Below are some comments by the author of the graph I linked to in the Fear Manach post recently. I thought you may have some opinions?

Religion National Identity 2011

Hi, I’m the ‘author’ (so to speak!) of that graph. I find it gratifying that others are interested in what it might be showing, so I’d like to expand on how I’d interpret it – particularly in relation to the ‘Northern Irish’ category.
The commenter who posted it on Slugger O’Toole seemed sceptical of my claim that most people choosing ‘Northern Irish’ were Catholic or of a Catholic background, but to me that fact seems obvious when a) British identity and Protestant background match so perfectly, meaning that aside from a small overlap (6%) with NI identity, the latter have to come from elsewhere, and b) ‘Irish’ identity only makes up a proportion of the Catholic population, so they also have to go elsewhere.
It’s important to remember there are two sets of figures, one with unique combinations of choices (i.e. Irish AND) in class 1, and the other being what I used for the graph which measures how many choose a particular identity alone or in combination (i.e. Irish OR British OR Northern Irish) in class 2. I feel safe using the latter because only one combination reached a significant level, that of British AND Northern Irish ONLY (e.g., ‘Irish’ and ‘Irish only’ are virtually the same)
Unpacking the choice of Northern Irish identity in class 2, I can see that while ‘British and Northern Irish only’ climbs significantly in the more Protestant-dominated areas, the ‘Northern Irish only’ figure drops somewhat also – hence the overall class 2 figure seems remarkably steady, and I would say doesn’t correlate with Protestant background (aside from the subset who choose it alongside British identity), or anything else really. I do also see that ‘Northern Irish only’ also drops in areas with high Irish identity (e.g. Derry, but also significantly Belfast which bucks the trend somewhat), which fits the idea that Northern Irish v. Irish is the choice amongst Catholic communities.
To address sammymcnally’s points, ‘Northern Irish’ is not correlated with British/Protestant identity (the highest figure for it, in class 2, is 34% in Down which is 63% Catholic, followed then by North Down at 33% which is 73% Protestant) and fluctuates within such a narrow range, statistically speaking, that it’s probably not worth divining any pattern in it. The combination of British and Northern Irish does increase along with the proportion of British only, but in a way it’s to be expected that the more people who are identifying as British anyway, the more would identify as both British and Northern Irish. The remarkable thing is really that the figure overall is so flat; even the ‘Northern Irish only’ figure which seems to exclude those of British/Protestant identities, while it is on a downward trend as those increase, does so much more shallowly than the Irish or Catholic figures – and even surpasses each at their 16% and 14% levels respectively. (although as I may have said the increase in ‘no religion’ in highly Protestant areas suggests perhaps a hidden Catholic figure).

One way of putting it could be that Catholic background = Irish + Northern Irish only while Protestant background = British only + British and Northern Irish only, although this breaks down somewhat as the Catholic figures drop below 25% – of course there are some people of neither background which would have to be factored in too. But it does seem that in the graph when Irish identity rises – relative to the non-Protestant population, that is – Northern Irish (only) identity drops and vice versa, within a relatively stable band or proportion of the total population, in diverse areas. Which is a little weird and I don’t fully understand how that could work in terms of either statistics or demographics, but it seems to be the case.