Ok, so after a brief holiday I have returned to the blog (much to the relief of my hundreds of fans...). I won't trouble you with the reasons why, other than to say, well it was fun and involved some plastic food.
So today, my soapbox is on the very titillating subject of statistical misrepresentation. I think it would be stating the obvious to mention I am partial to the odd stat now and then, however, I have no delusions of being a statistical whizz, and am currently stalking a professor who had a test examining type 2 error named after him. But, the glaring information bleep on an unnamed current affairs expose on the banning of the burqa has sparked my stats antenna. It's well acknowledged that many media and propaganda machines utilize statistics to gather support and engage the population. The Nazis did it particularly well, and in a subtle and innocent way at first. No-one started talking about a holocaust until the population was on board with the regime. Unfortunately, history is taking some time to sink in, and there appears to be a growing mechanism of fear, alienation and catastrophising around Arabic nationals both in Australia and around the world. US is known to have significant issues re this, Paris has controversially engaged a burqa ban, and as I found in japan recently, there is a strong sense of "Arabic=bad". Anyway, this isn't about stats. So...
*steps onto soapbox 'ahem'*
A very damning statistic was presented on said program saying 95% of Australians were supportive of banning the burqa. My interest was peaked, as frankly I was surprised that the Australian populace would become that outraged about anything, even football! although it must be noted that following the Four corners live export story there were reports showing consistently high concern re the immediate halt of live exports. These reports were consistent through many many channels and were seemingly quite robust.
It turns out that this 95% statistic re burqa was reported on the data of 35000 people. In my basic gestimating, with a population approaching 22 million, this is a representative sample of about 0.2%. Or probably the amount of people aged over 90 in australia currently.
Is this misrepresentation intentional? Maybe not. However, it is hard to imagine that a high level of transparency was thought out when this was presented with "scary" images of women wearing full face burqas, and very emotive language. I instead propose that this was reasonably blatant scaremongering technique (or another word Jonathan Holmes might use on media watch).
I am not claiming any kind of intellectual superiority, but I will acknowledge a reasonably good understanding of stats and their use. I enjoy working with numbers, and i am the first to jump up and down when I think i have found something. But I am also aware that I looked up when hearing a number like 95%, and questioned it. Did everyone watching that program question the methodology or was it touted in watercooler type discussions to engage and enable further segregation in the community? I'm not denouncing stats usage, the ideal program for me is chock full of graphs and statistical modeling. I'm just saying 100% transparency is expected. As I am expecting 100% of all readers to agree with me.
- Posted using BlogPress from my iPhone
