When the London media’s star journalists went in search of some post-referendum colour after the UK’s Brexit vote in the summer of 2016, they headed for the bleak towns along the North Sea coast, several of which had voted heavily for leave. London, in contrast, voted remain, as did Manchester and other prosperous cities. In places such as Clacton in Essex and King’s Lynn in Norfolk they found what they were looking for – xenophobic folks worried about immigration and happy to blame the European Union for their woes. Unlike cool metropolitan London.

Identification of the target location was easier for last week’s referendum on the eighth amendment: Roscommon was sure to be chock full of conservatives since it had been the only county to vote no at the marriage equality referendum a few years back. Sure enough, the BBC and Sky News despatched their finest to the north midlands in search of reactionaries. But the Rossies let them down and voted yes. They would have found a better backdrop to fit the script in Donegal, the only county to vote (narrowly) no on this occasion. A few reporters did indeed wend their way to Tyrconnell after the event but their Roscommon misadventure was their own fault. Election results are reported on a geographical basis but it is not location which drives voter behaviour.

After the Brexit referendum, academic political scientists in the UK got to work on the figures and rapidly discovered that older people tended to vote leave. And the cities have a younger age profile.

The people who live in Clacton, comparing like-with-like within age-groups, are not much more anti-European than their London counterparts. Actually, things are a bit more complicated – it is difficult to disentangle various compounding effects (younger people have higher educational attainment, for example) but you get the general idea.

People in Clacton did not vote to leave because they live in Clacton, but in large part because they are typical of their age group nationally and the age distribution in Clacton is older than in London.

RTÉ conducted an exit poll last Friday which showed dramatic differences in voting preferences by age group. The RTÉ exit poll over-estimated the Yes vote by 3 percentage points, at 69.4 versus the actual 66.4, a small enough error – The Irish Times exit poll was even closer. The RTÉ figures by age group are set out in the graph.

The percentages are a little high, presumably for all age-groups, since we know what the actual result was. We do not know what turnout was by age-group but we do know that overall turnout was pretty high for a referendum. More importantly, we know (from the 2016 Census) that there are significant differences in the population composition by age across the country.

As in the UK, the urban areas have younger populations. The differences in voter preferences by age group revealed by the RTÉ poll are so large that this factor alone could explain quite a lot of the apparent geographical variation in the actual result.

To keep it simple, the 11 Dublin city and county constituencies voted 75.5% yes, the 29 outside Dublin went 63.6% yes. How much of this apparent 12-point difference is due to the fact that the Dubs are different, rather than just younger? And they are younger, to a substantial degree.

The two youngest age groups with the highest propensity (nationally) to vote yes constitute 34.6% of the Dublin electorate but just 27.4% in the 29 non-Dublin areas. The two older groups, more likely to vote no, are just 36% in Dublin but over 42% outside.

A crude calculation applying the RTÉ percentages to the actual age-distributions shows that allowing for this factor closes most of the apparent 12-point gap. There are individual counties in the west (including Donegal) where the age composition is most heavily skewed to the older end of the spectrum and they had the biggest no votes.

Doubtless the political science people in the Irish universities will start burrowing into the data fairly soon and will come up with a more sophisticated analysis. But the tentative conclusion at this stage is that the Dubs and the culchies are not as different in political attitudes, age-group by age-group, as you might imagine – generational composition is more important than location. Helps with Gaelic football too!