Wednesday, October 19, 2022

We instinctively tend toward solutions that consist of adding something rather than subtracting something, even if the subtraction would be superior

Fillon, Adrien A., Fabien Girandola, Nathalie Bonnardel, and Lionel Souchet. 2022. “People Systematically Overlook Subtractive Changes (2021): Replication and Extension.” PsyArXiv. October 18. doi:10.31234/

Abstract: People systematically overlook subtractive changes and favor additive ones when generating new ideas. In a preregistered experiment conducted via the Prolific platform among French adults (N = 477), we replicated Experiments 2, 3 and 4 in Adams and colleagues (2021)’s study. We replicated the overlooking of subtraction, as participants generated 1155 additive ideas and only 297 subtractive ideas. Cueing participants (“Remember that you can add things or take them away”) increased the percentage who generated at least one subtractive idea (overall OR = 2.52, improvement condition, ϕ = 0.18, make-it-worse condition, ϕ = 0.24).

Results therefore provided empirical support for the overlooking of subtractive changes hypothesis. We also found that norms affected the generation of new ideas (descriptive OR = 7.49, injunctive OR = 6.86). Cues and injunctive (but not descriptive) norms were both related to the asymmetry.

Men do not find the bodies of apparent greater health and fertility the most attractive bodies; the attractiveness is linked to youth and low parity

The picky men: Men's preference for women's body differed among attractiveness, health, and fertility conditions. Chengyang Han et al. Personality and Individual Differences, Volume 201, February 2023, 111921.

Abstract: The health and fertility hypothesis suggests that low body weight young women are healthy and fertile, thereby judged as attractive to men. Although it has been widely accepted, few studies have tested the health and fertility hypothesis on a perceptual level, that is whether the most attractive female body is also perceived as the healthiest and most fertile. In the current study, we investigated young and older men's preferences women's body weight, using 3D human body models. With an interactive body preference task, men chose the BMI and body fat of women's body shapes perceived as most attractive, healthiest and most fertile. The results showed that both young and older men had similar patterns of preferences for women's bodies. For BMI, the most attractive body weight was not seen as the healthiest or the most fertile. Compared to the most attractive BMI, higher BMI was required to be seen as the healthiest and this figure was even higher for fertility judgements. Body fat generally showed similar patterns of results as BMI. Our findings challenge the health and fertility hypothesis and point to the alternative explanation that the judgement of women's attractiveness tracks cues indicating youth and low parity.


From an evolutionary perspective, sexual selection should have driven men and women to be attracted to cues that would maximize their reproductive success (Buss, 1988; Symons, 1979). For decades, researchers have found several traits that determine physical attractiveness in women, including Body Mass Index (BMI, weight divided by squared height, kg/m2), waist-to-hip ratio, waist-to-stature ratio, lumbar curvature, and leg length (see Lassek & Gaulin, 2016 for a review). Among all of these cues, BMI is argued to be one of the most crucial determinants of the attractiveness of women, as it plays a critical role in many cultures, especially in industrialized societies. Overall, lower values of BMI are found to be seen as attractive in women, although severe underweight is not attractive (Swami, 2015; Tovée et al., 1999).

It has been widely accepted that the preference for low BMI is adaptive because it indicates higher mate value with respect to health, fertility, fecundity, youthfulness, nulliparity and maternal investment (for a review see Bovet, 2019). One explanation that perhaps gets the most attention is the health and fertility hypothesis. Numerous studies have examined this hypothesis by exploring the relationship between BMI and health indicators (Lassek & Gaulin, 2018a) as well as fertility indicators (Lassek & Gaulin, 2018b). Although there is abundant physiological evidence both in support of and against this relationship (see Lassek and Gaulin, 2018a, Lassek and Gaulin, 2018b for reviews), evidence from the perceptual level is rare. Few studies have tested the perceived attractiveness, health and fertility of female bodies at the same time. Although the perceptual evidence is not mandatory for the hypothesis to be valid, it can provide additional support for or against it.

Singh (1993) published an influential study of female attractiveness claiming that attractive bodily features are indicators of women's fertility and health. Since then, most research in the field of female attractiveness has accepted this argument and built on it (Grammer et al., 2003; Marlowe et al., 2005; Weeden & Sabini, 2005). Not surprisingly, this hypothesis has been extended to low BMI which is assumed to signal optimal health and fertility (Tovée et al., 1999). However, recent evidence has cast doubt on these claims and proposed that the low BMI might serve as cues to nubility and reproductive value (Andrews et al., 2017; Lassek and Gaulin, 2019, Lassek and Gaulin, 2021).

In the past few decades, a great deal of research has investigated the BMI values associated with the most attractive female body. In well-nourished populations, it has been consistently found that the most preferred BMIs are around 18–20 (Crossley et al., 2012; Tovée & Cornelissen, 2001; Wang et al., 2015), which are far below the mean or modal values of typical young women in these populations (Lassek and Gaulin, 2016, Lassek and Gaulin, 2018b). This figure is even lower in East-Asian populations, which are 18.43 in Japanese (Swami et al., 2006) and 17.28 in Malaysian Chinese (Stephen & Perera, 2014).

According to the classification of BMI by WHO, the healthy BMI range is 18.5–24.99, BMI < 18.5 is classified as underweight and BMI >25 is classified as overweight. Abundant evidence from epidemiological studies has shown that overweight status is positively related to a series of cardiovascular diseases like diabetes, hypertension, heart attacks and stroke (Aune et al., 2016; Chen et al., 2013; GBD 2015 Obesity Collaborators, 2017; Khan et al., 2018) and some cancers like endometrial, breast, and colon cancer (Bhaskaran et al., 2014). Complementing these findings, a meta-analysis of over 10 million participants has found that all-cause mortality is lowest in the normal BMI range (20–25), whereas BMIs below (BMI < 20) or above the range (BMI > 25) was positively associated with overall mortality (Di Angelantonio et al., 2016). Similarly, a systematic review suggested that normal weight is associated with longer life expectancy compared to underweight and obese status (Bhaskaran et al., 2018). It should be kept in mind that most evidence suggesting the negative relationship between BMI and health is based on well-nourished populations like North American and Europeans.

In sharp contrast, studies of subsistence and forager groups found that plumpness was preferred in these cultures. Female bodies with a substantially high BMI (e.g., BMI ≥ 25) were judged to be attractive (Boothroyd et al., 2016; Swami et al., 2012; Tovée et al., 2006). This preference is especially strong in environments experiencing food scarcity (Anderson et al., 1992). These cultural variations have been attributed to psychological adaptations to local environments where lower BMI tends to be associated with poor health (Swami, 2015; Tovée et al., 2006).

In fact, evidence from subsistence populations indicates an inverse relationship between BMI and health, where the lowest mortality was found in women with high BMI (e.g. BMI > 25) and women with low BMIs have increased mortality rates (Hanson et al., 1995; Hodge et al., 1996; Sear, 2006; Wang & Hoy, 2002). This negative relationship between BMI and mortality might be due to the increased risks of getting infections in women with low BMI. Evidence from hunter-gatherer groups showed that the majority of female death in these populations was caused by infections in women with low BMIs (Strickland & Ulijaszek, 1993). Even in developed areas where people have access to antibiotics, women with low BMIs are more likely to get infections (Flegal et al., 2007; Milner & Beck, 2012). More recently, researchers found further evidence supporting the inverse relationship between BMI and health in women of reproductive age. Using a large U.S. sample dataset, Lassek and Gaulin (2018a) found that women of reproductive age with BMIs lower than 20 have worse health conditions than women with higher BMIs controlling for other factors that possibly affect health. Specifically, they were more likely to have infections, bed disability days, hospital days and more major disabilities.

To sum up, the existing evidence suggests that BMI does relate to health in many ways, where normal weight might be most protective. However, low BMIs, which are preferred by men in well-nourished populations, are linked to relatively higher mortality rates (Aune et al., 2016; Di Angelantonio et al., 2016) and poor health (Lassek & Gaulin, 2018a), not only in subsistence population but also in well-nourished population. This evidence may chanllenge the health and fertility hypothesis which posits that the preferred BMIs are indicators of good health (Tovée et al., 1999).

It is worth noting that nearly all the aforementioned studies are correlational or cross-sectional, one cannot be sure whether it is low or high BMI directly causes these health issues. Furthermore, one should be careful in interpreting the relationship between BMI and health because BMI conflates fat mass and muscle mass. At a given BMI, the body compositions vary between individuals. For example, people with low fat mass and high muscle mass will be classified as overweight by the BMI classification. Yet, higher fat mass is detrimental to health (Gómez-Ambrosi et al., 2011) while higher muscle mass is associated with enhanced fitness and health (Frankenfield et al., 2001; Johnson et al., 2015).

Perhaps the more relevant effect of BMI on health is its association to fertility health. Evidence from a US national longitudinal research showed that infertility rate is highest in underweight and obese women (Jokela et al., 2008). There is evidence suggesting that obesity is related to polycystic ovarian syndrome, a disease that could cause irregular periods, excess androgen levels and policystic ovaries, which in turn cause infertility problems (Barber et al., 2006; Lim et al., 2013; Vrbikova & Hainer, 2009). Compared to normal weight women, obese women are more likely to have miscarriages whether they conceived naturally or following ovulation induction (Lashen et al., 2004; Metwally et al., 2008). Furthermore, obese women have increased risks of birth defects such as neural tube defects, anencephaly, spina bifida, cardiac septal anomalies l and hydrocephaly (Rasmussen et al., 2008; Stothard et al., 2009).

Nonetheless, low body weight is detrimental to fertility health as well. Indeed, considerable evidence has shown that being underweight adversely affects menstrual function, pregnancy outcomes, perinatal outcomes and neonatal outcomes. To begin with, numerous studies have reported that being underweight, or states of energy deprivation like rapid weight loss and excessive physical activity can lead to menstrual dysfunctions like hypothalamic amenorrhoea, then cause infertility (Frisch, 1987, Frisch, 2004; Stokić et al., 2005; Støving et al., 1999). Women who have low BMI (e.g. BMI <19/20) like athletes and those with eating disorders are more likely to develop amenorrhoea, which is due to endocrine alterations, such as lowered estrogen levels (Ackerman & Misra, 2018; Hamilton-Fairley & Taylor, 2003; Ledger & Skull, 2004; Ziomkiewicz et al., 2008). When putting on weight, resumption of menstrual cycles is observed (Ackerman & Misra, 2018; Arends et al., 2012; Swenne, 2004).

More importantly, pregnancy rates were found to decrease with BMI. One study found that in women with BMI < 21, a one-unit BMI decrease was related to 3 % lower pregnancy rate (Van Der Steeg et al., 2008). Wang et al. (2000) showed that underweight women have a lower chance of getting pregnant with assisted reproduction treatment than women with normal BMI.

Even after conception, women with low BMIs are more likely to have negative pregnancy, perinatal and neonatal outcomes. A meta-analysis including different types of conception indicated that prepregnancy underweight is associated with increased risks of miscarriage (Balsells et al., 2016). Complementing that, a large-scale study with 3854 nulliparous women found that the average duration of pregnancy was shorter in underweight women compared to normal weight women (Hoellen et al., 2014). Consequently, preterm deliveries were significantly more common in underweight women. This gestational age difference was even more evident in extremely underweight women (pre-conceptional BMI < 16). As a result, low birth weight is positively linked to prepregnancy underweight status. Similar results were observed in women undergoing in vitro fertilization. Compared to normal weight women, implantation, clinical pregnancy, and ongoing pregnancy rates were lower in underweight women (Tang et al., 2021).

Despite the close relationship between weight and fertility, BMI may not be as closely related to fertility as body fat percentage (BF). Evidence has shown that a certain amount of BF is necessary for maintaining reproductive function. For example, Frisch, 1987, Frisch, 2004 stated that 22 % BF is necessary to maintain normal periods, get pregnant, and lactate. On the contrary, women who have very low BF like female athletes were found to be more likely to experience fertility-related problems like amenorrhoea, anovulation, irregular menstrual cycles, and delayed menarche compared to women with normal BF levels (Klentrou & Plyley, 2003; Redman & Loucks, 2005; Torstveit & Sundgot-Borgen, 2005; Zanker, 2006). In addition, studies have shown that poor nutrition which possibly relates to low BF is linked to delayed menarche (Gluckman & Hanson, 2006; Thomas et al., 2001). In other words, the reproductive span is shorter in women with low BF levels compared to women with normal BF levels.

One study that included healthy reproductive-aged women found that women with very low BF (<22 %) have lower levels of estradiol compared with women with average levels of BF (Ziomkiewicz et al., 2008). Additionally, this study found that a 10 % increase in BF was associated with an increase in estradiol levels. Estradiol plays a critical role in reproductive functions. Lower levels of estradiol decrease pregnancy rates both in healthy naturally conceiving women (Lu et al., 1999; Venners et al., 2006) and in women undergoing in vitro fertilization (Blazar et al., 2004; Chen et al., 2003). Consequently, low BF adversely affects fertility health.

The studies cited above mainly focused on the relationship between BMI and physical health as well as fertility health. Only few studies investigated these relationships from a perceptual perspective, which is how women's body weight affects their perceived health and fertility. Furthermore, to our knowledge, no study has examined the relationship between women's BF and their perceived health and perceived fertility. As noted above, BF might be more important to women's fertility than BMI. Hence, the current study aimed to test the health and fertility hypothesis from a perceptual perspective. Using 3D female body models with varying BMI and BF levels, we attempted to investigate men's preference for women's bodies in three dimensions: the BMI and BF levels found most attractive, the BMI and BF levels perceived to be healthiest, and the BMI and BF levels perceived to be most fertile. The evidence cited above may not support the widespread assumption that men prefer low BMI in female bodies because it indicates better health and fertility. Hence, here we predict that men will choose lower BMI and BF when optimizing attractiveness, than when optimizing the healthy and fertile appearance of women's bodies.

In addition, we also attempted to explore the age effect on body weight preferences under attractiveness, health, and fertility conditions. Previous studies showed mixed results regarding the age effect on the preferences of women's BMI. For example, Sorokowski et al. (2014) found older men preferred higher BMI of women's body than young men do, whereas George et al. (2008) found no effects of age on attractiveness preferences of BMI of women's body. In the literature, older men are generally less picky than younger men when judging women's facial attractiveness (Han et al., 2022; Marcinkowska et al., 2017), which may also present in body weight preferences. Moreover, social media plays an important role in shaping individual's perception of attractiveness, especially the most attractive body weight (Grabe et al., 2008; Swami, 2015). Young adults use social media more than older adults (Sharifian et al., 2021, also see surveys of U.S. adults present by Pew Research Center, 2021). One recent study showed that higher social media use was associated with more negative effects in young adults, but not in older adults (Sharifian et al., 2021). This may indicate that young adults are more influenced by social media than older adults. Slim bodies are frequently exposed in social media. Therefore, we predict that younger men will show stronger preferences for lower BMI and BF of women's body than older men.