Understanding "Absolute Risk"

[Mtngrl]

Popular cancer lore tends to both overstate an individual's risk of getting breast cancer and understate the risk of dying from it. This is good for fundraising, but it's not so good as truth-telling. For example, giving a five year survival statistic of 90-something percent makes it sound like early detection and treatment = cure. Since many breast cancers recur a long time after initial diagnosis and treatment, the five-year survival rate is virtually meaningless. But it makes a nice story about "survivorship."

Then there's risk. The "absolute risk" of getting breast cancer over the course of an 80-year lifespan is 12%. That does not mean that 1 in 8 women will get breast cancer. One in 8 women who live to be 80 will get breast cancer. Risk goes up every decade of life.

Think about it logically. There are about 320 million people in the US. Slightly more than half are female, but let's say 160 million. Roughly 80 million are adult women. If 12% of them had breast cancer, that would be over 9 million women. In fact, 2.8 million people have or have had breast cancer in the US in 2014. That's 3.5% of 80 million.

Here's a brief explanation of what absolute risk means in relation to breast cancer: http://www.breastcancer.org/symptoms/understand_bc/risk/understanding

Another page on the same site uses this "1 in 8" = 12% language. Amazing. http://www.breastcancer.org/symptoms/understand_bc/statistics

[Lani]

https://ia700802.us.archive.org/31/items/HowToLieWithStatistics/Huff-HowToLieWithStatistics.pdf

[thinkpositive]

Amy,

Thanks for the posting. I had drafted a response earlier but didn't hit the "post" button. I thought that my draft might have had too much of a negative connotation to it. However, after reviewing Lani's post I see that I am not alone in my feelings about stats. It seems that the writer can generally find a stat to support their point of view. The reader must be diligent in finding out just what the stat really means! It isn't that the referenced stat is wrong, it is just that it may not mean what the reader is lead to believe it means.

Take care,

Brenda

[Mtngrl]

Thanks, Brenda and Lani.

I didn't read all 129 pages of Lani's posting, but I downloaded it. A few years ago I took intro to statistics, and it covered some of the techniques for "lying with statistics."

If I were the Queen of Everything, I'd have every 8th grader take a course in practical skepticism. We should never take anything at face value, but question all of it. What are the premises? Do they have data? Do the data support the conclusion they're drawing? It's especially easy to lie with statistics by distorting graphs and bar charts to make small differences appear larger or vice versa.

Don't ever let anyone else do your thinking for you.

[PeaceMomma]

I completely agree with you, Amy - if I could, I would add skeptical thinking to every grade, starting at about 3rd. I have talked to my daughter since she was quite young about making sure she knows what someone is talking about and definitely not taking numbers at face value.

My father used to explain to me how easy it is to lie with numbers. And people think that you tell them numbers, it makes it somehow more "accurate", but it's really just the spin they want to have...

Is it 25% less than the other, or is the other 33% more expensive? It's the same difference. ;)

[Carol Ann]

Thanks for these articles and discussion, Amy!

Carol Ann

[Hopeful]

Don't ever let anyone else do your thinking for you.

Best possible advice.

Hopeful

[Saygoon]

fund raising vs. truth telling - now that's another bag of worms! The lies statistics show makes a lot of people a lot of money particularly in the BC business.
If all of us bought into the grave stats none of us would be here now