Nilufar saw a job advertisement from a small design studio in Samarkand. It promised "an average salary of 3,000 dollars a month", which was more than she earned at the time. Before the interview, she did what her statistics teacher had taught her: she asked for the median salary.
The answer surprised her. The studio had nine people. Eight designers earned 2,000 dollars a month each. The director earned 11,000 dollars. Eight times 2,000 is 16,000; with the director's 11,000 the total was 27,000; and 27,000 divided among nine people is exactly 3,000. The advertisement was accurate.
But the median told a different story. If the nine salaries are written in order, the fifth one, in the middle, is 2,000 dollars. Eight of the nine people earned exactly the median and less than the mean. One large salary had pulled the average up by 1,000 dollars without changing what a typical designer was paid.
Nilufar went to the interview anyway, because she liked the studio's work. When the director offered her 2,000 dollars, she was ready. She explained, politely, that the advertised average did not reflect what designers earned, and she asked for 2,300 dollars, pointing to her experience with clients abroad.
The director laughed and agreed to 2,200 dollars. He also changed the next advertisement. It now said "typical designer salary: 2,000 dollars", which was less attractive but more honest.
Nilufar's rule for any data about pay is simple: when one number is far from the others, the mean is a marketing figure and the median is the real one.