Statistics
- Thread starter JLM
- Start date
Statistics are as reliable as the data they're based on and the knowledge and honesty of the person creating them. Somebody once said that statistics don't lie, but liars use statistics, and that's true. Statistics is not a simple subject and most people don't have a good understanding of it, very little of it is intuitively obvious, especially when it comes to the probability component. It's actually fairly easy to produce deceptive statistics and deceptive claims based on them. A poster here, for instance, recently remarked in another thread that the suicide rate is anomalously high among homosexuals and implicitly used that statistical fact to argue that homosexuality is a corrupt and depraved lifestyle choice that leads people to depression and suicide. That's probably the most common statistical fallacy, confusing correlation with causation. Reality just isn't as simple as many people would like to think it is.
yeah, that's the figure I was quoted and of those 44.39% didn't have a clue what they were talking about. :smile:
One example of statistics not holding up are the Stanley Cup playoff results based on the statistics during the regular season.
They can calibrate the number of liars by asking questions they know the statistical distribution of the answers to.
Statistics are a lot like anecdotes. A good anecdote leaves out the unimportant information and focuses on the interesting things. Probably the anecdote doesn't feature the color of my clothing, so I would leave that out, but maybe it is important. Just like with anecdotes, you don't get all the information, because there is too much information. "There was man walking his dog." Who cares?
Similarly, a good anecdote builds context. If you walk in on the tail end of the story, you probably don't find it so amusing: you are lacking the context. Statistics, taken out of context are uninformative or worse, misleading.
Yet I would still say an anecdote tells the whole story, because who cares about secondary information?
You might not be interested, but I am, and if these forms of research can not give me what I think are important factors in their final conclusions or the factors on how they came to such a conclusion, then they're a complete waste of time, are about as useful as snot on bread and are only given out to the public via news reports or "Scientific Studies" to sway public opinion on a matter, either this way or that.
Case in point:
Cannabis smoke 'worse' than tobacco
BBC NEWS | Health | Cannabis smoke 'worse' than tobacco"Smoking pure cannabis is more harmful to lungs than tobacco, a health charity is warning. A study by the British Lung Foundation found that just three cannabis joints a day cause the same damage as 20 cigarettes.
And when cannabis and tobacco are smoked together, the effects are dramatically worse.
Evidence shows that tar from cannabis cigarettes contains 50% more cancer causing carcinogens than tobacco."
Dr Mark Britton, chairman of the British Lung Foundation, said: "These statistics will come as a surprise to many people, especially those who choose to smoke cannabis rather than tobacco in the belief it is safer for them.
"It is vital that people are fully aware of the dangers so they can make an educated decision and know the damage they may be causing.
Which of course was later found as false by other "Statistics"
In the above "Study" they don't exactly explain how they came to their conclusions....
• Were the studies based on a joint vs. a cigarette or piped weed vs. piped tobacco?
• What about bongs or vaporizers?
• Were the joints without any filters, used a rolled-cardboard filter or a filter similar to that of a cigarette?
• they later claim when mixed with tobacco, the results are worse, but once again, it falls on the filter information.
• what kind of papers were used?
• How long did someone hold in the smoke compared to tobacco?
• How much did someone take in compared to tobacco?
It states none of these and any or all of these factors being answered can very easily tell a much different story.
Take note that the creator of this statistic is talking to the reporters that his little study is without question and an un-deniable "FACT" considering the very limited information provided.
Besides the flaws above, I've smoked weed for the last 13 years.... smoked tobacco for the last 4 and then when I actually took up tobacco smoking, and only then, did I begin to immediately have my lungs filled with crap, did I begin to hack crap all over the place..... while during the whole time I smoked weed (and at times in my younger days, it was a lot more then just three joints a day) I never once had the problems I have had smoking tobacco...... my lungs were clear, I wasn't hacking anything up, no problems with my lungs..... and basically my entire first hand, long-term experience with marijuana counters everything the above statistic and the little pompous fella claims to be true.
One would think I should be dead by now, based on his above vague statistic..... Cheech & Chong, Willie Nelson, Stephen King, Arnold Schwarzenegger, Paul McCartney, and a pile of others should have died long ago as well.
Now I'm not trying to bring this into a weed debate, as there's plenty of those already around.... my point being is how many (usually those who conducted the study) try to make Statistics out to be un-questionable facts, when in reality, they're not facts and only give a very vague and very limited answer to a much bigger situation.
You may not care about all the extra details and are just happy taking what people tell you to be the truth, so long as what you're told has the rubber stamp of approval "Statistically Proven" that's all you need..... but I want more information..... and if they can't provide the necessary information/evidence to further back up their studies, they're not worth two lumps of chicken crap to me.
A tool to in-accurately sway ignorant public opinion. Look at many of the incidences over the years talking about gun-related crimes, drugs, global warming and many other common social issues that have statistics revolving around them everywhere you look. We have politicians and other government bodies trying to tell us one thing, using statistics, which later end up being something totally different down the road and us being told after the fact that they forgot to calculate something, or they never looked very far into the study and just use it for a good sound bite to sway public opinion through fear of the world heating up and we're all going to die unless we start buying mercury-laced light bulbs.
Don't tell me, tell buddy in my above example who seems to think his study is oh-so-holy true. I take stats for exactly what they are..... useless.
That's debatable..... it could also be easily argued that if people relied less on faulty and limited statistics, we could have progressed even further then we currently have, due to avoiding certain research or studies based on previously flawed statistical probabilities that said something was too dangerous or perfectly safe, when in reality and after the fact, that something wasn't what the statistic claimed.
But neither argument can be proven or dis-proven since what is done in the past is done, like debating what life would be like if Hitler was never born..... it's all assumption.
I think most of people's problems with stats is that they either get misinterpretations of some set of stats or else they misinterpret the set of stats by themselves. Data is simply data.
I just swished my spoon in my tea 5 times. That is a statistic. Misinterpretation would be, for instance, that I swished my fork in my tea 5 times. Or that I swished my spoon 5 times in my coffee. Or that I swirled my spoon in my tea 5 times. Ot that I swished my spoon in my cold tea.
You have to pay attention to what the details are.
I just swished my spoon in my tea 5 times. That is a statistic. Misinterpretation would be, for instance, that I swished my fork in my tea 5 times. Or that I swished my spoon 5 times in my coffee. Or that I swirled my spoon in my tea 5 times. Ot that I swished my spoon in my cold tea.
You have to pay attention to what the details are.
Or a lot of times they look for the statistics that bear out their preconceived notions. Statistics like longevity in certain countries mean next to nothing on their own. For instance before you put too much faith in that statistic you might want to look at the one for the "reliability of keeping records" or how many of the deceased had a birth certificate. When you study the output you also have to look at how much bullsh*t is in the mix. :lol::lol::lol:
Whether something is believed by you to be worthwhile or not is not an absolute measure of its worth. Statistics answer questions in an unambiguous fashion. Whether those are the questions you would ask or not is not important. Probably people that actually have to make policy decisions from a seat which is not an armchair are interested in the answers.
The most common misinterpretation is probably: you always stir your tea 5 times. Error bars (themselves a statistic) are important and often overlooked.
That would be the ± factor (AKA the margin of error).
That would be another example, yes.
The ± makes a difference alright.
So that would be a no then. Sometimes it's more helpful if you just answer a question rather than blather on.
Since you missed my point I'll spell it out for you. If you want to perform an experiment, you have to control all of the factors you are interested in. We never know all of the factors, so we can't control all of the factors. A large body of statistics is devoted to how to deal with knowns, known unknowns, and unknown unknowns.
More than that, some factors won't make any detectable difference on the outcome.
And lastly, it's incredibly expensive to conduct research. If you want to control every variable in every experiment, then there will be very few experiments. Who would pay for that? We wouldn't have the technology and knowledge we have today if we narrowed the research to only those studies which investigate everything simultaneously...
They reviewed available information at the time. They weren't running experiments, it's a review article.
A study by the British Lung Foundation found that just three cannabis joints a day cause the same damage as 20 cigarettes.
Cannabis tends to be smoked in a way which increases the puff volume by two-thirds and depth of inhalation by one-third.
I answered two of those questions above...by doing something perhaps unconventional to you. I actually read the news story, and the publication in question.
How do you figure that? The actual finding is that one form of pot smoking is more harmful than cigarettes. The pot smoke contains more carcinogens. And if you actually read the report, the intention is to synthesize what is known, and distibute that information. It also calls for more studies.
Like maybe, bong smoking and vapourizors remove most of the harmful byproducts of combustion, and then effects like this (listed in the actual review document) wouldn't be negated:
If you read the actual reports, maybe you wouldn't come away with that notion. No scientist I know believes that what they find is absolute fact. A little scientific literacy would inform people like yourself that there ARE caveats. I believe I said that already in this thread.
Strawman.
What is useless is your editorializing.
Debatable perhaps, but unlikely. However, your assumption isn't a given; that people rely on faulty and limited statistics has to be shown with evidence. You've not shown that!
The improvements in agricultural yield? Not possible without statistics. Statistics which have shown one variety of wheat works better under certain conditions than another wheat variety. That one nitrogen amendment works better than another. And when there is no difference in yield between treatments, qualitative reasons can be used (cheaper ingredient) or quantitative reasons (contributes less to runoff pollution).