(Ed: A followup can be found at: The Straw Stereotype, Redux)
One thing that really annoys me is straw stereotypes.
Many people misinterpret, perhaps wilfully, statements that take the form "Group A possess attribute B" (or similar ones) to be absolute declarations (i.e. "All of Group A possess attribute B").
They then find one counter-example and crow triumphantly that the "generalisation" is false.
Let us take one example to see how ridiculous this is:
"Singapore is hotter than Anchorage, Alaska"
This is a pretty uncontroversial statement.
Yet, the highest temperature ever recorded in Anchorage, Alaska was 29 degrees Celsius (data from 1916-present, temperature recorded from Jun-Aug).
Meanwhile, the record low for Singapore was 19.4 degrees Celsius one January (data from 1929-1941, 1948-2011 data).
Therefore Singapore is not hotter than Anchorage, Alaska.
This sort of trivial "proof by contradiction" is not only not very interesting but alien to the way normal people process such statements.
When people say that "Singapore is hotter than Anchorage, Alaska" they do not mean:
"Singapore is always hotter than Anchorage, Alaska, has always been hotter than Anchorage, Alaska and will always be hotter than Anchorage, Alaska - come rain or shine, climate change or the End of the World",
But rather:
"Singapore is almost always hotter than Anchorage, Alaska" or even
"Singapore is usually hotter than Anchorage, Alaska"
In reality no one except people firing cheap shots at "stereotypes" and "generalisation" takes such a totalising and reductive approach to general statements; the only statements of the form "All of Group A possess attribute B" which are always true are tautologies (e.g. "All bachelors are unmarried").