Scottistics

[Guest]

A few authors, most notably Niall Ferguson in his “The Pity of War 1914-1918” and E Spiers in “The Scottish Soldier at War” in “Facing Armageddon” have claimed that 26.4% of Scottish recruits died in WWI. On the back of this figure some authors and newspapers have also implied that the sacrifice of the Scottish nation was proportionally far greater than any other allied nation and only less than that of Serbia. I believe this figure and the associated claims to be greatly exaggerated. It is a figure based on extremely poor analysis. The real numbers are tragic enough, without any need to exaggerate the enormous sacrifice made by the Scottish people.

The common statistics used to produce the headline 26.4% come from two sources: the denominator is based on the 557,618 men and women who were recruited in Scotland during WWI (as recorded by the War Office in 1920). This does not include the standing Army (regular and TF) at the outbreak of the War. Secondly the numerator is based on 147,208 men and women recorded by the Scottish National War Memorial (SNWM) as died in WWI. To save you the calculation, 147,208 divided by 557,618 is 26.4%.

Since the books were published, the SNWM figure had adjusted to 148,706 which would make the ratio of Scots recruits who died marginally higher at 26.7%. I have analysed the whole SNWM database with help from Geoff’s search engine and the SNWM website. (many thanks Geoff). There is one primary issue with the integrity of the statistic derived from the SNWM. The database includes 16,592 duplicate entries (triplicates and quadruplicates in some cases). Adjusting for this, the number of Scots died would be 132,113.

Analysis of the residual 132,113 reveals that 72,321 of these men were born in Scotland (54.7%). This figure is remarkably consistent with the War Office’s initial 1920 estimate of 72,000 Scots died in the Great War (later revised to 74,000). Of the remaining 59,792 some 19,975 were born outside Scotland of which 16,555 were born in England, presumably including an unidentified number of men with no Scottish heritage. The 132,113 also includes 39,817 men who have no recorded place of birth – some 30.1% of the residual sample. A fair proportion of these served in the Canadian, Australian, NZ and South African forces.

It is important to recognise that the SNWM does not claim that the men and women on its database are only Scots, merely that they served with a Scottish regiment, or were expatriate Scots serving with Canadian, Australian forces etc., or were Scots recruited elsewhere serving in Irish, Welsh or English regiments and Corps. This latter group also includes the London, Liverpool and Tyneside Scottish. The point being that the 132,113 includes tens of thousands of men (and women) who were not part of the 557,618 men (and women) recruited in Scotland. The breakdown of these expatriate Scots dead as recorded by the SNWM is as follows:

Canadian Expeditionary Force......................................................................7,661
Australian Expeditionary Force.....................................................................4,308
South African Scottish..................................................................................1,178
New Zealand Expeditionary Force...................................................................888
Scottish Casualties Extracted from English Regiments.................................9,498
Scottish Casualties Extracted from Irish Regiments......................................2,775
Scottish Casualties Extracted from Welsh Regiments.....................................283
London Scottish*..........................................................................................1,554
Liverpool Scottish*........................................................................................1,070
Tyneside Scottish*........................................................................................2,671

Total ‘Expatriate Scots’...............................................................................31,886

* SNWM data differs slightly from CWGC and SDGW in all three cases. The variations are so small compared to the database that they would have an insignificant impact on the conclusions. It is worth noting from the samples above the number of Scots born recruits in the London Scottish, Liverpool Scottish and Tyneside Scottish as a per cent of recruits whose birthplace is known is respectively 12.0%, 8.1% and 2.9%. Despite this, for the purposes of the exercise I have assumed 100% have Scottish heritage.

Many might argue that these men should not be classed as Scottish and should be stripped out of the calculation. If we did this, the number would reduce to 100,227. This is within 0.2% of the 100,000 estimated in a white paper when the SNWM was first established in the 1920s, it being generally accepted that the then official figure of 74,000 was too low. What defines a Scot is highly subjective, and can be an emotionally charged debate, so for arguments sake, let’s assume all these 31,886 men and women were part of the Scottish diaspora. If these are to be included in the ‘Scottish’ dead then we need to include them in the base of the calculation too as they were not part of the 557,618 men and women recruited in Scotland. Critically the denominator of the calculation should include their ‘Scottish’ comrades who also signed up in the same units but survived. This would generate a more accurate figure of the number of ‘Scots’ died as a per cent of all the ‘Scots’ recruited.

Example. For every one of the 7,661 expatriate Scot who joined and died with the Canadian Expeditionary Force according to the SNWM data, many more would have joined and survived. We can make a good estimate of those numbers as the Canadian Government has kept records. The CWGC records a total of 64,976 Canadian dead and the Canadian Government records 611,711 men and women signed up during the Great War. This suggests that 10.6% of Canadians who signed up died in the Great War. It is a reasonable assumption that warfare is indiscriminate and Canadian Scots died in similar proportions to their Canadian comrades of non-Scottish origin. If the 7,661 Canadians Scots died represents 10.6% of all Canadian Scots recruits, this implies roughly 72,273 expatriate Scots would have signed up. If expatriate Scots died are to be included, then the denominator of the calculation should be adjusted to include the 72,000 Canadian Scot recruits.

A similar calculation from CWGC figures and the official figures from the various Govts can be made for the Australian, South African , New Zealand etc. Using Govt data for recruits and CWGC data for the dead. For the British troops we can use the numbers from ‘Statistics of the Military Effort of the British Empire during the Great War, London HMSO, 1922 part V (War Office) which, adjusting for the Scots recruited in Scotland and associated dead, and the standing Army at the outbreak of the Great War, the remainder of the British Isles suffered 11.7% dead.

Following the arithmetic through, the calculations would need to add 471,025 to the 557,618 in the denominator, This adjusts for the 80,687 Scots in the standing Army (Regular and TF) serving at the outbreak of the Great War (i.e. not ‘recruited in 1914-1918) and the 390,338 ‘Scots’ serving outside the Scottish Regiments with the Canadians, Australians, English etc. (80,687 + 390,338 = 417,025) The grand total is 1,028,643. (Calc: 417,025 + 557,618).

Taking the SNWM figure of 132,113 (adjusted to eliminate the duplicate entries) and the larger recruiting base to include all expatriate Scots of 1,028,643 we get a figure of 12.8% - a figure less than half of the original claims. This figure is still slightly inflated as it does not adjust for the non-Scots serving in Scottish Regiments. Of the 81,530 Scottish dead in who served in the Scottish infantry regiments recorded in the SNWM data, 80.9% of those men whose birthplace is known were born in Scotland. Implicitly this suggests 19.1% were not born in Scotland, but subjective arguments can be made about men born outside Scotland to Scottish parents who returned. Being born outside of one’s country does not preclude claims of nationality However, it is clear that some men with absolutely no Scottish heritage are included in the 132,113 and I would suggest at least 10% would be a conservative estimate (roughly half of the men born outside Scotland who served in Scottish regiments). Due to the subjectivity of this assumption I have deliberately not included it in my calculation.

The calculations are not that complex. There are a few sources for the numbers necessary for the adjustments. Typically the War Office figures are consistent with the Government figures, but often slightly conflict with the CWGC data on the numbers of dead. The CWGC data is in every case the higher number, perhaps reflecting decades of post war adjustments to the original data immediately collated after the War. In every case I have used official figures for the numbers recruited. Due to the sight differences between the CWGC and Govt data, it is possible to derive slightly different final numbers . As we are analysing very large numbers, comparing results from the extremes of the sources would vary the conclusions by less than a per cent.

There is no nationalistic agenda. My aim is merely to establish a more accurate figure than the one being propagated through various publications and newspapers. I fully accept that there are assumptions in my approach, but I believe them to be very conservative. Even if one was to make very generous assumptions, it is impossible to get a figure anywhere close to the 26.4%. Others may differ in their opinions and it is not my aim to convince anyone, merely to highlight a different methodology. If anyone wants to see the detailed calculations I would be happy to share them. I spoke at length with the Secretary to the Trustees of the Scottish National War Memorial and I am indebted to them for their advice and guidance.

Any mistakes are entirely mine.

PS. Numbers have been edited for a slight error in the database. The net effect on the conclusion is an upwards revision of 0.2%. (20 May 12)

[Tunesmith]

You've evidently done a massive amount of work on this, Martin. A very interesting read. Many thanks for sharing it.

[WilliamRev]

This is a fascinating article - many thanks. For a while I have wondered about the two regular battalions of the Royal Scots Fusiliers, with between a quarter and a third of their number recruited in England, many on the South Coast, and whether it is fair to count all of the casualties as "Scottish" as seems to have sometimes been the case.

William

[phil andrade]

Well done, Martin !

This is a disturbing indication of how grosly misleading statistics enter folklore. It's all the more worrying when it emanates from the pen of a historian of Niall Ferguson's calibre and renown.

People find statistics unpalatable, especially casualty statistcs. It seems callous and unhealthy to study tabulations which relate to human agony, in which the destruction of flesh and blood and bone is reduced to ciphering. I myself have antagonised one or two people on the Forum by excessive fixation in this regard. but I do believe that it's very important that we try and get things right ...... especially when it comes to the fate of millions of people.

Phil (PJA)

[hazelclark]

That has been a massive study Martin. I did not think it was possible, and as you say it is an emotional subject for many including myself. You seem to have conducted the study in a dispassionate and impartial way and I thank you for that. As you have shown, it is possible to look at the figures without denigrating any particular group and thus should not cause distress to anyone.

Hazel C.

[phil andrade]

Might it be that popular perception of Scotland's loss of life in the Great war is dominated by Loos, September 25th 1915, just as that of the United Kingdom is set by the First Day of the Somme ?

Phil (PJA)

[geraint]

Excellent research Martin. I've often wondered how figures are compiled. A relative of mine from Cerrigydrudion, a rural village, speaking only pigeon English went to work at Cammel Laird Birkinhead. He became friendly with a couple of Scots lads, then in August 14 went with them up north to join the A+S Highlanders and was killed at Gallipoli. Is he a Scottish casualty?

[nigelfe]

The whole 'Scottish heritage' thing is very flaky and meaningless. How do you count someone who had one grandparent each from England, Wales, Ireland and Scotland? Are they counted four times in four different sets of 'national' data?

Afternote: this whole matter should be called 'fictistics'.

[phil andrade]

It is bad, though, isn't it ?

A very reputable historian - who predicates his arguments on tabulations of statistics - states that the death rate was virtually double the true figure.

I have to say, that I was suspicious about that claim from the moment that I saw it, and I am so grateful to Martin for working hard and confirming my suspicions.

I wonder how many other erroneous calculations are being bandied about. Officers having a life expectancy of three weeks, a generation wiped out, sixty thousand men from the BEF being "killed" in a single day, and so on....

Phil (PJA)

[salesie]

It is bad, though, isn't it ?

A very reputable historian - who predicates his arguments on tabulations of statistics - states that the death rate was virtually double the true figure.

I have to say, that I was suspicious about that claim from the moment that I saw it, and I am so grateful to Martin for working hard and confirming my suspicions.

I wonder how many other erroneous calculations are being bandied about. Officers having a life expectancy of three weeks, a generation wiped out, sixty thousand men from the BEF being "killed" in a single day, and so on....

Phil (PJA)

And let's not forget, Phil, the Sanitatsbericht being previously bandied about by some on the forum (up until last year that is) as being the one true source (the Holy Grail) of German casualty stats!

As for the figures in this thread - Martin has done an excellent job in highlighting the inherent problem with statistics (especially when it comes to WW1 casualties) i.e. numbers are just numbers until compilation and interpretation, then those very same numbers are transformed into "Lies, damn lies, and statistics"; and which one of the three they are seen as is invariably a matter of point of view and/or hidden agenda and/or poor maths.

　

Cheers-salesie.

[phil andrade]

Have I gone mad, or was there an elaborate graphed presentation by Martin about Scottish regiments and their true racial provenance, which has just disappeared ?

Phil (PJA)

[Guest]

Pulled it as it was not quite finished. Will post it again later..... Have stumbled on something really quite interesting that should I think show that a significant number of the Scottish diaspora who make up the numbers of the SNWM database don't appear to have particularly Scottish names. ...so you are not going mad. = : o

[hazelclark]

Might it be that popular perception of Scotland's loss of life in the Great war is dominated by Loos, September 25th 1915, just as that of the United Kingdom is set by the First Day of the Somme ?

Phil (PJA)

Or the Welsh at Mametz Wood or the Canadians at Passchendaele?

[phil andrade]

There is something about the Scottish reputation and Loos which stands out more. Perhaps because it is more associated with a single day, as opposed to days and weeks of persistent attacks. Maybe I'm wrong to say this, but whenever I think of bagpipes, kilts and other Scottish caricature from the Great War, the name Loos comes straight to my mind.

I hope you don't mind me writing in bold format like this : I find the text hard to discern otherwise.

Is there a reckoning of how many Scotsmen died on September 25th, 1915 ?

I remember reading in Vera Brittain's Testament of Youth that the British people commemorated three days from the war that were conspicuous for enormous loss of life : July 1st 1916 ; March 21st 1918 ; and September 25th 1915.

It also comes to my mind that the French sneered at the English, and maintained that they left the heavy fighting to their Dominion troops and to the Scots !

Edit : Niall Cherry's book tabulates the casualties of 44th and 46th Brigades - which consisted entirely of Scottish battalions - as 4,312 for the 25th September 1915. That is shocking : an average of 539 per battalion. There were plenty of other Scottish units : the six Scottish battalions of 9th Division suffered 1,708 casualties that day, and there were several others that suffered very heavily. It looks as if the best part of ten thousand casualties were incurred by Scottish regiments that day ; enough to account for a legendary reputation.

Phil (PJA)

[Guest]

While crunching the data of the 148,000 names on the Scottish National War Memorial over a number of weeks it was apparent that there were subtle differences between some of the data sets. The data was attacked from two entirely different standpoints to ensure that any particular methodology didn’t skew the conclusions when eliminating duplicate (and triplicate etc) entries. The data was sifted by unit, and separately by surname of the individuals on an alphabetic basis. As a secondary step, the alphabetical tranches of names were subsequently sub-divided among the unit headings. During this process it became very clear that the distribution of surnames was very uneven, and patterns were evolving.

One glaringly obvious feature in the lists of Scottish surnames is the high frequency of the letter M due to the high incidence of names starting with Mc and Mac. It struck me that some frequency analysis might throw something up. When comparing the frequency of each group of surnames starting with each letter of the alphabet (measured as a % of each unit total) it was immediately apparent that there were distinct patterns, and sub-groups. This might suggest there was some underlying fundamental difference in the distribution of names across different units. The samples were just too large and the variations were too wide for this to be random. Certain units had significantly higher frequency of surnames starting with particular letters than others. Not only were there distinct patterns between units, there were distinct patterns within sub-sets within units. An analysis of the data gives us clues as to the country of origin of other samples within the SNWM data where the country of birth is not known.

The most pronounced difference between the various subgroups is between the surname distribution of the Scots-born men compared to the English-born men from the SNWM data. The respective samples are 72,000 and 16,500 which is hopefully large enough to be statistically relevant. The Scots-born-born sample can be considered as a benchmark sample. The chart at the bottom shows the frequency as a % of the total in each sample. They are generically known as a “Manhattan” charts due to its similarity with the NY skyline. Of note, the two groups have some distinctive characteristics that act as a signature profile for the two sub-sets:

1.The expected spike in names beginning with the letter M in the Scots born-sample is a major signature of the distribution of Scots-born samples. It is two-and-a-half times higher than the English reading at 22.2% v 8.8%. This spike is the most distinctive characteristic of Scottish data samples and is more pronounced the further north a regiments’ recruiting area is based. An analysis of each Scottish infantry regiment reveals distinct differences in the regimental signatures; the highland regiments having significantly higher spikes in M than the lowland regiments and the Scots Guards. This perhaps reveals the growing underlying influx of English recruits as the data sets move gradually south. The M spike of the Cameron Highlanders is 27.4% which is in stark contrast to the Tyneside Scottish’ M spike of only 8.1%, the lowest of any regimental group.

2. The similar height of the ‘six pillars’ in letters B, C, H, P, M and W in the English-born names are distinctive.

3. The very high English H is a particularly important signature as it does not appear in Scottish samples (or Irish samples). Its presence might indicate a local concentration of English names. Similarly the higher W than in Scottish samples is notable and is distinctive feature of English names.

3. The English B is more frequent than the English C while the Scottish B is less frequent than the Scottish C. This English ‘step-up’ and Scottish ‘step-down’ is an important differential.

4. The twin towers B and C and the outlier S, all similar frequencies is distinctively Scottish.

5. A dip in the sequence J,K L in the English sample v. a straight increase in J,K, L in the Scottish sample is notable

6. Both samples show a steady increase in frequency in the series E, F, G, H.

7. The down-and-up in the sequence M,N,O is common to both but more pronounced in the English sub-set

These discrete differences can be used as templates to compare other samples where the place of birth is not known. The samples of Canadian Scottish, Australian Scottish, as well as the large blind spots in the Scottish Casualties Extracted from English Regiments…Irish Regiments …Welsh Regiments. Food for thought.…… Any mistakes are mine ….. More to follow…..MG

[WilliamRev]

This is fascinating - I'm not sure why, but it is!

William

[hazelclark]

There is something about the Scottish reputation and Loos which stands out more. Perhaps because it is more associated with a single day, as opposed to days and weeks of persistent attacks. Maybe I'm wrong to say this, but whenever I think of bagpipes, kilts and other Scottish caricature from the Great War, the name Loos comes straight to my mind.

I hope you don't mind me writing in bold format like this : I find the text hard to discern otherwise.

Is there a reckoning of how many Scotsmen died on September 25th, 1915 ?

I remember reading in Vera Brittain's Testament of Youth that the British people commemorated three days from the war that were conspicuous for enormous loss of life : July 1st 1916 ; March 21st 1918 ; and September 25th 1915.

It also comes to my mind that the French sneered at the English, and maintained that they left the heavy fighting to their Dominion troops and to the Scots !

Edit : Niall Cherry's book tabulates the casualties of 44th and 46th Brigades - which consisted entirely of Scottish battalions - as 4,312 for the 25th September 1915. That is shocking : an average of 539 per battalion. There were plenty of other Scottish units : the six Scottish battalions of 9th Division suffered 1,708 casualties that day, and there were several others that suffered very heavily. It looks as if the best part of ten thousand casualties were incurred by Scottish regiments that day ; enough to account for a legendary reputation.

Phil (PJA)

Can't figure out the "quote" thing as i am technologically challenged but it is interesting that you have also read about the French view of the English versus the Dominions and the Scots last night, I finished Richard Holmes book "The Western Front" in which he makes the same statement quoting a French officer who said "The English are useless, it's the Scots the Australians and Canadians who do all the work".(page223) Anyway, if that feeling was general one can see how the exaggeration of casualties came about. And remember the Newfoundland Regiment that was virually wiped out.- 733 out of 801 casualties at Beaumont Hamel. Newfoundland's population was relatively small and those kinds of things tend to stick in people's minds.

By the way, I am NOT casting aspersions on any one group or another. Anybody who was in any way involved in the hell that was WW1 deserves a medal in my opinion, and probably most of the heroics went unobserved and the soldiers concerned dead.

Hazel C

[phil andrade]

Newfoundland and Beaumont Hamel came to my mind, too, Hazel.

The thing about Loos that makes it transcendental as far as Scotland is concerned is that not only were the Scotish casualties that day extreme.....it might also be the case that, quite literally, the preponderance of British dead that day were Scottish. If this is the case, then we have to countenance the possibility that that single day rather distorted perception for posterity.

Another thing : there had been a Victorian love affair with Scottish soldiers for generations : The Scots Greys at Waterloo; The Thin Red Line at Balaclava...think of those pictures by Lady Butler.

This persisted in a sense for generations after the Great War. Heck, I can remember listening as a kid about fifty years ago to Andy Stewart singing some sickly dirge " There was a soldier, a Scottish soldier, who wandered far away....".

Maybe Niall Fergussson was captivated by it, too !

Phil (PJA)

[nigelfe]

Of course English surnames may indicate a Scottish mother.

[phil andrade]

There is a "Flowers of the Forest" syndrome at work in the perception of Scottish losses in the Great War.

If I'm right in my suggestions about Loos, then it's all too understandable.

It goes beyond the Great War : it's even entered American folklore.

The historian Grady McWiney argued that the excessive casualty rates of the American Civil War can - to a degree - be attributable to the Scottish provenance of so many of the soldiers, particularly in the South. There was, he suggests, a tradition of the "Claymore Charge" instilled in them, which gave them a thirst for the attack and a determination to engage in close quarters combat. They paid a lavish price in blood for this.

Phil (PJA)

[trajan]

Martin G,

An excellent piece of solid scholarship! No doubt it wil be reanalysed countless times, but you have put a lot of data there for us and given many food for thought.

Trajan

[hazelclark]

Newfoundland and Beaumont Hamel came to my mind, too, Hazel.

The thing about Loos that makes it transcendental as far as Scotland is concerned is that not only were the Scotish casualties that day extreme.....it might also be the case that, quite literally, the preponderance of British dead that day were Scottish. If this is the case, then we have to countenance the possibility that that single day rather distorted perception for posterity.

Another thing : there had been a Victorian love affair with Scottish soldiers for generations : The Scots Greys at Waterloo; The Thin Red Line at Balaclava...think of those pictures by Lady Butler.

This persisted in a sense for generations after the Great War. Heck, I can remember listening as a kid about fifty years ago to Andy Stewart singing some sickly dirge " There was a soldier, a Scottish soldier, who wandered far away....".

Maybe Niall Fergussson was captivated by it, too !

Phil (PJA)

This is getting to be a bit off topic but if you think about it, life was hard in Scotland at the turn of the century and the tougher, more adventurous members of families went to the "new world". to make something of their lives and in many cases to pave the way for younger siblings and friends from the "old country". Scotland's main export was her people. I would therefor suggest that many of those men in Dominion armies, fighting for their "homeland" were some of the best and toughest that Scotland had produced. Of course there were lots of Irish, Welsh and English in this category also.

As you say, there has always been this mysticism about the Scottish soldier and when you think of the numbers who became mercenaries in European armies (particularly French) after the '15 and the '45 rebellions it is not surprising. Don't forget also how many fought for the English with Wolfe on the Plains of Abraham. "No great Mischief"

[Guest]

Here is an interesting Manhattan chart comparing the A-Z surname frequency distribution of the Northumberland Fusiliers (ex Tynside Scottish) compared to the Tyneside Scottish. Both samples include English-born, Scots-born and men whose birthplace is unknown. The Northumberland Fusiliers sample is 15,000 men. Note that the chart's Y axis is on the same scale as the first chart on post No.15.

What struck me is the close match with the distribution of the large Northumberland Fusilier sample which is dominated by English-born. As a check I have compared the Tyneside Scottish to a random English Line Infantry Regiment (the Worcestershires) and the charts are very similar. All the key distribution characteristics of samples dominated with English-born surnames are in the Tyneside Scottish sample.

1. Lack of Scottish M spike

2. B-C step-down

3. EFGH rise, culminating in a high 'English' H

4. The six pillars of B,C, H, M, S and W

5. The fall-and-rise of N, O, P with a higher P

6. J,K,L step-down-and-up

Etc.....

Note the complete absence of the Scottish M and other typically Scottish distribution characteristics: 8.1% v All Scots-born SNWM data 22.2%. The Tyneside Scottish sample is smaller at 2,671, but (I think) still is statistically meaningful. 1% would be 27 men in this sample. It is statistically very unlikely that all the Tyneside Scots had English fathers and Scottish mothers etc... Food for thought. I have done similar exercises with the Liverpool Scottish, London Scottish etc and am now working on part III which will look at the 100 most common surnames.

Nothing conclusive yet, but I think the data is beginning to reveal some of its demographic secrets.Many thanks again to Geoff and his amazing search engine. Any mistakes are mine. MG

[dundeesown]

Must admit you have me stumped,with Scottish M spike and English H spike are these charts of just Scottish and English names ? what are you classing as a Scottish or English name ?

and I assure you I am baffled and not just at it.

Gary

[Guest]

Gary, they are merely the distributions of surnames from 2 samples of men died in the great war (Post 15). One sample is of men born in Scotland, the other is of men born in England. All it does is measure the proportion of men who have surnames starting with the letters A, B, C etc..... when we look at the distribution of these sub-sets, the two sets of data have very different distribution profiles. The main differences are listed in post 15. It proves nothing other than show these two samples have different distributions. If you did this exercise with data from different countries from 1914-1918, each would have different distributions that would be characteristic of that particular nation. The Irish distribution is different still, as is the Welsh. German , Spanish, French etc would all be different in their own way. When looking at other samples (Tyneside Scots, or Liverpool Scots for example) we can see if the distribution of the surnames is more similar to English-born or Scots-born. and possibly make some assumptions.

There really isn't any name that can be classed as Scots or English, almost all names are found in both population samples from 1914-1918. What is different is the proportion that various names appear. The most common surnames in the English language are common to both England and Scotland - Smith and Brown for example* - but some surnames that are common to both occur in significantly larger proportions in Scotland than in England - the Mc and Mac for example - there are hundreds (thousands?) of people with Mc or mac born in England, but there are significantly higher proportions born in Scotland. This is not a debate, it is a simple fact. It is part of the allure as it eliminates a significant part of the conjecture. The higher occurrence of certain surnames in the Scottish-born groups makes certain groups in the alphabet stand out - M in the Scots samples and H in the English samples for example, so it follows that a random third sample of men (with no recorded birthplace) that shows a particularly high proportion of names starting with the letter M might (I stress might) suggest it is a Scottish group of men

This is all very preliminary, but the results (so far) surprised me a lot. A more refined study would look at, perhaps the 100 most common surnames among English-born and Scots-born and probably try and identify surnames that were very common in one group and not common in the other. The appearance of these names in other samples in large numbers might suggest an high proportion of men from one or other communities. Put simply, if you went to America and found a large random sample of men with a very high proportion with name starting with Mc or Mac, you might reasonably assume some Scottish heritage, although you might not be able to prove it.

MG

* I hope to soon be able to show some analysis on this...