02 April 2012

Revisiting the Census

I'm afraid the census is still bugging me, notably Figure 36. Looking at it the other day, and the related tables 36 and 37 at the back of the CSO's publication, left me utterly baffled. Here's the offending figure, to begin with:

Now, unlike good charts, graphs, and figures, which are meant to make data clear, this just bamboozles. It seems to have been in response to this, moreso than to the claim on figure 35 that much of the 179,889-strong increase in the number of Ireland's Catholics was due to recent immigration, that the Irish Times claimed that:
'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.'
My italics, obviously. As we've already seen, this is clearly rubbish.  Just under 39% of the increase was driven by non-Irish immigration, some of which was from such countries as the Philippines and Brazil. Fine, we know this. The Irish Times screwed up. Or, putting it another way, the increase in the number of people describing themselves as Roman Catholic was most definitely not driven by Eastern Europeans moving here.

Still, what's going on in Figure 36? The other day I stared at it till I got exhausted, thought I understood it and then in exhausted frustration cast aside my momentary comprehension. It's been pointed out to me in a comment that I was wrong to have done so, and true enough, exhausted as I was, my maths were all over the shop.

Perhaps, then, what's going on in Figure 36 is a chart that tries to do a lot, effectively compressing eighteen separate charts into one, each bar being an independent set of data points, the horizontal dark bars measuring the percentage increase of Irish members of a particular religious group, and the horizontal light bars measuring the percentage increase of non-Irish members of a particular religious group.

Obviously it's deeply misleading to compress this into one chart, as 1% of the 2006 figure for Irish Catholics would be about 34,000, 1% of the 2006 figure for Irish Orthodox Christians would be about 29, 1% of the 2006 figure for non-Irish Muslims would be 216, and 1% of the 2006 figures for non-Irish people of no religion being 684. These figures simply cannot be compared in any meaningful way.

Nonetheless, let's take a look at a compressed version of table 37, from the appendices, drawing also from the 2006 figures; this should shed useful light onto figure 36. Let's put aside the figure for Apostolic and Pentecostal Christians, as I can't see a 2006 figure for them, and likewise ignore the 'Christians' figure, as though that probably means 'all other Christians', I can't be sure.

On the face of it, this seems to explain everything. It's quite plausible that the dark and light bars for Catholics tally with the 3.4% and 32.5% increase figures we can see from the table; that looks about right. Likewise, I'll buy the 5.3% and 7.4% increases for Presbyterians. The 'Other religions' look fine too, as do the 'No religion' bars.

Unfortunately, the rest of the figure doesn't work. Look at the Muslims on figure 36: the chart looks just about right for the 35% increase in non-Irish Muslims, but it seems to be rather overstating the 60% increase in Irish Muslims.

The Church of Ireland are curiously rendered on figure 36. I'll buy that the dark bar represents a 7% increase in Irish membership, but there's no way that very tangible light bar represents the 2.3% decrease in non-Irish membership that actually took place. The chart makes it appear as though non-Irish membership rose, when it actually fell.

What about the Orthodox, then? We know their Irish membership increased by 194% between censuses, but figure 36 seems to suggest the increase was by no more than 120%; meanwhile their non-Irish membership rose by 107%, but the chart appears to be saying there was a mere 45% rise or thereabouts.

No, unless I'm even more confused than I was, I'd be pretty confident in saying that the CSO has made a pig's ear of this chart. 

Well, really.

No comments: