29 March 2012

Torturing the Census Data

Back in my misspent youth, I enjoyed a brief dalliance with Fine Gael membership -- yes, I know, but I think we can all be forgiven our childhood errors. Besides, I rightly suspected John Bruton had real potential and suspicions I had about Labour were to be vindicated a couple of months later when they merrily snuggled up against Fianna Fáil in the governmental bed.

Anyway, I went to meeting one evening in UCD in anticipation of Michael Noonan, the finance spokesman of the day, addressing us; Noonan couldn't come, alas, but Austen Currie, one of my own TDs, stepped into the breach. Speaking at length about the situation in the North, Currie described how he'd gotten tired over the years with people saying that in time Northern Irish Catholics would outbreed Northern Irish Protestants.

'Rabbit theory,' he called this, dismissing it as unhelpful and crude; his standard response to people in the North who'd taken this line had been to advise them to go home and start breeding, as they obviously weren't interested in discussing things and working in accord with justice and charity towards the common good of the Northern Irish people as a whole.

Unfortunately, to judge by comments I've seen earlier today in response to the national census figures, we're not immune to our own form of rabbit theory in the Republic.


'Increasing' doesn't mean 'big'
'No Religion up 45% in Irish census, despite census question that favoured religion,' say some.

Number of Catholics in Ireland up by 4.9%, say others. Ah, the first crowd tend to counter, but most of the new Catholics are immigrants, they say, as though immigrants don't count. The rise in disbelief is an Irish phenomenon, they insist: atheism is on the rise. Huzzah!

People need to catch on to themselves.  'Increasing' does not mean 'big', just as 'decreasing' doesn't mean 'small'. What's more, we need to recognise that while numbers matter, truth matters more. Societies shouldn't be dictated to by their minorities, but they should most certainly be judged by how they treat their minorities. A State which doesn't act in accord with justice is no more than an organised crime syndicate.

Yes, there were about 80,000 more people in Ireland in 2011 who were willing to say they've no religion -- whether because they're atheist, agnostic, or apathetic -- than there had been five years earlier. Big deal: remember when people used to get excited about the Green Party vote doubling, when it had just risen from a risible 1.4% to a laughable 2.8% of the national vote? Indeed, do you remember the Greens? Or the PDs, who averaged just under 6% of the vote in each of the six campaigns they fought? 


Some pictures really are worth a thousand words
More substantively, the number of people willing to self-identify as Christians had risen by about 230,000 in the same period. Yes, I know there's a reasonable case to be made that it's not fair that children are probably listed by their parents as Catholics or whatever, regardless of what they believe themselves, but we should be careful about assuming that religion is about what you think rather than -- as so often -- what you do or who you are.

180,000 of those new Christians are Catholics, and most of the increase in Catholics in particular and Christians in general is due to natural increase rather than immigration.

Yes, really. Pay attention at the back.

Let's look at the broad national picture, as shown in Figure 35 of the Census 2011 Highlights, published this morning.


Ignore the second sentence for now; it's overblown nonsense and I'll deal with it in a minute. Instead focus on the main story, as taken for all in all, it's quite striking: 84% Catholic; 6.25% other Christians; 6% atheists, agnostics, and those of no religion in whatever sense; 2.5% everybody of religions other than Christianity; and 1.5% all those who declined to answer the question.

Granted, other than that more than 90% of the Irish population self-identify as Christians, with more than 90% of those self-identifying as Catholic, we don't know what these figures mean; there are already people arguing that Mass rates suggest that many Catholics are Catholic in name only.

I'm not sure about this, as leaving aside the issue of sacramentality, I'm always uneasy with attempts to second-guess people's self-identification. It always seems to be those who are quickest to claim the 'no comment' and 'no religion' answerers as secret atheists who are also quickest to claim that most religious people are secret atheists too. You'd almost think they had an agenda, the way they go on.

So all we have really go on are the numbers.


Pay more attention to the numbers than to the pictures
Now, what about the relative increases? Here, unfortunately, there seems to be a false narrative already in danger of slipping into the mainstream. Here's the Irish Times, for instance:
'The result show Ireland is still a predominantly Catholic country. Some 84 per cent of people described themselves as Roman Catholic, an increase of almost 5 per cent which was driven mainly by Eastern Europeans moving here.'
This is obviously misleading in the sense that it suggests that in 2006 our population was 79% Catholic and is now 84% Catholic, whereas it's the number of Catholics that's risen by almost 5%, not the percentage of the Irish population espousing Catholicism, but more importantly, the article perpetuates what seems to be an error in the CSO document itself.

It's an error that gives the impression that growth in Irish Catholicism has stagnated, and that the Irish Church is dependent wholly on immigration for new blood, and an error that needs to be challenged.

Look at Figure 36 here, which aside from the kind of disingenous chart that'd give Edward Tufte heart failure, purports to show the percentage change in religion by nationality.


It's not very clear what this chart purports to show, and it's hardly surprising that the Irish Times has read it as saying that Ireland's Catholic increase is almost exclusively due to lots of Catholics having moved to Ireland since 2006, mainly from mainland Europe. That's certainly the most natural reading of it.

And that's not true.


Are alarm bells ringing yet?
They should be. How high do you think our immigration rates have been since 2006? They were massive before that, of course, such that there were Polish signs at Luas crossings and in the Tax Office, but since then? Immigration's not been insignificant since the economic downturn, but does anyone seriously think it's been high enough to explain away the greater number of 180,000 new Catholics? High enough to explain it being possible to fill Croke Park twice over with new Catholics? Really?

Table N, below, should cause us to wonder further. Seemingly since 2006, roughly 52,000 Poles have moved to Ireland, along with 10,000 Lithuanians, 6,000 Latvians, 4,500 Brazilians, and 4,000 Filipinos. That's 76,500 newcomers from countries with largely Catholic populations.You'd fill Old Trafford with that. You wouldn't fill Croke Park once, let alone twice.


154,000 more immigrants, more or less, in that five year period, including Romanians and Indians. But Figure 36, as we've seen, makes it seem as though almost all the 180,000 or so new Catholics were immigrants. What's going on?

Well, putting it bluntly, it looks as though the likes of the Irish Times has taken the CSO's summary and messed up whatever it was that the CSO was clumsily trying to communicate. In 2006, there appear to have been 213,412 non-Irish Catholics in Ireland. By 2011, this number had increased to 282,799. In other words, only 69,387 of Ireland's new Catholics are people who've moved to the country since 2006.

Now that's not a small number, of course -- it'd almost fill Old Trafford, to put it mildly, and constitutes more than 55% of the last five years' non-Irish immigrants, contrary to popular narratives about how we have to change our education system to accommodate the massive numbers of immigrants who aren't Catholic -- but let's not exaggerate its importance if doing so means we've to ignore the natural increase that'd comfortably fill Old Trafford again, and Goodison Park or Stamford Bridge too.

According to the CSO, in 2006 there were 3,644,965 Catholics in Ireland; by 2011 there were 3,831,187, such that it seems that Ireland's Catholic population rose by 186,222 in that period. 37% of this figure can be accounted for by immigration during those five years, but that still means that 63% of it is natural increase.

It looks to me as though somebody  looked at the total figures showing how many Catholics in Ireland are from the rest of Europe -- just over 162,000 -- and mistakenly treated them as though they'd all arrived since 2006, whereas the vast majority arrived before then and have since become well-established in Ireland.

It's disappointing to see such sloppiness from our Central Statistics Office. Disappointing too to see people recycling supposed data without crunching numbers for themselves. People should get beyond the graphs.

Torture the data. It'll tell you what you need to know.




Update: I no longer understand this at all. I went through a phase later this evening of thinking I'd misunderstood the chart and been unfair to the CSO, but now I'm left baffled as to what it means by any definition. That it's a mess with the Catholics is clear, but look at the Orthodox, for instance. It looks from Figure 36 as though three quarters of the 24,000-strong Orthodox increase are Irish, but if the chart's saying that it's clearly wrong as we know from the tables at the back that just under 27% of the increase was Irish. So it obviously can't mean what it most naturally would be taken as meaning.

Having squinted at the chart for a bit longer than the Irish Times has evidently done, though, I thought I'd figured out what's going on, reckoning that it was a chart that's trying to do a lot, but in a murky way and not comparing like with like in any sense. My guess was that it was intending to show two data sets, the dark bars showing the increases on a percentage basis of Irish people in each religious group, and the light bars showing the increases on a percentage basis of non-Irish people in each religious group.

I suspected that it would have been easier to decode had it been a grouped vertical bar chart, with nine pairs of columns showing the increases or decreases in each religious group by national and non-national criteria. What's more, I thought it'd have been more honest if it were based on total numbers, rather than by percentages, as though percentage increases are in any way comparable: a sect with five members gaining ten new ones would show a 200 per cent increase, but it'd still count for nothing!

The problem is that that doesn't work either. Again, take the Orthodox as an example. Given that there were 2,881 Irish Orthodox Christians in 2006, and that this rose to 8,465 in 2011, the chart should represent this with a 294% increase in the dark bars. Likewise, there having been 16,845 non-Irish Orthodox in 2006, rising to 34,854 of such in 2011, the chart should likewise show a 207% increase in the light bars...

Except that's not what's what's happened. The chart shows roughly a 120% increase in the dark bars representing Irish nationals, and a 45% or so increase in the light bars representing non-Irish nationals.

I have no idea whatsoever what this chart is meant to be telling us.

5 comments:

Lmbsaflavin said...

What's going on with 20K new Orthodox? I would have thought people walking away from Irish Catholicism to another denomination would be more likely to go to the CofI, but if the pull is to Orthodoxy then the idea that it's people on the liberal sides of various issues seeking another option seems quite wrong.

Pavement Trauma said...

I don't consider Figure 36 very enlightening but it is not wrong about the growth in Catholics.

It is kind of hard to make out given the scale used but to my eye it looks like there was about a 30% increase in the number of non-Irish Catholics and about a 3% increase in the number of Irish Catholics. This is about right when you crunch the numbers.

The CSO says *much* of the increase in Catholics came from newcomers - and about 39% of it has. The problem comes when the Irish Times then translates *much* as *mainly*.

The Thirsty Gargoyle said...

That might be right for Catholics, but I think it falls apart when you look at the likes of the Orthodox etc. I really am utterly baffled by this.

The best advice I think anyone can have on this to look at tables 36 and 37 at the back, compare them with the 2006 data that's online, and disregard what the paper of vinyl says on the subject.

Liam said...

36 is indeed confusing and utterly meaningless without the real numbers. Not a statistician meself but I have been involved in projects producing data in graph form and am very aware of how they are read. That is appallingly and unnecessary bad.

Reflective Living said...

"The problem is that that doesn't work either. Again, take the Orthodox as an example. Given that there were 2,881 Irish Orthodox Christians in 2006, and that this rose to 8,465 in 2011, the chart should represent this with a 294% increase in the dark bars. Likewise, ..."
The increase would be 294% if there were an additional 8,495.
However, it rose to 8,495 from 2,881, so the rise is 5,584, which is just below 200% - you do the math.