Turn hell hound, turn

I don't have a PhD, nor am I a Postdoctoral Research Fellow at the Saltzman Institute of War and Peace Studies at Columbia University. But I know a self-important windbag when I see one. Richard Hanania brings all those attributes to the party in his article

It Isn’t Your Imagination: Twitter Treats Conservatives More Harshly Than Liberals 
Source

The gist of his beef (if that's not too sinewy a metaphor) is:

Until now, conservatives have had to rely on anecdotes to make their case. To see whether there is an empirical basis for such claims, I decided to look into the issue of Twitter bias by putting together a database of prominent, politically active users who are known to have been temporarily or permanently suspended from the platform.

He continues  with a description of the massive database he has used to put flesh on the whinging bones of neo-con "argument" :

My results make it difficult to take claims of political neutrality seriously. Of 22 prominent, politically active individuals who are known to have been suspended since 2005 and who expressed a preference in the 2016 U.S. presidential election, 21 supported Donald Trump.

"My results", he says – pretty impressive. As I write, his piece has attracted more than 140 comments, mostly unreadable and full of references to a mixture of acronyms and slang that frankly isn't worth the bother of decrypting. But there is the occasional nugget, like

...And the statistical analysis? Unless we know the proportion of liberals versus conservatives versus apoliticals in the base population from which the 22 cases were drawn, the analysis is meaningless. Maybe conservatives are more likely to use Twitter? That could account for some or all of the bias. How would we even know what the political composition of the base population looks like?

Count 'em: 22; that's the sample size. The same (very long) comment concludes:

And even in the unlikely circumstance that the base population has exactly equal numbers of liberals and conservatives, using 5% as the significance level versus 1% requires further justification, as does the use of what is apparently a ‘one-tailed’ analysis [the author is vague on this point].

As I said at the outset, I have no statistical expertise. But I think I can detect here a statistician  who doesn't buy the "analysis".

But this comment provoked a huge backlash, which the author "Jack B. Nimble" (wish I'd thought of that) batted away with arguments  such as

Do Twitter users skew liberal? Your linked data are from 2012! I can suggest two reasons why old data are unreliable:

1. [HD: my numbering and formatting]  Trump effect: Trump’s use of Twitter may have drawn in hordes of his supporters to Twitter starting in 2015, so they can follow his tweets in real time.

and

2. Obama effect: Obama may have encouraged a generation of young adults to become liberals starting in 2009, and we all know that young people adopt new tech sooner than old people.

Are Twitter users in 2019 more likely to be liberal than the base population of average citizens? I don’t know and you don’t know.

 This is what paedocracy looks like

And in unrelated news, I was reminded recently of a post that I wrote a  few years ago, when the government were renewing their relentless attacks on the young (this was before they wasted 2½ years arguing the toss about Brexit and doing little else). Their latest wheeze at the time was to restrict Housing Benefit so that the little blighters could stay at home until they were 21.

This was the latest in a series of reforms (that's politician-speak for retrograde step) that started much earlier:

If I had my time again I'd be feeling extremely paranoid. The pressure starts being heaped on at primary school, where – as I mentioned last time – not only are the hoops you have to jump through getting smaller and higher, they are held by fools (or lions led by donkeys: look at the comments to that David Crystal blog I cited, and you'll see a good and conscientious teacher being forced into the goons' short-sighted bidding by an inflexible marking scheme). 

Then there's secondary school, where the hoops are not only smaller and higher, now they're ringed with flame. Where Victorian schools had notices saying Boys and Girls, they should now say

Lasciate ogni speranza, voi ch'entrate 

(commonly mis-translated as 'Abandon hope all ye who enter here'. It's 'all hope'.) 

And when the kids get through the hoops anyway, there's the ritual annual decrying of standards. 'More of them should be failing' snarl the hounds of hell (oh yes, I'm still working on the Dante theme).

I went on to talk about the woeful imposition (and raising "until the pips squeak") of tuition fees. Generally, my feelings about the way young people were being screwed over were not optimistic.

We're filling the streets with angry young men. And somehow I don't think it's just a revolution in theatre we're fomenting. Today's Jimmy [HD 2019: Jimmy Porter] is armed not just with an ironing board but with the power of the Internet.

But the streets are being filled by angry young men and women armed with nothing more threatening than disgust at the mess adults have made of the world, chanting 'This is what democracy Looks Like':

There's hope for the planet yet, if these good people have  anything to do with it. The comparison between these children's righteous anger and the feeding frenzy of bile accompanying Dr Hanania's whinging is heartening.

But prepare to repel time's wingéd chariot.

b

Wringing out the old

Another of my occasional State of the Blog posts giving you a chance to see a stat display that is normally reserved for the blogger...

<so_whats_new>
(not unlike all the other stats I've blogged about over the years... But what makes these different is that the others have been the default Overview,  whereas these indicate which posts rank in the top 10 of all time [Year Zero being 2012].)
</so_whats_new>

Far and away the most visited, in spite of its age, is one about Latin phrases. It's coming up to six years old, and last time I looked the screen capture it was based on was missing. I've no idea what  makes it nearly nine times as popular as no. 2 on the list. I expect some Influencer has spread the word; maybe it's on a syllabus somewhere perish the thought.

The remaining nine fall into 4 broad groups:

• Two that use the same Pedants of the world unite joke (2 and 6)
• Three on various philological points  (3-5)
• Three relatively recent ones (7-9)
• One that, being on its own at ^no 10, is in no way a broad group; so sue me :-)

In an update (after the New Year's dust has settled) I'll add some links (though in the meantime you can search using dates (or guessed themes, if you're feeling really adventurous)  :-) But I want to get this out there before December 2018 goes down as The Month of the Solitary Post,

Happy 2019!

b

Update – 2019.01.02.16:55: Here are the links and a few descriptive pointers. I've also fixed a pretty gross typo, in blue.

• ^no 1 The web page this originally pointed to is gone now but here‘s the jpg it used.
• ^no 2  and  ^no 6 These are the two on pedantry.
• ^no 3 ,  ^no 4, and ^no 5 These are quite old (but rather fun,  TISIAS) philological stories.
• ^no 7  , ^no 8 , and ^no 9   These three, despite their recency, make the top 10 because I plugged them in an MFL Teachers' group that, for reasons best known to Facebook (and I've given up beating my head on that cyber wall), I can no longer access.
• ^no 10 On a machine translation boo-boo.

Quips and quiddities

I've been thinking about seasons – specifically about adding an s. Taking as an example the frame

"a <season> day" 

I thought it was simple (if arbitrary): you can have a winters day or a summers day, but you can't have a springs day or an autumns day.

But I was forgetting the importance of collocation – what comes next  (in this case).

I was led to uncover this when I wanted to put numbers on this pattern. BNC doesn't happen to include an instance of a winters day.  And this made me use the "any noun" search syntax; which led me to conclude that the pattern is even more uneven. All season names are much preferred without the s (I've no  idea what its syntactical status is – some kind of possessive, I suppose; stay tuned for an update.)

By the wonders of BNC, the following seven links all run a BNC query; depending on line speeds, processor speeds, and all sorts of other techy imponderables, you may have to wait a second or two after clicking for the full story to unfold

winter [*n]  1590  winters [*n] 19 

spring [*n]  1086  springs [*n] 90 

summer [*n] 23163  summers [*n] 21 

autumn [*n]   774...

<tangent>
Why so many more cases of summer? About 30 times as many as of autumn, 20 times as many as of spring , 15 times as many as of winter. (Those numbers are wobbly; I didn't use a cuaculator. I could've just said an order of magnitude greater, but that expression has been sadly debased.) Hmm...  Meanwhile, back at the a <season>s <noun> pattern...
</tangent>

 ...But autumns doesn't follow the pattern of feast and famine. In that case it's feast and starvation; absolute starvation. The string autumns [*n]  just doesn't occur in BNC. And precious few occur in the much bigger (520 million words – more than five times bigger) COCA; just three (of which one is a mistake, resulting from a mistaken parsing of the word back):

autumns [*n] 3

I'm  not sure what this shows, except that when you put numbers on something you end up finding that things aren't quite as clear-cut as they seemed at first.

And here are two numbers that have only the very faintest soupspoon of a connection (and even that is arguable – it's just that the number of refugees world-wide can only increase when nation-states take more than their fair share of the planet's resources).

The first comes from a UNHCR report. CNN reported it like this

 The second comes from the ONS:

But the numbers reached out and grabbed my attention by dint of their similarity. I suppose another way of looking at it is that we have reached a tipping point: the number of refugees world-wide now exceeds the population of the UK (and that's before any readjustment that Nicola Sturgeon might have in mind ).

But time's wingéd chariot could do with some 3-in-ONE.

b
PS A couple of clues:

Affect brilliant fedora without an emergency jump-starter (13)
Such a way of arguing makes him a demon (2, 7)

Update 2016.09.16.11:45 – Added PPS and PPPS

PPS And I meant to add, justifying my subject-line (this is more of a quiddity than a quip), earlier this week I heard something of interest to a one-time translator of songs (a bit of background covered in this old post) . In Monday's Woman's Hour (the song starts at about 18'30")  Kizzy Crawford sang about a pond skater (no, really). She sang in English, but she sang the last chorus in Welsh. One word leapt out at me from the Welsh. Earlier (English) choruses had involved the word lily pad (see? – it really was about a pond skater); and the word that jumped out at me from the Welsh was lily pad.

I looked this up in an online Welsh dictionary: pad lili. What was up? Was Kizzy fibbing?

Of course not. She does say, in the interview that precedes the song, that Welsh is more appropriate to singing about nature. And one of the ways it has of being more flexible may be the choice of translation sources.  By chance I found this alternative translation (in Google Translate): lilypad  – note the lack of word-break.

Marvellous things, dictionaries. But they have their limits.

PPPS The promised update on syntax: the jury's still out. In some cases, it's obviously a possessive; summer's lease is the metaphorical lease held by belonging to a metaphorical summer. There's an apostrophe, and its aptness is not in question. But in summers day there's no apostrophe* and no sense of possession. I'll ask some teachers.

Update 2016.09.18.10:45 – Added P⁴S

Yes – it's an attributive genitive (also called descriptive). Some people omit the apostrophe, and in cases where the idea of possession is weak (or absent) this tendency is stronger.


Update 2016.09.26.11:15 – Added P⁵S

*Nonsense – I just misunderstood the BNC's search algorithm's treatment of the apostrophe.  I think  it's true, however, that the apostrophe , which usually marks possession, is less widely insisted on (and, from the point of view of a language historian, more likely to  be dropped   – rather like the apostrophe marking omission before words such as bus, cello or phone – when the sense of possession is weaker.

Update 2016.10.29.15:05 – Answers to those clues: DEFIBRILLATOR and AD HOMINEM.

Anything for the weekend?

Harry S. N. Greene, pathologist, cancer researcher, and Yale professor, when testifying in 1957 to a Congressional committee, disputing an interpretation of a statistical study,  famously said

It was noted long ago that the front row of burlesque houses was occupied predominantly by bald-headed men. In fact, such a row became known as the bald-headed row. It might be assumed from this on statistical evidence that the continued close observation of chorus girls in tights caused loss of hair from the top of the head. 

See more

I think of this wherever a politician mentions the 'Weekend Effect'  (painfully often of late).

Whereas people in ivory towers (no – considering the state of academic funding I should make that stucco-clad breezeblock towers) may talk about "The Post hoc ergo propter hoc fallacy", I think of bald men and naughty ladies.

A BMJ study wisely says (my emphasis)

The weekend effect is real, concludes Helen Crump in her review of the evidence (doi:10.1136/bmj.h4473). Paul Aylin confirms this in his Editorial but explains that we are left with a range of possible explanations (doi: 10.1136/bmj.h4652). These need to be scrutinised before assumptions and suggestions harden into policy. 

Here are a few obvious yeah-buts that have occurred to me without the benefit of any training in statistical analysis (beyond O-level maths).

• People don't practise many extreme sports during the working week; they save their death-defying stunts up for the weekend. Even practices as gentle as rambling (risking exposure, hypothermia...) can lead to weekend emergency hospital admissions.
• People who start feeling dicky during the week don't go straight to hospital. Come 5 o'clock Friday though, and nobody's picking up the tab for their misfortune, they high-tail it to A&E (in the absence of a weekend GP service)
• Elective surgery is done during the working week. Emergency surgery is done from Monday through to Sunday (sorry – can't bring myself to say "Twenty-f..."; see – just no can do) . As a result, the average surgery patient is automatically nearer death at the weekend.
• etc etc ...

Last Wednesday, Inside Health went into this in much more (and more persuasive) detail. Gov.uk published a round-up of some research last year, but the list of 8 papers was compiled in October 2015, and the earliest 2 date from 2010. And only the most recent 2 date from 2015.

All of  which reminds me of the lady mentioned on Midweek (?) last Wednesday who always packed a hand grenade when flying, to reduce the possibility of anyone else having one. Or the driver who, on learning that most road-traffic accidents happen near junctions, automatically put his foot down whenever he saw one. As the title of an early BMJ article warned at the time of the first paroxysm of Jeremy Hunt's madness:

Seven day working: why the health secretary’s proposal is not as simple as it sounds

And as archy said, in archy and mehitabel (rough quote),

whenever a politician 

does get an idea 

he usually gets it wrong

Well, must go. Time for More or Less.

b

PS – A crossword clue:

Model worker? Hardly, arriving at THIS time. –  (8)

Update 2016.03.11.10:40 – Added PPS

Time's up: TEMPLATE (with apologies to people whose brand of English doesn't include the word temp.