Lies, Damn Lies, And Statistics

by Sherry Knight Rossiter at

In June 1882, Arthur James Balfour, first Earl of Balfour, was quoted in the Leeds Mercury as saying, “There are three kinds of falsehoods: lies, damn lies, and statistics.”

While this phrase is often interpreted to mean that statistics lie, that is not entirely correct.

Statistics represent, in a numerical manner, the interpretation of the data or information collected.

 The interpretation of the data depends on how it is presented statistically

However, it is how the data collected is presented that can sometimes misrepresent or obfuscate the true meaning of the data. In other words, the interpretation of the data depends on how it is presented statistically.

Over the years, I have read many allegedly “scientific” research studies that came to erroneous or questionable conclusions simply because of the researcher’s assumptions or biased presentation of the data. Perhaps the most recent and egregious examples of this appear in journals and media discussions about human-caused climate change. Much of what passes for scientific inquiry these days comes from computer modeling, but the outcomes of such studies frequently yield inaccurate results because the researchers either used faulty data to begin with or they used false (i.e., incorrect) assumptions. The computer industry’s mantra – garbage in, garbage out (GIGO) – is still a valid statement.