# Histograms - mean or median?

Is it better to use the mean or median when looking at averages in a histogram? I'm thinking median, but I'm not quite sure why.

-
The answer is subjective. It entirely depends on what you wish to report and what the data looks like. If possible, report both. – Inquest Dec 13 '12 at 20:18
My teacher asked us this question and we had to decide on one or the other, not both; the data is a bunch of speeds of cars at a point. (10-15mph, 40-45mph, etc) The data is a bit all over the place, but I'd have thought that wouldn't change the rationale of choosing mean/median. – kerbtr Dec 13 '12 at 20:26
Depends on what you are trying to determine. Why do you care about the speeds of the cars at that point? – Thomas Andrews Dec 13 '12 at 20:32
We were given the data, and we're merely trying to get the average of the speeds. No reason is given; just meant to say why the estimate of the mean or the median is a better representation of the average speed. – kerbtr Dec 13 '12 at 20:37
Just to get your thoughts running, Consider this: If I am calculating the average cost of cars and I walk into a suburban showroom and come out. My average would probably be a good estimate for how much you would pay in general for a car. Here, averaging makes sense. Now, consider that I go into a suburban showroom and then to a showroom in Beverly Hills where I am told prices of Ferrari and Porsche. If I now report the average, it wouldn't be a good estimate for how much I expect to pay for a general car. (Imagine the data: 15k 15k 15k 100k) The median however would be a robust statistic. – Inquest Dec 13 '12 at 20:58
show 1 more comment