Talk:Generation time

Untitled

The current article has (at least) three problems:

• It only deals with the existing definitions and formulas for the generation time, and does not discuss its importance. It would be nice to add some content showing how it is used.
• It is underlinked. For instance, "age-structured population dynamics" should probably be linked, but I wasn't able to find a relevant link on Wikipedia. The same goes for a few other technical terms.
• It doesn't have enough in-links. I tried to add a few ones (eg. in the "average familial generation length" of the Generation article), but this is probably insufficient.

Malparti (talk) 20:36, 22 May 2014 (UTC)

Definition of generation time

About 1 month ago I added a sentence stating that another definition of generation time was "the average age of the mother when she has her first offspring". This was based on the original definition of generation time as "the average time between two consecutive generations in the lineages of a population", since my interpretation was that a new generation started when the first members of that generation were born. An anonymous user stated that this definition was inaccurate and unrelated to the original, but upon further research and a look at the Wikipedia article "Generation", I found a source that stated something very similar to what I did: a generation can be defined as "the term of years, roughly 30 among human beings, accepted as the average period between the birth of parents and the birth of their offspring." (Dictionary.com). I am of the opinion that my original definition was correct, but perhaps it should be changed to this second one? TROPtastic (talk) 06:13, 4 August 2017 (UTC)

Hi TROPtastic. I am the anonymous user who reverted your edit, and I have reverted it again. I will explain why below. Please let us not start an edit war and let's agree on what should be done on this talk page without going back and forth between the two versions in the meantime.

It is common to assign a precise name to a fuzzy concept. If you are familiar with population dynamics, this will not surprise you -- think about carrying capacity, or fitness. If you are not familiar with these notions, think about the notion of sweetness of a food: we both know what it means ("the sensation you get when you put sugar in your mouth"), and this is very convenient to talk about food. However, if we were to do some science about food, we might have to quantify it. And to do so we would have to define it, which would prove very challenging: natural definitions (e.g, the sugar content) are usually too naive to be useful; and there will often have to settle somewhere between a practical (e.g, "how much does it have to be diluted till the average person can't taste its sweetness anymore?") and a theoretical definition (presumably involving the affinity with the sweet receptor(s?)).

I took the time to develop this toy-example about sweetness because I think it reflects what our problem is with the generation time: we know what we want it to mean, but there is no "best" way to define it mathematically. Plus, there is an extra complication caused by the fact that (1) it is used by people from different fields (mostly biologists and demographers) and (2) in both fields, it is used both for theory and in applications.

I am the main author of the current version of the article and, admittedly, I'm more on the "theoretical biology" end of the spectrum, so that might bias the choice of definitions presented in this article. However, on the theoretical end of the spectrum biology and demography tend to converge, and formulas used to compute the generation time in practice are often approximations of the ones presented in the article. These are the reasons that made me choose to focus more on the "theoretician's" definitions of generation time. I have given the three main ones, used by both demographers (see Coale 1972 in the bibliography) and biologists (Charlesworth 1994 is in population dynamics, and I have also seen demographers cite it). To be honest, I don't think the third definition is so useful, and judging from the articles I've read it seems that it is less used by researches. Anyway, might point is that for many researchers, these are "the" three main definitions.

Now remains the question of whether we should add the definition you mention. In populations dynamics textbooks it is usually omitted (I'll try to give some arguments for this), but I have seen it used: (1) in demographic tables (2) in few biology articles where the average age of mothers should have been used instead (for instance, in the very famous 1993 paper by Martin and Palumbi, Body size, metabolic rate, generation time and the molecular clock I think they use this definition -- but the average time between two meiosis in the genealogy of a gene is the average age at mothers at birth, and not the average age at which they had their first child). Generally speaking, let's say that the age at birth of first offspring is not very useful in population dynamics and evolution. Because of the reason I mentioned above, it seems less suited for studying molecular evolution. And in most applications, individuals, though they can have different traits, are born "identically distributed" so that there is nothing special about the first offspring.

A case might be made that the age of birth of the first child could be more relevant for some applications related to sociology. I have no doubt that this is probably the case, but the current version of the article is not concerned with that so we cannot just use this definition in the introduction and then forget about it. We would need first to write a dedicated section about that measure, ideally with some examples of applications. And then we could talk about changing the introduction. However because I do not think that this measure is superior to the other, I would object using it in the introduction -- as said earlier, the generation time is a vague concept, so the introduction should reflect that and remain vague, rather than try to give a precise (but arbitrary) definition.

Finally, in your edit you had changed the range given for the generation time. Judging from the figures you used, I assume you have been looking at chart SF2.3.B for the year 2014, whereas the figure I chose are based on chart SF2.3.A, for all years. This gives a range of 24-32 years. However, this is only for OECD countries, and does not take into account the fact that the generation time is shorter in less developed countries -- and might have been even shorter at earlier times in the history of mankind. This is important because in many applications (e.g, evolutionary biology) it is the generation time at the paleolithic more that in modern society that is going to be relevant. I chose 22 arbitrarily because it gives a 10 year range. Nevertheless it's probably a poor choice because the second digit gives the impression that this is a precise figure. 20-30 would probably be much better, but I was bother by the apparent inconsistency with the OECD charts...

Best, Malparti (talk) 00:33, 8 August 2017 (UTC)

Hi Malparti,

First of all, I'd like to sincerely thank you for the really thoughtful and informative reply. It would have been much easier for you to have said something like "I am highly knowledgeable about theoretical biology, your 'corrections' are wrong" and left it at that. The fact that you took the time to write out your response and explain your position when you didn't have to (and in a civil and non-condescending manner to boot) says a lot about your character.

Secondly, I'd like to say that I actually do agree with everything that you say in your message. The "toy example" of trying to accurately define sweetness really helps to illustrate your point and makes the sources that you cite in your argument (which are rather more rigourous than the dictionary entry I cited) more understandable. It now makes a lot of sense to me to say that the definition of "generation time" in this article should be one (or more) of the main definitions that are actually in use by professionals in the relevant fields instead of a less-accurate "common language" definition like the one I suggested. I also agree that an "age of mother at birth of first child"-type definition would require its own section explaining its significance if it were to be included in the article, given that it's used less often than the other definitions. Finally, your clarification as to why you changed the "24-32" age range back makes perfect sense. I too had some misgivings about OECD countries not representing developing nations, but I thought it was safer to cite the age range strictly and qualify it with something like "as found in OECD countries." However, after reading your explanation on the purpose of the article I think that it makes more sense to keep the range as general and broadly-applicable as possible. 22-32 makes sense as something that matches up approximately with the OECD data while reflecting the different demographics in poorer countries, but I agree that it does sound more specific than it actually is and that could contribute to some misconceptions.

The one thing that I will suggest adding is some sort of qualifier to the introduction summarizing what you've said to me. I imagine that most people who come from a position of inferior knowledge and read your explanation would quickly understand and agree with the rationale behind the definitions you chose. However, a casual observer who stumbles upon the article and only reads the introduction without any background knowledge or delving further into the topic may misunderstand the purpose of the article and think "oh, there's a common definition that I can add here, and look, the age range doesn't even match up with the source", make the 'corrections', and go off thinking that they've made a positive contribution (like I did). A few sentences explaining that this is an article about the scientific definition of generation time and that the resulting definitions are necessarily less precise (but also less arbitrary) than laymen's definitions could really go a long way to discouraging 'unintentional defacing' of this article in the future. The age range especially is something I could see being changed again in the future without some sort of clarification (although perhaps this talk page conversation will be sufficient?). If you agree with my suggestion, I would definitely leave it to you to come up with a way to add this information to the introduction, since I think you would be able to leverage your greater knowledge base to come up with something that improves understanding of the article, the concepts within, and the article's purpose without writing something that is incorrect.

Once again, I really appreciate your civil response and the thought you put into it. It really helped me to see how my position was wrong and further develop my understanding.

Warm regards, TROPtastic (talk) 01:50, 8 August 2017 (UTC)