Thursday, October 22, 2009

Numbers Don't Lie

So numbers don't lie. I'm sure we have all heard that before. I, for one, is a big fan of numbers. I judge whether an activity is safe by it statistics. If the fatality rate is 0.01%, it's probably pretty safe. If the fatality rate is 40%, you probably should give it a second thought before attempting said stunt. Then there are activities that regardless of what the statistics says, you have to stop and ask, "And why do you want to do that?" Example: bungee jumping. I just don't get it.

But I digressed. Whilst numbers don't lie, people certain can. Numbers can be fabricated. And if they are real, we still have to question how they are obtained. They may not be accurate. But the above two scenarios are not what I want to discuss here. I want to take this opportunity (?) to discuss the presentation of numbers. Given that most laymen are somewhat mathematically challenged, it's not that difficult to lead them to think what they want you to think, while presenting you with partial facts.

Here's an example that I've heard on the news recently: swine flu cases among children in the bay area had increased by 80%. "Wow!" Right? Wrong. Whilst 80% sounds like a lot, one must stop to ask: 80% of what? If we had a whopping five cases before, 80% of that is 4 cases. That only give us a total of 14 cases. "Wow!" really?

Another one that I've heard in the past: 90% of those who don't [use this product] had back problems after the age of 70. Again, WOW! Right? How can we not run out there and buy the product after that statistics? Now let's stop for a moment and ask this question: What's the percentage of those who uses the product and have back problems after 70? Of course the commercial wasn't going to give you that piece of info. My guess? Still 90%. Where did that come from? 90% of the general population has some sort of back problems after 70. The product can be anything.

And here's a good one that even I almost fell for: 15% of women who lived above the age of 100 had never been married in their lives. WOW! Yeah, that one got even the best of us -- even I said Wow! Probably the reaction they were looking for. 15% of women who lived above 100 never been married. The deduction from that would naturally be: women can live longer without the stress of being in a marriage.

But wait. Question #1: the other 85% had been married. Now if it's the other way round:85% of those over 100 had never been married, that'd truly be a reason for women to reconsider getting married. 15% is really only the minority. Question #2: is it even established that being married or not has anything to do with a woman's life span? What's the percentage of married women in the generally population? For an even clearer picture, we should divide them up onto two groups: percentage of married women between the age group of 25 - 65 and percentage of married women between the age of 65 - 85. Without those numbers to compare with, the statement tells me nothing more than what it presents on its face value: that 15% of women lived over 100 had never been married. We cannot infer that marriage has anything to do with a woman's life span.

So the next time you have random studies and random numbers thrown at you, stop and think about what it really means or are you just listening to what they want to you hear. Even do a quick search online to see how the data is gathered. It just might pay off!

1 comment:

D said...

You said it sister! We may be living in the age of information, but no one claimed it was accurate information! We are bombarded with stats all the time, and our happy-go-lucky news media doesn't check the statistics for us, leaving us with nothing more than propaganda.

For example, you and I have talked about global warming. That is one area that we are all being toyed with when it comes to statistics! For example: the Soviet Union fell in 1990-1991, and during the budget crisis of that tumultuous decade, the majority of the weather stations in Siberia were shut down, for (no surprise) lack of funding! What happens when you loose a sizable portion of statistical data from a primarily cold region? The overall stats start reflecting warmer temperatures. This is where the banner slogan "1990's ... The warmest decade on record!" came from. Thank you Al Gore and your impressive manipulations of the facts!

You are so right when you say that everyone should take statistics with a grain of salt, and check into them ourselves. People might be surprised what they find!