4 ms·

Some of the worst science reporting I've ever read. That said, after reading the journal article itself, I found plenty of false assumptions there as well. "..

Some of the worst science reporting I've ever read. That said, after reading the journal article itself, I found plenty of false assumptions there as well. "... if functional consequences of SR manipulation were to be found in mammals, then it would suggest that mammals (either in individual species, or in general), possess unknown physiological mechanisms to control birth SR." This sentence in the introduction assumes a level of choice on the part of the mammals that is pervasive in the article but never substantiated or even formalized to the point of a testable hypothesis. It is the type of sloppy science that takes away from the truth, and it's becoming way too common. It's as if scientists feel that if they just have some good hard numbers, they get the freedom to toss conjecture around as Theory for a few paragraphs. The study found that gender bias in birth rates can affect the number of grandchildren. That is the takeaway here. No evidence whatsoever of what caused that bias is present, nor do I accept their statement that the study taking place in a zoo means that their findings are somehow more likely to be present in the wild. That would require a different experiment on its own.

takluyver 13y ago
I think the interesting finding is that the biases in sex ratio were towards the more successful sex. That is, if A has more sons than daughters, each of those sons has more offspring on average than other males. But if A has more daughters than sons, each of those daughters has more offspring than other females. That implies that the bias is adaptive: whatever is causing it, it is maximising the number of grandchildren A has. Also, I don't see any problem with conjecture in scientific papers.
azakai 13y ago
> This sentence in the introduction assumes a level of choice on the part of the mammals that is pervasive in the article but never substantiated or even formalized to the point of a testable hypothesis. The field here is evolutionary biology, and specifically the area of parental investment and applications of game theory. It is very common in that area to talk about "choice" etc. while not meaning anything conscious. For example, it is acceptable to say that genes "want" to propagate copies of themselves, but of course this is just shorthand for something very different. In the field, this is not misunderstood, it is the norm. > The study found that gender bias in birth rates can affect the number of grandchildren. No, the point was that they found that gender bias in birth rates appeared to be not entirely random (50-50) and weighted towards what generates an optimal number of grandchildren. That exactly supports a hypothesis that has been around for decades that through female choice (again, choice is not meant in the sense that I can choose what to eat for lunch), animals can affect gender birth ratios in order to maximize the number of grandchildren. (Here you can see the connection to game theory.) It is very hard to study that empirically, which is why this study is receiving a lot of attention - after decades, it is among the first to actually give measurements. > nor do I accept their statement that the study taking place in a zoo means that their findings are somehow more likely to be present in the wild. That would require a different experiment on its own. This is indeed a weakness in their methodology, and surely being debated heavily.
- gcb0 13y ago
  So it's like comparative lit in biology? Joking aside, the grand patent post is right, all the article does is get one consequence (not 50/50 sex split) and immediately assumes parent control. in reptiles it's know that egg temperature determine offspring sex. But it's also believed that parents are ignorant of this fact.
  - azakai 13y ago
    It is not just a non-50/50 split. It is also that the bias, when present, is tuned to optimize for the number of grandchildren. That is the long-standing hypothesis which they have supported.

takluyver 13y ago

I think the interesting finding is that the biases in sex ratio were towards the more successful sex. That is, if A has more sons than daughters, each of those sons has more offspring on average than other males. But if A has more daughters than sons, each of those daughters has more offspring than other females. That implies that the bias is adaptive: whatever is causing it, it is maximising the number of grandchildren A has. Also, I don't see any problem with conjecture in scientific papers.

azakai 13y ago

> This sentence in the introduction assumes a level of choice on the part of the mammals that is pervasive in the article but never substantiated or even formalized to the point of a testable hypothesis. The field here is evolutionary biology, and specifically the area of parental investment and applications of game theory. It is very common in that area to talk about "choice" etc. while not meaning anything conscious. For example, it is acceptable to say that genes "want" to propagate copies of themselves, but of course this is just shorthand for something very different. In the field, this is not misunderstood, it is the norm. > The study found that gender bias in birth rates can affect the number of grandchildren. No, the point was that they found that gender bias in birth rates appeared to be not entirely random (50-50) and weighted towards what generates an optimal number of grandchildren. That exactly supports a hypothesis that has been around for decades that through female choice (again, choice is not meant in the sense that I can choose what to eat for lunch), animals can affect gender birth ratios in order to maximize the number of grandchildren. (Here you can see the connection to game theory.) It is very hard to study that empirically, which is why this study is receiving a lot of attention - after decades, it is among the first to actually give measurements. > nor do I accept their statement that the study taking place in a zoo means that their findings are somehow more likely to be present in the wild. That would require a different experiment on its own. This is indeed a weakness in their methodology, and surely being debated heavily.

gcb0 13y ago
So it's like comparative lit in biology? Joking aside, the grand patent post is right, all the article does is get one consequence (not 50/50 sex split) and immediately assumes parent control. in reptiles it's know that egg temperature determine offspring sex. But it's also believed that parents are ignorant of this fact.
- azakai 13y ago
  It is not just a non-50/50 split. It is also that the bias, when present, is tuned to optimize for the number of grandchildren. That is the long-standing hypothesis which they have supported.