Monday 06 January, 2014
Statistics are impressive. There’s nothing like a bunch of numbers and statistical terminology to make you feel like a website knows what it’s talking about. They look authoritative, but they could very well be meaningless. At worst, they could be misleading and deceptive.
Don’t be fooled when you see figures. Always be skeptical and read the stats closely. To help you train your eye for statistical shenanigans, here are some of the most common ways marketers and others lie with statistics.
The Small Sample
What do you think when someone tells you their sales have doubled in the last week? Sounds pretty great, doesn’t it? They must be doing something right if they’re doubling their sales.
But what if actually they made one sale last week and two this week? Not so impressive now. Small samples are unreliable. In order to spot any kind of trend, you need large numbers. Look at the numbers and don’t take claims at face value.
This is especially important with surveys. If a survey says that 90% of people believe such and such, find out who those people are and how many there are of them.
Correlation Ain’t Causation
Even if you spot a trend, it doesn’t necessarily mean there’s a cause and effect relationship. Correlation does not mean causation. Take the example of how disco caused a rise in terrorism. Once disco got popular, so did hijacking and terrorist attacks. It must be the dance craze!
As much as you’d love to blame disco for any kind of tragedy or catastrophe, there’s simply no connection. Two things just happened at around the same time.
This fallacy is often used in medical claims. A clever marketer will tell you that cancer rates have increased because of our use of plastic products – and then try to sell you their non-plastic products as a substitute.
Most people understand basically what ‘average’ means, but what about mean, median and mode? There are actually several types of averages and they don’t all mean the same thing.
Let’s consider a fictional multi-level marketing scheme. A company claims that its average sales person makes 4.9 sales in their first two weeks. That doesn’t sound too bad but what they’re not telling you is that 9 out of 10 sales people sell 1 measly product while one hotshot sells 40.
Finally, you can convince somebody of nearly anything if you lump things together the right way. Consider crime statistics telling us the number of people killed each year by cocaine, heroin, crystal meth and marijuana. The only problem is that marijuana has never killed a single person in known history. It just keeps bad company with more lethal drugs.
By the same logic, I could just as well tell you how many people are killed each year by heart attacks, accidents, terrorist attacks and herbal tea. You’ll think twice the next time you want a nice, hot cup of chamomile tea.
Whenever you see statistics, use some skepticism and common sense before you take them at face value.