Are Statistics Good or Bad?
June 2, 2010 | 2 Comments | Science
Mathematician Des MacHale did the math:
The average human has one breast and one testicle.
MacHale’s statistic is true, but it is also absurd depending on your perspective when you read it. Statistics are bad when people think they mean something else. Imagine how professionally devastating it would be for an organization to blindly believe in MacHale’s average human and create for that target audience.
If math is an exact science, then why do statistics have a reputation for rubbish? I blame human nature. Many of us treat statistics as straightforward and easy to understand nuggets of truth that mean whatever we first assume they mean. Unfortunately, with stats our initial perspective can be wrong and the data over-generalized or even manipulated.
Personally, I love statistics. I think they are wonderfully good. But I also recommend the following tips for healthy stat consumption:
- Understand the data.
For credibility, look at the research’s parameters (e.g., sample size, methodology, who was surveyed, how they were asked, and who did the asking). For knowledge, look at the raw data, the scope of responses, and the individual percentage of each response. - Understand the context.
Most statistics are skewed by cultural biases. Online polls are skewed by the psychographics of the Website’s audience (plus they exclude the preferences of non-computer users). San Francisco is different from Memphis. USA is different from China. Lifestyle surveys from lad mags will naturally usually have different statistics than those of First Baptist Church congregations. - Understand the generalization.
Statistics are flawed generalizations that make the complexities of life simple. Our minds aren’t strong enough (or sometimes we don’t have the time required) to grasp the full scope and depths of something. Statistics bend the truth by simplifying the complex into simple bite size statements. The greater the generalization, the greater the ease of understanding. The greater the generalization, the greater the lie? For example, 84% of Americans are satisfied with their personal life. But these numbers change by state and city and even city block. They change by race, religion, politics, income, and marital status. They change for you and for me. It changes by the year, day, and hour. The true answer is a complex moving target. The question is how much “truth” are we willing to sacrifice to make it easier to understand? - Understand the agenda.
Everyone has a bias. Some are malicious. Some are in the unconscious mind. Whether for good or evil, a person’s bias will influence how they convert raw data into statistics. Understanding the researchers behind a statistic can offer great insight into the statistic itself. - See it as a limited insight.
Obviously, statistics have their limitations, but that is certainly no reason to give up on them. Statistics are far better than blind assumptions. They are limited insights that create a loose framework for reality. Don’t make statistics the law, but don’t ignore them either. - Mash it up.
One of the best things you can do with statistics is mash them up. Cross-analyze statistics. By having a constant intake of stats, you will gradually see a clearer truth. By comparing similar stats, you can better identify what is true and what is false. Most importantly, you will begin to see the big picture. - See the big picture.
Although bite-size statistics can be quite myopic, hundreds of them overtime can help you begin to see the big picture. In great numbers, over-simplified generalizations can reveal complex truths. Don’t miss the opportunity to see the big picture.
So tell me…
Are statistics good or bad?
What percentage of statistics do you believe? (Me: 90%)
