In the recent thread about towel(l)ing, Peter Duncanson quoted a
passage from /The Times/:
"Children under 14 are being hit hardest and the NHS was told to
plan for a worst-case scenario of up to half of all children being
infected during a first pandemic wave."[1]
This is unremarkable, if we assume that 'children under 14' is a
subset of 'all children.' But I hear time and time again 'half of
all' or 'n percent of all' et sim. where no such comparison exists.
If I google "percent of all", I get a staggering number of hits.
Scanning the first page, I reckon that "all" is redundant in most of
them. A few examples from this and similar searches:
* 75 percent of all phishing emails target PayPal or eBay users
* 22 percent of all traffic accidents in the United States are
caused by drivers running red lights.
* More than half of all deaths of people of working age in Russia
are caused by alcohol
* Half of all British servicemen say they want to quit
* One third of all oceanic shark species at risk of extinction
* A quarter of all the largest public-sector database projects,
including the ID cards register, are fundamentally flawed
* half of all men and three-quarters of all women who have chlamydia
have no symptoms
There are many cases where the "all" is required, and others where
it might help, but I can't help getting annoyed at (particularly)
journalists who routinely stick one in where it's not needed. Any
statisticians out there?

Signature
Noel
[1] There was a typo in Peter's quotation which seems to have been
corrected since, with a recommendation turning into a diktat:
http://groups.google.co.uk/group/alt.usage.english/msg/ec16fa669c2f1c43?hl=en ,
cf. sixth para at
http://www.timesonline.co.uk/tol/life_and_style/health/article6716477.ece .
Bertel Lund Hansen - 22 Jul 2009 13:25 GMT
Ildhund skrev:
> There are many cases where the "all" is required, and others where
> it might help, but I can't help getting annoyed at (particularly)
> journalists who routinely stick one in where it's not needed. Any
> statisticians out there?
I think you will see that in 62 % of all cases. (Oh well, a
pretty obvious remark)
How would you write instead?

Signature
Bertel, Denmark
Ildhund - 23 Jul 2009 22:13 GMT
Bertel Lund Hansen wrote...
> Ildhund skrev:
>
[quoted text clipped - 5 lines]
> I think you will see that in 62 % of all cases. (Oh well, a pretty
> obvious remark)
Omit 'all.'

Signature
Noel
CDB - 22 Jul 2009 13:27 GMT
> In the recent thread about towel(l)ing, Peter Duncanson quoted a
> passage from /The Times/:
> "Children under 14 are being hit hardest and the NHS was told to
> plan for a worst-case scenario of up to half of all children being
> infected during a first pandemic wave."[1]
> This is unremarkable, if we assume that 'children under 14' is a
> subset of 'all children.' But I hear time and time again 'half of
> all' or 'n percent of all' et sim. where no such comparison exists.
> If I google "percent of all", I get a staggering number of hits.
> Scanning the first page, I reckon that "all" is redundant in most of
> them. A few examples from this and similar searches:
> * 75 percent of all phishing emails target PayPal or eBay users
> * 22 percent of all traffic accidents in the United States are
[quoted text clipped - 7 lines]
> * half of all men and three-quarters of all women who have chlamydia
> have no symptoms
In all the examples above, the word seems to be used to emphasise the
large size of the complete set. (Not *"14.3 percent of all Snow
White's dwarfs".) You could call it an idiom.
> There are many cases where the "all" is required, and others where
> it might help, but I can't help getting annoyed at (particularly)
> journalists who routinely stick one in where it's not needed. Any
> statisticians out there?
Jerry Friedman - 23 Jul 2009 04:15 GMT
...
> * half of all men and three-quarters of all women who have chlamydia
> have no symptoms
[quoted text clipped - 3 lines]
> journalists who routinely stick one in where it's not needed. Any
> statisticians out there?
...
Sure. According to my calculations, 83% of all people who write "half
of all" also write "one of the only times" and "first ever".
--
Jerry Friedman doesn't know why he associates those expressions.