2. The Biased-Sample Fallacy The Fallacy of the Biased Sample is committed whenever the data for a statistical inference are drawn from a sample that is not representative of the population under consideration. The data drawn and used to make a generalization is drawn from a group that does not represent the whole. Here is an argument that commits the fallacy of the biased sample: ln a recent survey conducted by Wall Street Weekly, 80% of the respondents indicated their strong disapproval of increased capital gains taxes. This survey clearly shows that increased capital gains taxes will meet with strong opposition from the electorate. The data for the inference in this argument are drawn from a sample that is not representative of the entire electorate. Since the survey was conducted of people who invest, not all members of the electorate have an equal chance of being included in the sample. Moreover,persons who read about investing are more likely to have an opinion on the topic of taxes on investment different from the population at large. |