How to Lie with Statistics

Topics: Arithmetic mean, Sample size, Lie Pages: 5 (1969 words) Published: June 4, 2012
How to Lie with Statistics Summary

Sampling can completely distort data and mislead the reader. The sample is supposed to represent the general population, however this is rarely the case because of the biases that lie with in sampling. For instance, the people that you interview could tend to lean towards one specific group of people. In the Yale example on page sixteen, the people that did not make a lot of money could be harder to find and interview than the rich people that have been successful. The richer people are going to be more likely to be found and answer the questionnaire, which will therefore skew the data. In addition, people could also lie about their income; some may overstate it and others could understate it. Furthermore, this was also the case in the example of the Literary Digest, their poll with regards to the election was not accurate, because the only people that they could reach to poll were the rich, because they had telephones and magazine subscriptions, and that particular group of people was biased towards the Republican Party. In many other cases, biases can be created when the person that is being interviewed is not telling the truth. We have no way of telling if the reports are from honest people. Moreover, people that are polling others could also manipulate data, because they are more likely to lean towards a certain group of people when choosing whom to give the questionnaire. There are several biases that could leave the reader to believe something that is not true. The presenter may state that the average of the general population is x, however it may only be represent able of a certain group in the general population.

Biases within sampling are not the only way people lie with statistics, they also can choose a certain type of average. Using the same data points and calculating a different average and not specify which average was used can completely...