Hello everybody, the short article below (see link) is a great example of how statistics can be as dangerously misleading as glow worms in leading us away from a precarious foggy path.
The piece quotes a report showing that 91% of 18-to-24-year-olds say they are stressed at work– compared to 84% on average. The author then continues to focus on the stress suffered by Generation Z.
Bouncing my forehead up and down on the desk a few times seemed to shake an idea loose from my few remaining dendrites. It might be fun to explore the statistical hype behind clickbait, 'shock' headlines and pulp fiction? We could also look at how pejorative reporting turns useful data into weaponry, which tends to encourage tropes like sexism, racism and general bigotry. Things should be so much better.
Copycat style and no substance
The takeaway from the linked article (above) should not be that 91% of young workers are stressed, that is just a relatively small deviation from a trend. It is that 84% of all workers are stressed with minor differences in the data. 84%! That is what should be exercising us; that four out of five of us, across all age groups and genders and everything else, experience some kind of mental difficulty within the workplace (assuming that this report is correct).
It is almost always the case that those who spoon feeds us statistics go on to tell us how to interpret the numbers as they drip down our bibs. These interpretations tend to build barricades rather than help us understand what is really happening (Gen Z V The Rest; Gay V Straight; Left V Right). One of the most execrable examples of this was the book Men are from Mars, Women are from Venus. This global best seller took tiny differences between men and woman, and made them seem like unspannable chasms of genetic certainty:
‘Men want to feel useful, women want to feel loved’
‘Women assign different meanings to words than men’
‘Women want to listen, men want to solve’
Clearing the fog
Piffle, drivel and codswallop! This A grade guff is almost never about meaningful variations but instead is predicated on weeny statistical differences between means.. Or, putting it another way, the small differences between the average scores of populations (men and women in this case).
See Figure 1 below. It shows a red population and a blue population. They can be boys and girls if you like, or people of different ethnic backgrounds, or - and let’s stick with the following idea for a moment for a moment – heads versus tails on a coin.
Figure 1. The difference between sample means (e.g., Heads V Tails; Females v Males)
Flip a coin often enough and you will find that, say, out of every 1,000 tosses your currency will emerge heads up 498 times and tails up 502. This difference will be significant because a very similar result occurs across a number of trials. And here is the rub; in statistical terms 'significant' does not have to mean large, it just means consistent. The actual difference may be close to irrelevant but don’t let that stand in the way of a misleading best seller or newspaper article.
So while on average men may have better spatial awareness, just as the average women is a better communicator, in reality these comments verge on the worthless. The differences between groups tend to be trite compared to the similarities they share. You will struggle to find a woman who doesn’t want to help solve a significant other’s problems, or a man who doesn’t want to be loved. Look again at figure 1. Note how small is the gap between the averages and how wide the overall spread of data. This is typical. And once this generally immutable detail is properly noted, we have a nuclear weapon to fire against sexism, racism etc.
Attacking the insidious
As an example, let us consider the noxious argument of white supremacy. The whole premise is that white is cleverer than black (an argument, which if it is to hold just a teaspoon of water, also needs to recognize that the same methodology finds Asian to be smarter than white. It doesn’t seem to do this, but never mind…).
Well, give into it for moment. Let the racists keep the point. You will convince few of them otherwise because racism is strongly correlated with low IQ. At the worst, the very worst, you will have conceded the statistical width of a coin toss. Then, by the law of averages it becomes instantly likely that your white supremacist is of lower intelligence than the next passing person of colour.
Meanwhile no graph – even back to Professor Hans Eysenck’s discredited research in the 1970s – will show a gap of greater than 5% between ‘races’ (as per Eysenck’s description). Discrimination is therefore utterly pointless. Again we note that tiny differences between your two populations are overpowered by
(a) huge differences within those groups, and
(b) overwhelming similarities between them.
If we are to confine any selection to skin colour, then – to use another common argument that all the best athletes are black – heavyweight boxing -that apex of athleticism and power - would never have seen Marciano, Klitschko or Fury, some of the best heavyweights ever to lace up gloves. Just as the world would have been deprived of the genius of Mandela, Greer (Bonnie) , Simone and King were intellectual and linguistic genius said to be confined by skin colour. Discrimination, prejudice, bigotry whatever you want to call it, is always stupid. Consider everybody and you have the widest pool of talent available.
So, whenever phrases such as ’women are like this’ or ‘millennials do that’, occur, they are almost always wrong and based on the statistical semantics magnified by the Men are from Mars School of Tosh. This is precisely the discriminatory language that, in different guises, kept the patriarchy in power for centuries. It is disheartening in the extreme to see the same dangerous superficiality popping up in the female lexicon as the scales shift today. What unites us is almost always far more telling than that which separates us. Remember, everybody deserves the chance to be an idiot.
What is going on?
Once we accept that we are largely the same, we can explore the anomalies that augment understanding rather than magnify difference. To refer to the earlier article again, are the levels of stress greater in the under 25s, or is there some other reason for the small discrepancy?
The new generation tends towards more emotional disclosure than has been the case historically. The stiff upper lip is somewhat looser and more inclined to vent. But again, only by degree. Do these degrees of difference overlap? Is this the explanation? If not, and the coin does have a tiny tendency to land head down, then relevant, micro corrections can be made. The use of a megaphone, exaggeration and the drawing up of discrete camps can be avoided.
Humans are largely similar. While social evolution rattles ahead with the brakes off, our physical selves change far more slowly. It is normal that our reactions to a gamut of stimuli should be largely the same regardless of generations, genders and nationalities with group-specific variations appearing only at the fringes.
When we magnify small differences and treat them as immutable and absolute, we ingrain ill-informed and skewed perception at the expense of collaboration and understanding. We also become inured to the whole concept of variance. So that when we do find those, rare as a unicorn, large differences between groups, we will fail to discriminate between important schism and the usual trumped-up, magnified pavement cracks so beloved by those pushing their own weak storylines.
All of which brings us back to what are we going to do about the hideous degree of stress in the workplace? This is the real story.
Thanks for reading. If I can help, just shout. C
Comments