How Companies, Organisations and People can use Maths and Statistics to be Cleverly Deceptive
Posted: Monday, November 30, 2009
by Connor Davidson
My Books and Articles (under construction)
If I were to show you a graph showing a great increase in sales but yet sales had been on a very slow increase for a few years - what would you think? Or Sales have pumped over the last few years but, yet, the graph shows an exponential increase - is this a contradiction? Or told you that we are on average the most expensive shop around but yet in a survey were the cheapest. You'd be inclined to ask how that is possible. It is actually very simple. Manipulate. The. Maths.
The second one is an extreme case and I have only ever seen it done once in my life – the creator was a smartass to be blunt. What you do is plot the graph using with the negative values which, show a decrease (you know where the 0 is one the graph, just below there is the negative values for the up and down axis) and then flip it. But so that you are being correct: put a little note at the bottom saying -f(x). This tells all the people who have studied equations of function graphs that the Y values (up and down axis numbers) have been inverted. While everyone else in the room is looking impressed. That is just being cruel. Besides, I could plot a graph of their morality with negative values of Y and I assure you that I'd not be inverting it.
The third one I'd like to draw attention to is surveys. There is a saying that I am fond of that goes: "you can get a survey to say anything depending on how you ask the question". For example, if you asked 1000 heterosexual British people if they thought that being gay is pure evil I'd be surprised if more than one person said yes – I'd be intrigued to know why. But if you asked 1000 if they like the idea of being gay then most of them would say no. Not because they were against gay people but because it sounds like you are asking them if they want to be gay. Then you could word overall results something like, "90% of Brits don't like the idea of people being gay". This is entirely misrepresentative but not lying per se.
This reminds me very much of the adverts that supermarkets do where they show that "an independent price check" found them to be the cheapest around. This seems fair enough but when I saw several different supermarkets running the same ad; I got very suspicious. The answer, I suspect, is that the independent price checks were taken on different products. Two totally opposite statements can't be equally true. You can't have two cheapest supermarkets – two equally cheap but not two cheapest. The "cheapests " is not a word for a reason.
The bottom line is, as with so much of my writings, question everything. I hope in future that you will be more sceptical and thus less susceptible to cheap tricks.
This Article has been viewed 1,180 times. (Not updated in real-time.)
More commentsGreat article Connor, you bring up some good points on how people will blindly accept swift marketing.Please log in to respond to this comment.Thanks for the comment.Please log in to respond to this comment.
You're absolutely right about being able to skew the statistics to prove whatever case you want to prove. That's why I don't believe statistics without further examination.Please log in to respond to this comment.Thanks for the comment.Please log in to respond to this comment.
I find this article very interesting. Keep writing! =)Please log in to respond to this comment.Thanks for the comment.Please log in to respond to this comment.
Great article! This will make me pay attention more when confronted with statistical analysis'.Please log in to respond to this comment.Thanks for the comment.Please log in to respond to this comment.
Great article, Connor! I had no idea but you are so correct about how the question is phrased.Think something similar has been done with the global warming stuff! (Your variety of subjects never ceases to amaze me. Keep them coming.)Please log in to respond to this comment.I love variety.Thanks for the comment.Please log in to respond to this comment.
NIce article Connor. Reminds me of what my dad always used to say. "Figures don't lie, but liars can figure." And something a college professor of mine said; you need to know HOW a piece of research was done before judging it's accuracy.Please log in to respond to this comment.You've got it bang on. I've just learned a wonderful saying: "Figures don't lie, but liars can figure.". Thanks.Thanks for the comment.Please log in to respond to this comment.
Thank you for writing this, Connor. Hardly a day goes by where I don't hear or read someone skewing the facts by some kind of false representation. I find it incredibly frustrating.My rule of thumb is that if it isn't stated plainly, then something is being hidden. For example, when a politician says, "I unequivocally and categorically deny the charges" instead of just saying "I didn't do it," this has historically shown that the politician is guilty. (Presidents Nixon, Regan, Bush 41, and Clinton have each made this statement and were later found to be guilty of what they were denying.)Please log in to respond to this comment.Good point. I recall Clinton saying something along the lines of I did not have a with that woman. Not long before admitting the opposite.Thanks for the comment.Please log in to respond to this comment.
9 out of 10 doctors say, "This is a very good article." I'm not a doctor, but I play one on TV.
I enjoy statistics very much, and will probably be using them in the future, as I have in jobs past. We had an old expression in Experimental Psychology: "Statistics (or graphs) don't lie, but liars use statistics (or graphs)."As for Vince Robertson's statement, I used to be frustrated as hell by White House Press Managers who couldn't give a simply yes or no. Although I loved the guy very much, the worst offender was Clinton's spokes people. By the time they were done answering the question, you had no idea of what they said.Another colorful character, God bless his heart, who tried to persuade us with graphs and charts was Ross Perot.Graphs are very useful if used honestly. Too often when the statistics don't quite rule out the null hypothesis, the statistician will keep refining standard deviations and Pearson r's or coefficients until, finally, "yes, there is a correlation."- GPlease log in to respond to this comment.Since standard deviation is based on totals of data one small change can totally change the answer. For example, 500^2 = 250000 to 510^2 = 260100.Thanks for the comment.Please log in to respond to this comment.
Mark Twain said that there are lies, damn lies and statistics!Besides,74% of all statistics are fabricated and made up on the spot!Please log in to respond to this comment.Ha.Thanks for the comment.Please log in to respond to this comment.
Very informative article. Gives me lots of food for thought. Thanks for sharing.Linda DPlease log in to respond to this comment.Thanks for the comment.Please log in to respond to this comment.
More comments
We want your comments! If you can read this, you don't have javascript enabled, so you can't use this comment system. Please enable javascript.