Strength and competitiveness are more strongly predictive of retrospective sport participation than 2D:4D in university-aged women

Lifetime sport and physical activity participation provide many physiological and sociological benefits; consequently, understanding this behaviour is important to the health and well-being of women and girls. However, there is a discrepancy between leisure time physical activity (LTPA) and sport participation rates of males and females worldwide. As a population, North Americans do not meet the guidelines of daily recommended activity in both sexes, but males are more likely to participate in LTPA, including sport participation, than females (Stephens et al. 1985). Moreover, while there is a general decline in physical activity and sport participation starting in the early teens, the largest drop in participation rates occur between the mid to late teens for females. This drop in participation could be due to many factors including, but not limited to, intrapersonal (i.e., stress (Crane and Temple 2015)), economical (i.e., low familial social economic status (Walters et al. 2009)), or environmental (i.e., proximity of facilities (White and McTeer 2012)) factors. Generally, the implementation of programs to overcome barriers are promising, but not completely effective. Youth sport participation in Canada (both male and female) has been fairly consistent over the past decade with approximately 75% (81% of boys versus 70% girls) of youth aged 5–17 years participating in sport at some point (Canadian Heritage 2013), a sex difference that is even greater in other parts of the world.

A primary difference in development between males and females is the exposure to prenatal and pubertal androgens. These androgens may impact various traits beneficial to sport performance such as: increased skeletal muscle mass (Yu et al. 2014), a major determinant of strength and power; regulation of erythropoiesis (Beggs et al. 2014), thereby increasing the oxygen carrying capacity of blood; increased cardiac size (Gray 1918) resulting in a larger stroke volume and cardiac output; and fat loss (Usui et al. 2014), which is highly beneficial in most sport settings. Further, the second to fourth digit (2D:4D) ratio, a proxy for prenatal androgen exposure (Manning et al. 2002) and circulating testosterone, has been correlated with competitiveness and aggressiveness (Book et al. 2001; Coyne et al. 2007), which may predispose individuals to success in sporting contexts. This success in sport may subsequently lead to an increase in physical practice, extrinsic motivation (awards, praise, etc. (Vallerand and Losier 1999)), intrinsic motivation (fun, winning, etc. (Vallerand et al. 1987)), and (or) a belief in physical abilities (self-efficacy (McAuley et al. 1991)), that may increase success in sport in a positive feedback cycle.

There are logistical, time, and financial restrictions to performing large-scale studies including the hormonal analyses of androgen status. As an alternative to hormonal analyses, the 2D:4D ratio has been examined as an indirect marker of prenatal androgen exposure within a variety of studies testing its relationship with sport performance (e.g., a cycle sprint test (Crewther et al. 2015)), sporting ability (i.e., highest level achieved of all sports played from 11 years old (Paul et al. 2006)) and sport-related psychological variables (i.e., mental toughness, optimism, goal orientations, aggression, and coping style (Golby and Meggs 2011)). Although studies have indicated that prenatal testosterone can influence sport interest (Giuliano et al. 2000; Manning et al. 2002) as well as participation in competitive sport (Hönekopp and Schuster 2010), to our knowledge, no studies have directly addressed androgens as a potential influence on recreational and competitive sport participation.

Exogenous anabolic androgen administration is one of the biggest doping concerns for male and female athletic competitions because of the effectiveness of these hormones in enhancing sport performance (Sjöqvist et al. 2008); however, the extent to which androgens impact sport participation remain unclear. Combined with the potential sport-related benefits of androgens on performance, we predicted that 2D:4D would correlate with female sport participation through adolescence (9–18 years) and young adulthood. It was hypothesized that early exposure to androgens, as inferred from a lower 2D:4D, would predispose girls toward sport participation.

Cross-sectional analyses of indirect prenatal androgen concentrations using 2D:4D were obtained from a sample of 92 females (18–30 years) recruited from the University of Windsor, Windsor, Ontario, and the Windsor-Essex community. Participant demographic, anthropometric, sport-specific behaviour tendencies, and retrospective sport participation information were collected. All procedures were cleared by the Research Ethics Board at the University of Windsor and consent was obtained from all participants prior to undertaking any of the methods described below.

Anthropometric and demographic data of the participants are outlined in Table 1. All participants had, or were in the completion of, a postsecondary education. Only one participant indicated that she had not participated in sport in any year in the retrospective timeline. Descriptive statistics for the main variables are displayed in Table 2.

Our purpose was to determine if 2D:4D was associated with female sport participation throughout adolescence (9–18 years) and to determine whether this indirect measure of prenatal androgen exposure could predict female sport participation in those ages. We predicted that early exposure to androgens, as inferred from 2D:4D, would predispose girls toward sport participation. In fact, this has been shown to some extent in males and females, where lower ratios appear to relate to fitness and level of sport achieved (Sjöqvist et al. 2008; Ranson et al. 2015). 2D:4D was not significantly correlated with TSP_9to18 alone (r = −0.065, p = 0.538, CI = −0.277 to 0.146) or when the potential for covariates and confounders were added in regression analyses. However, in examining the 95% CI for this relationship, it is important to note that the majority of the CI were within the negative range as predicated. Nonetheless, whether positive or negative, the CIs suggest moderate to weak associations at best in either direction. From this we can infer that 2D:4D might not play a large role in predisposing girls towards sport participation as predicted. However, given that all but one of our participants participated in sport at some point throughout our age range, we do not have an adequate comparison of 2D:4D against individuals who did not participate in sport. While previous research has found an association between the 2D:4D of elite female fencers (n = 87) and current and highest past rankings (Bescos et al. 2009), the athletes in that study had all achieved world-class ranking in their past, and this association was true only when ethnicity and the rankings were taken into account. Similarly, Paul et al. (2006) analyzed the 2D:4D of 607 females (53.8 ± 8.5 years) who were asked to rate their highest level achieved (on a scale of 1–5, ranging from social participation only to national level) for a total of 12 sports. When adjusted for age, the highest level achieved in any sport was significantly, negatively associated with average 2D:4D (b = −4.93, p = 0.01) (Paul et al. 2006). Participants in the current study were younger (22.32 ± 2.68 years) than those in the study by Paul et al. (2006). Therefore, some of the participants in the current study may not have reached their full athletic potential (i.e., highest level) within the timeframe analyzed (e.g., the average age of female Olympians is in the mid to late 20s (Patel 2012)). Given the above findings, it is possible that the relationship between 2D:4D and sport success might manifest in the differentiation of elite level athletes from recreational athletes and (or) those who do not participate in sport. Nonetheless, one of the central aims of the current study was to determine the relationship between prenatal androgen exposure and sport participation, not simply highest level.

Moreover, Paul et al. (2006) suggested that the performance of childhood athletes may not correlate well with best adult performance due to variability in development. Therefore, it may be more valid to look at adolescent sport participation as a sum, because of ranges in age of development (Paul et al. 2006). In the current study, TSP was divided into TSP_early (9–13 years) and TSP_late (14–18 years). Age of menarche was included as a covariate in regression analyses to determine whether physiological influences may have had a greater impact at later stages or for girls who began development later. In all cases, 2D:4D was not significantly related to sport participation. Traits that would otherwise correlate with sport participation, such as youth fitness, have been shown to correlate with 2D:4D in young boys and girls (Hönekopp et al. 2006); however, that and recent evidence suggests that this effect appears much stronger in boys (Ranson et al. 2015).

One of the strongest correlations found in the current study was the association of TSP_9to18, TSP_early, and TSP_late with a marker of upper body strength (i.e., HG_max). Not only was HG_max significantly correlated on its own but, according to structure coefficients, it also contributed 35% and 43.3% to the predicted sport participation score in the regression models for TSP_9to18 and TSP_late, respectively. However, it is likely that by participating in sport an athlete is exposed to environments in which strength will increase. Conversely, a person who is stronger initially might do better in a sport, thereby increasing their sport success and subsequent participation. It has been proposed that androgens help to develop grip strength in men and women, although this relation may be subdued in females (Isen et al. 2014). Interestingly, HG_max was removed from the final model in the prediction for TSP_early. This may suggest that strength is a result of sport participation given that TSP_late would have been closer in time to participation in this study. Nonetheless, the relationship between HG_max and sport participation was clear even without a categorization of sport type, which confirms that strength is an important component of all sport participation and achievement (Fink et al. 2006; Ribeiro et al. 2016).

A strong relationship between competitiveness in sport situations (SOQ) and TSP_9to18, TSP_early, and TSP_late was found (r = 0.473, p = 0.000, r = 0.362, p = 0.000, and r = 0.545, p = 0.000, respectively), suggesting that the higher a female’s measure of competitiveness, the higher their sport participation score. There are three subscales (i.e., win orientation, goal orientation, and competitiveness) associated with the SOQ and the correlations of the subscales with TSP_9to18 were as follows: SOQ_win (r = 0.324, p = 0.002), SOQ_goal (r = 0.331, p = 0.001), and SOQ_comp (r = 0.544, p = 0.000). The original purpose of the SOQ in the current study was to aid in predicting sport participation (i.e., did 2D:4D correlate with competitiveness and consequently, could this explain sport participation?). Much like HG_max, whether competitive girls were drawn toward sport or sport participation, increased competitiveness in sport is still unclear; however, it is important to note that all participants (excluding one) had or currently were participating in sport and exhibited SOQ scores similar to those found in collegiate athletes (Finkenberg et al. 1998).

Taken together, we did not find evidence that prenatal androgen exposure is significantly correlated with female sport participation at young ages. While this is not surprising given the complexities of behaviour that include, but are not limited to, familial and peer influences, socioeconomic status, positive and negative experiences, and chance, the hormonal advantages to sport participation and achievement associated with androgens remain intriguing.

We would like to thank the individuals who participated in the study and the undergraduate students at the University of Windsor who volunteered their time to help with data collection and analyses. Lastly, we would like to thank the Centre for Human Performance and Health and Physical Activity and Cardiovascular Research Lab in the Faculty of Human Kinetics in which data collection took place.