Average data - means and medians

We calculate two averages - means and medians - for many fields in our database across countries, regions and globally for market sets and sector sets.

Outlier data can have a big impact on the value of mean averages. For example, one company with an anomalously high P/E ratio of 5000 could increase the mean P/E ratio of the S&P 500 by a full 10 points. We winsorize our averages at the 3% level in order to minimise this effect.

The difference between means and medians can be large due to skewed data sets, which is normal in financial data. Given that many individual investors prefer to invest in small caps, medians can often be far more appropriate as a measure of central tendency.