Environmental adversity is associated with lower investment in collective actions.

Environmental adversity is associated with a wide range of biological outcomes and behaviors that seem to fulfill a need to favor immediate over long-term benefits. Adversity is also associated with decreased investment in cooperation, which is defined as a long-term strategy. Beyond establishing the correlation between adversity and cooperation, the channel through which this relationship arises remains unclear. We propose that this relationship is mediated by a present bias at the psychological level, which is embodied in the reproduction-maintenance trade-off at the biological level. We report two pre-registered studies applying structural equation models to test this relationship on large-scale datasets (the European Values Study and the World Values Survey). The present study replicates existing research linking adverse environments (both in childhood and in adulthood) with decreased investment in adult cooperation and finds that this association is indeed mediated by variations in individuals' reproduction-maintenance trade-off.

1. Introduction

Environmental adversity is often conceptualized as the product of two different dimensions: harshness (i.e., externally caused levels of morbidity-mortality that an individual cannot control) and unpredictability (i.e., spatial-temporal variation in harshness) [1]. In humans, adversity is closely aligned with economic deprivation and resource scarcity [2-4] and is commonly associated with a wide range of detrimental biological outcomes and behaviors, ranging from reduced lifespan [5], detrimental health behaviors [6-10], suboptimal financial decisions or lower education [11-17]. Adversity has also been targeted as a potential candidate causing variability in a number of social behaviors seemingly unrelated to those mentioned above, such as social preferences [18-21], social learning biases [22-26], social trust [27-29] and cooperation [30]. However, whether cooperation is decreased or increased in adverse environments is still debated.

In theory, the relationship between environmental adversity and cooperation could go either way [30]. On the one hand, adverse environments might be associated with increased cooperation as a means to mitigate direct survival threats, such as pooling material resources to reduce the impact of external events like famine [31,32]. On the other hand, adverse environments might be associated with decreased cooperation because adversity constrains people to focus on trying to meet their individual needs [30].

Whether adverse environments are associated with decreased or increased cooperation seems particularly influenced by the way cooperation is measured. The main distinction is between self-reported questionnaires and field experiments on the one hand and behavioral experiments (i.e., economic games) on the other hand. Self-reported measures of cooperation and economic games differ in several important ways: (1) self-reports are more influenced by social desirability bias, (2) they are based on people’s limited insight on their own behavior and (3) self-reports are based on attitudes while economic games are based on actual behavior. It is important to note that some of the studies cited as support for the association between adverse environments and increased cooperation [31,32] use agent-based models, which are distinct from self-reported measures, field experiments and behavioral experiments.

The existing empirical evidence is currently inconclusive. When cooperation is measured via self-reported questionnaires or in field experiments, studies usually report an association between adverse environments and decreased cooperation. Notably, Korndörfer et al. [33] and Schmukle et al. [34] found a negative correlation between lower social class and cooperative behaviors (e.g., donating, volunteering, helping) using questionnaire data from large and representative international samples (average N over 9 studies = 20.243). Using field experiments, Andreoni et al. [35] and Nettle et al. [36] also found that lower social class correlates with decreased cooperation.

When cooperation is measured using behavioral experiments (i.e., economic games and actual behavior in various settings under scrutiny of the experimenter), the results are much more mixed. Some studies report an association between adverse environments and decreased cooperation [33,34,36-38], but other studies find the opposite pattern [39-41] or no effect at all [42,43].

The reason why these behavioral experiments give mixed results is unclear, but several explanations have been put forward. First, some of these studies have small sample sizes, which may explain, for example, why Piff et al. [41], who found a positive relation between environmental adversity and cooperation, was not replicated in a high-powered pre-registered study [42]. Second, these studies usually only use one economic game, which may lower their generalizability [44]. In accordance with this idea, McAuliffe et al. [45] found that single economic games do not correlate with real-life cooperation, but a general factor based on several economic games does. Finally, there is increasing evidence that economic games lack ecological validity [46].

Beyond establishing the correlation between adversity and cooperation, the channel through which this relationship arises also remains unclear. In a large review paper, Pepper and Nettle [47] put forward the idea that there might be a common origin to the constellation of behaviors observed in adverse environments, which is that they reflect a contextually appropriate need to favor immediate over long-term benefits. Recent studies indeed reveal that people living in adverse conditions prefer immediate rewards over delayed rewards, discount future rewards more, are more pessimistic about their future, and are more impulsive than those living in more affluent conditions [48-53]. In sum, people’s living environment appears to affect the present-future trade-off, with adverse conditions promoting present-oriented behaviors, and affluent conditions promoting future-oriented behaviors.

Evolutionary models [54,55] and empirical evidence [56-60] suggest that variations in cooperation might also depend on variations in the present-future trade-off, and ultimately on the environmental factors that shape it. Cooperation is defined as “a behavior that provides a benefit to another individual (recipient), and the evolution of which has been dependent on its beneficial effect for the recipient” [61,62]. Cooperation has two subcategories: altruistic and mutually beneficial behaviors [62]. We are particularly interested in mutually beneficial cooperative behaviors, where the actor incurs immediate costs and the benefits are delayed. In this framework, cooperation is a future-oriented strategy [63,64]: in the short term, it is more advantageous to reap immediate, smaller, but more certain benefits by cooperating less but in the long term, it is more advantageous to invest in large-scale cooperation so as to reap longer-term benefits such as increased social reputation.

The way the present-future trade-off is instantiated at the psychological level might partly result from how it is embodied at the biological level [47,65], notably through the way organisms optimally allocate their limited stock of energy between reproductive goals and maintenance goals [1,65-68]. Similarly to the present-future trade-off observed at the psychological level, the shape of the reproduction-maintenance trade-off is partly conditioned by the environment, and therefore subjected to plasticity during an individual’s life. In line with this phenotypic plasticity perspective, correlational research in humans living in industrial and post-industrial societies suggests that experiencing adversity during childhood accelerates reproduction while decreasing investment in health [69-71]. By contrast, safe childhood conditions are associated with delayed reproduction [1,66,72-75] and are negatively associated with a number of morbidity risks [76]. In sum, prioritizing reproduction is classically viewed as a present-oriented life-history strategy because it comes with direct and indirect survival costs that might otherwise be minimized by delaying reproduction. The reason is that adverse conditions make life uncertain, so that one cannot be sure that an investment made in the present (e.g., health efforts) will pay-off in the future (e.g., living longer). Counteracting these risks requires prioritizing immediate but certain benefits (i.e., reproduction) at the expense of long-term but uncertain benefits (i.e., somatic investments and health) [77-80].

Taken together, these observations raise the possibility that the effect of environmental adversity on cooperation might be conditional on the shape of the reproduction-maintenance trade-off. More specifically, individuals who experienced environmental adversity–both during childhood and adulthood–might invest less in cooperative activities, and this effect might be conditioned at least partly on the priority they give to reproduction at the expense of health. Decreasing the costs of cooperative interactions might be contextually adaptive in the most deprived societies where social trust is lower and interpersonal violence is higher, and where people feel more exposed to risks of exploitation outside their kinship network [81-83].

In this paper, we test the relationship between childhood and current environmental adversity and people’s propensity to invest in collective actions, which we use as a proxy for cooperation. Collective action can indeed be defined as “multiple individuals pay the cost of contributing to a group project and share in the resulting benefit” [84,85]. We further test the hypothesis that the relationship between adversity and collective action is conditional on an individuals’ reproduction-maintenance trade-off. Our specific hypotheses were that (1) childhood environmental adversity is associated with lower investment in collective actions, and that this effect is mediated by variations in individuals’ reproduction-maintenance trade-off; (2) current environmental adversity is associated with lower investment in collective actions, and that this effect is mediated by variations in individuals’ reproduction-maintenance trade-off.

To test these hypotheses we applied multivariate models (structural equation modeling) on two independent datasets containing information about the values, the beliefs and the behaviors of more than 295.000 participants representative of more than 100 countries all over the world (the European Values Study and the World Values Survey). The robustness of the models was assessed by employing a discovery/ replication procedure. First, each dataset was randomly split into two subsamples of similar size: a discovery sample and a replication sample. Second, the discovery sample was used to design and test the model. We then pre-registered the analysis plan in the Open Science Framework. Finally, the replication sample was used to check the robustness of the model. The present study replicates existing research linking adverse environments in childhood and adulthood with decreased investment in adult cooperation [33,34,36-38], and finds that this association is mediated by variations in individuals’ reproduction-maintenance-trade-off. Furthermore, it adds to the consistent pattern of results (i.e., a relationship between adverse environments and decreased cooperation) found by existing studies that use self-reported questionnaires to measure cooperation [33,34].

2. Study 1: Childhood adversity and collective action in the European Values Study

We tested the association between childhood environmental adversity, the reproduction-maintenance trade-off and adult investment in collective actions by applying structural equation modeling on the European Values Study dataset [86]. The European Values Study has relied on social scientists since 1981 to collect data from large representative samples of the populations from 47 European countries. In this study, we used the fourth wave of the survey for which data was collected between 2008 and 2010. We focused on the fourth wave because it includes the same collective action variables as the World Values Survey, which gives us the opportunity to replicate our initial findings on an independent sample.

2.1 Methods

2.1.1 Respondents

66.281 respondents living in 46 different countries are included in wave 4 of the European Values Study. However, some respondents could not be included in our analyses. Two of the variables used to model the reproduction-maintenance trade-off were only relevant for respondents with children, which led us to focus on the 45.624 respondents who report having children. We randomly split this sample into a discovery subsample (22.570 respondents) and a replication subsample (23.054 respondents). In both subsamples, we excluded respondents with too much missing data (the criterion was set to more than two standard deviations from a respondent’s respective subsample mean) which led to 20.755 respondents for the discovery sample and 21.156 respondents for the replication sample. The final discovery sample after listwise deletions included 16.790 respondents (females: N = 9.837; mean age = 52 ± 15 sd; mean number of respondents per country = 382 ± 126 sd) and the final replication sample after listwise deletions included 17.087 respondents (females: N = 9.957; mean age = 52 ± 15 sd; mean number of respondents per country = 380 ± 136 sd). The final replication sample after listwise deletions consists of 45 countries (descriptive statistics for each country are detailed in S1 Table).

2.1.2 Variables

We selected variables tapping three constructs: childhood environmental adversity, the reproduction-maintenance trade-off and collective action. Childhood environmental adversity was modeled as a formative latent variable including the following items: “death of the father before age 16” (scale 0–1), “death of the mother before age 16” (scale 0–1), “parental education” (scale 1–8: the higher the score the lower the educational level of the parents), “parents had problems replacing things” (scale 1–4: the higher the score the more difficulty replacing things) and “parents had problems making ends meet” (scale 1–4: the higher the score the more difficulty making ends meet). Based on the recommendations of Brumbach et al. [69], childhood environmental adversity was modeled as a formative latent variable rather than a reflective latent variable. This was done because adverse childhood events can be seen as risk factors that are not necessarily correlated with one another. However, they all contribute to the cumulative probability of developing a particular outcome (in our case an increased investment in reproduction or somatic maintenance). For example, having been exposed to the death of a parent, lower parental education or lower parental socioeconomic status are three events that increase the probability of experiencing an adverse childhood environment, but that might occur independently.

The reproduction-maintenance-trade-off was modeled by constructing a reflective latent variable based on “number of children”, “age at first birth” (calculated from the birth year of the respondent and the reported birth year of their first child) and “subjective health status” (scale 1–5: the higher the score the better the respondents’ reported health). These items were chosen because they approximate the reproduction-maintenance trade-off [1]. For example, recent studies report that some health costs are induced by increased investment in reproductive functions positively associated to fertility–in both women (e.g., high number of pregnancies, high parity) and men (e.g., high level of testosterone). These health costs include increased production of pro-inflammatory cytokines, insulin resistance, increased production of triglycerides and LDL cholesterol, increased oxidative stress, altered immune function, cardio-vascular diseases [66,75,87] and have long-lasting health correlates, including on people’s lifespan, which has been shown to be well captured by self-reported health [88].

Cooperation was modeled by constructing a reflective latent variable. In articles using the European Values Study and the World Values Survey [89,90], cooperation is often assessed using questions tapping cooperative morality (e.g., the respondent’s opinion about claiming government benefits to which you are not entitled; avoiding a fare on public transport; cheating on taxes if you have a chance). Such questions do not test actual cooperation, but rather people’s moral judgments. Importantly, moral judgments are not systematically aligned with actual behavior. For instance, individuals who claim that it is wrong to cheat on taxes may actually cheat themselves. Individuals living in low trust societies are indeed more likely to cheat because they think that others are cheating [91] and they are also more likely to be in favor of harsh punishments for cheaters because they do not trust that cheating can be avoided by means other than harsh punishments [37].

To address this problem, we chose to focus on respondents’ actual investment in collective actions. We scanned the European Values Study to identify variables that had previously been found to successfully predict real-life cooperation [33] and that tapped actual behaviors involving concrete costs and delayed benefits for the agent and benefits for others. Two sets of variables met our a priori defined criteria: how much someone volunteers (scale 0–15) and how often someone participates in political actions that might benefit the community, without receiving financial compensation (scale 0–5) (items are detailed in S2 Table). People who volunteer or take part in political actions indeed incur concrete costs, including time and energy, to engage in actions that benefit others. Furthermore, volunteering and political actions also have delayed benefits for the actor. Volunteering, for example, is positively associated with concern for social reputation or image [92] and is a means of obtaining human capital that will yield returns in the labor market [93,94]. Political action, for example voting, has also been shown to correlate with higher social reputation [95,96]. Although volunteering and political action are limited in their own way (for instance, many cooperative interactions happen outside the activities described in these items), these items have the advantage of being the most concrete instances of true investment in collective goods available in the European Values Study. Volunteering was calculated by taking the sum of 15 separate items that indicate if respondents volunteer for different groups (e.g., religious organizations, trade unions and sport groups). Political action was calculated by taking the sum of five separate items that indicate if respondents take part in different political actions (e.g., signing a petition, joining boycotts and attending lawful demonstrations).

2.1.3 Model specification and fit

First, the reproduction-maintenance trade-off was regressed on childhood environmental adversity. Second, collective action was regressed on the reproduction-maintenance trade-off and childhood environmental adversity. Third, childhood environmental adversity was modeled as a formative latent variable. These variables need to be scaled for identification purposes by fixing the coefficient of one of the causal indicators [70]. Childhood environmental adversity was scaled by setting the path from “parental education” to 1. Fourth, the reflective latent variable reproduction-maintenance trade-off and the reflective latent variable collective action were scaled by fixing their variance to 1. Fifth, the correlation between the residual errors of “age at first birth” and “number of children” was included in the model. The reason for this is that, all else being equal, the sooner you have children, the more children you can have. This correlation is not captured by the latent variable reproduction-maintenance trade-off and is therefore included separately; these two variables indeed covary mechanically [70].

Hooper et al. [97] recommend determining model fit by using χ2, Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR). A good fit is assumed if the indices are close to the following values: χ2 p-value > 0.05, CFI > 0.95, RMSEA p-value < 0.07 and SRMR < 0.08. It is important to note however, that the large sample size of our study prevents us from interpreting χ2 as evidence for a discrepancy between the sample and the model-implied covariance matrix. The chi-square statistic is indeed known to be particularly sensitive to sample size, which can lead models fitted on large samples to be systematically rejected [98]. Furthermore, the χ2 assumes normality [99] and the deviation from normality characterizing some of our variables might result in the rejection of the model even when the model is properly specified. Therefore, only the scaled versions of the CFI, RMSEA and SRMR are reported, which eliminates the issues of sample size dependency and non-normality of continuous and categorical data [100].

2.1.4 Covariates

Given that cooperation is affected by age, we included age as an auxiliary variable to control for its effect on the collective action variables. Freund & Blanchard-Fields [101], for example, found that older adults report valuing contributions to the public good more positively and are more likely to behave altruistically than younger adults. Furthermore, age is also used as an auxiliary variable to control for its effect on the “subjective health status” variable and on the “number of children” variable. The reason for this is that, all else being equal, the older you get the poorer your health is likely to be and the more children you are likely to have.

Childhood environmental adversity and current environmental adversity are moderately correlated [51]. Therefore, we ran an additional analysis which was not pre-registered where we controlled for the effect of current adversity on the reproduction-maintenance trade-off and collective action variables. Current adversity was modeled using the respondent’s current income level (scale 1–10: the higher the score the lower the income level).

2.1.5 Analyses

We first carried out our analyses on the discovery sample; we then pre-registered our analysis plan on the Open Science Framework (https://osf.io/642hd); finally, we applied the pre-registered analyses on the replication sample. The results of the replication sample are presented below. The results of the discovery sample are detailed in S1 Text; the significance and direction of the key associations are consistent with the replication sample. Although, two additional indicators for childhood environmental adversity reach significance: “parents had problems making ends meet” and “death of mother”.

All statistical analyses were carried out in R 3.4.4 (https://www.r-project.org/) with R Studio 1.1.456. The structural equation model was fitted using the R package lavaan [102]. The WLSMV estimator was used for its robustness to departures from normality [102].

When data was missing, we conducted listwise deletions. Although our variables of interest showed overall low percentages of missing responses (ranging from 0 to 13%), an additional analysis was done on the replication sample with imputed data. The results of this analysis are largely consistent with the results obtained using listwise deletions (see S2 Text). The main difference is that childhood environmental adversity is significantly–albeit weakly—associated with increased adult involvement in collective action–a result that we will discuss in more detail below.

For the mediation analyses the bootstrap method developed by Preacher & Hayes [103] was used, which is recommended by MacKinnon et al. [104]. This is a non-parametric resampling test. The main feature of this test is that it does not rely on the assumption of normality. Bootstrapping estimates the upper limit and the lower limit of the confidence intervals of an indirect effect. We computed bootstrapped 95% confidence intervals (1000 bootstrap samples). The direct effect, indirect effect and total effect can be found in S3 Text.

2.2 Results

2.2.1 Correlation matrix and descriptive statistics

Descriptive statistics and the correlation matrix for the variables included in the structural equation model can be found in Table 1.

Table 1

European Values Study descriptive statistics and correlations.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
(1) Parental education	-
(2) Parents problem making ends meet	0.28*	-
(3) Parents problem replacing things	0.24*	0.71*	-
(4) Death of father	0.07*	0.11*	0.10*	-
(5) Death of mother	0.02*	0.01	0.01	0.03*	-
(6) Health	-0.16*	-0.16*	-0.18*	-0.06*	-0.02*	-
(7) Age 1st birth	-0.06*	-0.08*	-0.08*	-0.02	-0.01	0.11*	-
(8) Number of children	0.19*	0.10*	0.07*	0.02	0.03*	-0.08*	-0.18*	-
(9) Volunteering	-0.03*	-0.02*	-0.03*	-0.01	0,00	0.1*	0.05*	0.04*	-
(10) Political action	-0.1*	-0.07*	-0.11*	-0.02*	0,00	0.15*	0.13*	-0.01	0.12*	-
(11) Age	0.31*	0.16*	0.13*	0.11*	0.05*	-0.33*	0.1*	0.24*	0,00	-0.01	-
Mean	5.49	2.40	2.34	0.05	0.02	3.56	25.18	2.21	0.38	0.61	51.96
SD	1.99	1.17	1.17	0.22	0.13	0.96	5.18	1.12	1.21	0.96	15.00
Range	1–8	1–4	1–4	0–1	0–1	1–5	12–83	1–13	0–15	0–5	18–108

* p < 0.05.

2.2.2 Model fit

The scaled CFI value (0.900), the scaled RMSEA value (0.040) and the scaled SRMR value (0.018) are fairly consistent with a close-fitting model. Therefore, the approximate fit indices reveal no strong misspecification for this model.

2.2.3 Measurement model

The standardized regression weights can be found in Fig 1 (the full results can be found in S3 Table). Childhood environmental adversity is driven by two main contributors: “parents had problems replacing things” (UnStd c = 2.37 (0.26), z = 9.01, p < 0.001, Std c = 0.69) and “death of the father before age 16” (UnStd c = 1.65 (0.70), z = 2.37, p < 0.05, Std c = 0.09). “Death of the mother before age 16” and “parents had problems makings ends meet” did not load significantly on childhood environmental adversity.

Fig 1

European Values Study standardized parameter values estimated by the structural equation model.

Significant paths at the 5% level are represented with a continuous arrow.

“Subjective health status” (UnStd c = -0.39 (0.02), z = -25.97, p < 0.001, Std c = -0.43) and “age at first birth” (UnStd c = -1.60 (0.07), z = -24.63, p < 0.001, Std c = -0.33) loaded significantly on the reproduction-maintenance trade-off. The pattern of covariation follows our predictions: poorer reported health and a younger age at first child’s birth. “Number of children” did not load significantly on the reproduction-maintenance trade-off. However, inspection of the estimated covariance shows that “number of children” is not independent but covaries with “age at first birth” in the expected way (UnStd c = -1.15 (0.04), z = -31.63, p < 0.001, Std c = -0.22), specifically a higher age at first child’s birth correlates with lower number of children, an observation that is in line with existing findings [70]. Hence, the reproduction-maintenance trade-off is fairly consistent with prior studies [69,70].

“Volunteering” (UnStd c = 0.34 (0.02), z = 17.12, p < 0.001, Std c = 0.52) and “political action” (UnStd c = 0.40 (0.03), z = 14.90, p < 0.001, Std c = 0.61) loaded significantly on the collective action latent variable, whose greater values indicate higher investments in both volunteering and political activities.

2.2.4 Structural model

Fig 1 shows that the direct effect of childhood environment adversity is significantly–albeit weakly–associated with decreased adult involvement in collective action (UnStd c = -0.03 (0.01), z = -3.50, p < 0.001, Std c = -0.08). This implies that an adverse childhood environment is associated with less investment in collective actions later in life, relatively independently of variations in individuals’ reproduction-maintenance trade-off. Furthermore, an adverse childhood environment is associated with variations in individuals’ reproduction-maintenance-trade-off (UnStd c = 0.08 (0.01), z = 10.46, p < 0.001, Std c = 0.29), specifically an increased investment in reproduction and a decreased investment in somatic maintenance. The reproduction-maintenance trade-off is itself associated with lower adult involvement in collective action (UnStd c = -1.07 (0.10), z = -10.87, p < 0.001, Std c = -0.73). In line with our first hypothesis, a significant part of the effect of childhood environmental adversity on adult involvement in collective action is mediated by the reproduction-maintenance trade-off (indirect effect: UnStd c = -0.0027 (0.0004), bootstrapped ci lower = -0.0035, bootstrapped ci upper = -0.0020, z = -6.92, p < 0.001).

The above mentioned effects remain stable after the inclusion of the respondent’s current adversity as an additional covariate (scaled CFI = 0.897; scaled RMSEA = 0.042; scaled SRMR = 0.017), and the magnitude of key parameters stayed significant and in the expected direction (the full results can be found in S4 Table).

Given that the mediation effect is strong, the direct effect of childhood environmental adversity on collective action is to be interpreted with caution. Our analyses show that it is indeed a small and unstable effect, which reverses in the dataset that includes imputed data (UnStd c = 0.04 (0.02), z = 2.23, p < 0.05, Std c = 0.13). To confirm that this coefficient instability can be attributed to the presence of a mediator capturing a large part of the effect, we ran an additional model on the replication sample with imputed data, but without the mediating variable (i.e., reproduction-maintenance trade-off). This analysis revealed that adverse childhood environments are indeed associated with decreased collective action (UnStd c = -0.07 (0.01), z = -7.09, p < 0.001, Std c = -0.21).

The other regression coefficients are larger and more stable across analyses and datasets (i.e., the dataset implementing listwise deletions and the imputed dataset). This is the case for the regression between childhood environmental adversity and the reproduction-maintenance trade-off (listwise deletions UnStd c = 0.08 (0.01), z = 10.46, p < 0.001, Std c = 0.29; imputed data UnStd c = 0.07 (0.01), z = 6.92, p < 0.001, Std c = 0.39), for the regression between the reproduction-maintenance trade-off and collective action (listwise deletions UnStd c = -1.07 (0.10), z = -10.87, p < 0.001, Std c = -0.73; imputed data UnStd c = -1.50 (0.34), z = -4.42, p < 0.001, Std c = -0.88), as well as for the mediation effect (listwise deletions indirect effect: UnStd c = -0.0027 (0.0004), bootstrapped ci lower = -0.0035, bootstrapped ci upper = -0.0020, z = -6.92, p < 0.001; imputed data indirect effect: UnStd c = -0.0023 (0.0004), bootstrapped ci lower = -0.0030, bootstrapped ci upper = -0.0016, z = -6.64, p < 0.001). Overall, these results converge on the idea that the relationship between childhood environmental adversity and collective action is largely conditional on the reproduction-maintenance trade-off.

Finally, the range for “age at first child’s birth” is from 12 to 83 years old. This range is theoretically possible but unlikely (especially for women). We suspect that some of these data points are in fact entry errors. However, we kept the full sample for two reasons: first, it is impossible to tease apart entry errors from true late first births and second, very few individuals reported to have had their first birth at a later age (only 14 women out of 17.087). To ensure that the inclusion of these 14 participants did not alter our results we ran the model again without these participants and found the same overall pattern.

2.2.5 Conclusions

Our prediction was that people who grew up in adverse environments invest less in collective actions later in life and that this effect is mediated by modulations of the reproduction-maintenance trade-off. Data from the European Values Study partly confirm this hypothesis and show that the association between childhood environmental adversity and adult involvement in collective action is partly mediated by variations in individuals’ reproduction-maintenance-trade-off.

3. Study 2: Current adversity and collective action in the World Values Survey

We then had the objective of replicating the pattern of associations found in study 1 on an independent sample extracted from the World Values Survey dataset [105]. The World Values Survey has relied on social scientists since 1981 to collect data from large representative samples of the populations from almost 100 European and non-European countries. In this study, we focused on the fourth wave (1999–2004) for two reasons: first, it includes non-European countries; second, unlike the other waves, the fourth wave includes the same collective action variables as the ones included in the European Values Study sample analyzed in study 1.

It is worth acknowledging that using the World Values Survey fourth wave also comes with several limitations. First, unlike the European Values Study fourth wave, the World Values Survey fourth wave (and the other waves as well) does not contain any data about the respondents’ age at first reproduction. Therefore, the reproduction-maintenance trade-off can only be modeled on two indicators: the respondents’ number of children and their self-reported health status. Nevertheless, as we point in section 2.1.2 of the present article, the existing literature provides substantial evidence that the number of children and the self-reported health status can be considered as fairly good proxies for the reproduction-maintenance trade-off. The second limitation is that the World Values Survey fourth wave contains no data on the respondents’ childhood environment. We therefore focus on respondents’ current income as an indicator of the level of adversity they are currently exposed to. This choice is primarily motivated by a number of studies showing that income correlates with almost all forms of morbidity and mortality [2-4], such that economically deprived people also experience greater exposure to disability and death. The second motivation comes from the fact that economic status in adulthood generally correlates with economic status in childhood [106,107].

3.1 Methods

3.1.1 Respondents

59.030 respondents living in 40 different countries are included in wave 4 of the World Values Survey. However, some respondents could not be included in our analyses. First, one of the variables (i.e., number of children) used to model the reproduction-maintenance-trade-off is only informing for respondents with children. The reason is that people who do not have children now fall into three categories: (1) people who do not have children yet but will do so in the future, (2) people who do not have children yet and will choose not to do so in the future and (3) people who do not have children yet but would like to and can’t. Since we have no way of distinguishing these three groups, we focused on the people who do have children, which left us with the 41.799 respondents who report having children (this also makes it consistent with the first study, where we also removed respondents without children). Second, we removed countries for which the collective action variables had not been filled out, which reduced the dataset to 25.670 respondents living in 27 different countries. We randomly split this sample into a discovery subsample (12.962 respondents) and a replication subsample (12.708 respondents). In both subsamples, we excluded respondents with too much missing data (the criterion was set to more than two standard deviations from a respondent’s respective subsample mean) which led to 12.083 respondents for the discovery sample and 11.888 respondents for the replication sample. The final discovery sample after listwise deletions included 10.829 respondents (females: N = 5.633; mean age = 45 ± 15 sd; mean number of respondents per country = 417 ± 150 sd) and the final replication sample after listwise deletions included 10.617 respondents (females: N = 5.589; mean age = 45 ± 14 sd; mean number of respondents per country = 408 ± 157 sd). The final replication sample after listwise deletions consists of 26 countries (descriptive statistics for each country are detailed in S5 Table).

3.1.2 Variables

We selected variables tapping three constructs: current adversity, the reproduction-maintenance-trade-off and collective action. Respondents’ current adversity was represented by their current income level (scale 1–10: the higher the score the lower the income level) [2-4,51].

The reproduction-maintenance-trade-off was modeled by constructing a reflective latent variable based on “number of children” and “subjective health status” (scale 1–5: the higher the score the better the respondents’ reported health).

The reflective latent construct “collective action” was estimated from the very same indicators as in study 1. The only difference is that scores in “volunteering” and “political action” were calculated on the basis of 14 and 4 items respectively (instead of 15 and 5 items in study 1, respectively) (items are detailed in S6 Table).

3.1.3 Model specification and fit

The model’s structure was the same as the one specified in study 1, with two notable differences. First, childhood environmental adversity was replaced by current adversity. Second, “age at first birth” is no longer an indicator of the latent variable reproduction-maintenance-trade-off. Age was again used as an auxiliary variable to control for its effect on the collective action variables and “subjective health status” variable and “number of children” variable. Model fit was determined in the same manner as in study 1.

3.1.4 Analyses

We first carried out our analyses on the discovery sample; we then pre-registered our analysis plan on the Open Science Framework (https://osf.io/37bs2/); finally, we applied the pre-registered analyses on the replication sample. The results of the replication sample are presented below (the results of the discovery sample are detailed in S4 Text; the significance and direction of the key associations are consistent with the replication sample).

When data was missing, we conducted listwise deletions (as in study 1). Although our variables of interest showed overall low percentages of missing responses (ranging from 0 to 13%), an additional analysis was done on the replication sample with imputed data. The results of this analysis are consistent with the results obtained using listwise deletions (see S5 Text).

All statistical analyses and the method for mediation analyses were similar to those reported in study 1. The direct effect, indirect effect and total effect can be found in S6 Text.

3.2 Results

3.2.1 Correlation matrix and descriptive statistics

Descriptive statistics and the correlation matrix for the variables included in the structural equation model can be found in Table 2.

Table 2

World Values Survey descriptive statistics and correlations.

	(1)	(2)	(3)	(4)	(5)	(6)
(1) Current adversity	-
(2) Health	-0.18*	-
(3) Number of children	0.13*	-0.09*	-
(4) Volunteering	-0.04*	0.03*	0.06*	-
(5) Political action	-0.13*	0.06*	-0.05*	0.09*	-
(6) Age	0.03	-0.21*	0.24*	-0.05*	0.06*	-
Mean	6.56	3.78	2.93	1.03	0.57	44.53
SD	2.35	0.87	1.78	1.88	0.94	14.45
Range	1–10	1–5	1–8	0–14	0–4	16–98

* p < 0.05.

3.2.2 Model fit

The scaled CFI value (0.960), the scaled RMSEA value (0.047) and the scaled SRMR value (0.018) are consistent with a close-fitting model. Therefore, the approximate fit indices reveal no strong misspecification for this model.

3.2.3 Measurement model

The standardized regression weights can be found in Fig 2 (the full results can be found in S7 Table). “Subjective health status” (UnStd c = -0.15 (0.03), z = -4.38, p < 0.001, Std c = -0.24) and “number of children” (UnStd c = 0.22 (0.05), z = 4.39, p < 0.001, Std c = 0.17) loaded significantly on the reproduction-maintenance-trade-off. As in study 1, the pattern of covariation follows our predictions: poorer reported health and higher number of children. Hence, the reproduction-maintenance-trade-off is consistent with prior studies [69,70].

Fig 2

World Values Survey standardized parameter values estimated by the structural equation model.

Significant paths at the 5% level are represented with a continuous arrow.

“Volunteering” (UnStd c = 0.20 (0.04), z = 5.69, p < 0.001, Std c = 0.12) and “political action” (UnStd c = 0.63 (0.16), z = 4.01, p < 0.001, Std c = 0.75) loaded significantly on the collective action latent variable, whose greater values indicate higher investments in political activities and, to a lesser extent, higher investments in volunteering activities.

3.2.4 Structural model

Fig 2 shows that the direct effect of current adversity on adult involvement in collective action is not significant (UnStd c = 0.12 (0.11), z = 1.12, p = 0.26, Std c = 0.25). Furthermore, more current adversity is associated with variations in individuals’ reproduction-maintenance trade-off (UnStd c = 0.42 (0.10), z = 4.36, p < 0.001, Std c = 0.70), specifically an increased investment in reproduction and a decreased investment in somatic maintenance. The reproduction-maintenance-trade-off is itself associated with lower adult involvement in collective action (UnStd c = -0.48 (0.19), z = -2.49, p < 0.05, Std c = -0.60). In line with our second hypothesis, the effect of current adversity on adult involvement in collective action is mediated by the reproduction-maintenance trade-off (indirect effect: UnStd c = -0.003 (0.001), bootstrapped ci lower = -0.004, bootstrapped ci upper = -0.001, z = -2.67, p < 0.01).

Despite the fact that the association pattern reported above is identical to the one found with the European Values Study (EVS) data, one should note that the effects are smaller (by comparing the z-values). This is the case for the regression between environmental adversity and collective action (EVS UnStd c = -0.03 (0.01), z = -3.50, p < 0.001, Std c = -0.08; WVS UnStd c = 0.12 (0.11), z = 1.12, p = 0.26, Std c = 0.25), between environmental adversity and the reproduction-maintenance trade-off (EVS UnStd c = 0.08 (0.01), z = 10.46, p < 0.001, Std c = 0.29; WVS UnStd c = 0.42 (0.10), z = 4.36, p < 0.001, Std c = 0.70), between the reproduction-maintenance trade-off and collective action (EVS UnStd c = -1.07 (0.10), z = -10.87, p < 0.001, Std c = -0.73; WVS UnStd c = -0.48 (0.19), z = -2.49, p < 0.05, Std c = -0.60) as well as for the mediation effect (EVS indirect effect: UnStd c = -0.0027 (0.0004), bootstrapped ci lower = -0.0035, bootstrapped ci upper = -0.0020, z = -6.92, p < 0.001; WVS indirect effect: UnStd c = -0.003 (0.001), bootstrapped ci lower = -0.004, bootstrapped ci upper = -0.001, z = -2.67, p < 0.01).

3.2.5 Conclusions

Our prediction was that people who currently live in adverse environments invest less in collective action and that this effect is mediated by modulations of the reproduction-maintenance trade-off. Data from the World Values Survey confirm this hypothesis and show that the association between current adversity and adult involvement in collective action is mediated by variations in individuals’ reproduction-maintenance trade-off.

4. Discussion

The present study replicates existing research linking adverse environments in childhood and adulthood with decreased investment in adult cooperation [33,34,36-38], and finds that this association is mediated by variations in individuals’ reproduction-maintenance-trade-off. Furthermore, it adds to the consistent pattern of results (i.e., a relationship between adverse environments and decreased cooperation) found by existing studies that use self-reported questionnaires to measure cooperation [33,34].

Even though the findings reported above are consistent with the existing literature, it is important that we highlight several limitations. First, we used people’s involvement in collective actions as a proxy for cooperation but cooperation obviously encompasses more than collective action behaviors. The reason we chose these items is that, unlike self-reported attitudes about cooperative morals often used as proxies for cooperation in other survey studies [90], these were the only variables that met our a priori defined criteria (actual behaviors involving concrete costs and delayed benefits for the agent and have benefits for others that had previously been successful in predicting real-life cooperation) in the European Values Study and the World Values Survey. However, people might cooperate in different ways than in large-scale collective action. For example, they might actively cooperate within more restricted social networks centered around kinship [108]. Thus, it remains an open question whether people cooperate less under adverse conditions in the broadest sense or simply invest less in collective actions.

Second, our goal was to investigate if part of the variability in people’s involvement in collective action was mediated by differences in individuals’ reproduction-maintenance-trade-off. Although the reproduction-maintenance-trade-off is well captured in the World Values Survey model, it is less well captured in the European Values Study model. Specifically “number of children” is not significantly correlated with the reproduction-maintenance trade-off. This may reflect the fact that the reproduction-maintenance trade-off indicators are influenced by multiple causes beyond the reproduction-maintenance trade-off itself, including cultural factors like contraceptives. Such cultural factors might account for the fact that “number of children” is not significantly correlated with respondents’ reproduction-maintenance trade-off in the European Values Study [109].

Third, the large sample sizes used in our studies raise the possibility that the statistically significant associations reported are primarily due to high statistical power. Therefore, the reporting of p-values is less informative. Standardized regression coefficients can be used as an effect size index [110]. Standardized regression coefficients refer to how many standard deviations the dependent variable will change per a standard deviation increase in the independent variable, holding all other variables constant. Standardized regression coefficients can be interpreted similar to correlations, where <0.2 is considered a weak effect, between 0.2 and 0.5 as a moderate effect and >0.5 as a strong effect [111]. Looking at our standardized path coefficients; the effect between adversity and collective action is weak (EVS Std c = -0.08; WVS not significant), the effect between adversity and the reproduction maintenance trade-off is moderate for the European Values Study and strong for the World Values Survey (EVS Std c = 0.29; WVS Std c = 0.70) and the effect between the reproduction maintenance trade-off and collective action is strong (EVS Std c = -0.73; WVS Std c = -0.60).

Fourth, it is important to note that life-history theory has been hotly debated for theoretical, empirical and methodological reasons in recent years [112-116]. Of particular relevance to this paper is the debate questioning whether life-history theory can be applied to explain individual differences in humans. Life-history theory was initially developed to explain differences between species [117] and later applied to individual differences in humans by invoking genetic factors, early environment conditions, or phenotypic plasticity [1,73,77,118], the latter two concepts being central to the present paper. One criticism laid out by Zietsch & Sidari [116] is that the processes that create between species trait covariation (i.e., Darwinian evolution) and within species trait covariation (e.g., developmental plasticity) are fundamentally different. In response to this, Del Giudice [119] has argued that the two processes can in fact be functionally related. He proposed the term “ecological gambit”, where trade-offs in the context of life-history strategies form the basis for the functional link between population and individual differences, as is the case in our analyses focusing on the reproduction-maintenance trade-off. In addition, there is evidence that phenotypic plasticity is the dominant source of inter-individual and inter-population variation in reproductive traits and associated health costs [66,120].

More generally, the correlational nature of our data hinders the possibility of producing inferences about the causal role of adverse environments on cooperation. For example, it is possible that the relationship between environmental adversity and cooperation is reversed: adverse environments might be associated with lower cooperation because individuals who cooperate less are less likely to do well in life and therefore more likely to end up living in a adverse environment. This alternative is plausible albeit a little harder to reconcile with the fact that we also found an association between an adverse childhood environment and adult cooperation. Another possibility is that there is an unknown variable that influences both environmental adversity and cooperation. Finally, the relationship between environmental adversity and cooperation might not be monotonic. For example, Lazarus [121] has proposed that cooperation may be an inverted-U function of environmental adversity.

One way to demonstrate the causal effect of environmental adversity on cooperation would be to experimentally manipulate acute stress [13], for example by administering hydrocortisone (a synthetic form of cortisol). Using this methodology, Riis-Vestergaard et al. [122] found that administering hydrocortisone increases temporal discounting over the short term. Another possibility to induce stress is to expose people to mortality cues [51], e.g. in the form of news articles stating that in the local area there is an increase in violence. However, it is important to note that there are large differences between the effects of acute stress and chronic stress [123]. For example, acute stress has been found to increase physical arousal, learning, motivated behavior and sensation seeking. By contrast, chronic stress has been found to impair attentional control, behavioral flexibility and it can promote anxiety, depression and learned helplessness.

Another option to identify causal relationships is to analyze longitudinal data involving exogenous shocks to the individual’s environment (e.g., sudden increase of income or in contrast famine, war, etc.). There have been a handful of longitudinal studies testing the effect of stress during childhood and social behavior in adulthood using exogenous shocks of positive or negative income on different behavioral outcomes. These studies have shown that higher levels of income tend to make individuals more ‘agreeable’ and more trustful [27,124] while higher levels of violence during childhood make people more antisocial and more violent [125,126]. One of the most noteworthy pieces of research in this area is a quasi-experimental study led by Akee et al. [124]. In this study, American Indian children were exposed to a positive income shock following the opening of a casino that decided to allocate part of its revenue to Cherokee households. Particularly interesting is that the increase in income led to an increase in agreeableness, which correlates with cooperation and unselfishness [127,128]. Hence, at least part of the effects of childhood environmental adversity on cooperation might capture long-term plastic responses to the environment.

It should be noted that our results are compatible with a range of mechanisms. The standard mechanism put forward in the literature (and in our introduction) is that cooperation is adaptively reduced in adverse environments because individuals have a shorter life expectancy and should therefore be more short-term oriented [47]. However, the idea that people focus more on the present because they have a higher probability of dying at age 60 instead of age 80 is problematic. Why should this affect their interest in cooperation today? The fact that someone is alive at age 60 or age 80 indeed has little impact on his or her probability of being alive tomorrow [129,130]. A recent theoretical model demonstrates that waiting costs (i.e., the cost paid by an individual even when the benefits are guaranteed) are another mechanism by which adverse environments can be associated with higher discount rates [131], and therefore decreased cooperation. When an individual has to choose between obtaining a resource immediately or a larger resource later she must take into account all the benefits she could have earned, during the delay, by wisely using this resource. Resources can be transformed into capital which allows individuals to better exploit their environment. When an individual chooses a delayed resource over an immediate one, she therefore renounces all the benefits this supplementary capital would have generated during that delay. These forgone benefits are independent of mortality or other collection risk. They are paid even if the individual is certain to obtain the resource eventually. They constitute the waiting cost per se. The more an individual can gain by investing new resources wisely into capital, the higher the waiting cost. The models shows that individuals who have already accumulated a lot of personal capital, and are high up on their pyramid of needs, should be relatively patient because each supplementary unit of capital makes a small difference to them anyway, whereas those who have accumulated little and are still at the bottom of their pyramid of needs should have a short time horizon as each additional resource can make a big difference to them.

Another mechanism by which environmental adversity can have an effect on cooperation is risk management. For example, Amir et al. [39] put forward an uncertainty management perspective, where adverse environments are associated with risk averse, present-oriented and more cooperative preferences. Although, they do not specifically test whether risk is a mediator between environmental adversity and cooperation. Even in the uncertainty management perspective proposed by Amir et al. [39], adverse environments are associated with a present-orientation, which is in line with the framework of our study. Amir et al. [39] see cooperation as a form of risk pooling which would reduce uncertainty. Therefore, in adverse and uncertain environments, this would be associated with increased cooperation. We fully agree that in adverse environments people will want to reduce uncertainty, but we are not entirely convinced that the primary means to achieve this is increased cooperation. One possibility is that this might be the case in specific contexts, for example, where population size is low and where the likelihood of interacting with a third party outside the kinship community is small. However, the pattern might be reversed in places where the population size is larger and where the likelihood of interacting with non-kin increases, because the likelihood of being cheated also increases as a result [132-134], and this risk might further increase in adverse conditions. In these environments, trusting other people less and cooperating less can be viewed as a means to reduce uncertainty, a view which is in contradiction with Amir et al.’s findings [39], but which fits a large set of work in the social sciences [81,83,135] as well as our own results. We think that risk management strategies in the context of social behaviors is an interesting topic for further research.

Whether cooperation is decreased or increased in adverse environments is still debated. The present study contributes to this debate by reporting two pre-registered studies applying structural equation models to test this relationship on large-scale datasets (the European Values Study and the World Values Survey). The results show that adverse environments in childhood and adulthood are associated with decreased investment in adult cooperation. Furthermore, the channel through which this relationship arises is also unclear. We proposed that this relationship is mediated by a present bias at the psychological level, which is embodied in the reproduction-maintenance trade-off at the biological level. We indeed found that the relationship between environmental adversity and cooperation is mediated by individuals’ reproduction-maintenance trade-off.

European Values Study countries descriptive statistics.

Values are unstandardized coefficients with standard deviations in brackets.

(XLSX)

Click here for additional data file.

European Values Study collective action items.

(DOCX)

Click here for additional data file.

European Values Study listwise deletions—Full results.

(XLSX)

Click here for additional data file.

European Values Study listwise deletions—Controlled for current adversity.

(XLSX)

Click here for additional data file.

World Values Study countries descriptive statistics.

Values are unstandardized coefficients with standard deviations in brackets.

(XLSX)

Click here for additional data file.

World Values Survey collective action items.

(DOCX)

Click here for additional data file.

World Values Survey listwise deletions—Full results.

(XLSX)

Click here for additional data file.

European Values Study discovery sample—Results.

(DOCX)

Click here for additional data file.

European Values Study imputed data—Results.

(DOCX)

Click here for additional data file.

European Values Study listwise deletions—Mediation effects.

(DOCX)

Click here for additional data file.

World Values Survey discovery sample—Results.

(DOCX)

Click here for additional data file.

World Values Survey imputed data—Results.

(DOCX)

Click here for additional data file.

World Values Survey listwise deletions—Mediation effects.

(DOCX)

Click here for additional data file.

11 Mar 2020

PONE-D-20-00738

Environmental harshness is associated with lower investment in collective actions

PLOS ONE

Dear Mr Lettinga,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Apr 25 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Valerio Capraro

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Additional Editor Comments (if provided):

The reviewers like the paper but suggest several improvements. Therefore, I would like to invite you to revise your manuscript according to their comments. Needless to say that all comments must be addressed.

I am looking forward for the revision.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Environmental harshness is associated with lower investment in collective actions

Lettinga, N., Jacquet, P.O., André, J-B., Baumard, N., Chevallier, C.

Reviewer: John Lazarus

Limitations of review. I am not expert in structural equation modelling and I do not have experience working with either of the data sets used here.

General comments

1. This is a valuable study in a currently lively and somewhat controversial area of research. It is particularly useful in measuring associations of prosociality with both environmental harshness and measures predicted by life history strategy (LHS) theory.

2. The paper tests the hypotheses that prosociality is reduced in adults both in a harsher current environment and when raised in a harsher environment during childhood. The paper is clearly written, the Introduction covers the relevant literature well and the Discussion section is commendable for the way it considers limitations of the study; problems of causal inference from correlational data; and alternative evolutionary explanations for the findings.

3. Throughout the paper the authors describe various behavioural measures examined by others, and by themselves in this paper, as ‘cooperation’. In evolutionary biology cooperation is defined as a social action resulting in a benefit to both (all) parties in the interaction. The behaviours analysed here are volunteering and political action, behaviours that would be defined in evolutionary biology as ‘altruism’; that is, actions resulting in a benefit to the recipient(s) and a cost to the self, and the authors agree that these two behaviours do have this cost-benefit structure (lines 187, 189-90 and 475 [later numerical references are also to line numbers]). Similarly some behavioural measures in previous studies cited here as cooperative (e.g. reference 30) are actually altruistic (e.g. volunteering and charitable giving).

Since this paper is testing hypotheses within the discipline of evolutionary human studies it is important that the terminology is correct. This is far from being a dry terminological matter; rather it is central to the hypotheses tested here, as I shall go on to argue. The authors are testing the idea that in harsh

2

environments there is a ‘need to favour immediate over long-term benefits’ (54) and they state that ‘cooperation is a future-oriented strategy: in the short term, it is more advantageous to reap immediate, smaller, but more certain benefits by cooperating less’ (62-64).

I have a number of comments related to this point: • (a) The behavioural measures employed here – volunteering and political action – are altruistic not cooperative as defined above. Consequently they are suitable for testing the hypothesis that they will be less evident in harsh environments, but this is because they are immediately costly, and not, as argued in the paper, because benefits from them are delayed. • (b) I do not agree with the claim quoted above (62-64, also 19-20) that cooperation is necessarily a ‘future-oriented strategy’. It is futureoriented in the case of reciprocal altruism where, after an exchange of altruistic acts between two individuals, both have gained a net benefit (but is only future-oriented if the focal individual is the first altruist). (Volunteering and political action, in themselves, do not fall into this category – rather they are altruistic – although one can imagine scenarios in which they are followed by a reciprocal benefit.) But where the cooperation involves joint simultaneous activity there is an immediate benefit to both parties (compared to not cooperating). • (c) In consequence of this last point, and following the arguments of the authors, simultaneous cooperation, as defined here, would be predicted to be more common under harsher conditions since it provides an immediate benefit in comparison with the decision to refrain from cooperation. • (d) In using a ‘proxy’ measure for cooperation (e.g. 95-6) it is necessary that the proxy satisfies the definition of cooperation given above. I do not mean here the distinction between behavioural and attitudinal measures (472-3), which is, of course, important. Rather, I mean that whether it is a behavioural or an attitudinal measure it should satisfy the above definition of cooperation (which for an attitude would refer to an imagined cost-benefit structure). • (e) I suggest that the authors look again at their use of the terms ‘cooperation’ and ‘altruism’, and their derivatives, in light of these comments. At the same time it would be useful also to check use of the terms ‘prosociality’ and ‘collective action’. ‘Prosociality’ refers to a social action providing a benefit to the recipient and therefore includes both altruism and cooperation, as defined above. I would like to see a

3

definition in the paper of ‘collective action’ since it is used a number of times. My next point argues, more generally, for the importance of clear definitions and distinctions between the different types of prosocial behaviour for this field of research.

4. Following on from comment (e), it seems to me that the discrepant findings in the literature about the effect of environmental harshness on prosociality (the authors refer to this at 42-6) may be partly a function of the literature not distinguishing carefully enough between different types of social interaction, with potentially different cost-benefit structures for the players; broadly cooperation versus altruism.

I declare an interest here. I have a paper (Lazarus 2017, and included with this review), one not cited by the authors, in which I review evidence from humans and other taxa that cooperation (defined, as above, in evolutionary biology) may be an inverted-U function of environmental adversity (i.e. harshness). The review is accompanied by an explanatory evolutionary model that differs from the ideas discussed in the present paper (i.e. selection for present-focussed decision making in harsh conditions). The model is intended to supplement, and not to replace, explanations based on present-future trade-offs. The authors may wish to refer to my paper.

My paper also discusses the point made in the paragraph before last and I quote the passage here (Lazarus 2017, section 6, p. 590):

6. Other forms of prosociality Cooperative behaviour shares motivational and emotional features, and evolutionary and cultural origins, with related prosocial behaviours, beliefs and attitudes, and the influence of adversity has been studied here too. At the level of individual differences those with more adverse life experience exhibit more empathy, compassion and generosity in charitable giving and aid to a stranger (Lim and DeSteno 2016; Lim 2017). The influence of social class on prosociality is controversial, with recent conflicting findings for a range of measures: utilitarianism, empathy, feelings of entitlement, narcissism, theft, lying, cheating, helpfulness, generosity, trust, trustworthiness, volunteering and charitable donation. Some studies find greater prosociality in higher classes (Korndörfer, Egloff, and Schmukle 2015) but others find the opposite class effect (Côté, House, and Willer 2015), to mention just two of the most recent reports. In understanding these apparently conflicting findings it will be helpful to develop more tailored predictions for each measure, as attempted here for

4

cooperation. If the inverted-U relationship proposed here between environmental quality and cooperation should hold also for some of these other forms of prosociality this might resolve some of the (apparent) inconsistencies in the data, since conflicting studies could be situated on opposite sides of the environmental inflection point. As others have noted it is also important to take account of the methodology by which these results have been obtained, from behavioural observations and experiments to self-report, in both the lab and the real world, since they each have their own psychological influences.

I invite the authors to engage with these ideas.

Detailed comments

Title: Since the LHS variables play a major role in the association between environmental harshness and the prosociality variables I suggest that LHS is added to the title.

76. ‘By contrast’ with the previous sentence these references are expected to deal with the effect of the experience of safe conditions during human childhood. However, two seem not to be relevant: [54] is a comparative mammalian paper and [60] is a study of guppies.

121 and 338. Provide references or web links to the two data sets.

123. The countries in the survey: it would be useful, in the supplementary material, to provide lists of the countries in each study uniquely, and in both studies. Country differences may help to explain some differences in the results of the two studies, and will also be useful for comparison with future studies.

134. ‘in’ should read ‘into’.

135-7 & 370-72. ‘In both subsamples, we excluded participants with too many empty cells (more than two standard deviations from their respective subsample mean)’. I find this unclear; ‘empty cells’ seems to refer to missing data whereas the passage in parentheses seems to refer to outliers.

151. ‘parent’s’ should read ‘parents’. ‘end’ should read ‘ends’, both here and in the figures.

5

166. This should probably read ‘health costs are induced’.

177. ‘no’ should read ‘not’.

248 and 403. Discovery and replication samples in the two studies. In the Results section of each study there is analysis of just one set of data. It is not stated whether this is the discovery sample, the replication sample, or the two combined. If it is one of the two samples then the second should be added, or comparison made with the first. If it is the two combined then any differences between the two samples should be described.

255-9. Additional analysis with imputed data, to deal with the problem of missing responses. ‘The results of this analysis are consistent with the results obtained using listwise deletions (see S1 Text in the Supporting Information)’ (257-9). In contrast to this statement in the paper there is an important difference in the results of the two analyses. In the main text analysis the authors conclude from a significant negative association between childhood harshness and collective action (309 and Figure 2) that ‘This implies that a harsher childhood environment is associated with less investment in collective actions later in life’ (309-10). However, the same conclusion is drawn from the additional analysis although now the same correlation is significant and positive (S1 Text, p. 3, lines 3-4 and Figure S1). This requires Discussion in the main text, particularly since it represents a core prediction of the paper.

282-7 and S1 Text. Discussion of the effect of the remaining independent variable, parental education, is missing. Is this because of the standardisation procedure?

291. There seem to be words missing at the end of this line.

29. State what the ‘expected way’ is.

303 & 444. ‘bold arrow’ would probably be better as ‘continuous arrow’, in contrast to the dotted arrows. Certainly in the typescript one cannot tell bold from normal lines.

364-7. Excluding respondents without children. Further explanation would be useful here since the variable ‘only relevant for respondents with children’ (365-6), age at first birth, is not used in this second study (but is in the first). So

6

were childless respondents excluded so that the two studies would be more similar in their respondent characteristics?

390. Insert ‘respectively’ after ‘study 1’?

480. ‘whether people cooperate less under harsher conditions . . .’?

517. ‘administrating’ should read ‘administering’.

531. Delete ‘shocks’?

537-9. Either ‘correlates’ should read ‘correlated’ if the correlation was within the study or a reference should be given if it was a correlation with another study.

560 ‘Their model’ would read better as ‘The model’.

830. Add journal name.

Reference

Lazarus, J. (2017) Cooperation in adversity: An evolutionary approach. Global Discourse 7 (4): 571–598, doi: 10.1080/23269995.2017.1402426

Reviewer #2: This manuscript aims to build on previous research showing a negative association between “environmental harshness” and cooperation in contemporary human societies. Specifically, the authors argue that this association is mediated psychologically by a “present bias” (or future discounting) and is embodied biologically by altering the tradeoff between reproduction and maintenance in favor of reproduction. The manuscript tests this hypothesis through analysis of data from two large, cross-cultural studies: the European Values Study (EVS) and the World Values Survey (WVS).

Overall, the manuscript is clearly written and accessible, both to specialists working in this area as well as a more general scholarly audience. The theoretical and empirical motivations are linked directly to the hypotheses. The steps used to produce the data samples for analysis and the methods of analysis are both well documented. The results of the analysis are linked back to the hypotheses, and the discussion includes a helpful review of some limitations of the study and unresolved theoretical debates.

There are several additional strengths of the research featured in the manuscript. The use of large, cross-cultural surveys presents opportunities to test associations between “environmental harshness” and cooperation in a variety of contexts with measures that arise from a common survey methodology, which should enhance the quality of comparisons. The large sample sizes in these data sets also allow the authors to develop their models using one half of the sample, then validate these models with the other half. The analysis explores direct associations between “environmental harshness” and cooperation, as well as mediations via the trade-off between reproduction and maintenance, which unites two strands of research on environmental harshness that are often pursued separately. The discussion of analytic methods and interpretation of results are also sufficiently nuanced to take into account both the strengths and weaknesses of the data used in the study, which helps clarify the contributions of this research as well as opportunities for future studies to build on them. I also appreciated the choice to focus on “collective action” variables from these particular data sets, rather than moral attitudes about cooperation, which have been the focus of previous research.

Along this these strengths, there are several areas where the manuscript could be improved. I outline these below:

1. The manuscript uses the term “environmental harshness” to refer to a broad range of conditions, defined by the “rate at which external morbidity and mortality causes disability & death in a population” (p.3). The authors go on to explain that this often manifests in contemporary human populations as “economic deprivation and resource scarcity.” I found the use of the term “environmental harshness” to be somewhat confusing in this regard, as I was expecting the research to focus more on ecological conditions rather than social conditions tied to “harshness.” Considering the countries included in these data sets, the real-world challenges reflected in the variables used to operationalize the concept of “harshness” are more directly tied to social conditions and dynamics than they are to ecological conditions and dynamics. The manuscript is fairly clear in how it defines the term “environmental harshness,” so this is not a major concern, but I do wonder if it would be closer to the spirit of the analysis to replace this term with something like “economic deprivation and scarcity,” which probably better reflects the specific measures used in the study and the social contexts where these surveys were conducted.

2. Related to the concern about the definition of “environmental harshness” described above, I felt the manuscript could do more to consider recent research on the evolution of cooperation that shows a positive association between environmental harshness and cooperation. For example, several papers by Smaldino and colleagues (2013a; 2013b; 2013c ) use agent-based models to show that increasing environmental harshness leads to increasing cooperation, as does research by Aktipis, Cronk and colleagues (2011; 2015; 2016; 2019) These positive associations run in the opposite direction as most of the research cited in the manuscript, which is presented as the current consensus from an empirical standpoint. The studies showing positive associations also focus more directly on ecological conditions linked to “harshness,” so perhaps this is one reason they diverge from those cited in this manuscript, which share the focus on economic deprivation and scarcity. Regardless, I would like to see this manuscript engage more directly with this work showing positive associations between environmental harshness and cooperation, both in terms of its theoretical framework and interpretation of its results.

3. The authors suggest one of the manuscript’s primary contributions to research in this field is to test the idea that the association between environmental harshness and cooperation is mediated by a present bias at the psychological level that is in turn embodied at the biological level in terms of reproductive strategies. This is an interesting idea that builds on previous research on variation in human life history strategies in relation to environmental harshness, but there are some limitations in the analysis featured in this manuscript that I think should be discussed. First, the analysis does not examine present bias or other psychological mechanisms directly. Instead, it infers these from specific strategies for reproduction (e.g. early reproduction, more children) and cooperation (e.g. fewer contributions to collective action). These inferences are generally logical, but there may be alternative explanations that don’t involve present bias at the psychological level. For example, research by Amir and colleagues on this topic (2019) has suggested that risk management is an alternative explanation, which is not cited nor discussed in the manuscript. Considering the possibility that reproductive and cooperative strategies may also be shaped by risk management, I feel the authors should address this alternative explanation (and the empirical research supporting it) in their manuscript. Indeed, Amir et al (2018) have empirical evidence that environmental harshness is positively associated with cooperation, connecting this issue to the previous one. Perhaps there are even variables in the EVS and WVS that could be used to test this alternate hypothesis. Also, it’s worth noting that some of the work by this research team also provides support for some of the associations presented in this manuscript (Amir et al 2016)

4. There are several missing details about the EVS and WVS data sets that are important for interpreting the results featured in the manuscript. First, the manuscript mentions the number of countries included in each data set, but does not explain which countries are involved. This is more important for the WVS than the EVS, because part of the justification for including analysis from the WVS was to expand the sample of countries considered beyond Europe. I recommend the authors list these countries included in the samples used for analysis for both data sets, either in the main text or in a table included in the main text, perhaps along with descriptive stats at the country level for each of the main variables. Second, the manuscript should engage with or at least acknowledge the extent of variation in social and ecological contexts that are and are not reflected in these samples. Does environmental harshness (or economic deprivation and scarcity) manifest in similar ways in all countries, or are there variations that should be accounted for? More details along these lines would help develop some key parts of the manuscript where this issue arises. For example, in the discussion, the manuscript notes the fact that “number of children” did not substantially contribute to the construct of the reproduction-maintenance tradeoff in the EVS, but it did in the WVS (p.22). The discussion suggests this may be due to the use of contraceptives and other “cultural factors.” This would be a good opportunity to provide more details about the countries included, specifically variation in demographic profiles and other factors linked to reproduction, as well as factors linked to the measures of environmental harshness and collective action. Finally, the modeling strategy doesn’t appear to account in any way for differences at the country level when estimating the associations among the key variables. A multilevel modeling approach might help clarify the extent that the associations reported in the manuscript are general across all countries.

5. The manuscript’s presentation of results and discussion focuses mostly on the statistical significance of the documented associations, but there is little discussion of effect sizes, aside from a brief mention that the reported effect sizes are generally small (p.14), as well as how the effect sizes for the direct association between environmental harshness and collective action decrease when accounting for mediation via reproductive strategies (p.15). This left me wondering about the substantive significance of these associations in relation to observed variation in the key variables. The bivariate correlation coefficients appear to suggest relatively little co-variance in the two composite collective action measures (volunteering and political action) and the variables representing the two predictor constructs. The large sample sizes used in this study raise the possibility that the statistically significant associations reported are due primarily to the high statistical power enabled by the large sample sizes used in this analysis, rather than effect sizes. More detailed discussion of effect sizes and substantive significance would help balance the interpretation of full significance of these associations.

6. Finally, the Discussion section ends rather abruptly with a discussion of limitations and unresolved issues. I suggest the authors consider including at least a short paragraph or two to return to the main contributions of this work and perhaps suggest some important directions for further research in this area, which would help return to the general motivations of this work.

Addressing these issues should help to highlight the strengths of this manuscript and its valuable contributions to our understanding of the links between environmental harshness, reproductive strategies, and cooperation in contemporary human societies.

Citations

Aktipis, C.A., Cronk, L. & de Aguiar, R. Risk-Pooling and Herd Survival: An Agent-Based Model of a Maasai Gift-Giving System. Hum Ecol 39, 131–140 (2011). https://doi.org/10.1007/s10745-010-9364-9

Aktipis, A., de Aguiar, R., Flaherty, A. et al. Cooperation in an Uncertain World: For the Maasai of East Africa, Need-Based Transfers Outperform Account-Keeping in Volatile Environments. Hum Ecol 44, 353–364 (2016). https://doi.org/10.1007/s10745-016-9823-z

Amir D, Jordan MR, Bribiescas RG (2016) A Longitudinal Assessment of Associations between Adolescent Environment, Adversity Perception, and Economic Status on Fertility and Age of Menarche. PLoS ONE 11(6): e0155883. doi:10.1371/journal. pone.0155883

Dorsa Amir, Matthew R. Jordan, David G. Rand. 2018. An uncertainty management perspective on long-run impacts of adversity: The influence of childhood socioeconomic status on risk, time, and social preferences,

Journal of Experimental Social Psychology, Volume 79, Pages 217-226, ISSN 0022-1031

Amir, D., Jordan, M. R., McAuliffe, K., Valeggia, C. R., Sugiyama, L. S., Bribiescas, R. G., Snodgrass, J. J., & Dunham, Y. (2019). The developmental origins of risk and time preferences across diverse societies. Journal of Experimental Psychology: General.

Cronk L. et al. (2019) Managing Risk Through Cooperation: Need-Based Transfers and Risk Pooling Among the Societies of the Human Generosity Project. In: Lozny L., McGovern T. (eds) Global Perspectives on Long Term Community Resource Management. Studies in Human Ecology and Adaptation, vol 11. Springer, Cham

Yan Hao, Dieter Armbruster, Lee Cronk, C. Athena Aktipis, (2015). Need-based transfers on a network: a model of risk-pooling in ecologically volatile environments, Evolution and Human Behavior, Volume 36, Issue 4, 2015, Pages 265-273, ISSN 1090-5138,

Paul E. Smaldino, Jeffrey C. Schank, and Richard McElreath (2013a), "Increased Costs of Cooperation Help Cooperators in the Long Run.," The American Naturalist 181, no. 4 (April 2013): 451-463.

Smaldino PE, Newson L, Schank JC, Richerson PJ (2013b) Simulating the Evolution of the Human Family: Cooperative Breeding Increases in Harsh Environments. PLOS ONE 8(11): e80753.

Paul E. Smaldino, (2013c). Cooperation in harsh environments and the emergence of spatial patterns. Chaos, Solitons & Fractals, Volume 56, 2013, Pages 6-12, ISSN 0960-0779,

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: John Lazarus

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: Lettinga et al 2020-to authors.pdf

Click here for additional data file.

Submitted filename: 2017 Lazarus Cooperation in adversity an evolutionary approach.pdf

Click here for additional data file.

24 Apr 2020

Subject: Resubmitted Manuscript PONE-D-20-00738 in Draft

Dear Dr. Capraro,

Thank you for your helpful review of our manuscript and for providing us with the opportunity to revise and resubmit our work. We wish to thank the two reviewers for their positive evaluation of the work, their helpful comments and advice. In the “response to reviewers” file we provide a point-by-point reply to the reviewers’ comments, with changes to the paper placed in italics. We addressed all the reviewers’ comments and made significant changes to the introduction, clarified several topics and provided additional supportive data.

We hope you will find this version of the paper acceptable.

With kind regards,

Niels Lettinga, Pierre O. Jacquet, Jean-Baptiste André, Nicolas Baumard, Coralie Chevallier

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

18 Jun 2020

PONE-D-20-00738R1

Environmental adversity is associated with lower investment in collective actions

PLOS ONE

Dear Dr. Lettinga,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Aug 02 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Valerio Capraro

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Reviewer: John Lazarus

My replies to the authors’ responses are headed by the numbered points from my first review (or by line numbers in the first reviewed paper, following ‘pp’) in square brackets, followed by page number in the Track Changes version of the revised paper. (In their response the authors use the page numbers of the revised clean version but in checking the revision the Track Changes version is preferable.)

I appreciate the authors’ constructive engagement with my comments and their detailed and considered responses. None of my comments are major enough to require another look at the paper.

Looking at this again I see that figures 1 and 3 are redundant and could be removed; figures 2 and 4 repeat them, respectively, while adding the statistics.

[3] 98-103. Definition of cooperation. Now that the authors provide a source (West et al., references 61, 62) for their definition, and explain it, this point is satisfactorily resolved.

[3, particularly 3a, b, e] 136-7, 236-49. The authors are now explicit about the potential delayed benefits of volunteering and political action (with supporting references), and have defined ‘collective action’. So this is OK.

[4] ‘the discrepant findings in the literature about the effect of environmental harshness on prosociality (the authors refer to this at 42-6) may be partly a function of the literature not distinguishing carefully enough between different types of social interaction’. This is a general point that I invited the authors to engage with but it is not essential for their paper that they do so. Although they did not take up this point they did add an interesting discussion (54-83) of a second point I raised here that, ‘. . . it is also important to take account of the methodology by which these results have been obtained, from behavioural observations and experiments to self-report, in both the lab and the real world, since they each have their own psychological influences.’

Detailed comments on first submission

[pp 255-9] 392-413. The new analysis and comparisons between existing analyses in response to my comment are a useful addition to the paper. However, the authors should state at 396-9 whether the new analysis is (a) on the replication or the discovery sample, and (b) on the listwise deletion or imputed data set.

All other comments have been answered satisfactorily.

Comments on new text & new supplementary material

57, 60. ‘correlate’ not the best word here since causation is proposed.

310-12 & S1 Text. Largely ‘consistent with the replication sample’ yes, but point out that there are two additional significant associations with Childhood adversity in the discovery sample and name them.

393, 399, 405, 412, 560, 564. To stick to the variable name in the studies, as elsewhere in the paper, ‘cooperation’ should read ‘collective action’.

456-7. Please explain the logic of ‘because not having children now does not mean that you don’t want children in the future’.

634. ‘linear’ should read ‘monotonic’.

707. ‘contest’ should read ‘context’; ‘are’ should read ‘is’.

687-707. With reference to this discussion on uncertainty the authors might be interested in the following theoretical paper, co-authored by myself, but there is no need to cite it here: Andras, P., Lazarus, J., & Roberts, G. (2007). Environmental adversity and uncertainty favour cooperation. BMC Evolutionary Biology, 7, 240. doi: 10.1186/1471-2148-7-240.

Reference 1 has a subtitle, to be added: ‘The impact of harsh versus unpredictable environments on the evolution and development of life history strategies’.

References 61 & 62. The journal names in the doi of each reference have been switched.

Reference 133 is perhaps incomplete.

S1 Text (and check other text and tables for same comment). As I’m sure the authors know, p values aren’t actually zero, statistics packages merely print them this way; ‘0.00’ should read ‘<0.005’.

Reviewer #2: In addition to the strengths of the initial submission, my review highlighted six issues for improvement. I have read the authors’ reply to each issue and reviewed the relevant changes made to the revised manuscript. I am satisfied with the responses and revisions and support publication of this manuscript. I offer more specific replies on each point below:

1. Use of the term “environmental harshness.”

- The authors have replaced the term “environmental harshness” with “environmental adversity.” I support this change and feel it helps clarify the specific social and environmental conditions that are the focus of this manuscript in relation to variation in reproductive strategies and cooperation.

2. Positive or negative association between environmental adversity and cooperation?

- The revised manuscript has an expanded review of published literature on this question and effectively highlights studies that find both positive and negative associations. I particularly like the way the authors explore possible methodological factors that might help explain these diverging results and link this idea to the results featured in their manuscript. My only suggestions for improving this aspect of the manuscript further would be: a) to reflect on and explore possible differences between self-reported measures of cooperation and behavioral measures (especially via economic experiments or “games”) and b) to add a note explaining that several of the studies cited as support for a positive association between environmental adversity and cooperation use agent-based models as a methodology, which is distinct in important ways from both self-reported and behavioral economics-based measures.

3. Alternative explanations/mechanisms underlying associations between environmental adversity and present bias.

- The revised manuscript features an extended discussion of alternative explanations/mechanisms, including uncertainty management. I appreciate the way the authors consider these alternatives and discuss how the results in this manuscript do and do not align with them.

4. More information on EVS and WVS data sets.

- The supplementary files associated with the manuscript now provide more detailed data about the countries included in the EVS and WVS data sets, as well as descriptive statistics for the variables used in analysis. I did notice one possible issue (hopefully minor) in Table S1 for the EVS, the range for the “Age1stBirth” variable goes from 12-83. My understanding is that this variable represents the age when each person responding to the survey first became a parent. Does this mean that the oldest parent in the data set was 83 years old when their first child was born? This would be very old for a male (but not unprecedented) and would, as far as I can tell, be a world record for a female. If the authors have not already looked into this, it might be good to double check the extreme old/young values for this variable in relation to other variables (gender, age) to make sure there are no errors in this important variable.

5. Effect sizes for SEMs.

- The authors have added some additional information about effect sizes for the associations reported in the SEMs and include some discussion of effect sizes in their assessment of the relationships between environmental adversity, reproduction, and cooperation. The manuscript’s conclusions continue to emphasize results of statistical significance tests the direction of associations, but given the large sample sizes, I still feel the discussion of effect size could be expanded to establish links to the substantive significance of these associations. Perhaps this could be done by linking the estimated coefficients to units of measurement for each variable, then discussing what would constitute a meaningful change in levels of cooperation and how large a change in environmental adversity/reproductive strategy is required to produce this change, according to the SEMs. Perhaps the authors may find this recent article useful in this regard:

Brenna Gomer, Ge Jiang & Ke-Hai Yuan (2019) New Effect Size Measures for Structural Equation Modeling, Structural Equation Modeling: A Multidisciplinary Journal, 26:3, 371-389, DOI: 10.1080/10705511.2018.1545231

6. Summary paragraph.

- The revised manuscript contains a summary paragraph at the end of the Discussion section that nicely states the key contributions of this study.

I thank the authors for their considerate responses to the issues raised in my initial review.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: John Lazarus

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

10 Jul 2020

Our response can be found in the file "Response to Reviewers".

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

14 Jul 2020

Environmental adversity is associated with lower investment in collective actions

PONE-D-20-00738R2

Dear Dr. Lettinga,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Valerio Capraro

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

17 Jul 2020

PONE-D-20-00738R2

Environmental adversity is associated with lower investment in collective actions

Dear Dr. Lettinga:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Valerio Capraro

Academic Editor

PLOS ONE