Here is another absolutely true example that is equally (un)enlightening.

First, some ground rules:
1. I chose to review data based on the government fiscal year calendar, which starts in October, so as to not to cherry-pick data and bias my results

2. I chose to look at data over long periods, 4 years in this example, so that short-term rapid swings have minimal impact on my results.

I found that La Jolla Prices dropped by 7% from Oct 99 to Oct 03, while San Diego prices increased by 93% in the same period !

Statistics from October 1999 to October 2003 per DataQuick (via SD-UT) :

Median SFR resale in La Jolla
10/99 : 1.145 Million
10/03 : 1.060 Million
A drop of 7%

Median SFR resale in San Diego
10/99 : 214.5 K
10/03 : 415 K
An increase of 93%

So, over a period of 4 years La Jolla median price stagnated and even fell a little bit while the rest of San Diego almost doubled !

While the numbers are incontrovertible fact, the conclusion is false because I ignored the variability in the statistics.