[Zenst]

A few ways when you use averages and only use one type of average - mean average without including mode or median average types.

The age of on the road vehicles can be skewed by new cars. The age of scrappage can equally be skewed by a few new cars written off and as such, scrapped. Then we do not know average time of the road, the transition time of sitting in a garage/off-road and then scrapped a few years later. That could be a short period or decades. That again would skew averages.

As you can see - I have a pet hate about averages when they just use the mean average. Without knowing the other two averages or complete access to the data-sets - a biased perspective can and will be the outcome more often than not.

It's like a company employing 100 people - 99 earn 10k a year and one earns 1 million a year. The company can and often will say that their employees earn on average 19.9k a year. When the median and mode forms of averages would both in this instance yield an answer of 10k.

I'd be most happy if any use of averages has to include all 3 forms of averages - after all, we do teach them in school's, let's use them and save so much confusion, bias and statistical abuse.

https://en.wikipedia.org/wiki/Average

[billforsternz]

I like your thinking. I've always thought mean, median and mode should be thought of as different possible "average" definitions. But common usage always seems to have average == mean, and the other two as slightly less important ideas. This Wikipedia article pleasingly implies my (and your) preferred model is actually the correct one.