I’m starting my series on the Irish in the 19th century Black Country by looking at Wolverhampton. This is familiar ground for me, or at least should be – so I’m broadening my normal hyper-local view of Carribee Island out to look at some quantitative data on Wolverhampton as a whole. My hope is that this will give me opportunity to place the Irish in my apparent “Irish quarter” in some sort of context; but also to have a look at how far one can get using solely quantitative sources.
Sources of data
Digital sources are an unquestionably fantastic thing. They enable me to gather vast amounts of data from my office chair, but they aren’t without their difficulties. I love the Online Historical Population Reports (HistPop) website for example – it publishes most of the census enumerators statistical tables for the whole of the 19th century and beyond, but it dates from c.2003, and the web technology shows it. More up to date is the awesome I-CeM project, which draws its data from individual census returns provided by FindMyPast and enables users to drill down into anonymised returns as actual, computable data. Its drawback? It’s subscription only.
The whole population
The first thing I noticed is basic counting differences. On HistPop, the original enumerators’ tables show an 1851 total population of Wolverhampton of 49,985, with an Irish-born population of 3,491 – that’s 6.98%. On I-CeM, calculated from individual census returns rather than these tables, the total population is 49,798, the Irish-born population 3,783, and the proportion 7.56%. This is point one, perhaps: no matter what your historical source is there is always human intervention somewhere along the line – there is no such thing as unbiased data.
The Irish population
I’ll continue with I-CeM’s data because of its flexibility, but bearing that in mind. Either way, the Irish-born population is by far the biggest minority in the town. Of that 3,763, most are young – the 26-30 bracket is 824 – and married (3,318 compared to 286 widowed and 86 single). Occupational categories are notoriously unhelpful, but we’ll have a look. The most common category is “farm labourers, general farming” with 533. In Wolverhampton? My theory here is something of an uncertainty on the part of the respondents; perhaps it’s comforting to define yourself by what you have come from in Ireland – what you hope to return to – and not by the way you are struggling by on one evening in March. Indeed, the Irish-born are way over-represented in this occupation – just 283 English, 16 Welsh, 2 Scottish and 16 others are in farming (and 10 English and 2 Welsh are “farmers’ sons”).
This map appears to confirm the basic thesis [all maps can be opened out to larger versions by the way – just click on them]. The concentrations of Irish are on the NE side of the town centre. Producing this map threw up a few basic issues. This is a choropleth map – i.e. showing statistics by area – showing the 1851 census enumeration districts (imperfectly, I might add – it’s surprisingly difficult to judge the districts from their written descriptions alone). You can see the massive ones on the outside and the smaller near the centre – that’s to be expected really, but although the enumerators tended to try and divide up population areas reasonably equally, there are massive differences – some districts are 21 pages, others are 50. So this is a bit inflexible. They’re also tricky to map accurately without a digitised version of an 1851 map – which there currently isn’t. The other thing to notice is the lack of nuance in these categorisations – it’s interesting to note that every census district in town had at least one Irish-born individual, but we can’t distinguish here between 1 and 185. Perhaps if we had more categories?
This map shows ten categories and… it’s not all that different. We still can’t distinguish between 1 and 74 in the large Southern and Western districts of Penn, Finchfield, Merridale and Goldthorn Hill, for instance. We can get a bit of a better breakdown in the NE section – the densest districts are still those around Carribee Island and Stafford Street (which comes as something of a relief, seeing as I’ve predicated my thesis on the idea that there was a densest Irish area in the town). We can see that the Irish population extends up Stafford Street and across the town centre a bit as well – these are all areas which are densely populated and constructed with court housing to a large extent, which confirms my suspicions in terms of the homes of the Irish. We can draw some basic conclusions from here, but I still think it’s missing something, don’t you?
The next map is categorised into non-equal areas. Happily, ArcGIS has a number of different breakdowns to switch between – this one is the cartographer George Jenks’ method of finding the natural breaks in a dataset and exploiting that for the best visualisation. You can find the full methodology here. We now see the density map in a lot more detail, showing with a lot more clarity clusters of other Irish populations in Wolverhampton – along Walsall Street and St James’ Square, for instance, and along Salop Street. This shows better what I’ve been expecting based on, for example, the Watch Committee minutes, which regularly mentions these areas as Irish hotspots. It throws up some unexpected results too – there’s a cluster on Merridale Street in the SW of town, and a high proportion in the as-yet-not-hugely-developed Springfield area, to the far NE. I would have expected this in 1881 or 1891 more so than here – the area was developed as part of the slum clearance programme that knocked down Carribee Island. There’s an interesting gap to note as well – despite being surrounded by heavily Irish areas, Faulkland Street (just to the North of the darkest green area) is very light. Certainly this includes St Mary’s church and Terrace, but what was it about this street that stood it out? It’s still a street of mid-century court housing.
The next option is Standard Deviation, which is no longer a measure of population numbers but a statistical measure showing deviance from the mean. So we can’t calculate people numbers – just difference. It’s a useful visual device though, particularly to show unusual results. The average number of Irish-born residents in a Wolverhampton census district in in 1851 is 50.75. In some areas of Wolverhampton this would be ludicrously high: the semi-rural area to the South of Chapel Ash has just one return, for instance. That compares to the aforementioned Merridale Street, not too distant, with 45 – enough to put it in a higher bracket in the Jenks distribution, but still below average. Standard Deviation ignores that sort of thing, but is really useful for showing us the outliers. From this, for instance, we can note immediately that the area to the West of North Street (where the town and civic halls will later appear) is unusual – and looking more closely, there are 117 Irish-born residents, right at the upper limit of the Jenks bracket. So – there’s a result I wouldn’t have necessarily clocked. Carribee Island, we see again, is more than 2.5 SDs outside the mean – in fact there are 743 Irish-born residents here.
A perhaps more useful SD visualisation is achieved using 0.25 SDs. Standard Deviation (he says, plumbing the distant depths of A-Level Statistics and failing to remember a single thing) is the square root of the average deviation from the mean. Got that? Basically, we now have a range of 16 categories instead of 4. We can see from this that while that district we just mentioned (W of North Street)is an outlier in some respects, it fits into a broader distribution of Irish around NE Wolverhampton, appearing to radiate out from the darkest blue – Carribee Island.
You can see that each of these maps tells a different story. All of the them are true in a basic respect – they are all based on the same raw data. But different statistical tools and measurements show quite different sides of the argument. There will be times when showing the basic, equally-categorised data will be enough. At other times, a statistical tool reveals a lot more.
It ought to be borne in mind that there are all sorts of other issues with using census data – try Higgs’ key Making Sense Of The Census for this. There are even more issues with just taking one snapshot day as representative data – especially because, like a modern geographer or social scientist, I didn’t design the research tool myself. For a historian, contextualising such data is the point of what we do. The biggest contextual point is that 1851 saw the first unblighted potato crop in Ireland since 1844. For the previous six years, the great famine had sent something like a million people fleeing their homeland; and a million dead of starvation. Many went to America, but many came to the UK and a fair few of those found themselves in Wolverhampton on census night 1851. Many will have been in transit – who knows how many? I’d recommend reading John Herson here – his description of Stafford as a transitory town for Irish immigrants is really useful. Immigration only increased after this point too. Many settled in the town too – but what is settled? Comfortable? Economically stable? Residentially secure? Not likely.
A bit of geographical context would be handy too. Roads, canals, public functions – all play their part in the choice (or lack of) of where to settle. What I’ve not mapped is industry, residential districts, commercial districts… all play their part in a much bigger picture.
Choosing how to present data visually is going to be one of the biggest methodological challenges I’ll face in my PhD, I think. There’s a lot of pros and cons, limitations and extrapolations, and so on. Hopefully, this post shows a little of my workings – what I don’t think it does is to show fully the situation of the Irish in Wolverhampton by itself. Much more context and argument is needed. You can get a basic idea of the residential distribution of the Irish in the Victorian city from these maps – the biggest Irish populations lived where the houses were less pleasant, nearer to industry, between the shopping centre and the suburbs, typically to the NE. You can see, too, that the Carribee Island area – and the districts surrounding it – really was the hotspot of Irish population in Wolverhampton at this time. I could add topographical data, longitudinal data, more detailed general population data, other data sources, and so on – but you’ll just have to wait til my thesis is completed.