Have you ever wondered how long is a generation? Learn how long a generation is by looking at various historical populations and better understand how generation length may affect your DNA matches in this article from Jayne.

Can you think of any unusual generation time situations in your family? Any cases where the age of a parent when a child was born was outside of what might be considered the norm?

How long is a generation?

Here’s what I mean. Often we might think of a new generation being born about every 20-25 years. But on an individual level, this varies quite a bit. My mom is the “surprise” youngest in her family of siblings by 11 years. Her oldest brother is 20 years older than she is. My grandma had her first child at 22 and her last at 42. Taking the average of her ages at the time of birth of each of her 5 children, my grandma’s average generation time is 29.2 years. That’s skewed quite a bit older from that conventional wisdom of a 20-25 year generation time. 

This gets even more interesting as we look at subsequent generations. My mom has a niece who is older than she is, which is pretty unusual.

Figure 1. An instance of “negative generation time” where members of the next generation are older than members of the previous generation.

At the second cousin level, the age differences between my children and many of my cousins’ children look like they could span an additional generation, but they really are 2nd cousins.

Figure 2. Differences in age within the same generation can magnify and become inflated over subsequent generations in specific cases.

You may begin to contemplate generation time yourself as you investigate your DNA match list. Trying to narrow down your potential relationships can be influenced by your age difference. You share about the same amount of DNA with your great-grandparent as you do your first cousin, but an age difference of 70 years would all but rule out that first cousin relationship.

In genetics, there’s another space where the effect of an estimated generation time becomes considerable. Researchers use generation time to estimate how far in the past major events in the development of the human family occurred. Through changes in DNA seen in branches in the human family in different parts of the world, researchers can detect the number of generations ago that things like continental migrations occurred. 

For example, the peopling of the Americas by way of the Bering Strait is estimated to have occurred by some figures between 700-1000 generations ago. To translate those figures into a date, we need to have a good estimate of generation time. The effect of using a figure that is off by 10 years can be huge– it moves the migration date by 10,000 years! So for good reason, scientists have put some effort into making an evidence-based figure for generation time that represents the human experience well.

How many years are in a generation?

So, does that 20-25 year per generation figure hold up? Researchers have used a couple of strategies to vet this:

1. Observation. By combing through historical records, or cataloging the ages of parents at childrens’ births in various kinds of modern populations
2. Inference from genetic data. An understanding of the types of mutations that can happen when DNA is transmitted forward from parent to child, and when in a parent’s lifespan they are more likely to occur, has been used to estimate generation times over the long course of human history (which is pretty darn cool!).

Let’s take a look at some of the findings of the different efforts to query historic human generation time, starting with observational research. In each of these studies researchers cataloged the age of men and women at the time of birth of their child(ren). Data was compiled from written histories or longitudinal studies covering decades of contemporary families.

Table 1. Estimates of human generation time in the literature. References for each population are listed at the end of this article.

Some things to notice here. Men are able to participate in child-bearing deeper into their lifespan than women, and this is reflected in the skew to an older average generation time for men than women in all these geographically diverse populations. 

What about that 20-25 year generation time figure? For both men and women, 20-25 years appears to be an underestimation. In some cases, observed figures exceed 30% longer generation times for men, and 10% longer average times for women than the originally assumed 20-25 years. And to be certain, these numbers reflect an average, as even for the same person the age of their parenthood at the time of their first- to their last-born child will vary significantly from the numbers shown here.

Apparent also may be a difference in average generation time among different cultures. A general thought may be that as societies’ economies develop from forager to agrarian and to stages beyond, the generation time lengthens. That trend can be detected somewhat in this smattering of studies, albeit with some cases that disrupt that pattern. Perhaps notable is that for contemporary hunter-gatherer societies the generation time for females tracks to the 25 year generation time, a younger average age than for females in other groups. However, male generation times seem to be similar in hunter-gatherer groups, around 31 years, to many other populations examined around the globe.

Examining contemporary population patterns can be useful in its own right, but caution should be drawn to apply generation time figures obtained today as generally applicable extending into deeper history. Scientists have detected changes in the rate of growth of the human family over hundreds of thousands of years that can be correlated to environmental, cultural, and demographic factors that can change rapidly. In the relatively recent history of the United States, for example, researchers demonstrated a modest “marriage squeeze” for women extending within a decade of the Civil War when 620,000 men died, leaving somewhat of a vacuum of marriageable partners. However, this trend self-corrected quickly and was no longer detected within 10 years of the end of the war.

Beyond observing contemporary groups whose reliable records extend perhaps several hundred years at best, geneticists have developed new tools for estimating generation times throughout the history of the human family through its deep 250,000 years. Within our DNA itself is a record of mutations that have occurred since our most ancient origins, and are a gold mine for this type of question. 

Figure 3. A SNP is a single-base-pair mutation that happened once in history, and can be dated.

The kinds of mutations that scientists examined for this question are a change in a single base pair on the DNA strand–really a very tiny change. Through studying each of these tiny changes and how they show up in populations across the world, geneticists can tell that the mutation happened only once (it doesn’t mutate back and forth readily at all) at a certain point in time and can actually date it.

There is a large repository to which scientists all over the globe contribute their work to track and date these small DNA mutations– it’s called the Atlas of Variant Age. 

Another key piece of this puzzle builds on previous findings that demonstrates that certain types of mutations are more likely to happen depending on the age of the parents. For instance, there are types of mutations that are rare when parents conceive a baby in their early 20s, but become much more common in their offspring when parents conceive in their late 30s and 40s. 

By examining DNA mutations that occurred at certain dates in the past, together with the findings that parental age influences the type of mutation, researchers have been able to predict the average generation time at many points in the last 250,000 years of human history. Analyzing this information also allowed them to separately estimate average parenthood ages for men and women throughout antiquity.

Take a look at this graph and its astounding discoveries (Figure 4). I marvel that we can learn so much about the past by deciphering the clues left behind in the DNA of living people. Let’s walk through the findings together.

Figure 4. Human average generation time over 250,000 years

First, as we go left to right on the graph we go farther into the past, back about 250,000 years. As we move upward from the bottom of the graph, the average generation time lengthens. There are several things to notice.

1. Today we exist at the time in the entire history of humans when average generation time is at its longest, approaching 30 years. 
2. Moving back in time, the average generation time was at its all-time shortest (24.9 years) about 6400 years ago, followed by a rapid increase. This change is attributed to the rise in human civilizations occurring at approximately this time.
3. Back in time further still, a peak is seen about 38,000 years ago followed by decline. This is just before the beginning of a world event known as the Last Glacial Maximum, or ice age. In this extreme environmental episode that lasted for about 15,000 years, ice sheets that were miles thick covered much of North America, Europe and Asia. Paradoxically in other areas of the world, continental swaths that weren’t locked in by ice became overly arid. A significant constriction in world populations was seen during this time, and findings from this study also show a coinciding drop in generation time as well.

In previous studies, certain mutations have been attributed distinctly to male or female ancestors. This information was leveraged to separately estimate male and female average generation times at different points in history (Figure 5).

Figure 5. Male and female average generation times over 250,000 years

The blue and red points, respectively, estimate average male and female generation times at different points in the past. We can see that as expected, the parenthood age of men skews older than women, with the narrowing and widening of that gap apparent in trends over time. The generation time of both men and women has rapidly increased to the current day, closing the gap somewhat compared to previous ages.

A major takeaway of this study is that it provides figures for average generation time overall for humans across the 250,000 year course of history, and separately for men and women (Table 2).

Table 2. Human average generation time over 250,000 years. Standard error for each average: overall 26.9 ± 3.4 years, women 23.2 ± 3.0 years, men 30.7 ± 4.8 years.

How do the average generation times obtained through genetic analysis compare to times estimated by observing different contemporary cultures? In all cases, they are shorter!

And they get even shorter when looking separately at generation times by geographic area: East Asia, South Asia, Europe (Figure 6). While humans expanded from Africa into these areas well before 10,000 years ago, researchers discovered that generation times were on average less than 20 years for populations outside of Africa until the most recent 5,000 years (rise of early civilizations) when they began to increase steadily to today. In contrast, generation time in Africa has remained more steady through the last 10,000 years, right around 27 years.

Figure 6. Average generation time in the last 10,000 years by continental groups: AFR- Africa, EAS- East Asia, EUR- Europe, SAS- South Asia.

So how is this all relevant, and what do we do with this information going forward?

Many genetic genealogists and other researchers have recommended using generation times approaching 30 years for females, and 30-35 years for males. It may not be surprising that those longer generation time numbers came through in their research, because they were observing the relatively recent history of the human family: the last few hundred years.

For you, as a genetic genealogist, applying generation time knowledge to your own mysteries, you may choose to use those longer generation times because your people likely come from the very recent past. And for people studying archaic humans, the shorter generation intervals are likely more appropriate than longer figures obtained from modern populations.

I imagine, like me, you have a DNA match list with various unknown DNA cousins who you are trying to sort out and identify. Having a round-about figure for generation time can be a significant tool in figuring this out. As mentioned early in this article, taking into account the age difference between you and another match can eliminate relationships that are compatible by shared DNA, but nearly impossible by time (grandchild/grandparent share a similar amount of DNA as half-siblings, but if two matches are about the same age the grandparent relationship is very unlikely). 

The cM Explainer tool developed by MyHeritage interprets age difference considerations by automating that calculation and incorporating it into a probability that includes not just shared DNA but also age difference. This brings to your attention relationships that are more likely with that mystery cousin in your match list. Their technology is based on extensive research into family tree data that examined generation time (age of parents at the time of child birth), and age differences among other types of relationships. Here’s what MyHeritage saw for generation times in their pedigree data.

Figure 7. Generation time data for MyHeritage aggregated pedigrees

It looks like they observe similar figures to other studies of recent pedigrees and populations– late twenties for an average generation time of combined ages of men and women. From the research we’ve reviewed above, this is similar to other pedigree and modern population studies, and longer than generation times in more ancient trends. If MyHeritage split this out separately by male and female parents, we would expect to see a longer generation time for men than women.

The age differences observed for parent-child and other relationships form a scaffolding to help you interpret your DNA match list and figure out that mystery match. But you’ll still have to use your trusty old noggin to weigh what this algorithm presents to you.

For example, my Mom is about the same age as her niece and shares a typical amount of DNA for that relationship. When putting their data into cM Explainer (a generously free standalone tool), it all but completely obscures their true relationship. 

Figure 8. Output from cM Explainer for an aunt/niece relationship, where they share a similar age

This effect can become even more confounding in subsequent generations from an unusual generation time event, like my Mom’s. 

Figure 9. cM Explainer output for 2nd cousins with a typical amount of shared DNA but an unusually large age difference

My mom’s descendants and those of her siblings have unusually large age differences. These second cousins have a 32-year age difference. They share a typical amount of DNA for second cousins, and cM Explainer all but eliminates their true relationship.

One of the takeaways from what I’m trying to demonstrate to you, dear reader, is not that the cM Explainer tool is suspect, but the nature of probability is that even very low likelihood relationships are the right answer sometimes! You cannot truly eliminate any of the relationship probabilities on this list when trying to figure out who that mystery match is. You need to continue to entertain all of them, and bring in all the supplementary information you can find in addition to DNA evidence to determine that accurate relationship.

Well, it’s been fun to go on this long journey through the 250,000 year history of humans to discover how generation time influences what we can learn about each other, whether modern or ancient. I hope you have an increased sense of awe at what an incredible chronicling of events we carry around with us in every cell of our body. 

Now that you have a better understanding of how long a generation is, you may be interested to learn more about other generations before you in your own family. Get our free guide to find those ancestors using DNA.

Download the FREE Guide: Finding Ancestors Using DNA

References for data in Table 1:

Botswana Dobe !Kung: Nancy Howell, The Demography of the Dobe !Kung (1979; second edition New York: Walter de Gruyter, 2000).

French Canadian families 1850-1990s: Tremblay M. Bezina H. New estimates of intergenerational time intervals for the calculation of age and origins of mutations. American Journal of Human Genetics 2000; 66: 651-658

Icelandic genealogies 1742-2002: A Helgason, B Hrafnkelsson, JR Gulcher, R Ward, K Stefansson. A populationwide coalescent analysis of Icelandic matrilineal and patrilineal genealogies: evidence for a faster evolutionary rate of mtDNA lineages than Y chromosomes. American Journal of Human Genetics 2003; 72: 1370-1388.

Hunter-gatherer societies: Fenner JN. Cross-cultural estimation of the human generation interval for use in genetics-based population divergence studies. American Journal of Physical Anthropology 2005; 128: 415-423

Less developed nations: Fenner JN. Cross-cultural estimation of the human generation interval for use in genetics-based population divergence studies. American Journal of Physical Anthropology 2005; 128: 415-423

Developed nations: Fenner JN. Cross-cultural estimation of the human generation interval for use in genetics-based population divergence studies. American Journal of Physical Anthropology 2005; 128: 415-423

North Greenland families: Matsumura S, Forster P. Generation time and effective population size in Polar Eskimos. Proceedings of the Royal Society of Biological Sciences 2008; 275: 1501-1508

English males: King TE, Jobling MA. Founders, drift, and infidelity: the relationship between Y chromosome diversity and patrilineal surnames. Molecular Biology and Evolution 2009; 26(5):1093-110K2

Pedigree of Confucius over 80 generations: The longest family tree. GenealogyInTime Magazine, accessed 11 January 2015