3.1. Study Selection
Following a systematic search of PubMed (www.pubmed.ncbi.nlm.nih.gov) and Web of Science (www.webofscience.com), this search has identified 5869 and 5058 titles, respectively, of which 9654 have remained after removing duplicates (see Figure 1). Following a title screening, 9459 studies were excluded, leaving 195 for abstract screening. After this screening, 85 publications were chosen for full-text evaluation based on inclusion criteria, resulting in 31 included studies, while the other papers were discarded for being irrelevant and not fulfilling the inclusion criteria. The PRISMA flowchart in Figure 1 illustrates the selection process. seven authors were contacted to provide further information on their studies, to which two authors replied. For five articles, no information is provided about the age of the subjects, though authors were contacted several times. Two researchers were involved in screening the full texts for in- and exclusion criteria, and in case of disagreement, the study was discussed among all authors. The result of Cohen's kappa shows that they are in virtually perfect agreement (K = 0.89) about the decision of in- and exclusion of studies times.
3.2. Study characteristics
The 32 articles included in this systematic review have been published between 1997 and 2021 and contain data from 2721 males and 1385 females (total: 4.106 healthy adults; M pooled age = 22.14 SD pooled age = 3.11).
3.3. Results of individual studies
Here, the most consistent data describing hormonal impacts on several categories have been compiled to summarize the reviewed findings on the interaction between social hierarchy and hormones through decision-making and are presented in Tables 2–5. Paradigm descriptions can be found in Table 1.
3.3.1. The effect of hormones on decision-making.
As presented in Figure 1, 14 studies evaluated the effect of hormones on task-based decision-making (some studies tested multiple hormones, and the result of each hormone is reported as a separate study in the relevant section). Taken together, these studies provide data on 1617 males and 804 females (total: 2421, M pooled age = 24.21, SD pooled age = 3.71; one study were missing age-related information).
3.3.1.1. Testosterone. Twelve studies explored the impact of testosterone on decision-making paradigms via 17 experiments.
Four of these studies employed the Iowa gambling task (IGT), which includes two phases (see Table 1), uncertainty (first block), and risk (last block), and evaluated its connection with endogenous testosterone. However, all four studies discovered no link between endogenous testosterone and uncertain decisions in the uncertainty phase (Evans & Hampson, 2014; Reavis & Overman, 2001; Singh, 2021; Stanton et al., 2011a). Among these, in the risky phase, three of the studies observed riskier decision-making in individuals with higher endogenous testosterone levels who preferred more disadvantageous card sets (Evans & Hampson, 2014; Reavis & Overman, 2001; Stanton et al., 2011a). On the contrary, the fourth study (Singh, 2021) discovered that males with higher endogenous testosterone had a better decision performance in the risky phase, i.e., choosing the advantageous card set and making less risky decisions than males with lower endogenous testosterone.
Nofsinger, and Caleouge (2018) designed three experiments containing financial tasks (Portfolio Allocation and Portfolio Rebalancing, PORT; see Table 1) to investigate the association between endogenous testosterone and decision-making. The first experiment, by assessing investment asset allocation decisions, discovered that individuals with higher endogenous testosterone levels tended to select less risky options. In the second experiment, a similar version of the PORT was used and participants' positions within a group were revealed to increase competition. The finding showed a positive association between higher endogenous testosterone levels and investment risk selection. The third experiment involved long-term investment decision-making and showed that individuals with higher endogenous testosterone tended to choose riskier portfolios. Interestingly, no significant differences were found between males and females in any of the experiments (Nofsinger et al., 2018).
Two studies, comprising three experiments, investigated the impact of testosterone (endogenous levels: n=1, administration of testosterone: n=2) on risky decision-making using the Balloon Analogue Risk Task (BART) (Stanton et al., 2021; Wagels et al., 2017). One study, consisting of two experiments, found no significant effects of testosterone on risky decision-making. Specifically, the first experiment showed no significant sex difference in the association between endogenous testosterone and risk-taking behavior. The second experiment, which involved testosterone administration, did not reveal a significant difference in risky decision-making between the administration and placebo groups (Stanton et al., 2021). In the other study, which also involved testosterone administration, males with higher testosterone demonstrated riskier behavior, as indicated by a higher mean number of successful inflations (Wagels et al., 2017).
Two studies examined the relationship between testosterone and loss aversion paradigms (which includes the relative weighting of potential economic gains and losses), through three experiments (Stanton et al., 2011b; Stanton et al., 2021). In a study by Stanton et al. (2021), males exhibited different outcomes in two separate experiments. In the first experiment, testosterone administration reduced loss aversion, while in the second, it increased loss aversion. However, no significant differences were found in risk-taking levels compared to placebo. Also, Stanton et al. (2011b) reported no linear association between endogenous testosterone and loss aversion in both males and females. They report that the individuals with intermediate endogenous testosterone levels were risk- and ambiguity-averse compared to individuals with low or high testosterone, and showed females than males were more risk-averse and ambiguity-averse (Stanton et al., 2011b). Sapienza et al., (2009), using Holt and Laury's algorithm (risk aversion paradigm), observed a negative correlation between risk aversion and testosterone in females, not males.
Three studies investigated the impact of endogenous (Derntl et al., 2014) or administered (Cueva et al., 2015; Nadler et al., 2018) testosterone on decision-making. Derntl et al. (2014) used Haegler's risk game (HRG) and found no significant effect of endogenous testosterone on behavioral performance in both males and females. There was also no significant difference in the influence of testosterone on decision-making between males whose testosterone levels were assessed in the morning versus the ones evaluated in the afternoon. Females with low testosterone concentration had faster risk selection reactions. In contrast, Nadler et al. (2018) and Cueva et al. (2015) used different trading tasks and reported that testosterone administration compared to placebo significantly increased financial risk-taking, specifically in males (see Table 2).
3.3.1.2. Progesterone. Only one study has investigated the impact of endogenous progesterone on decision-making, using data from naturally cycling females (Derntl et al., 2014). The study found a negative correlation between progesterone concentrations during the luteal menstrual cycle phase (compared to the follicular phase and oral contraceptives) and risky decision-making performance, as measured by the HRG task. Furthermore, females with lower progesterone concentrations had faster reactions in high-risk situations but tended to select low-risk conditions more often than females with higher progesterone levels.
3.3.1.3. Estradiol. No studies were found that investigated the relationship between estradiol and decision-making.
3.3.1.4. Hormonal ratio
Several studies have investigated the effects of testosterone/estradiol (Goudriaan et al., 2010; Reavis & Overman, 2001), estradiol/progesterone (Derntl et al., 2014), ratios which are reviewed in the supplementary section (Table S2).
3.3.2. The effect of hormones on social decision-making
Eight articles evaluated performance-based social decision-making. Some studies tested multiple hormones, and the results of each hormone are reported as a separate study in the relevant section. The studies identified information from 273 male and 133 female participants (total: 406, Mpooled age = 22.38, SDpooled age = 3.24; two studies did not report age information) in various sex hormone conditions.
3.3.2.1. Testosterone. Seven studies investigated the association between testosterone and social decision-making paradigms. Using the ultimatum game (UG), it was found that there is a positive correlation between higher endogenous testosterone levels and the tendency to reject unfair offers (in males and females (Mehta & Beer, 2010); male-only-sample (Dreher et al., 2016), and increased rejection of low offers ((male-only sample (Burnham, 2007)). Also, other studies showed that testosterone administration decreased generosity with the money they controlled in males (Zak et al., 2009), increased the probability of accepting unfair offers, and decreased the probability of rejecting them in both males and females compared to placebo (Kopsida et al., 2016). In addition, Dreher et al. (2016) used a modified version of the UG. They observed that males who were administered testosterone were found to be more inclined to reward more offers and punish unreasonable offers, as well as choose greater rewards for the proposer. Wright et al. (2012) found that although testosterone administration interfered with groups' ability to make decisions and learn from each other, it did not affect their ability to make individual choices. In contrast, Zak et al. (2009) found no significant association between testosterone levels after administration and offers made in the Dictator Game (DG). Additionally, Mehta et al., (2017) found that higher testosterone levels were associated with choosing more hawkish decisions, and neither sex nor its interaction was a significant moderator of the effects of testosterone on decision-making.
3.3.2.2. Estradiol. Dreher et al. (2016) conducted a study exclusively on males to examine social decision-making using a modified version of the UG. They revealed that males with high administered estradiol levels were less likely to punish or reward offers, opting for lower amounts. In the placebo group, estradiol levels did not have an effect on the amount of punishment chosen.
3.3.2.3. Progesterone. No study investigated the association between progesterone and social decision-making.
3.3.2.4. Hormonal ratio.
One study reviewed estradiol/progesterone ratio (Strojny et al., 2021), and results are reported in the supplementary section (see Table S3)
3.3.3. Effects of hormone levels in social hierarchy
Seven articles (some studies tested multiple hormones, and the results of each hormone are reported as a separate study in the relevant section), reporting information on 514 male and 394 female participants (total: 908, Mpooled age = 20.78, SDpooled age = 2.39; two studies did not present age information in the text), were included in this section investigating how hormones affect social hierarchy, dominance, and competition.
3.3.3.1. Testosterone. Seven studies examined the link between testosterone and social hierarchy, with five suggesting an association between testosterone levels (pre- and post-administration) and social status, particularly in males (Casto et al., 2020; Edwards et al., 2006; Grant & France, 2001; Vermeer et al., 2020; Zilioli & Watson, 2014).
In detail, in a study by Zilioli and Watson (2014), two days of league matches of a video game were used to test whether testosterone reactivity predicted task performance and the impact of competition. Males in unstable hierarchies had higher testosterone levels than those in stable hierarchies, and testosterone levels decreased the most in those who lost both events on the second day. Grant and France (2001) found a significant correlation between serum testosterone and dominance scores, but no correlation between testosterone and scores in both female groups (hormonal contraceptive users and non-users). Casto et al., (2020) explored the link between endogenous testosterone and competitive endurance using the competitive will task. Both sexes were tested individually or with an opponent (first experiment), or as a team with or without an opponent (second experiment), with social presence and status mediating the relationship. Results showed that increasing testosterone predicted better performance among males, particularly those who won or ranked first while decreasing testosterone predicted worse performance. However, only females who lost in dyads showed the effect. Using the real effort paradigm, Vermeer et al. (2020) demonstrated that testosterone administration boosted motivation to compete for status, but only in those with an unstable, low status. Another study found that post-game testosterone levels in males were significantly correlated with player ratings and social connectivity, while pre-game testosterone levels in females were positively associated with teammate ratings and self-ratings of social connectivity (Edwards et al., 2006).
However, Pranjal H. Mehta and Josephs (2010), using a leadership task, and Casto et al., (2019) using a status hierarchy ranking task found no statistically significant association between endogenous testosterone and performance in both sexes.
3.3.3.2. Progesterone. No studies were found that investigated an association between progesterone and social hierarchy.
3.3.3.3. Estradiol. No studies were found that investigated the relationship between estradiol and decision-making.
3.3.4. Effects of social hierarchy and hormone levels in decision-making
Three studies (please Table 5) (some studies tested multiple hormones, the results of each hormone are reported as a separate study in the relevant section) investigated the association between social hierarchy and different hormones in decision-making with information about 317 males and 54 females (total: 371, M pooled age = 21.2, SD pooled age = 3.09).
3.3.4.1. Testosterone. Three studies examined the link between testosterone and social hierarchy in decision-making. Inoue et al. (2017) found that male senior players often offered less value, and regardless of seniority, individuals with greater testosterone levels offered more money in the UG. Mehta et al. (2015), using the dot estimation competitive task, found that females were more competitive during the social competition after testosterone administration than after the placebo, and high-dominant individuals were more likely to compete again after winning. Smith and Apicella (2017) found no relationship between position type, endogenous testosterone levels, risk, and power in male-only samples.
3.3.4.2. Progesterone. No studies were found that investigated the effect of progesterone and social hierarchy on decision-making.
3.3.4.3. Estradiol. No studies were found that investigated the effect of estradiol and social hierarchy on decision-making.