Thursday, November 15, 2012

How long is a generation?

kw: analysis, genealogy, history

Many years ago I noticed that when I drew out my family tree as I knew it then, there was a distinct skew. My father had an ancestor on the Mayflower, who was born just before 1600, in the 13th generation (counting myself as Generation Zero). My mother had an ancestor from Nantucket, whose forebear came to the New World in 1626 on one of the later "fleets", who was born in 1608, but was in the 10th generation. This made the tree longer on my father's side, to reach approximately to the year 1600.

Recently someone asked how many generations there are in a century, another way of asking how many years there are in a "typical" generation. I decided to look into it in my own pedigree. I scouted all the ancestors born between 1590 and 1610 (nearly all were immigrants, and of course, for those who were not, their child or children immigrated). I found 13 on my father's side and 13 on my mother's side. I determined their generation. They are all in generation 10, 11, 12, or 13. I also found an ancestor, born in 1628, in generation 9 (nearly all of that generation were born from the 1640s to the 1660s).

I have three brothers, and the average of our birth years is 1952. Of the 13 paternal ancestors scouted, the average birth year was 1603, and for the 13 maternal ancestors it was 1601. I suppose I could have found averages for each generation, but we'll see that this is sufficiently accurate. The year spans of either 349 or 351 years, divided by the number of generations, produced these results:
  • On my father's side, 349 divided by 10, 11, 12 and 13 yield 34.9, 31.7, 29 and 26.8 years per generation. The grand average is 31.1 years.
  • On my mother's side, 351 divided by 10, 11, and 12 yield 35.1, 31.9 and 29.25 years per generation. The grand average is 32.8 years. Also, that one maternal ancestor born in 1628, in generation 9, leads to a figure of 36 years per generation (this one is not counted in the average).
The overall range of between 26.8 and 36.0 can now be treated as the 2-sigma limits of a normal distribution, centered on about 32 years, with an approximate standard deviation of 2.3. Turning this into a 3-sigma range, I conclude:

For European immigrants to the New World, a generation for them and their descendants, averaged over 10 generations or so, is 32 years, plus or minus 7 (25-39). This ought to cover 997 out of 1000 cases.

It will be interesting to see if I hear from avid genealogists who have 10+ generation cases outside this range. Of course, single generations will be more variable, ranging from 12 to about 50. My wife and I were both just over 40 when our son was born, for example, while the ages of my brothers and their wives when their children were born range from 25 to 35.

No comments: