6.1 An example

It is often said that you can prove anything with statistics. And this seems to be true. Let us look at a few examples to illustrate our point. Say for instance we wanted to know what the religious composition of South Africa was. A quick search on the Internet and one book gave me the following results (they will probably be different by the time you read this, but we are trying to establish a principle here):

http://www.geocities.com/TheTropics/8240/stats.htm

• Christian 72 %
• Hindu 2 %
• Traditional religion 20 %,
• Secular 5 %,
• Muslim 1 %

http://www.phrasebase.com/countries/South%20Africa.html

• Christian 68 %
• Muslim 2 %,
• Hindu 1.5 %,
• Indigenous beliefs and animist 28.5 %

http://www.southafrica.net/index.cfm?SitePageID=355

• Christian – 80 %
• Other major religious groups are the Hindus, Muslims and Jews.
• A "minority" regard themselves as traditionalists or of no specific religious affiliation.

South African Christian Handbook (1999–2000)

• Christian 54 %
• Other faiths 11.4 %
• No religion 12.9 %
• Not stated 21.6 %

When we look at the above statistics from four different sources, we are not much the wiser. This, however, illustrates how difficult it is to get accurate statistics about religion (and for that matter about any phenomenon). Statistics should never be swallowed whole! One should always look carefully at the criteria statisticians used to obtain their data. If the manner in which the data was gathered is not reliable then the results should be discarded.