– by Yohsuke Ohtsubo

Men are riskier than women, which is one of the most robust findings in psychology. When compared to women, men tend to report that they seek more risks in psychometric studies. In laboratory studies, men tend to show riskier behaviors, such as expanding a computerized balloon to the point of exploding to earn more money. Outside of the lab, men tend to drive faster (often causing more traffic accidents), engage more in extreme sports (e.g., cliff diving), and more often become involved in homicidal conflicts.

A common evolutionary explanation for this robust sex difference is sexual selection: Men’s risk proneness was selected for because it conferred upon men (but not women) fitness advantages in intrasexual competition and/or intersexual mate choice. However, sexual selection is associated with at least two “contradictory” predictions. Men in good conditions who are able to take risks (e.g., cliff diving) appeal to women by taking the risks. Thus, it is expected that risk-seeking men enjoy higher reproductive success (i.e., having more children). Alternatively, it is possible that men in poor conditions who have less to lose are more likely to take risks (e.g., crimes). In this case, risk-seeking men, due to their poor conditions, should have fewer children than men in better conditions, but they are still better off taking the risks because they would otherwise end up childless.

In our study, we were interested in whether risk-prone men tend to have more offspring than risk-averse men in modern environments (i.e., contemporary Japan and the US). However, we were aware of the possibility that we could end up with null results (i.e., no association between men’s risk proneness and reproductive success) because sexual selection operated in the past—our species’ ancestral environments. If modern environments are so much different from the ancestral environments, risk proneness may no longer confer any fitness advantage on men.

Given the possible evolutionary mismatch between the modern and ancestral environments, one may find this investigation itself a risky endeavor. However, as we acknowledged in the paper, we had preliminary evidence—our unpublished study involving middle-aged Japanese men showed that their retrospective reports of risk proneness were positively correlated with the number of children they had. That said, this investigation might have been risky because, partly due to the low reliability of recall data, the correlation was quite small (approximately 0.15).

Nevertheless, our curiosity outweighed the fear of null results. We collected data from 1,205 Japanese individuals (601 men and 604 women) aged 45 to 55 years old. Among other things, we measured their retrospective risk proneness during their 20s and 30s and the number of children they had. We assumed that the number of children reported by this age group comes very close to their lifetime reproductive success because less than 5% of fathers and only 0.19% of mothers of newborns in Japan are 45 years and older. We successfully replicated the small correlation for men. The retrospective risk proneness × number of children correlation was small but significant in men (0.123), while it was virtually zero in women. More importantly, the difference between men’s and women’s correlations (0.123 vs. 0.001) was significant (many thanks to an a priori power analysis!).

Encouraged by the Japanese results, we went ahead with another round of data collection in the US. This successfully replicated the Japanese results—men’s retrospective reports of risk proneness were significantly correlated with the number of children they had (0.143), while the same correlation for women was not significant (−0.002). And, again, the men’s correlation was significantly larger than the women’s!

In both countries, we used the domain general risk proneness score in the main analysis. Sample items included “I took risks regularly” and “I preferred to avoid risks” (reverse-coded item). However, for exploratory purposes, we also included domain-specific risk measures, which measured respondents’ willingness to take different types of risks: recreational risks (e.g., rock climbing, scuba diving), health risks (e.g., smoking), career risks (e.g., quitting a job without another to go), financial risks (e.g., gambling), safety risks (e.g., fast driving) and social risks (e.g., publicly challenging a rule or decision). Interestingly, only recreational and safety risk-taking consistently exhibited a similar pattern as the domain general risk proneness: men’s, but not women’s, recreational and safety risk scores were significantly correlated with reproductive success.

These results are generally consistent with the notion that men in good conditions take more risks, and their risk-proneness is associated with higher reproductive success. However, isn’t it possible that risk proneness is associated with higher reproductive success because only successful risk takers are involved in our data? Unsuccessful risk takers might have died in risky activities, such as rock climbing and cliff diving. The unmeasured cost of dying and the measured reproductive benefit may cancel each other out. Our unpublished follow-up data collected in Japan may have some clues on how to address this problem. It shows that men with good records in high school and college sports reported being more risk-prone during their school days than a control group of average achievers. Physically fit athletic men may engage in risky behaviors, such as extreme sports, while experiencing a lower risk of injury. Obviously, this issue needs more studies.

Despite certain limitations, the evidence from our research is useful for evolutionary psychologists to respond to a common criticism against evolutionary psychology. It is sometimes criticized as being a “just-so-story”—there is no direct evidence that this behavioral trait underwent sexual selection. That’s true. However, it is certainly more persuasive to say, “We know this is the case in modern environments, so perhaps it was like this in the past” than to say, “Perhaps it was like this in the past.” Although our research is admittedly a small first step, we think it still increased our knowledge about the role sexual selection has played (and seemingly continues to play) in men’s risk proneness.

Read the original article: Sakamoto, R., & Ohtsubo, Y. (2025). Men’s but not women’s risk proneness in early adulthood is associated with lifetime reproductive success: Evidence for sexual selection in modern environments. Evolution and Human Behavior, 46(1), 106654.

– by Luke Glowacki

When I set out to write my recent article The Controversial Origins of War and Peace: Apes, Foragers, and Human Evolution, I hoped to provide clarity on the heated debate about the role of war and peace in human evolution. Is warfare part of our evolutionary legacy, or did it emerge from the societal shifts brought about by agriculture and permanent settlements? This isn’t just an academic puzzle—it’s a question that cuts to the core of how we understand ourselves, our past, and the future of human coexistence. My goal was to peel back the oversimplification that commonly characterizes both sides and dig into the messy, complex reality of our evolutionary journey.

The debate often falls into two camps: “deep rooters” and “shallow rooters.” Deep rooters argue that warfare has ancient roots, embedded deep in our evolutionary history, possibly stretching back to our last common ancestor with chimpanzees and bonobos. They point to the brutal, coordinated violence among chimpanzees as evidence that our lineage is steeped in conflict. Deep rooters take the fact that hunter-gatherers often have war as evidence that war would have occurred among our Pleistocene ancestors. Shallow rooters, on the other hand, see war as a recent human invention, a consequence of agriculture, sedentary life, and hierarchical societies. War emerged only recently they argue, once we stopped foraging and started living in settled agricultural communities. They highlight the peaceful, cooperative lives of bonobos and some hunter-gatherers, painting a picture of a more harmonious human past.

But here’s the rub, the debate has often overlooked that both war and intergroup cooperation are parts of our evolutionary legacy, and the evidence is not nearly as clearcut as both camps sometimes assume. Simplifying human evolution into a binary of “war” or “peace” misses the fascinating complexity of our story. Chimpanzees do engage in deadly raids but using them as a direct blueprint for early human behavior ignores millions of years of evolution that followed our split. And while bonobos are often held up as paragons of peace, their societies aren’t entirely free of aggression. While lethal raids have not been reported for bonobos, they do appear to have rates of aggression that can rival chimpanzees. And neither species may be useful for understanding the last common ancestor humans shared with other apes. Further, we’ve had seven million years of evolution since sharing a common ancestor with chimpanzees and bonobos, raising questions about how much of our behavior is really inherited from a last common ancestor.

When I turned to the archaeological and ethnographic records of hunter-gatherers, the picture is  even more complex. It’s tempting to believe that warfare only appeared with the advent of agriculture, but that’s not what the evidence shows. Small-scale intergroup violence existed long before humans settled into farming communities, yet larger-scale violence appears to emerge only more recently. Just as often though, hunter-gatherers engaged in cooperative trade and alliances. And while many of them have war, war may be intermittent and even rare. Evidence of intergroup trade spanning hundreds of kilometers extends deep into the Paleolithic.

Our ancestors weren’t locked into one mode of interaction; they were adaptable, capable of both conflict and collaboration depending on the circumstances. And like humans everywhere, they likely used conflict and cooperation to obtain their goals depending on the circumstances. Because both cooperation and conflict have strong fitness relevance, both were likely important selective features in our species’ history.

What struck me—and what I hope strikes readers—is the dual legacy of war and peace in human evolution. Both violence and cooperation have been powerful forces shaping who we are. This duality resonates with the world we live in today, where acts of war and moments of profound cooperation coexist. We’re equipped for both, and it’s our environment, culture, and social structures that shapes which path we follow.

This understanding isn’t just intellectually satisfying—it’s hopeful. If peace is as much a part of our evolutionary toolkit as war, then striving for harmonious coexistence isn’t some naive dream; it’s a fundamental part of who we are. Recognizing this can shape how we approach conflict resolution and peacebuilding today. It reminds us that while conflict may be a part of our past, it is not an inevitable feature of our future.

Read the original article: Glowacki, L. (2024) The controversial Origins of war and peace: apes, foragers, and human evolution. Evolution & Human Behavior, 45(6), 106618.

– by Amy Zhao and Brendan Zietsch

Facial attractiveness studies have typically relied on asking people to rate facial photos of real-life participants or images of computer-generated faces. However, these ratings can be subjective and affected by rater biases. More recent studies (such as our own) have attempted to avoid subjective biases through the use of facial landmarks to derive objective measures of facial traits. However, these landmark-based measures ignore features thought to be relevant to face perception such as skin colour and contrast, hair, and eye colour. Here, we introduce deep neural networks as a method that combines the strengths of both approaches while addressing the limitations of facial landmarks.

Facial recognition neural network models are designed to extract abstract facial features from images. Each image input yields one set of multidimensional coordinates in feature space — a space representing the compressed version of the original image, with the number of dimensions representing the number of abstract facial features. The distance between two points (i.e. faces) in feature space reflects facial similarity, with similar faces represented by points that are closer together. While these coordinates lack direct interpretability, they effectively quantify abstract facial qualities that can be used to calculate facial traits relevant to facial attractiveness research.

We applied an existing facial recognition neural network model (VGG16) to facial images from our speed-dating study (n = 682). We used the extracted feature space coordinates to calculate traits such as facial averageness, similarity, and masculinity to predict in-person ratings of facial attractiveness and kindness. We then compared this neural network method to traditional manual (and automatic) landmark methods.

An issue that has been alluded to in past studies is that landmark measures of masculinity could be influenced by facial pitch (upward or downward tilt of a face). In our images, men tended to tilt their heads upward compared to women (there was a significant difference in facial pitch angle between genders). We found facial pitch was highly correlated with landmark measures of masculinity (-.17 ≤ r ≤ -.73). In contrast, there was little to no correlation between facial pitch and neural network measures of masculinity (.00 ≤ r ≤ -.23). Likely, gender differences in the way that men and women pose for photos might bias typical landmark masculinity measures. Here, we demonstrate that neural networks can extract facial information without being affected by limitations associated with landmarks.

Overall, facial measures derived from neural networks predicted in-person ratings, largely replicating what we found in our previous study using manual landmarks. Some differences were that neural network measures of masculinity robustly predicted facial attractiveness in men, whereas there was only context-dependent evidence for this in our previous study. We also found novel evidence for assortative preferences for facial masculinity. For example, participants with sex-atypical faces (a masculine woman or feminine man) revealed stronger preferences for a partner who was sex-typical (i.e. they rated sex-typical partners more attractive) than those with sex-atypical faces. We believe that we saw such effects in the context of neural network measures of masculinity due to increased visual information that was extracted from images of participants as well as decreased noise from participant facial pitch.

Neural network-derived measures had small to moderate correlations with landmark-based measures (.11 ≤ r ≤ .33), while manual and automatic landmarks were moderate to strongly correlated as expected (.29 ≤ r ≤ .86). Neural network masculinity measures were more accurate when it came to classifying the sex of the participant (95.6 % ≤ accuracy) compared to landmark measures (75.3% ≤ accuracy ≤ 88.8 %). Both low correlations between neural network and landmark measures as well as relatively higher sex-classification accuracy from neural network measures suggest that there is relevant information from facial photos that is uncaptured by landmarks. However, we did not find that neural networks were better (explained more variance) at predicting in-person ratings compared to other landmark measures.

While we found that neural network-derived measures do indeed predict in-person ratings, the underlying method is not well understood. Unlike landmarks, where we understand that the “average” male is one that most resembles the facial structure (as described by landmark coordinates) of the average male face, we are unclear as to what the “average” neural network face resembles – that is, what aspects of the face are contributing more or less to its position in feature space. While there are some ways in which we can use landmarks to describe (and visualise) shape variation, we are unaware of any straightforward way to visualise variation in feature space coordinates. (We note that we did create composite images of the top 20 participants for each trait.) While we controlled for ethnicity variables, this does not mitigate any systematic biases that may arise from imbalances in training face recognition models used by automatic landmarks and neural networks.

Given the lack of transparency behind neural network models, we suggest that researchers use caution when employing these methods. However, we also believe that neural networks are a fast, reproducible, and powerful way to extract visual information without the limitations associated with landmarks. A link to instructions and code for obtaining feature space values using neural networks is available in the full text of this paper.

Read the original paper: Zhao, A.A.Z., & Zietsch, B. (2024). Deep neural networks generate facial metrics that overcome limitations of previous methods and predict in-person attraction. Evolution & Human Behavior, 45(6), 106632.