Several minor edits

paper.qmd

@@ -53,7 +53,7 @@ Abstract
 
 <!-- Theory and evidence suggest that the mating benefits of muscle mass in human males trade off with costs of increased energy intake and decreased measures of native immunity, likely due to an evolutionary history of sexual selection. It is unknown if females experience a similar tradeoff. Using data from the 2013-2014 phase of the National Health and Nutrition Examination Survey (NHANES), a large nationally representative sample of the US (N = 4384), we will test whether grip strength, a proxy for upper body strength and physical formidability, is a positive predictor of self-reported measures of adult mating success (lifetime number of sexual partners, past year number of sexual partners, age at first intercourse, and current partnership status) in both males and females, controlling for numerous anthropometric, socioeconomic, hormone, health, and physical activity related potential confounds. We will also test if there is a positive relationship between grip strength and dietary energy and protein intake, and a negative relationship between grip strength and native immune function. This study replicates an earlier study of US males [@lassek2009] with numerous additional control variables, and it additionally tests these relationships in females. -->
 
-Theory and evidence suggest that the mating benefits of muscle mass in human males trade off with costs of increased energy intake and decreased measures of innate immunity, likely due to an evolutionary history of sexual selection. Lassek & Gaulin (2009) demonstrated a positive association between male fat free mass and limb muscle volume and mating success, but did not investigate women. It is therefore unknown if females experience a similar tradeoff. Using data from the 2013-2014 phase of the National Health and Nutrition Examination Survey (NHANES), a large nationally representative sample of the US adults (N = `{r} nrow(designsH2$d.design.adults)`), we tested the prediction from the sexual selection hypothesis that the effect of upper body strength, proxied by grip strength, on mating success is significantly positive for males and significantly less so for females. We found a main effect of strength on mating success proxied by lifetime number of sexual partners and current partnered status, but not past year number of sexual partners or age at first intercourse. We found consistent evidence for a $grip\ strength\times sex$ interaction on partnered status, such that strength was significantly more important for male partnered status than female (but no significant interaction for lifetime number of partners). We also tested for tradeoffs of upper body strength with immune and dietary intake and found a positive relationship between grip strength and protein and energy intake, but no significant relationship between grip strength and innate immune function. Our results suggest that sexually dimorphic upper body strength might have evolved, in part, by increasing male long-term mating success.
+Theory and evidence suggest that the mating benefits of muscle mass in human males trade off with costs of increased energy intake and decreased measures of innate immunity, likely due to an evolutionary history of sexual selection. Lassek & Gaulin (2009) demonstrated a positive association between male fat free mass and limb muscle volume and mating success, but did not investigate women. It is therefore unknown if females experience a similar tradeoff. Using data from the 2013-2014 phase of the National Health and Nutrition Examination Survey (NHANES), a large nationally representative sample of the US adults (N = `{r} nrow(designsH2$d.design.adults)`), we tested the prediction from the sexual selection hypothesis that the association of upper body strength, proxied by grip strength, with mating success is significantly positive for males and significantly less so for females. We found a main effect of strength on mating success proxied by lifetime number of sexual partners and current partnered status, but not past year number of sexual partners or age at first intercourse. We found consistent evidence for a $grip\ strength\times sex$ interaction on partnered status, such that strength was significantly more important for male partnered status than female (but no significant interaction for lifetime number of partners). We also tested for tradeoffs of upper body strength with immune and dietary intake and found a positive relationship between grip strength and protein and energy intake, but no significant relationship between grip strength and innate immune function. Our results suggest that sexually dimorphic upper body strength might have evolved, in part, by increasing male long-term mating success.
 
 # Sexual dimorphism in modern humans
 
@@ -277,7 +277,7 @@ Our methods and models remained identical to those reported in the pilot study,
 
 ## Deviations from Stage 1 methods
 
-Models and R code for three of our four mating success outcome variables remain identical to those used in the Stage 1 pilot study. However, although the Poisson regression model of lifetime sexual partners in the pilot study included age as a control, it is more appropriate to instead include age as an exposure variable (an offset) in the model, which we now do. We also treat this exposure as years since sexual maturity (defined as age - 12 years because 95% of participants first had sex at age 12 or older). The changes  in the grip strength and $grip\ strength\times sex$ coefficients and their standard errors from this change in model specification are minimal (see @sfig-comparison).
+Model specification and R code for three of our four mating success outcome variables remain identical to those used in the Stage 1 pilot study. However, although the Poisson regression model of lifetime sexual partners in the pilot study included age as a control, it is more appropriate to instead include age as an exposure variable (an offset) in the model, an adjustment we apply to both the pilot and confirmatory models. We also treat this exposure as years since sexual maturity (defined as age - 12 years because 95% of participants first had sex at age 12 or older). The changes  in the grip strength and $grip\ strength\times sex$ coefficients and their standard errors from this change in model specification are minimal (see @sfig-comparison).
 
 As described in the methods section, a different set of weights is required in models that utilize dietary data. In one of the models of pilot data using dietary variables (the expanded controls immune model) we used the MEC weights instead of the dietary weights, an oversight we correct here in both the pilot and confirmatory models.
 
@@ -383,21 +383,21 @@ Unlike grip strength, arm, leg, and total (but not trunk) lean masses were all n
 
 ## Discussion
 
-The purpose of this study was to replicate, as closely as possible, the @lassek2009 study with new data that used combined grip strength as a proxy for upper body strength, as well as the lean masses of arms, legs, trunk, and total body. Unlike @lassek2009, we included women to test if the effects of upper body strength and lean masses on mating success were greater for men than women, as predicted by the sexual selection hypothesis. We also included numerous additional control variables, and used a registered report format with pre-specified models developed using pilot data in Stage 1 and then fit on held-out data in Stage 2. 
+The purpose of this study was to replicate, as closely as possible, the @lassek2009 study with new data that used combined grip strength as a proxy for upper body strength, as well as the lean masses of arms, legs, trunk, and total body. Unlike @lassek2009, we included women to test if the associations of upper body strength and lean masses with mating success were greater for men than women, as predicted by the sexual selection hypothesis. We also included numerous additional control variables, and used a registered report format with pre-specified models developed using pilot data in Stage 1 and then fit on held-out data in Stage 2. 
 
 In the Stage 2 Confirmatory Study, we found consistent support for a positive relationship between both sex-specific combined grip strength and sex-specific arm lean mass, and two measures of mating success: being currently partnered, and lifetime number of sexual partners for men, controlling for numerous possible confounds. We found no consistent association between either sex-specific combined grip strength or any sex-specific lean masses and age at first sex or number of past year sexual partners. 
 
-We also tested for interactions between sex and strength on these mating success outcomes. Across both pilot and confirmatory data, we found a consistent interaction between sex and sex-specific grip strength on partnered status, such that sex-specific grip strength was more important for predicting males' partnered status than females. Although in the pilot data we did see evidence of an interaction between sex and sex-specific grip strength on the outcome lifetime number of sexual partners, such that sex-specific grip strength was more important for predicting males' lifetime sexual partners, this effect was not significant in the confirmatory data. 
+We also tested for interactions between sex and strength on these mating success outcomes. Across both pilot and confirmatory analyses, we found a consistent interaction between sex and sex-specific grip strength on partnered status, such that sex-specific grip strength was a better predictor of males' partnered status than females. Although in the pilot analysis we did see evidence of an interaction between sex and sex-specific grip strength on the outcome lifetime number of sexual partners, such that sex-specific grip strength was a better predictor of males' lifetime sexual partners, this effect was not significant in the confirmatory analysis. 
 
-The lack of a sex difference in the significant positive effect of sex-specific grip strength on short term mating success (proxied by number of lifetime sexual partners) is puzzling. Men have greater strength than women and their average partner numbers are higher. Still, women with higher sex-specific grip strength and sex-specific arm lean mass unexpectedly reported more lifetime sexual partners on average than women with lower grip strength or arm lean mass, even after controlling for a host of potential confounds. A meta-analysis of 77 species, including humans and 7 other mammal species, found a positive Bateman gradient for females, i.e., a positive correlation between mating success and reproductive success, although authors and commentators both caution that correlation is not causation [@fromonteil2023; @kokko2023]. It is also not clear why women's strength would be positively associated with partner numbers.
+The lack of a sex difference in the significant positive effect of sex-specific grip strength on short term mating success (proxied by number of lifetime sexual partners) is puzzling. Men have greater strength than women and their average partner numbers are higher. Still, women with higher grip strength and arm lean mass unexpectedly reported more lifetime sexual partners on average than women with lower grip strength or arm lean mass, even after controlling for a host of potential confounds. A meta-analysis of 77 species, including humans and 7 other mammal species, found a positive Bateman gradient for females, i.e., a positive correlation between mating success and reproductive success, although authors and commentators both caution that correlation is not causation [@fromonteil2023; @kokko2023]. 
 
-It could be that there was selection for more formidable men to prefer more partner variety, and stronger women have a similar preference as a byproduct of selection on men. It could be that there is assortative mating on strength: if stronger men are motivated to switch partners more frequently, their (stronger) mates would also likely have more mates. It might be that stronger women require less male investment and so instead benefit from greater partner numbers through, e.g. genetic bet-hedging, forging relationships with multiple males, ability to conduct a more extensive search for a high quality long term mate, or through avoiding costly long term partnerships. It might be the case that there are some sex-specific confounds that we failed to control for, or that our models are otherwise misspecified. There also might be reverse causation, e.g., women who are interested in greater partner variety keep in better shape. Greater partner numbers might indicate mating failures rather than mating successes for males or females, although why strength would be associated with mating failures is not clear. Since a sex difference was significant in the Pilot Study (2011-12), but not in the Confirmatory Study (2013-14), it is possible that cultural changes influencing sexual behavior occurred between the data collection cycles. Or these results might simply be noise.
+Still, it is not clear why women's strength would be positively associated with partner numbers. It could be that there was selection for more formidable men to prefer more partner variety, and stronger women have a similar preference as a byproduct of selection on men. It could be that there is assortative mating on strength: if stronger men are motivated to switch partners more frequently, their (stronger) mates would also likely have more mates. It might be that stronger women require less male investment, or can take more physical risks, and so instead benefit from greater partner numbers through, e.g. genetic bet-hedging, forging relationships with multiple males, ability to conduct a more extensive search for a high quality long term mate, or through avoiding or leaving costly long term partnerships. It might be the case that there are some sex-specific confounds that we failed to control for, or that our models are otherwise misspecified. There also might be reverse causation, e.g., women who are interested in greater partner variety keep in better shape. Greater partner numbers might indicate mating failures rather than mating successes for males or females, although why strength would be associated with mating failures is not clear. Since a sex difference was significant in the Pilot Study (2011-12), but not in the Confirmatory Study (2013-14), it is possible that cultural changes influencing sexual behavior occurred between the data collection cycles. The sexual sexual hypothesis also might be wrong. Or these results might simply be noise.
 
 Finally, we found consistent evidence for a tradeoff between both grip strength and all lean masses, and immune investment and energy and protein intake. Grip strength and all lean masses were positively associated with energy and protein intake in our confirmatory models, and our proxy for immunity (WBCC) was significantly negatively associated with all lean masses (but was not significantly associated with grip strength). 
 
 ## Limitations
 
-Our registered report had one aim: to test the prediction from the sexual selection hypothesis that the effect of upper body strength on reproductive success is significantly positive for males and significantly less so for females. Our aim was not to test all possible byproduct hypotheses, nor was it possible to do so. We were unable to control for in-utero testosterone exposure, or other conditions such as Polycystic Ovarian Syndrome (PCOS) that might influence the relationship between grip strength and our measures of mating success. Likewise, NHANES does not collect data on sociosexuality. As described in our methods, in industrialized populations like the U.S., widespread access to contraceptives uncouples reproductive success from mating behavior for both men and women. Measures of mating success, including partnered status, number of sexual partners, and age at first sexual intercourse, are used as proxies as they are assumed to have been strongly correlated with reproductive success under ancestral conditions [@perusse1993]. Finally, all pilot results were obtained after considerable exploratory analysis, with a high risk of overfitting the data, and therefore our confirmatory results were critical to testing our predictions.
+Our registered report had one aim: to test the prediction from the sexual selection hypothesis that the effect of upper body strength on reproductive success is significantly positive for males and significantly less so for females. Our aim was not to test all possible byproduct hypotheses, nor was it possible to do so. We were unable to control for in-utero testosterone exposure, or other conditions such as Polycystic Ovarian Syndrome (PCOS) that might influence the relationship between grip strength and our measures of mating success. Likewise, NHANES does not collect data on sociosexuality. As described in our methods, in industrialized populations like the U.S., widespread access to contraceptives uncouples reproductive success from mating behavior for both men and women. Measures of mating success, including partnered status, number of sexual partners, and age at first sexual intercourse, are used as proxies as they are assumed to have been strongly correlated with reproductive success under ancestral conditions [@perusse1993]. In addition, in confirmatory analyses of mating outcomes (but not immune, energy, or protein outcomes) we removed `{r} sum(designsH$d.design.adults$variables$sex_partners>=100, na.rm = T)` participants with 100 or more lifetime sexual partners on the grounds that these might involve evolutionarily novel patterns of behavior that would not conform to predictions of the sexual selection hypothesis. Ideally, however, analyses would account for the full range of variation in sexual behavior. Finally, all pilot results were obtained after considerable exploratory analysis, with a high risk of overfitting the data, and therefore our confirmatory results were critical to testing our predictions. However, even our confirmatory results were based on generalized linear models, yet there might be important non-linear relationships that would alter interpretations.
 
 ## Conclusion 
 
@@ -437,7 +437,7 @@ Power to detect grip strength and sex X grip strength coefficients that range fr
 
 ::: {#stbl-stats}
 ```{r}
-pilot_summary_stats <- summary_tables(designsG$d.design.adults)
+pilot_summary_stats <- summary_tables(designsG2$d.design.adults)
 tbl_stack(list(pilot_summary_stats$con, pilot_summary_stats$cat))
 ```