Statistical special pleading is a logical fallacy which occurs when the interpretation of the relevant statistic is "massaged" by looking for ways to reclassify or requantify data from one portion of results, but not applying the same scrutiny to other categories.

Examples:

  • Studies have shown that 5% of patients with foobar disease die within a year. We have shown that when patients are treated with supermed, only 3% of those aged 60 or younger die within a year. Therefore supermed is an effective treatment.
  • A study of teen-age gang members has shown that 40% of those who are atheists have been convicted of a violent offence. This is hundreds of times the number of people in the general population who are convicted of such offences. This clearly shows that atheism leads to violence.

See also faulty generalization for other fallacies involving generalization.