In the traditionally male‐dominated sport of horseracing, it has been commonly assumed that female jockeys cannot compete equally against their male counterparts. Physical strength, body shape and tradition were reported by Roberts & MacLean as perceived reasons of restricted opportunities for female jockeys . Similarly, another study outlined the intimidating nature of the weighing room, which may represent a substantial barrier for female participation in the sport of racing . Thus, in 2018, the French racing jurisdiction France Galop, implemented a 2kg weight allowance for female jockeys, resulting in many more racehorses being ridden by female jockeys . Nevertheless, a minority of female jockeys remains, and is surprising given the high media and public interest in horse racing as a spectator sport, and the high levels of participation of women in other amateur and elite equestrian activities [28, 29]. No study has directly, and objectively, compared the performance of female versus male jockeys in terms of effects on important characteristics of the racehorse (e.g. speed, stride length, cardiovascular response). Here, in an observational study reporting on aspects of racehorse speed, stride and cardiovascular responses we show no overt effects of rider sex on any aspect of racehorse speed or stride length, during various training intensities. However, when the converse training intensity was conducted on the opposite training surface (e.g. slow canters on turf or fast gallops on sand, both of which were unusual) then small, but significant physiological effects – reduced or greater rates of recovery of racehorse heart rate, respectively – were observed. We also show that female jockeys have very similar racing success, compared to male jockeys. Thus, we can conclude with confidence that in our study we show no overt effects of rider sex on racehorse physiology in training, and on performance when racing.
Descriptive data of the cohort; number, type and intensity of training sessions by sex of work-rider
In this study, there was a clear difference in the proportion of training sessions completed by male or female work-riders. Despite a lower number of female (n=37) compared to male riders (n=64), female riders completed a higher number of training sessions (52.4%) than male riders (47.6%). This contrasts with a previous study, and our current data, where in actual races, there are fewer female riders getting proportionately fewer race rides . Thus, despite female riders completing many more work-rides in training, fewer progress to the professional ranks, securing race-rides. This observation questions equal racing opportunities between female and male jockeys. Indeed, even in training rides then some sex-bias is apparent: as the intensity of the training session increases, far more male riders are used, with female work-riders completing most slower, canter sessions. This variation may be explained by an unconscious bias made by trainers; the assumption that male riders with greater strength are more suited to higher work intensities. However, our data suggest that there is no evidence to support this contention; for all training intensities as directed by the trainer, we found no difference in the fastest 200m between racehorses ridden by a male or female work-rider (across 3,568 training sessions).
Could the faster training sessions be preferentially ridden by registered professional jockeys of which many more are men? Yes. In our dataset, a far greater proportion of ‘gallop’ sessions were conducted by past or current race registered male professional jockeys (male, 31.8% vs. female, 4.5%). This could reflect trainers choosing to use riders with greater race experience to recapitulate a race environment in training. In Australia, the majority of high-intensity workouts (i.e. gallops) were conducted on turf, while canters were completed on sand. This is consistent with the findings of Morrice-West et al. , who surveyed Australian trainers on the use of track surfaces for training. Sand or synthetic ‘all-weather’ surfaces were commonly used for slow workouts, while gallop work was conducted on turf. It is likely, therefore, that racehorses anticipate certain training intensities according to the track surface. Supporting this contention, in our study, heart rate at trot prior to galloping on grass was 124 ± 26 beats/min (mean ± S.D.) versus 114 ± 26 beats/min prior to a gallop on sand, where most canters occur. Racehorses no doubt anticipate the type of training session they are about to conduct.
Racehorse speed, stride length and training surfaces by sex of work-rider
To increase speed from canter to gallop, horses increase frequency of their stride to an extent, but predominantly speed increases due to increased stride length, as reported here and previously by others . Our results showed that near maximal speeds (55-60kph) were attained at a training intensity of ‘soft-gallop’ and only minimally increased thereafter. We acknowledge that speed can be affected by the track condition (i.e. a harder turf or relatively softer sand ), which was not measured in this study, but the size of our dataset would limit these relatively small effects. In our study, the mean time taken for a horse to cover a furlong (200m) did not differ according to the sex of the work rider. As expected, the longer the training session was (for example, from 1500m to 3000m), then the average time taken to cover each furlong gradually increased (see Figure 2a,b), suggesting a gradual slowing. Interestingly, riding style can influence racehorse speed; a crouched posture reduces aerodynamic drag and can improve racing times by up to 5-7% . Whilst this was not measured in the current study, we assume that male and female work-riders, usually with worn body protectors, do not adopt different postures in training.
Stride length increased with training intensity but was not affected by the work rider’s sex. Information on the riders’ weights was not obtained for the study. Nevertheless, we appreciate that an increment in the rider’s weight has the potential to reduce the stride length of horses and alter performance . The rider’s experience, registered race professional versus non-registered race professional was accounted for and no marked effect on speed or stride was highlighted. These findings run counter to those of Kapaun et al , who found that horses ridden by the professional rider had the highest trotting speed and the longest stride length compared to horses mounted by a hobby-rider. The effect of accumulated training sessions on heart rates of horses on both turf and sand, revealed greater fitness levels (i.e. lower heart rate for the same intensity) in horses having completed 11-20+ workouts. Age of the horse was included as a cofounder, which couldn’t have influenced our results for this observation. Indeed, heart rates during submaximal exercise provide a means of monitoring the adaptation of the cardiovascular system to chronic exercise, commonly referred as the ‘fitness” status of an athlete . Foreman  drew similar conclusions in Thoroughbreds undergoing exercise testing at different intensities; heart rate was lower following a conventional training program.
Racehorse cardiovascular responses to incremental training intensity by sex of work-rider
Racehorse training implies the cooperative effort of two distinctive individuals: horses and humans. Indeed, in any equestrian discipline including racing, positive interaction between horse and rider is paramount to cope with the emotional and physical challenges of the demands of training. However, no study has reported any equine cardiovascular responses as a result of an interaction with the rider. In this study, no effect of sex of the rider on racehorse heart rate, nor the peak of heart rate, during training at differing intensities was noted. That is, when male and female work-riders are instructed to exercise the racehorses at tempo (canters) or race-pace (gallops) then both do just that. There is no measurable difference between rider sex on racehorse speed or on racehorse cardiovascular response. The outcome is indistinguishable whether ridden by a male or female work-rider. However, it has been shown that rider ‘emotion/nerves’ can be faithfully transmitted to the horse. For example, when exposed to a novel stimulus  or mounted for the first time by a novel rider , horses respond with an increase in heart rate. When the rider, but not horse, knows in advance that a novel stimulus/object, that is known to cause the horse to startle, is about to be encountered, the increment in rider heart rate in anticipation is matched by an increment in the horses heart rate . Such acute fight-or-flight responses, most likely mediated by catecholamines and the stress hormone, cortisol, will facilitate improved training responses via energy mobilisation  and activated behavioural responses .
Racehorse recovery of heart rate after incremental training intensity by sex of work-rider
In our study, when assessing each individual horse in a repeated-measures analysis from peak heart rate through cardiovascular recovery to 15, 30 minutes to the last recorded heart rate that was back to, or below, baseline then a small signal suggestive of some effect of rider sex was noted. After sessions of slow canter on turf – the surface where nearly all gallops occurred – then heart rate area-under-the-curve or overall recovery was higher with a male rider onboard. After sessions of hard gallop on sand – the surface where nearly all canters occurred – then heart rate area-under-the-curve was lower with a male rider onboard. Could male work-riders, more so than female, anticipate the ‘expected’ training-intensity on a given surface and their higher or lower heart rate be transmitted faithfully to the horse? Without simultaneous measurement of work-rider heart rate, we cannot answer this question but it is intriguing and has been noted in other circumstances . Perhaps male work-riders ‘push’ the horses more (i.e. reaching the horses peak HR sooner) on turf, despite instructions to exercise horses at low speed. However, we observed no difference in average speed between rider sex at differing training intensities on each surface. It is also possible that for faster sessions on sand (it is unlikely to have happened for slow canters on turf) that changes in work rider for the main galloping session may have provoked changes in the horses’ heart rate. Nevertheless, this is only likely to have occurred for faster gallop sessions on sand, where we observed relatively faster recovery of heart rate with male riders aboard.
Using data publicly available online and expanding our analyses to two countries in order to increase the sample size, we questioned whether there was any overt difference in strike-rate during actual flat races between male and female jockeys, when accounting for the large difference in number of jockeys registered and number of races entered. Our results confirmed the existing trend of sex disparity between jockeys that is currently observed within the racing industry. When accounting for a larger proportion of male jockeys, a small but significantly greater winning strike-rate (~1%) was found in male compared to female jockeys in Australia, but not in the UK. This may be partly explained by the inability to focus solely on flat racing results, as no distinction could be made online for the available data in Victoria state, Australia. Regardless, the difference disappeared in Australia when considering if the horse/jockey achieved a first, second or third placing. Therefore, on balance, according to objective data then male riders do not win many more races than female jockeys. The ~1% difference is likely attributable to male riders getting more rides on better horses more likely to win in the first place. Furthermore, the quality of the race could not be taken into account and we were not able to get all the starting odds for the 52,000+ races we had data for and thus cannot adjust for this factor.
The study has several limitations, which should be acknowledged. First, there is a risk of selection bias due to the method of convenience sampling. However, the large sample size should minimise any slight effects of possible selection bias. Secondly, the study design is retrospective and descriptive. As a result, there are missing data for some riders and horses, along with other variables that have not been considered such as rider’s weight. Nevertheless, using mixed-effect models to analyse the data with due incorporation of possible confounders such as racetrack and with individual racehorse and work-rider included as random effects, should mean that any missing data are distributed at random, minimising any inherent bias. Equally, whilst we have not accounted for differences in the ‘quality’ of each racehorse in our analyses, there are potentially other variables associated with the horses that we may not have been able to account for, that may influence an individual racehorse’s response to training such as the fitness level of each horse and differences in pedigree. Although, again, these effects are estimated to be minimal in our opinion. The number of training sessions available for each rider was not balanced; many riders had only participated in <5 training sessions. Ideally, future prospective studies of a similar nature should source more observations per rider, if possible. Although it should be noted that many of the riders in our dataset, both male and female, had >100 sessions. Additionally, future studies could include the horse’s body weight, plasma lactate and cortisol and aspects of the rider’s physical status which could provide more insight into the interaction between riders and horses on performance and physiology in response to training over time. In addition, we acknowledge that the choice of combination for exercise intensity and track surface may have been opted by the trainer to alleviate any mitigating risks for the horse i.e. choice of ‘soft’ surface for a horse presenting poor hoof quality. Again, however, these instances would be expected to be relatively rare and the size of our dataset would mask such rare bias.
Perspectives and Significance
This is the first study to explore the effect of the sex of rider on racehorse performance physiology during training. Using measurements from objective fitness tracking systems, no marked sex differentials between work riders were observed on racehorse physiological responses (cardiovascular and locomotory parameters) or speed during training. Further research could expand on performance profiling in both male and female riders across different racing nations and types of racing (e.g. National Hunt or jump-racing). Our analyses also explored existing disparity between sex for racing opportunities. The study’s implications may breakdown any discriminatory barriers, improving the access of female jockeys to quality mounts and racing events. A clear opening emerges for the growing inclusion of female riders in the racing profession, as their underrepresentation remains tangible at the elite sporting level. This data can now reliably inform trainers and owners on the equal athletic potential of female riders for racing endeavours. New perspectives arise to shift traditional male-dominated perception, attitude and behaviour of the greater public. By using a relatively small dataset, we were able to show the absence of an effect of rider’s sex on racehorse’s response to training, for example, with increased workload, speed appears to be relatively unchanged for both sexes.