• Mark Monfort

Fool me, once shame on you...fool me twice, I should have Googled


Not often do we get to utilise skills learnt in the workplace in our everyday lives but I've been lucky to have that with analytics. More specifically, working in analytics has given me the ability to detect the smell of bullshit that exists in online news stories, comments and other claims we are exposed to. However, one doesn't need a statistics degree or to be working in finance to do this type of stuff themselves. Before I continue I should mention that this is not a post about fake news or by any means meant to delve into political agendas so let's leave these at the door for now.

Let's take a step back to a few years ago when I read a small footnote of a story from the Sydney Morning Herald which had mentioned the finding of a new type of flying dinosaur with a wingspan of 4 metres. Another newspaper with much less rigorous editing checks (it seems), clumsily reported the same story but added a zero. The winged creature was now 40 metres long. Not only that, but they expanded on this mistake to infer that due to its size, the dinosaur was as big as a jumbo jet. I found that quite odd and it would certainly make someone with a bit of common sense question the story if they'd heard the latter one first. Importantly, this is the first way in which one should be mindful of news because if it sounds too good to be true (or too outlandish) it probably is.

When news isn't so obvious however, there are some simple tasks that one can do to investigate the claims. There are even a number of checklists online built for this very purpose but off the top of my head I do the following:

  • check if other news sites are reporting it

  • check the reliability of those sites and what their political sway tends to be (remember that not all established media is necessarily left-centric and vice versa)

  • check what the wisdom of the crowds have to say? For this I would usually look to Twitter, Reddit or other areas where the internet tends to congregate towards and engage in topical discussions. It's in areas like this that you get to see impassioned arguments from both sides of an argument which is important if we're trying to be a neutral party

The above covers some generic ways to check but sometimes I've seen people get thrown off by biased claims using graphs and statistics. Yes, the use of data, that most objective of objects, can be twisted to achieve nefarious gains but that does not mean we are powerless. Case in point was a particular video on YouTube discussing immigration/segregation in America and how certain people of colour should not be considered for certain job roles. This YouTuber showed his followers a number of statistics which highlighted the disparity between the SAT scores (US High School final exams) of caucasian and african students in particular. The use of statistics was used to warrant the strong beliefs that he had and even repeated by viewers in the comments section. Guess you cannot argue with statistics right? Wrong, because what was missing here was context. There are many factors contributing to the headline level of why certain students did better than others. Socioeconomic factors play a strong role in understanding why some fail and some don't but to generalise off a headline level was preposterous. I even had evidence from other articles and research papers that were easily found off Google. I didn't have to deploy those weapons however as the previously vocal commentators were now deathly silent. With a lot of statistics its important to not always take them at face value. A good way to look at this is from the following image by PHDComics.com which shows just some of the simple ways in which stories can be manipulated.


Another couple of things to remember is the inferences that can be drawn by adding correlation to the mix. It's been out for a while but the site Spurious Correlations (HERE) shows this well. There is a common phrase amongst statisticians that "correlation does not equal causation" and it can be true for the most part. Sometimes there is an important relationship being highlighted between two events within an industry. For example, there is correlation between the share price of certain stocks in a similar industry like the US DIY Retailing space (Home Depot and Lowe's have similar trends in price). Other times the correlations are high but relationships ridiculous as you can see below.


What's important is to recognise that just because numbers are being thrown at you, it doesn't mean they're always true or that they haven't been doctored or manipulated to hide other factors from you. Again, to combat this you need to use your wit and your trusty friend Google to investigate a bit more and it actually does not take long to do for the most part. You really should try it the next time someone shares something in your news feed or you see a news story. Do a quick Google to see what else is being said about it.

In closing, I still find it amazing that this piece of technology we call the internet could have had the potential to uncover truths so easily and yet it seems to be more popular for confirming our long-held biases. It's not our fault that this happens, we humans are emotional creatures and are swayed by subjectivity. In an age of data overload, the steps I mentioned above are just some of the simple ways for you to take back control and all it takes are some simple keystrokes.

Try it with me.

W

W

W

.

G

O

O

G

L

E

.

C

O

M

Do it for your sanity!

#datamining #dataanalytics #google #correlations #spuriouscorrelations #datamanipulation

14 views0 comments

Recent Posts

See All

Follow

  • Facebook
  • Twitter
  • LinkedIn

©2018 by Mark Monfort. Proudly created with Wix.com