Tag Archives: Statistics

Numbers can be almost as good at lying as words. From Gary Galles at mises.org:

Major international comparisons have long concluded that Americans’ ability to effectively utilize mathematics is inadequate. Such conclusions divide students, parents, teachers and administrators into camps that share little more than blaming others for the problems. However, it is unclear whether all the finger-pointing indicates a real desire to overcome our innumeracy. In fact, we systematically misuse numbers to distort reality because we want to fool ourselves, making our ineptitude no surprise.

One of today’s most obvious misleading number games is grade inflation. Teachers have accommodated student desires for higher grades to the point that the median GPA of graduating college seniors has risen around a full grade point since it was about 2.2 in 1965. At some schools, almost everyone now gets As and Bs, and who is valedictorian has become a question of how many “perfect” students will share that title. Students have also pushed to allow A+ grades that count more.

High schools have gone even further. Many make advanced placement or community college courses worth an extra grade point. This has created a competition among students to take as many such GPA-padding courses as possible, especially ones they discover are actually easier than the corresponding high school courses. These and other policies (e.g., statewide comparisons crafted to show that, as in Lake Woebegone, all children are above average) have, however, thrown away much of the useful information such evaluations once contained.

Price inflation is another form of ego-building by manipulating comparison numbers. For most of us, if we want to brag that, say, we make more than our parents did, enough years of inflation can make it so. On the other hand, older Americans use it to “prove” how much better things used to be (e.g., “I remember when bread was a nickel” or “I only paid $22,000 for my house”).

Statistics and percentages are subject to the same abuse. Statistics are routinely manipulated, as with attempts to make insignificant changes appear significant. Instead of saying some drug increases the probability of some form of cancer from 0.00001 to 0.00003, reports scream that it triples your risk (from almost zero to almost zero). And “giving it 100%” was once going all out, but that has now frequently been eclipsed by claims of giving it 110%, 150%, 200%, or even 1000%. I’m 1000000% sure such inflated hyperbole is misleading.

To continue reading: Misleading with Numbers: It’s Worse When the Government Does It