Mutations can occur in humans due to errors in DNA recombination, radiation, or due to spontaneous reasons. The effects of these mutations can be positive, neutral, or negative, depending on the alteration made. While there is broad agreement that mutations are disproportionately likely to have negative or neutral effects on the phenotype of individuals, there is little agreement on what their net effect is. Some estimate the rate of deleterious mutations to be
"In addition, the association between paternal age and outcomes such as the number of children between siblings has stayed stagnant in Sweden - whether that was before industrialization occurred or after it occurred. If anything, mutational load appears to have been more of an issue in historical Sweden than modern Sweden."
I really wish people would read this paper more thoroughly. The graph is the sibling control results, which is less appropriate for assessing the relative reproductive success of high-ML vs low-ML than for assessing the causal deleterious effects that ML (Mutational Load) has on phenotypes; this is smaller in modern than historical sweden which is just a cherry on top of the paper's two other findings, these being:
1. for the non-sibling-control results, high-ML parents had a higher absolute number of children in historical sweden than low-ML parents, but in modern sweden, low-ML parents had more children than high-ML parents
2. average paternal age is down; people nowadays start having kids later, but stop having kids earlier, and the latter is slightly more important.
People also seem to not understand the study design here. First of all, there are three generations, let's call grandparent generation g1, parent generation g2, and child generation g3. g1 has children at varying paternal age, and this determines which g2 parents are high/low mutational load, and so it's the relative reproductive outcomes of different g2 individuals that they're studying. Second note, there's a reason that "number of children" appears in the chart twice, and they do not mean the same thing:
"(b) Statistical approach: ...We analysed reproductive success for all offspring, including those who died in childhood or never married."
in other words, when looking at how the ML of g2 individuals affects how many children they have, infant mortality + etc is taken into account unless specified otherwise (i.e. in e4), making m1 the most comprehensive possible estimate of the amount of purifying selection that occurs (and again, m1 finds positive effects of ML on success in historical sweden while finding negative effects in modern sweden).
Another ML null finding:
This k=262 meta-analysis found no correlation between publication year and percent left handed
https://rpubs.com/JLLJ/SFE9122 Spanish data shows a negative Flynn effect that is invariant using a higher order model of g. The effect size is ~0.4 points per decade, so not really explainable by mutational load/dysgenic fertility(perhaps demographic changes in Spain?)
Woodley reduced his estimate for ML in his emails with Rindermann in November 2015(Cognitve Capitalism 13.2.4), also differences in generational span mean that the stimate of the dysgenic effect should be increased somewhat. Rindermann estimates the change is from -0.38(Woodley 2014, Rindermann and Thompson 2011) per decade to -0.87(-0.59 if you assume the population starts at 0 years old), with a admittedly basic simulation. This is probably fixed if you use completed fertility(this might explain the differences between the NLSY79 vs international results for the iq-fertility relationship in the US). I also think that at least part of the decline in innovation could be explained by a decline in g(other factors like decreased individualism, and increased regulation is at least a partial one for technological innovation), given that an effect of 0.4 points per decade would decrease the percentage above 130 IQ by almost 50%(with a slight decrease in the mean above this threshold ~0.25 points), and above 145 by almost 60%(with a smaller change in the mean above this threshold).
You and Emil Kirkegaard are my favourite social scientists 👨🏿🔬
"In addition, the association between paternal age and outcomes such as the number of children between siblings has stayed stagnant in Sweden - whether that was before industrialization occurred or after it occurred. If anything, mutational load appears to have been more of an issue in historical Sweden than modern Sweden."
I really wish people would read this paper more thoroughly. The graph is the sibling control results, which is less appropriate for assessing the relative reproductive success of high-ML vs low-ML than for assessing the causal deleterious effects that ML (Mutational Load) has on phenotypes; this is smaller in modern than historical sweden which is just a cherry on top of the paper's two other findings, these being:
1. for the non-sibling-control results, high-ML parents had a higher absolute number of children in historical sweden than low-ML parents, but in modern sweden, low-ML parents had more children than high-ML parents
2. average paternal age is down; people nowadays start having kids later, but stop having kids earlier, and the latter is slightly more important.
People also seem to not understand the study design here. First of all, there are three generations, let's call grandparent generation g1, parent generation g2, and child generation g3. g1 has children at varying paternal age, and this determines which g2 parents are high/low mutational load, and so it's the relative reproductive outcomes of different g2 individuals that they're studying. Second note, there's a reason that "number of children" appears in the chart twice, and they do not mean the same thing:
"(b) Statistical approach: ...We analysed reproductive success for all offspring, including those who died in childhood or never married."
in other words, when looking at how the ML of g2 individuals affects how many children they have, infant mortality + etc is taken into account unless specified otherwise (i.e. in e4), making m1 the most comprehensive possible estimate of the amount of purifying selection that occurs (and again, m1 finds positive effects of ML on success in historical sweden while finding negative effects in modern sweden).
Another ML null finding:
This k=262 meta-analysis found no correlation between publication year and percent left handed
https://not-equal.org/content/pdf/misc/10.1037.bul0000229.pdf
https://rpubs.com/JLLJ/SFE9122 Spanish data shows a negative Flynn effect that is invariant using a higher order model of g. The effect size is ~0.4 points per decade, so not really explainable by mutational load/dysgenic fertility(perhaps demographic changes in Spain?)
>0.4 points per decade
No, I could believe that is due to mutational load or dysgenic effects.
I agree after re-reading some papers in the week after I commented.
Woodley reduced his estimate for ML in his emails with Rindermann in November 2015(Cognitve Capitalism 13.2.4), also differences in generational span mean that the stimate of the dysgenic effect should be increased somewhat. Rindermann estimates the change is from -0.38(Woodley 2014, Rindermann and Thompson 2011) per decade to -0.87(-0.59 if you assume the population starts at 0 years old), with a admittedly basic simulation. This is probably fixed if you use completed fertility(this might explain the differences between the NLSY79 vs international results for the iq-fertility relationship in the US). I also think that at least part of the decline in innovation could be explained by a decline in g(other factors like decreased individualism, and increased regulation is at least a partial one for technological innovation), given that an effect of 0.4 points per decade would decrease the percentage above 130 IQ by almost 50%(with a slight decrease in the mean above this threshold ~0.25 points), and above 145 by almost 60%(with a smaller change in the mean above this threshold).
A demographic cause would be consistent with Finnish data and Woodley's meta analysis