# 41. Over-generalisation ## 41.1. Definition Over-generalisation occurs when a statistic is used to arrive at a conclusion about a whole population based on data from just a particular subset which is not a representative sample of the whole population. For instance, suppose a researcher walks past an orchard and all the apples he sees are red. It is summer. The researcher could conclude that all apples are red in summer. The statement that "All apples are red" would be an instance of over-generalisation because the original statistic was true only of a specific subset of apples (those in summer, of a certain variety, in a certain place), which is not expected to be representative of the population of apples as a whole. A real-world case of over-generalisation can be observed in modern polling techniques, which in some countries prohibit calling mobile phones for over-the-phone political polls. This restriction excludes an entire class of people who do not possess a fixed telephone line or are not at home very often, etc. For example, young people are much more likely to have a mobile phone than a fixed line. Young people are likely to be more liberal than the general population. Young people who do not have a fixed line telephone are also likely to be more liberal than their demographic group as a whole (less fixed to an address). So these polls effectively exclude many participants that are likely to be more liberal. People who are included in the poll have fixed lines. They may possibly be older, more permanent in terms of address, and more politically conservative than the "mobile-phone only" group. On the other hand, they may also be poorer and more transient than the fixed line telephone owners. At any rate, the two groups will be demographically and politically different. Thus, a poll examining political voting preferences of young people, using only fixed line telephone polling, could not claim to be representative of true voting preferences across a whole population. It would be an "over-generalisation" because the sample used is not representative of the population as a whole, having excluded a large segment of the population who don't possess fixed telephone lines. Over-generalisation often occurs when information is passed through non-technical sources. It is very popular with the mass media which prefer to ignore nuances in favour of good, snappy headlines. ## 41.2. Persistence Low to High. ## 41.3. Accessibility High - We can all over-generalise a very specific study result into huge tracts of the population. ## 41.4. Conditions/Opportunity/Effectiveness Over generalisation is not that effective because it is almost always possible to find and show obvious and absurd exceptions to an over-generalisation, thus undermining the manipulative assertion and the manipulator. Opportunities for over-generalisation abound. However, as a manipulative technique it only works properly when it is credible to the target audience. ## 41.5. Methodology/Refinements/Sub-species None known. ## 41.6. Avoidance and Counteraction Whenever someone attempts to generalise the particular, red lights should start to flash. Regardless of the case, anyone attempting to generalise had better have a really well-developed argument to hand and a set of data to demonstrate that a generalisation is logically and statistically valid and reasonable.