> According to one study, just 50 out of 20,000 people managed to make a correct judgement with more than 80% accuracy. Most people might as well just flip a coin.
Oh the irony surrounding statistics and how often it is used to lie - case in point: the mode is completely absent and some useless factoid is presented instead.
That's not to say that argument is false ("you can't use body language"); merely that because of the useless information we don't know if it's true either.
Without knowing the experiment details (number of trials, ratio of people laying vs telling the truth) we don't know if 50 is actually significant. What's the curve of success for participants assuming random guessing (in line with the proportion of liars)? Is 50 to be expected?
My first thought is that with 20,000 participants, sheer chance will give you a handful of outliers, and 50 at 80% accuracy doesn't sound very high unless there were a very large number of trials per participant.
By my math, if each participant faced 25 trials, and each participant has a 50% chance of success on each trial, then you'd expect 40 people to get a score of 20+ (80%) just due to chance. To give an idea.
I'm talking about the "study, just 50 out of 20,000 people managed to make a correct judgement with more than 80% accuracy" which was asked by the grandparent post, not the paper by the people interviewed in the article.
Oh the irony surrounding statistics and how often it is used to lie - case in point: the mode is completely absent and some useless factoid is presented instead.
That's not to say that argument is false ("you can't use body language"); merely that because of the useless information we don't know if it's true either.