# NOT a FUNNY Question .... in case Cecil doesn`t answer....[odds of dying from cancer]

Hi Cecil, this is not a funny question, but it is important to almost everyone on the planet according to statistics……

if you google "chance of getting (or dying from) cancer " you will find tons of websites (mostly cancer research or cancer charities) that tell you that there is a 1 in 2 chance of getting cancer (male) in your lifetime, and a 1 in 4 chance of dying from it.

but when I look at the approximately 100-200 people that I know (old folks included), I would estimate that maybe 10 - 20 people have had cancer and maybe 1-5 have died from it …
WHAT GIVES?

how are the statistics so far off "boots on the ground”? is everyone I know super healthy … am I in trouble? Or is there a group of the population that has 80% cancer rates and is skewing the stats ??

sincerely
Matt

Reported (wrong forum)

I think you answered your own question. Your extremely small sample size of 100-200 ppl will never be very accurate representation of the 7billion population as a whole.

Thread title edited to indicate subject.

Colibri
General Questions Moderator

What websites are telling you you have a one-in-two chance in getting cancer?

Have you known anyone with pancreatic cancer?It has a five-year survival rate of just 6 percent.

How about liver cancer? Multiple myeloma? Cancer of the brain? All of those are inoperable and all of them have poor five-year survival rates.

Don’t know that many people with cancer? Just wait a few years. That 10-20 people you know who’ve had cancer? I can name more than that I know who’ve died. That 1-5 people you’ve known who died from cancer? There are more than that just in my high school graduating class.

Hell, just among my immediate relatives (siblings, parents, grandparents, aunts and uncles) a half-dozen died of cancer, and several more have been treated for it.

A sample size that small can be “representative” (in the strictly nontechnical and arguably incorrect sense we mean it here) such that sampling a population that small can give the same answer within a certain range (the confidence interval) a certain percentage of the time (the confidence level); that is, you can get meaningfully repeatable results even if you only sample tiny numbers of a huge population provided your sample is representative and provided you define ‘meaningfully repeatable’ in a useful fashion.

For example, if you want a confidence interval of 10 and a confidence level of 95%, you need to sample ninety-six people out of a total population of seven billion. Admittedly, getting replicability only such that your survey results will be within 10% (plus or minus) of a given mean 95% of the time might not be your idea of “meaningful”, but it is a fairly non-intuitive result that such a tiny sample size will get you even that much predictive power.

Here’s a sample size calculator you can play around with.

Point taken. But I meant his particular sample of 100-200 direct friends and family is certainly not a cross section of all humanity. Sure, I guess it’s possible, but going by my own 100 direct family/friends, I’m sure my sample is heavily skewed towards white middle class american city dwelling adults in the 35-55 age range.

Right. In order for any of the math regarding sample sizes to work, the sample has to be randomly selected from the entire population.

There is such a thing as stratified sampling, where the entire population is broken down into sub-populations and samples randomly drawn from each of those, but it’s likely that a few hundred close friends of a given person wouldn’t be much use even in that case.

If there’s one thing statisticians and computer scientists can agree on, it’s that randomness is hard.

For the death component you need to sample people that have died. Everyone dies. 100% of them. But everyone you know isn’t dead. That doesn’t mean the statistics are wrong.

If one in four people die from cancer, it does not mean 1 in four people you know have died from cancer, only that when they do die, it will be a cancer that was the main cause.

The 1 in 2 chance is having cancer over their lifetime. Not “have got or have had, up to this point in their life”. The only people you can sample for this number is people who have reached the end of their life.

This is a problem with a lot of glib statistics about causes of death. Simply reducing the number of people who die from some cause - cancer - heart disease, respiratory disease etc, is to some extent a zero sum game. The age people die at matters.Reducing early death rather than deaths as a whole.

Cancer is primarily a disease of very old people. If your sample is a) alive and b) not particularly old, it’s not surprising you have not encountered very many cases of deaths from cancer. Most people who make it to very old age either die from cardiovascular disease or cancer. Younger people tend to die from a wider variety of things. like falling off ladders, diabetes, guns, and other assorted options.

Yeah, but most of them are still alive, and haven’t had, or died from, cancer YET. Let me know when they all die, and we can talk statistics.

## Hmm. You’re noticing that nearly everybody you know is still alive, and wondering why in the face of this empirical evidence the hypothesis that man is mortal holds up? It’s because you’re young. When you reach a sufficient age, you’ll start noticing that a larger and larger fraction of the people you’ve ever met are dead,* and the hypothesis will become more compelling. If you live long enough, almost everybody you’ve ever known will be dead, and the truth of the hypothesis will be inescapable.

• About 25% from cancer, incidentally.

No idea what the actual percentages are, but my understanding was that just about every male (or at least a huge percentage of them) will eventually develop prostate cancer if they live long enough. Of course, if they are old enough by the time they develop it, odds are they will die of something else before the cancer kills them.

Also, I’m in my 50s, with most of my friends at least my age or older, and I’d bet at least half of my friends have had skin cancers or precancers sliced off. Of course, an undue portion of my friends share my likes of gardening and golf, so have gotten more than their share of sunshine in the SVP-free days of the 60s and 70s.

Add in that 100% of my parents either had or died from cancer, and I’m not seeing 50% as grossly out of whack. So maybe my personal survey cancels your out!

Very good reference, thank you.

I often wonder what the odds are of developing more than one kind of cancer. I’ve had testicular cancer, what are the odds I will also develop liver cancer or stomach cancer? Can I just multiply the odds of two kinds to get the likelihood of getting both? Or, since my odds of getting testicular cancer are already 100% (I’ve had it), the odds of getting another form of cancer are the same as anyone else?

I keep hearing that 100% of people die, but I know, like, hundreds of people who haven’t died. How do you explain that, huh?

I think, unfortunately, once you’ve had cancer, your odds of getting some other type are generally much higher than the general population. The environmental and genetic risk factors that increase your risk of one type of cancer may also increase your risk of others.

Fear Itself, assuming that different cancers are independent, your chances of getting liver cancer or stomach cancer are the same as for anyone else. But that may or may not be a good assumption.

There’s one source of dependence right off the bat, in that once someone dies from one cancer, they’re certainly not going to get any other. But your use of the past tense for your testicular cancer suggests that that one, at least, isn’t going to be an issue for you (at least, one hopes not).

But there are many other possible confounding factors. On the down side, maybe there are some genes, or combinations of genes, which increase the likelihood of cancers in general. Or maybe some occupations, or lifestyles, increase the likelihood of cancers in general. The fact that you’ve had one form of cancer might mean that you’re more likely to be in such a higher-risk group, in which case your chances for other cancers might be higher. On the other hand, having survived one cancer, you’re probably more alert to the possibility of others, which might mean more frequent medical check-ups that could detect cancer or precancerous conditions in earlier stages where medical intervention is more effective, or might have motivated you to take action to decrease your risk factors, in which case your chances for other cancers might be lower.

IN working with the stats, remember that far more people die with cancer than die of cancer.