50% of all women are lesbians and 75% of all women are Caucasian. Is that true? Not exactly, it’s a projected statistic based on a sample group of 4 of my friends. Statistics aren’t precisely lies, but they can be extremely misleading when presented as if they were absolutes.
Statistics are estimates, not facts. It is disingenuous to use “is” or “are” with any statement that is not an absolute measurement. 50% of my my rollerskate wheels are purple (8 are purple, 8 are blue); from that, I could estimate that 50% of all rollerskate wheels will be purple, but I can’t rightly say that “50% of all wheels are purple” without any caveat.
There are two main ways that statistics can be skewed: one is to use a too-small sample group to represent a large population (for example, 1000 people may seem like a lot to you personally, but compared to the population of this country [about 304 million], it’s a drop in the bucket – only about 0.0003%), another is to use a non-representative population (for example, if you’re trying to gauge support of a Republican president for the entire country using a sample group from only Democrat-leaning cities like New York and San Francisco, you’ll get results skewed further against that president than you would if your sample group was more diverse and representative of the population as a whole).
These two facets allow for mistakes (bad statistics) and for intentional misrepresentation (manipulative statistics), especially when coupled with statements as “fact.” Properly done statistics, with large enough sample groups and properly represented sample groups can give reasonably accurate (but never 100% accurate) estimates of a given population group, but without knowing the source, it’s smarter to be skeptical. Critical thinking lesson: if you hear something that sounds like a statistic (especially an “opinion” percentage in this election season), question the source – find out how big the sample group was and compare to the population number – the closer the two are, the better the estimate will be, but never think of it as a “fact”.
Honest statistics will always provide a margin of error (for example, +/ 3%).
Also, White People Like Statistics.