Statistics, their use and their limits

The book

The Dutch Centraal Plan Bureau has a well deserved reputation. When political parties write their political programs, the CPB will do the maths on them. As they are independent of much political machinations, they get it "wrong" when their numbers are not the happy numbers that make up the propaganda.

Statistics at the Wikimedia Foundation is a happy story. Erik Zachte started as a volunteer to compile numbers, this became his job and the WMF boasts well presented statistics providing not only the basic facts about its projects, the report cards provide management information that allow for informed understanding of how its projects are doing.

In addition to Erik's work there are scientists and volunteers producing even more numbers. I am particularly grateful for the numbers produced by Magnus that help us build relations with GLAMs.

After such a glowing review, it seems strange that I feel a need to discuss our statistics and their use. I was told that statistics is like a bikini, revealing but not showing everything. Often statistics and consequently their interpretations start from a certain understanding. For the WMF strategy for the next five years the numbers were presented at the chapter meeting in Berlin. When you analyse those numbers, and interpret some of the accompanying comments, the conclusions can be quite different from our current understanding.

An example; the comscore numbers for Indonesia were not available. It is however easy to argue that money invested in growing its Wikipedias are likely to produce more traffic then money invested in the Indian or English language Wikipedias.

It is said that the growth in the English Wikipedia had flatlined. From an economic perspective, it makes more sense for the Foundation to invest in the Russian, the Spanish, the Portuguese or the Indonesian Wikipedias as they are growing. There are however excellent arguments to invest in the English language Wikipedia, they are not made in way that prevents criticism.

A lot of good work has gone in the preparing of the strategy for the WMF. It is sad that it is all too easy and too obvious to find fault. An open discussion of the issues and the strategy may somewhat change future activities. The investment in getting a clue about what to do in the next five years has been worthwhile because the least it offered is a clue of where we want to go.
Thanks,
      GerardM