Does Jeanson's model match "the population growth data from 1000 BC to the present"?

I honestly don’t have any idea of what you’re talking about - what does the time period have to do with anything? How can this possibly be true? Covid waves had very different shapes around the world at the same time. Show some kind of evidence of your assertions.

It’s kind of ironic anti-creationists are appealing to spurious correlations now. That isn’t the only example in the book…

It was an example.

Covid doesn’t follow an exponential growth curve – it has peaks and troughs – which an exponential growth curve lacks.

Unless Jeanson (or you arguing on his behalf) can demonstrate (i) a valid reason why each and every one of his 300 Y lineages should represent 1.6 million people, no more and no less, and (ii) a valid reason for assigning each successive lineage the dates he has given them, there is nothing “ironic” about it. I took a look at his spreadsheet underlying it (his “Supplemental table 4”) – it was a mess – approximately 100 columns wide (columns go up to “CS”), with repeating column headings, large numbers of columns showing only values, not formulae (meaning it is difficult, and on occasions impossible, to work out where these values come from) and no documentation within his paper of the overall formula or of the data underlying this calculation.

Given that Jeanson fails to come even remotely close to documenting his working, I can see no reason whatsoever for accepting it at face value.

It is very easily possible that this ‘smoke and mirrors’ would fool somebody with no background in mathematics and/or modelling – but to somebody with that background, it is blindingly obvious. This work would not pass muster even as an undergraduate report.


Not the only example of a spurious correlation? This seems like a cryptic note to make you sound mysteriously knowledgable. But it only makes you look as if you’re avoiding every point.


I didn’t mean it to sound like that. I meant that it’s easy to explain away one correlation as spurious with hand-waving. It’s harder to explain more. I can share the others from the book if you want. I’ve been thinking about your responses, just haven’t had the time to respond in depth until now.

Your question is how can a small sample of the population be representative of the whole?

@Nesslig20 my question to you regarding coalescence and the resulting trees is: Is there any evidence these trees are biologically accurate for humans? Non-random mating could completely change these patterns. It just reminds me of something Sabine Hossenfelder had said in one of her videos: physicists have created books and books of beautiful math. But if it doesn’t model reality, it doesn’t mean anything.

Could you humor me (assuming you used a program to create these graphs and trees?) and use a population of 18 million with a growth rate of 10%? I want to see the probabilities are resulting tree. Also 4 billion and 10% would be helpful :slight_smile:

The “Materials and Methods” section is one long explanation of the formula and data underlying the calculation. I worked at deciphering it, so hopefully I can save you time and simplify the first half or so of the Excel sheet for you - it’s getting late and that’s all I’ve deciphered for now… Forgive me for any misunderstandings of my own.

Jeanson is using the data published by Karmin et al. found here in Table S7 (page 22.) They were the ones who sampled the 300 men globally. They reported molecular distances as years (of evolutionary time). So in Jeanson’s Supplemental Table 4, his first groupings of columns is data from their table using their labels. The second grouping is his conversion to YEC years for each node. The third grouping is the YEC date for each node. You can see those three sets of columns are part of larger groups labeled with either 4206 or 4636 (post flood years).

No. We are talking Y-chromosome lineages (R1b, etc, etc), not individuals. This 322 is the population of identified lineages, not a sample of it. And, according to Jeanson, each of these lineages represents exactly 1.57 million (male) individuals, no more, no less.

That “method” (and I use that term loosely :nauseated_face:) makes no sense! Jeanson rejects Karmin et al.'s ages, but for some unaccountable reason accepts the relative ages (so that a if one branch point is twice as old as another under Karmin et al., it remains twice as old under Jeanson). If Jeanson believes Karmin et al.'s ages to be so hideously inaccurate then why is he using them as the basis for his own ages?

Also why isn’t Jeanson using Karmin et al.'s lower and upper age bounds that consider “also the uncertainty due to the confidence intervals of the mutation rate 0.63-0.95 x 10-9”?

To call all this half-arsed, or half-baked would be a ludicrous understatement. “Half-witted” would be a more accurate description.


In this case it’s simple. You must understand that any exponential curve can be made to exactly match any other if you just change the scales to fit. The number of branches on a coalescent tree happens to look roughly exponential if you graph it over time. If the other fits are the same, they’re similarly obvious, and they mean nothing. Again, the number of branches on a tree does not estimate population size. “Minimum population” is a useless measure.

It should be obvious that it can’t, when you consider that you would get the same tree if the population were expanding, contracting, or stable.


I find it odd, considering how the population curve is supposedly the affirming independent data set, that Jeanson did not use the entire set of points available. That is why he shows population as a straight line from 1000 BC to 500 AD. From McEvedy and Jones, the missing points are ( divide in half for males )

500 AD 190 Million
400 AD 190 Million
200 AD 190 Million
1 AD 170 Million
200 BC 150 Million
500 BC 100 Million
1000 BC 50 Million

So that looks more like a doubling in 500 years, around another doubling in 700 years, then flat lined for 300 years.

Of course, the bigger problem is that the McEvedy and Jones population table extends to 4 million people in 10,000 BC.

Yes, his spreadsheet is a felony. Cell referenced Excel sheets with duplicated columns are a peeve of mine. Jeanson, if you are going to misinform the world, at least do it with intelligibly named excel ranges.


One thing that I think is worth pointing out at this stage, is although we’ve been talking at times about “correlation” (whether spurious or valid), this correlation has simply been asserted not demonstrated (including by the graphs generated by Jeanson’s incompetent spreadsheets), as no formal statistical analysis backing up this claim has been presented to date.

More particularly we have no demonstration that Jeanson’s ‘model’ provides a better fit than simply applying a simple exponential growth curve to the population data. This in turn creates a large question mark over the degree it can be legitimately said to “match” it.

Addendum: it is also worth noting that Jeanson’s ‘model’ predicts that there was an enormous population jump in the late 15th/early 16th century (depending on whether you use his lower or upper bound). Taken from the upper bound, it appears to have been as much as a 35% increase in world population in only half a century. This does not seem remotely realistic.


So @John_Harshman @Herman_Mays @Nesslig20 @dsterncardinale I wasn’t trying to be dense, but I guess I was, because I get what you were saying now. :upside_down_face: Thank you.

For some reason I didn’t get that coalescense theory is using probability to determine the bias or difference in the tree from the actual resulting from sampling. Hopefully that’s a good way to explain it, if that’s understanding it correctly now.

To understand how sampling affects a growing tree, I had to create a growing population, see what the full tree would look like after drift, and then do some sampling. I used a pencil and paper (I have no idea what program to use for this :joy:) so my populations are obviously very small. My first try I had a very steadily growing population. I noticed several things with that example: The full tree will lag population growth. When sampling, if the samples reach most of the lineages, the curve will flatten, as you have later coalescense. If a very successful lineage is oversampled, you could have a curve that is actually sharper than population growth because you have early coalescense. I think Jeanson noticed that in his paper, though not for the right reasons I think - more sampling from a small community leads to a steeper curve.

Here I had a population that had a lot of kids and tried to have sons and then became more stable. My colors don’t correlate here and it’s a mess, but I just thought I’d show an example, so if you want to explain this to others, you can see what helped me figure it out, I thought it was interesting that sampling clusters more closely around the actual population size than the full tree does. I wonder if it would be the case that the correct root and timeline will have sampling that will more closely fit the actual population curve than the wrong root and timeline for a growing population but I need a computer program to check. :slightly_smiling_face: Obviously bigger population sizes and more realistic scenarios would be helpful than my pencil drawings.

Jeanson meant this as an independent test of the timeline, but it looks like he’d have to do much more complicated simulations to show it can roughly indicate a correct timeline, if it does. Or just drop this idea as a test of various timelines.


Valerie, I would like you to answer honestly for me one very simple question, and I realize it may not be an easy question to answer honestly and I very much appreciate that.

You very much want Jeanson to be right about this, don’t you?


You can display genetic diversity in a population as a branching tree even when the population is stable. Jeanson would every time mistakenly say the population is growing.

In actual populations that are growing, undergo bottlenecks, or are stable the trees look nothing like Jeanson’s method would predict.

These are simply facts and it has been explained ad nauseum, literally, I think I may be nauseous explaining it again.

It takes a Herculean effort to get creationists to understand these essential points that anyone in an introductory course on population genetics would be exposed to. Why is that? Why does it take literally months of posts to get a creationist to even begin to acknowledge these facts?

Easy. Because it doesn’t lead them to the conclusion they want to reach. ALL of this is about creationists using science to affirm a prior religious belief. They don’t care if what they are doing is valid scientifically or not they only care if it may be used to make their beliefs appear to be vindicated scientifically. Of course, they are going to be resistant to being taught the actual science so long as it contradicts their religious preferences which is precisely why it takes so much effort to explain why they are wrong.

Jeanson is wrong, dead wrong, and he has absolutely no idea what he is talking about. That should be crystal clear to you by now Valerie.


I’m not sure what you meant by that, but I don’t think it’s true. Coalescent theory says at least two relevant things here: first, it models the time to a common ancestor for a bunch of randomly sampled sequences. Second, it affirms that this coalescence will occur regardless of whether the population is growing, shrinking, or remaining constant. Thus there is no match between the coalescent tree and population size, not even a proportional match.


Sure, that would be great. But I just said he was wrong on this point.

Honestly I needed a visual. I first had to learn about and understand drift correctly, and I wanted answers about a growing population and I wasn’t getting them until @Nesslig20 very thorough reply. I appreciate that. Thinking about it some more I realized what I was missing. Without learning it step-by-step, it’s not intuitive. So I realized I really hadn’t tested a growing population correctly. I did. I was honest about how it came out. I’m probably more frustrated than you. I had to admit I was wrong. I did want to understand. I shared what I learned so it’s easier to teach others because I don’t think a stable population is a very realistic scenario - people will realize that if that’s the example you used and feel they’re being misled, like I somewhat did. But people talking past each other is part of life.

I think creationists should understand evolution better than people in the field. But most people aren’t interested in learning about something they think isn’t true. I think people really think the things they believe are true; and they just don’t usually see a reason to test them.

1 Like

The point is Valerie that if Jeanson applied his method to a tree that we know is from a stable population then he would every time mistake it for a growing population. That alone should cause you to question his claims.

Also, for the vast majority of human history populations have been relatively small and relatively stable. Exponential growth is a very recent phenomenon. Jeanson sets the reader up by only showing an X-axis on his graphs that already fit with what he wants to believe about the time frame for human history and pretends that the coalescent events on his tree fit within this time frame when they in fact are much older.

Jeanson is wrong. Jeanson has no idea what he is doing when it comes to population genetics and molecular systematics. Very basic flaws in what he is doing have been revealed over and over and over again. The only reason he has any audience for this stuff at all is that there are so many people who know nothing about the science but desperately want him to be right so that their religious beliefs are affirmed.


And why exactly would it be great if Jeanson were right?

1 Like

Speaking as somebody with a background in statistics and probability, this does not appear to be the case. The only “sampling” involved at the core of Coalescent theory appears to be to the identification of two alleles for which a common ancestor is sought, not sampling in terms of “Statistical Sampling”. Also determination of “the bias or difference in the tree” does not appear to be central to the model (beyond the fact that any scientific model would attempt to minimise error or bias as much as possible). Reduction of bias and error is a major focus of Statistical Sampling methodologies.

As with many words, “sample” has different meanings in different fields. I’m sure our colleagues in the sciences would be quite irate if Statisticians tried to trademark it as being usable only in the statistical meaning of the word. :slight_smile:


I have no idea where she is getting that description of the coalescent.

This is an evolutionary assumption.

So is this.

I’m not in 3rd grade. :roll_eyes:

No, I tried to explain several times that’s not a good rejoinder to this method.

Ok. What would you call what I was noticed was a difference in trees depending on who was sampled?

Not at all. Small and stable population doesn’t say anything about evolution. These are verdicts of history, anthropology, and archaeology. There are some evolutionary process to be noted, such as lactose tolerance and of course haplogroups, but that is peripheral to our reconstruction of population.

Time frames are supported as above, and by geochronology techniques such as dendrochronology and radiocarbon dating, which are not evolutionary.

While the usage of “evolution” does sometimes take on the meaning of any transformative process over time, it does not encompass all of science and it does not designate deep time. Ancient population is not an evolutionary assumption. Even Jeanson maintains that it is an independent data set, right? Nor does the record of population depend on any way on evolutionary thought, although anthropology does inform evolutionary principles.


You’re not the only one. I have to work things out on paper all the time. You have no idea how many piles of scribbles and stuff I’ve thrown out and still got laying around. To learn anything you have to try to work it out yourself. It’s extremely difficult to read a sentence and just get it without some sort of picture in mind.