Welcome to Netrider ... Connecting Riders!

Interested in talking motorbikes with a terrific community of riders?
Signup (it's quick and free) to join the discussions and access the full suite of tools and information that Netrider has to offer.

Are you suspicious of 'statistics' too???

Discussion in 'The Pub' started by hornet, Jun 15, 2009.

  1. Over the last years we've seen everything reduced to statistics. We always knew that alcohol and drug abuse had a 'community' cost, but now, thanks to some statistics we know it is x-billion dollars per year. I wonder how they arrive at this figure when it seems to me that there are so many variables in calculation that a true figure could never be arrived at? Do they take ten addicts, add together the total healthcare costs for them all and then multiply that by the number of addicts they think there are?

    And another thing. No-one seems to question these figures; hell, governments even factor them into their budgets, and we just nod and move on. Policies and the expenditure of billions more dollars, are based on these sorts of extrapolations.

    {I'm not talking specifically about alcohol and drugs; think about stats for total cost to the community of road accidents, or lightnings strikes, or mothers-in-law, to name a few.}

    And then there's the times when the whole house of cards seems very shaky indeed; this study of over 100 cases of so-called 'drink-spiking' concluded that in those 100 cases there was no such thing http://www.news.com.au/story/0,27574,25638996-421,00.html
    ..but governments spend hundreds of thousands of dollars advertising the danger :? :?

    Some scientific person may well provide a good answer to this, but it seems as if we are being driven by very rubbery numbers...
  2. hmm, thats interesting. I know Wesley Private Hospital is 600 a night. It would be similar for the public ones, but paid by medicare instead. I would imagine it starts adding up pretty quickly. IIRC, there was about 10 drug and alco people there when i visited. thats $6000 a night for one very small hospital. add some more hospitals that can take many more people, the cost would almost grow exponentially.

    then take into account advertising and all that jazz.

    While the actual cost might not reach x billion it would still be very high, and im sure the governments can put enough red tape in to make up the difference
  3. I also don't see how they can do it. Take a drug addict for instance. They pay money for the drugs, yeah, but then some people get off it, some people go to rehab, some people are serial patients (aka they're admitted to hospital on a regular basis) and some never go to hospital. Some just fade away and die in the end. Way to many variables. Then there's the psychological side of the drug abuse, how can you put a price on a meth addict beating on his girlfriend?

    I don't believe the, do I do sometimes quote them to get a point across ;)

    All I know is, I've had a couple of mates now who've had one drink, leaving it unattended at one point, and we've had to carry home or take to hospital.

    Dude, there are people whose JOBS are to manipulate statistics to make it say what they want it to.
  4. 87.5% of statistics are made up on the spot....it's true.
  5. Oh, people can come up with statistics to prove anything, Kent. 14% of people know that.

    But seriously i think a lot of stats are done in a way that just backs the government/lobby groups point of view. i.e the way the question asked worded, bias within the sample.

    I was told by some prominent cardiologists that the 2nd hand smoke issue was overstated as stopping the actual smoker was a major public health issue that needed to be addressed - so 2nd hand smoke was talked up.
  6. Statistics show that 90% of tyres are rubbery.
  7. Having studied and worked with statistics, I simply don't believe any of the stats I see reported by journos. They make all sorts of mistakes, confusing correlations with causal relationships, mixing up absolute versus relative data. (edit: and even peer reviewed studies that assume that just because a result is significant, it must be important)
    And these people are shocked when you tell them the real truth (and yes, this IS true):

    That HALF of Australians are below average intelligence!
  8. I'm a mathematician and I don't believe any stats I see - you can always find some test to back up what you want to say, or even simple graph manipulation is a biggie used - say A is 82% and B is 84% instead of showing graph with scale 0 - 100% they show 80-85% which makes it look like there is a big difference - fools most people especially when flashed quickly in an ad.

    Also for reference (because now I'm talking about maths I'm getting a little excited) just because there is a 'statistically significant' relationship between two variables DOES NOT mean that one is caused by the other. This is often abused in poor research and media. There are additional tests to detmine if one variable influences another not just the stupid basic correlation test people do and think its mathematically sound.

    And statistically significant just means that you are X% confident in your results. And if they don't quote this figure I reject instantly and if it's lower than about 98% I don't consider depending on raw data source. 90% is really too low to suggest relationship and especially to suggest a causal relationship.

    Ok off my little high horse but I really really hate the misuse of statistical techniques which is very widespread.

  9. I'm 98% sure that is a strange thing to get excited about :LOL:
  10. The cure for bad statistics, though, is not no statistics, it's good statistics.

    In other words, be as sceptical as you like, but then just deciding that no evidence means anything and you may as well just stick with your existing opinions would be a massive step backward.

    Instead, educate yourself about science and numbers, learn to question the answers and judge the methodology and spot dodgy numbers and arguments... and base your opinions and actions on quality evidence.
  11. Couldn't agree more. Unfortunately the media don't publish enough info ot raw data to immediately figure out the truth. Which means until you investigate further, you have to be very careful about believing any stats in the media.
  12. Stats are a means to an end! What is it that you want to prove or dis-prove??...ah!!...we have just the stats to support that!! 8-[
  13. That's kinda what I was suggesting. I mean, even with a sample of, say, 1,000 alcohol-effected people and the costs associated with care/treatment/rehab, etc, surely there is so much variation from one person to the next that the figure of 1,000 couldn't honestly be extrapolated out to say that, since there are (probably) 1,000,000 people with an alcohol problem, that the cost then to society MUST be the sample figure multiplied???
  14. In my bookshelf, I have a book written in the 70's called "How To Lie With Statistics". As a result of reading this book, I view all statistics with deep suspicion - even my own.
  15. yeh i agree with you hornet, but i still think the real figure would be very high, even if the one thrown at us as highly inaccurate
  16. This is the next issue. Even if the “Raw data†were published, in surveys the exact wording would be necessary, and understanding of the sampling technique (Ask the same question at Broadmeadows station as you do to the people travelling through the Law Quad and Melbourne Uni, and you will have very different results) also ask the same question with 20 different wordings, and only publish the “Raw data†for the one that has results matching what you intend
    And so on and so on…
    If some Lobbyist/Politician/Journalist/Shock Jock wants an outcome and has the resources to do enough research they can find the statistics they require. And if they want to ensure that only the results they want come to light, then do what the tobacco companies did get a law firm to do the research and bury results you don’t like under lawyer client privilege.
  17. I do agree, but I put that all under the umbrella of 'raw data' to simplify thing :wink:
  18. I'm not suspicious of statistics.... I'm suspicious of the people who use them them and manipulate them to suit their own agenda.

    To understand any presented statistical information, you need to know the answers to these questions.
    Who is providing the data?
    Why are they presenting the data?
    How did they come to interpret it in that way?
    How can they benefit from the information they've provided?
    What information do they fail to show?
    What do they stand to loose if other interpretations proove to be more valid?
    How accurate are they?
    Is there evidence that counters their point of view?

    The 'truth' and 'facts' are not the same thing.
  19. Oh I understand that, I wasn’t having a go, just pointing out the limitations even of the idea of “Raw Data†when people with agendas get there hands on it
  20. I think the best example of this is that documentary that (supposedly) disproved global warming.

    He showed a graph of global tempratures through thousands of years, and it stopped at 1900 before it rose dramaticaly.

    Another was the theory that cows farting put out more polution than transport. I believe he did the comparison with the template of a rather high output from a cow, and rather low output from a very small vehicle.

    Granted, most of us know that that Doc was a load of bull anyway :)