) compared to the rest. Left skew would be if the distribution tailed off to the left. Personally I prefer using graphics to look at skew, rather than numbers to measure skew.

That three way scatterplot is actually of the log of years and posts. Anyone who's spent less than a year shaving with a straight will have a fractional "years" value (eg 0.5 for six months etc.) When you log a fraction you get a negative result (eg. log(0.5) = -0.69)

James.

52 here. been at it less than a year.

Boy, this takes me back to Statistics 101 at University....and, if I remember correctly, this is where we decide that, from here on, NO ONE over the age of 42 or under the age of 34 should post...otherwise, it will screw up the model and we'll have to re-calc it....






....OK....let's drink beers!

-whatever

-Lou

Originally Posted by scarface

Boy, this takes me back to Statistics 101 at University....and, if I remember correctly, this is where we decide that, from here on, NO ONE over the age of 42 or under the age of 34 should post...otherwise, it will screw up the model and we'll have to re-calc it....






....OK....let's drink beers!

-whatever

-Lou

Don't drink beer yet - there's an exam on this at the end