Saying what you mean -- statistics

[TheDustyZebra]

From an article on Yahoo Shine, regarding the dangers of eating processed meats:

The research, conducted by scientists across 10 European countries, included tracking the health of more than 450,000 participants over the course of a decade or more. Of the 26,000-plus who died during those years, the ones who ate the most processed meats—5.5 ounces or more every day of bacon (so, five strips), sausage, hotdogs, salami, ham and the like—were 44 percent more likely to have died than those who ate little or none.

Ok, so, breaking it down: of the 26,000-plus who died, the ones who ate the most processed meats were 44 percent more likely to have died than those who ate little or none.

But they're all dead.

This goes along with my pet peeves about other scientific/mathematical phrasing which is incorrect and/or physically impossible, such as "50 times less than" and "up to 75% off -- and more!".

 

[Jo Zebedee]

yep, we need to know the baseline first - how likely we are to die, anyway - before 44% more likely means anything.

*pass the bacon butty, please?

 

[Sapheron]

Ok, so, breaking it down: of the 26,000-plus who died, the ones who ate the most processed meats were 44 percent more likely to have died than those who ate little or none.

Not quite. Of the 450,000 people in the survey, most are still alive, and the ones who ate lots of processed meat were 44% more likely to be among the 26,000 unlucky sods who died.

Though, whatever the odds of my death, I shall give up nothing. I'm not sure a British person is even allowed to give up bacon and sausages. If they did what could go with their HP Sauce?

EDIT: Also, since when is ham processed meat? It's just a slightly salty piece of pork. It's no more processed than smoked fish or a piece of lamb with mint sauce on it.

 

[Ursa major]

That's what was meant; but what was said on Yahoo Shine was:

Of the 26,000-plus who died during those years, the ones who ate the most processed meats—5.5 ounces or more every day of bacon (so, five strips), sausage, hotdogs, salami, ham and the like—were 44 percent more likely to have died than those who ate little or none.

i.e. a badly written sentence, which is what TDZ was poking fun at (appropriate, given that this a site about writing and reading rather than mortality statistics).

 

[Tecdavid]

Though, whatever the odds of my death, I shall give up nothing. I'm not sure a British person is even allowed to give up bacon and sausages. If they did what could go with their HP Sauce?

HP Sauce needs nothing to dress! The food is the condiment to the sauce!

 

[allmywires]

As a scientist I'm always hesitant to use statistics because most of the time it's an approximation, or, more simply, just bullsh*t.

Lies, damn lies, and statistics, as the old adage goes. I personally never pay any attention to stats in the news. Mainly because I get most of my news from the Daily Mail* who change their 'X will cause cancer!' storylines every few days.

*DON'T HATE ME their website is easy to use ok.

 

[Ursa major]

'X will cause cancer!'

If that's true, I'm sticking to a 25-letter alphabet from now on....

 

[Tecdavid]

As a scientist I'm always hesitant to use statistics because most of the time it's an approximation, or, more simply, just bullsh*t.

Lies, damn lies, and statistics, as the old adage goes. I personally never pay any attention to stats in the news. Mainly because I get most of my news from the Daily Mail* who change their 'X will cause cancer!' storylines every few days.

*DON'T HATE ME their website is easy to use ok.

The writers of Cracked had a salty few words to say about them, I do believe.

 

[allmywires]

If that's true, I'm sticking to a 25-letter alphabet from now on....

Probably for the best.

The writers of Cracked had a salty few words to say about them, I do believe.

If you mean the DM, I think pretty much every logical, sane-thinking person has something bad to say about them It's okay because I read it *ironically*, guys.

 

[AMB]

I have no problem with you reading the Daily Mail. I read it too, on occasion.

Of course I also read the Telegraph, the Guardian and a few others as well. Then I try to decide which one is lying least, or figure out the common thread between takes on various stories and hope I can get something slightly factual out of it all.

Then I give it up for a lost cause and read something that actually admits to being fiction.

 

[Mirannan]

To use a hackneyed phrase, correlation does not imply causation. In this specific case: It is quite likely that those who ate more processed meat ate larger amounts of other junk food, and smaller amounts of nutritious food such as fresh fruit and vegetables, as well.

From the information given there is no way to know, and so the study quoted (unless there is other information in the study that was not mentioned in the news report) is of little or no value.

 

[The Judge]

Er... is there a reason you posted this in GWD, TDZ? I can see that it will improve our credibility if we use statistics better in our books, but I still don't think it's a writing issue. I'll move this over to Science for the moment (well, it's kind of vaguely sciencey to us arty people!) but I can always shift it back if I'm missing the point.

 

[Bowler1]

Just to be clear, I've not read the report but I do watch the news. These unlucky sods who died (halfway through a good English breakfast, what a way to go!) had a few other bad habits, drinking and smoking, and were a little more over weight when compared to the rest of the surveyed group (this is as per the BBC). Some Doctors have come out and said, you can't just pick processed meat when these people have so many other life style choices affecting their health.

For example.
100% of road deaths occur on roads, roads are bad for your health!
If you don't drink you will die, drinking is 100% good for your health!

I think these stats are 95% correct, with a small margin of error, so all roads are bad and drinking is good for you - Beer!

 

[Harpo]

You can do anything with n% of statistics (where n=whatever number you choose)

 

[TheDustyZebra]

Er... is there a reason you posted this in GWD, TDZ? I can see that it will improve our credibility if we use statistics better in our books, but I still don't think it's a writing issue. I'll move this over to Science for the moment (well, it's kind of vaguely sciencey to us arty people!) but I can always shift it back if I'm missing the point.

Well, there was, because it was an issue with the writing to me, but it very quickly wandered off in another direction, so I figured it would be moved.

 

[steve12553]

And of course pure, clean water is 100% fatal. All recorded cases of water drinking have have or will result in death. Also, being born is a confirmed contributing factor to death.

The preceding statements are actually more true that the quoted statement in the first post. I wish I had a nickel for every misleading headline I've seen on one of the Yahoo sites.

 

[Dave]

I'm glad I wasn't the only one confused by this piece of news. I also never considered Bacon and Ham to be processed meat, even if they are smoked. I can't see where the proof is here. We already know that too much Red meat is bad for your health. We already know that too many calories are bad for your health. Common sense tells me that eating nothing but Chorizo, Salami and Kielbasa is going to be bad for my health, but are they really worse for me that eating the same amount of Beef steak. On the other hand, I never ate 99p frozen value meals and since the Horse meat scandal, I doubt that I ever will.

 

[hopewrites]

I think the amount of salt that goes into preparing them are what constitutions them as processed meat for statistical reasons.

I agree though that statistics and studies can be used to prove anything.

A fortune teller once left me with the impression that I was going to die younger than I currently am. So ether I'm going to be involved in a fatal time travel accident, or I dont have to worry about it because what ever was going to kill me hasnt yet.

 

[Mirannan]

Another issue with statistics is the misuse of numbers that are actually correct, but misleading.

One example is the almost universal use of the arithmetic mean as the average for such things as annual income. This is probably because of the various types of average, this one produces the highest figures and hence serves the purposes of various governments. I'll illustrate this with an example; the figures are probably wrong, but they will serve.

Consider the mean income of working-age residents of the Seattle area, with one specific person removed from the list. The mean might be (sake of argument) $40,000. Add this one person back into the list and the figure jumps to maybe $100,000. Why? Because the one person is Bill Gates.

Another one, somewhat more amusing, and completely accurate:

In the UK, more than 99% of the population has more than the average number of legs. True, but highly misleading!

(Think about it.)