March update and collecting statistics

 I've been carrying out research on several 'non-PARRY' branches of my own family during this past month, and there's been very little progress specifically about PARRYs, or for the wider One-Name Study, so this will be quite a short update.

On the 15th March, I attended the Malvern Family History Show.  It was lovely to catch up with numerous members of the Guild of One-Name Studies, who were either there helping with the Guild stand, or on other family history society stands, presenting talks, or just having a day out, as I was.  

But I didn't take any photos, not even the lovely view of the hills in the sunshine, as we queued just before the show opened, which I now regret!

One of the aims of a One-Name Study is to develop "added value", going beyond the stories of individual families, in order to draw conclusions about the surname itself¹. One topic that I had hoped to write about this month concerns the frequency and distribution of the PARRY surname in British censuses.  Back in 2005, when my web site was still available, I had collected some preliminary information relating to the numbers of PARRYs in the censuses from 1851 - 1901, and created the following image, giving some indication of how the surname distribution had changed over those years: 

I'd also produced an attempt at mapping a modern (2002) distribution of the surname, which is available on the PARRY study profile, at https://one-name.org/name_profile/parry/ .

In the years since those were produced, the data for the other British censuses (1841, 1911, and 1921), as well as for the 1939 Register, has become available, and I have made a start on collecting the information from those, so that I can produce a more complete sequence of map images.

But things are never simple, are they?

One would think that it should be a fairly straightforward task, to do an exact search for the surname PARRY in the various genealogical databases, obtain totals per county, and then create the maps.  

But, no, it isn't!

I knew there was likely to be some differences between the totals provided by different companies - it is rare that they have been able to use data for a census that has already been transcribed.  So they have all undertaken their own transcription process, and therefore show differences due to the way the surname has sometimes been transcribed.  

Even if they had been able to use an 'already transcribed' dataset, as might have been possible with the 1881 census, which was initially transcribed in a joint project between the Federation of Family History Societies (now the Family History Federation), and the Genealogical Society of Utah, over the years, as users submit corrections to whichever company they might have searched in, the datasets would develop differences. (Although one might hope that this would eventually lead to a coalescing of the totals, as all the mistakes get picked up and corrected.)

But when you discover that the individual county totals from a particular company, when added together, don't even match the overall total that company shows for the census, then you realise a lot more work is going to be needed to obtain even moderately accurate figures!

I think some of the issues are to do with registration districts that overlap county boundaries, and therefore get included within both individual counties.  There are also variations in how the companies treat individual counties (for example, why does TheGenealogist lump Huntingdonshire and Rutland figures together?  They might be small counties, but they don't even border on each other.) 

Clearly, I need to do a lot more investigation before I can draw any conclusions, and post some updated distribution maps.

Notes and Sources

1. About One-Name Studies: https://one-name.org/one-name-studies/

"Similarities and Differences" - a late challenge post!

 Back in July, at the start of this current Guild "Blogging Challenge", I wrote:

 "Don't let a long gap prevent you from posting something!" 

I'm reminding myself about that today because, having not posted anything during October, and seeing the end of November approaching rapidly, it would be very easy to give up and say, "that's it, I've failed the challenge." 

But, in an effort to get back on track, this is just a short note, to comment on what I might have written, regarding October's topic of "Similarities or Differences." 

One of my ideas for this was to compare the PARRY surname with that of PERRY, since the two spellings are often interchanged - although the names have very different origins and distributions.  

Two things prompted this idea.  The first was due to me clearing out some old paperwork and finding a letter from a university, which confirmed that I'd successfully completed a period of study with them - but spelt my surname as PERRY.  (Fortunately, that mistake didn't have consequences for what I did next!)  Then, secondly, the Guild webinar for September was Howard Benbrook giving a talk entitled "Most Surnames come from Somewhere." He illustrated this using the Surname Atlas program that was produced by Steve Archer, and which contains numerical data for all forenames and surnames in the 1881 census of England, Scotland and Wales.¹

Surely, I thought, a combination of map images, to illustrate the distributions, combined with some verbal descriptions about the origins of the two surnames, would make an interesting article that fitted the topic?

Of course, having an idea is only the first step - it then needs to be followed by action, putting the information together in order to write the post.  And that was when life got in the way!

When I sat down at the computer today, I wasn't actually going to write about this - I'd been planning to start on November's post, and just conveniently 'forget' about October.  But I happened to come across an interesting website called "Select Surnames," by Colin Shelley, which provides very similar information about the two surnames, along with some examples of name bearers.  You can see his page at Select Surnames.

Since there are no images there, the following shows the distributions derived from Surname Atlas.  In the top row, the images show total numbers of PARRYs and PERRYs per county.  This is followed by the frequency figures (number of PARRYs and PERRYs per 100,000 people.)  As you can see, for the PARRYs, this moves the northern concentration away from Lancashire and Cheshire, back into North Wales, in particular, Anglesey, and the South Wales concentration moves from Glamorgan up into Brecon and Cardiganshire.  While the PERRYs do appear strongly along the English side of the border with Wales, adjusting for population indicates their highest concentration is in Somerset.

The lower two rows of images are derived from the same figures (total numbers and then frequency), but displayed per census district.  

There are probably several points that could be made from these but, for now, I think it's sufficient to say that, although there is some encroachment of both surnames in opposite directions across the border, the images do illustrate the different distributions of the two surnames, and the fact that PARRY originates in Wales and PERRY in England. 

1. Howard Benbrook's talk, https://one-name.org/surnamessomewhere2/

Surname Atlas, by Steve Archer, website http://www.archersoftware.co.uk/satlas01.htm