How many is ‘many’?

There’s a nasty little trick being used to argue for ending – and extending – the lockdown.

How many is ‘many’? It’s not a silly question, it’s actually quite important, especially now. Over the next few weeks, you’re going to hear arguments for keeping and ending the lockdown. Commentators from both sides of the argument will want to provide authoritative weight to justify their position. They’re also going to suggest statistical significance to that weight.

And that’s where the word will crop up: ‘many’. Example: “Many scientists are saying that…” or “Many businesses are facing…”. On the face of it, there’s a degree of accuracy to the claims; but if you dig deeper, there’s a flaw. To uncover it, all you need to do is ask, “How many is ‘many’?”.

To explain my point: In a room of 100 people, 40 have Samsung mobile phones. Is that ‘many’? It’s 40%. Is that statistically significant? Given there are dozens of different brands of phones, 40% sounds a sizeable chunk. But what if, of the remaining 60 people, 40 have iPhones? Is that 40% with Samsung phones still ‘many’. As I’ve said elsewhere, ‘content is king, context is King Kong’.

Let’s dip into science journalism for a second: Research in a town of 100 000 people has found 10 women with acute lymphoblastic leukaemia – a cancer of the blood cells. Is it statistically significant? The answer is yes. Acute lymphoblastic leukaemia is particularly rare – there are only about 800 cases diagnosed in the UK each year – and it usually affects children and older men, especially over the age of 70. Such a figure could therefore suggest an environmental reason for the anomaly.

But is that number ‘many’? I’d argue not.

Back to that same room of 100 people. It is serving a buffet dinner with two options: chicken or vegetarian. Fifty-one choose vegetarian; that’s statistically a ‘majority’ of people. But is it ‘many’? It’s only slightly more than half. So then at what point can it be safely argued that ‘many’ people chose the vegetarian option? Fifty-five? Sixty? Sixty-eight? For argument’s sake, let’s say 68; what if the buffet was presented at a function for nutritionists and the choice by 68 was representative of the eating lifestyle of the population in the room? Is it still significant? Is it still ‘many’?

What if the buffet was for former rugby players, it offered five choices, and 68 (of the 100) went for the vegetarian option? Is that statistically significant? Would there be safe grounds to say, “Many of those at the Blixton Rugby Club Retirees Dinner chose the vegetarian option”?

Probably; but then statistics would also be available to say, “Of the 100 former rugby players who attended the dinner, 68 chose the vegetarian option in a five-choice meal” That tells a far more compelling story, and, importantly, allows the reader to make up their own mind as to the statistical significance of beefy former rugby player gobbling up a bowl of eggplant.

If a commentator for, or against, ending lockdown uses the word ‘many’ to support their argument, then either they have analysed the statistics and made a sound value judgement of the significance of the outcomes, or they’re just assuming as such. If it’s the latter, their invocation of statistical significance is lazy at best, disingenuous at worst.

Whichever it is, they have a cause for challenge; so put them on the spot: “How many is ‘many’?


Note: The same point argued above is applicable to the word ‘lots’, as in “lots of people believe the lockdown…”, “there are lots of reasons why the lockdown…”, etc. What constitutes ‘lots’? Is 5kg lots of sand? It is if you’re walking around with it in your pockets, not so much if you’re trying to put out a raging oil fire.

Re: the image – okay, perhaps it’s a little harsh, but, admit it, it caught your attention.