How Common is your Surname?

I guess I am stuck on surnames for a while. One of the major challenges of genealogical research is the relative with a very common name. Unique names, especially unique surnames are quite rare since in most of the world's cultures some form of the name of a parent is often passed on down to a child.

Let's explore the issue of the common or uncommon surnames with some hypothetical searches. Let's suppose that your ancestral line has a name such as "Smith." Most English speaking people would recognize that Smith is a very common name. But where is it common and how common is it and why, as a genealogist, would I like to know this information? Let's extend our hypothetical some more. Suppose you are searching for an ancestor named George Smith. What are the chances that you will find other unrelated people in the same area where your ancestor lived with that exact name? I you know the answer to that question, you then know how difficult it will be to separate out those people and you also know that your chances of choosing the wrong person are either high or low.

Now it is time for an example search. In this hypothetical situation, I am going to suppose that my "George Smith" ancestor came from England and that he was supposedly born in the county of Kent in 1880. OK, so what do I do next? I use Findmypast.com to make the following searches.

The first search is to get an idea of how common the name is for all times in all the records on the Findmypast.com website. So I search for the surname "Smith" in all the records without putting in a timeframe or being any more specific than limiting the search to England. This is what the search screenshot looks like:

This is what the search results look like:

I have just confirmed by previous supposition that Smith is very common name. But how many Smiths are there in Kent born about 1880 in Kent? I will now edit my search to see.

Here are the results:

The number drops considerably but is still unworkable. What about adding the given name George?

Now we are getting an idea of the challenge of finding our particular George Smith. I have a one in 4,184 of guessing which George Smith is mine if all I know is his name, his birth year and the county in England where he was born.

But what I constantly see in online family trees is exactly that. I see people identified with a name, a generally vague date and a non-specific place. What does this show? It shows that these guesses about the identity of their person are much more likely wrong than they are correct. Here is a real example from the FamilySearch.org Family Tree.

OK, here the date is 1715. The place is a little more specific but there are no sources. So, this person is not real. He is a guess. How accurate is that guess? Let's do the exercise from the hypothetical. First, how many people in England at all times have the Tarbutt surname? I won't show all the screens this time, but the number is relatively small if I search for all the name variants: 176,091. Now what if I put in a date and a county? The number drops to 103 a significant improvement in the odds that I have the right person. Now what if I add the name George? Here are the results.

What? There are no results? What does this mean? It means that in all the records on Findmypast.com, they do not have one George Tarbutt (with variations) that lived in Kent in 1715. So what are my chances of finding him? Almost exactly zero. Why?

Let's go a few steps further. Here is his family as it appears in the FamilySearch.org Family Tree:

Now, if I look at his son, William Tarbutt, I see that he has extensive documentation in form of 31 sources with three record hints waiting to be entered.

All is not lost, we may still find George Tarbutt. But we now know that we have to do some real research in real records and not simply rely on an unsupported entry in the Family Tree for an identity. Time to get to work. Our surname search has lead us to the methodology we should use to determine the accuracy of our work.