Birthplace and birthdate of Brazilian Olympic medalists Local de nascimento e data de nascimento de medalhistas olímpicos brasileiros

he aim of the present study was to analyze the association between birthplace and relative age with winning Olympic medals in Brazilian athletes. he sample consisted of 186 Olympic medalist athletes born in Brazil. Data analysis was performed through descriptive (incidence and percentage) and inferential statistics (Chi-Square to binary logistic regression). he association between population contingents and Olympic champions presented a signiicant result (p <0.05); however, no signiicant associations were found with relative age. It was concluded that most Olympic medalists were born in places that present better living conditions, with intermediate MHDI, which appears only in cities with more than 100 thousand habitants, and relative age is not an important criterion in winning medals among athletes investigated.


INTRODUCTION
Success in sport has been related to the birthplace of athletes and the proportion of their communities [1][2][3][4] .As described in specialized literature, it is important to point out that the birthplace provides an overview of the place where children spent their childhood and had contact with sports practice 4,5 .In the Brazilian context, few studies have studied this subject.In the available surveys, among these opportunities and experiences in certain environments, the national socioeconomic aspects related to demographic rates and Human Development Index (HDI) seem to condition the intrinsic and extrinsic motivation for sports practice 5,6 , the type of activities experienced and the quality of sports talent training 5 .herefore, the birthplace may limit or beneit the sporting performance.Studies have indicated that medium-sized cities tend to combine better sporting opportunities, while very small cities do not provide facilities with minimum conditions and in large cities, the diiculties are the costs and distances of athletes' displacement 4 .
Recent studies on the birthplace have suggested that there is no signiicant interaction of this variable with the birthdate 1,3,4 .However, the birthdate that corresponds to the relative age is another element that has shown to be relevant for the formation of athletes 1 .Relative age is deined as the chronological diference of individuals born in the same year 7,8 .In the research ield, several indings indicate a strong relation between relative age and sports involvement and progression, pointing out that athletes born in the irst trimester or semester of the year tend to have more opportunities in sports teams, a fact that seems to be associated with greater psychobiological maturation in relation to those born in the last months of the year [9][10][11][12][13] .
In assessing athletes' birthplaces, the indings of this study may foster further discussions on sport political issues on sports in the country in the case of diferences between populations and the winning of Olympic medals that can be answered according to ease of access, safety and improvement of sports facilities.hese conditions are often disregarded, although they may create a considerable disadvantage for practitioners who wish to achieve professional sport performance 13 .In addition, regarding relative age, data have demonstrated the need to follow the indications of Cobley, Baker 14 on reducing or eradicating sports inequalities, since annual age groups and the efects of relative age seem to limit the immediate and long-term participation of younger or late matured athletes.his occurs either for the positive development of young people through sports, or for the development of elite athletes.
hus, despite the possible inluence of birthplace and birthdate on athlete performance, few studies have investigated this issue in the Olympic context 15,16 , in the various individual and team modalities and in Brazilian athletes of both sexes.hus, the present study aimed to analyze the association between relative age (birth quartiles) and the birthplace (MHDI, region, population size) of Brazilian athletes with the irst medal at the Olympic Games from 2000 to 2016.

Sample
he study sample consists of all the athletes born in the national territory, who won their irst medal (gold, silver or bronze) in an edition of the Olympic Games from Sydney (2000); Athens (2004)

Procedures for data collection
Data referring to athletes -year of birth, hometown and year of the irst Olympic medal -were collected by the Federal Government's Oicial Portal on the 2016 Olympic and Paralympic Games 17 , considering only the irst medal won by athletes.
Relative age was obtained according to the birthdate of Olympic athletes and divided in quarters: 1 st quartile (January, February, March); 2 nd quartile (April, May, June); 3 rd quartile (July, August, September); 4 th quartile (October, November, December).
he values in relation to the Municipal Human Development Index (MHDI) and the population contingent of cities were extracted from the database of the United Nations Development Program 18 .To determine the MHDI, the Atlas of Human Development classiication was used in the Brazilian metropolitan regions.MHDI adjusts the global HDI for the reality of Brazilian municipalities and metropolitan regions 18 .Data were taken from the 1991 census because this census was the irst to be released by UNDP.Since the HDI varies from 0 to 1 and has the following classiication: Very low (0.00-0.49);Low (0.50-0.59);Intermediate (0.60-0.69);High (0.70-0.79);Very High (0.80-1) 18 .

Statistical analysis
Descriptive analysis (frequency, percentage, mean and standard deviation) in each category of athletes' variables was initially performed.he chi-square test was used to compare diferences among independent variables and to verify which categories of lower prevalence among dependent variables.hen, crude analysis of binary logistic regression was performed to estimate the odds ratio (OR) and the prevalence of being an Olympic medalist.he variables with p-value less than 0.20 were inserted together for analysis adjusted with the BACKWARD-WALD method.he power of the Phi () efect proposed by Cohen 19 (small: 0.1, medium: 0.3, large 0.5) was also calculated.In addition, the Hosmer-Lemeshow test (HL (df)) was used to verify the quality of the model for the exposed data.All analyses were performed in the SPSS version 18.0 program, adopting signiicance value of 5%.

RESULTS
Table 1 presents the descriptive data of variables involving Brazilian Olympic medalists from 2000 to 2016, such as sex, region, MHDI, birth quartile, sports modality and population.Table 2 shows the associations among variables, sex, region, MHDI, birth quartile, modality and population and whether or not the athlete was an Olympic champion.
here is a signiicant association between population and being or not an Olympic champion (p <0.05).
Table 3 shows the crude logistic regression analysis for the condition of the athlete winning a gold medal.A signiicant association between cities with population of 100,000-499,999 and OR of 176% of the athlete being Olympic champion was observed.In the analysis adjusted by categorical variables sex and population, it was observed that there is a signiicant OR of 182%, considering people of the same sex, and of cities with population of 100,000-499,999.he model presented good it to HL data (6) χ2 = 1.43, p = 0.96.Table 4 shows the crude logistic regression analysis for the condition of the athlete winning the silver medal and it was observed that the OR of an athlete to win the silver medal reduces in 64% when he/she comes from a city with population of 100,000-499,999.In the analysis adjusted by categorical variables sex and MHDI, it was observed that considering people of the same sex and low MHDI, the chance of winning the silver medal is reduced in 61%. he model presented good it to HL data (3) χ2 = 1.68 p = 0.64.Table 5 shows the crude logistic regression analysis for the condition of the athlete winning the bronze medal; however, it was not possible to observe any signiicant OR.In the analysis adjusted by categorical variable birth quartile and sex, no signiicant OR was observed.he model presented good it to HL data (7) χ2 = 7.19, p = 0.41.

DISCUSSION
he present study aimed to analyze the relationship between birthdate (birth quartile) and birthplace (MHDI, region, population size) of Brazilian athletes with winning the irst medal in the Olympic Games from 2000 to 2016.It could be observed that the majority of Olympic medalists have the following characteristics: I) males; II) from the southeastern region; III) cities with intermediate MHDI; IV) born in the second quartile of the year; V) cities ≥2,500,000 inhabitants; VI) team sports.According to the odds ratio, athletes from cities with population of 100,000-499,999 have greater chance of winning a gold medal.Additionally, same-sex athletes have a reduced chance of winning the silver medal. he results about the population size of the hometown of participants of the present study difer from some countries 2 and are in agreement with others 1 , and in Brazil, medium and large cities seem to ofer a greater development of athletes in sports.here was a signiicant odds ratio of 176% for same-sex athletes from cities with population of 100,000-499,999 and the possibility of the athlete being Olympic champion, a fact that difers from the population range presented by Balish, Rainham 2 .he authors 2 investigated the association between the proportion of cities and individual and team school physical education and in public policies 26 , which may inluence the participatory culture and the development of women in sports, with the possible achievement of an Olympic medal.However, categorical variable sex was included as a control, since the authors of the present study consider this variable to be important for controlling the others variables analyzed.In addition, it can be seen in the present data that the variable elucidated better some allusions about the Olympic medal by the test of regression models.
Greater depth in data was not possible due to the study limitations associated to the small number of census to identify the real MHDI of cities and that can accurately correspond to the years of birth of athletes investigated.It was also considered the probability of the migration of athletes from their hometown.However, athletes' city changes are likely to be essentially the same and opposite, and the number of athletes who migrated from larger cities to smaller cities is probably equivalent to the number of athletes born in smaller cities and moved to larger cities 4,5 .
Despite the limitations, the data found may elucidate further discussions about Olympic medalist athletes.Further studies should investigate the sports in an individual way, with greater number of participants considering women and men, because they present own characteristics.It is also suggested to consider social aspects, with studies focused on the quality of relationships among sports characters, opportunities for practice and the training context of elite athletes, which are indicated for a better understanding of factors such as MHDI and the process of preparation of athletes in the scenarios of sport development in the long term.

CONCLUSION
he present study pointed out that athletes born in cities with population of 100,000-499,999 inhabitants have higher odds ratio of becoming Olympic champions, as well as in the regression adjusted with variable male sex in relation to female sex and Olympic silver medalists (p ≤0.05).In relation to categorical variable birth quartile referring to birthdate, no signiicant results were found.It was concluded that Olympic medalist athletes are mostly men, possibly because of the sports culture in the country of promoting and disseminating primarily men's sports.In addition, athletes are from places that present better living conditions, with intermediate MHDI, which appears only in cities with more than 100 thousand inhabitants.

Table 1 .
Absolute and relative frequency of athletes according to gender, birthplace (region, MHDI and population), birthdate (birth quartile) and sports modality of Brazilian Olympic medalists from 2000 to 2016.

Table 2 .
Associations among variables sex, birthplace (region, MHDI and population), birthdate (birth quartile), sports modality and whether or not the athlete was an Olympic champion.

Table 3 .
Crude and adjusted logistic regression for association between Olympic gold medal and variables sex, birthplace (population, MHDI and region) and birthdate (birth quartile) OR: Odds Ratio.CI: Confidence Interval.* ≤0.05.

Table 4 .
Crude and adjusted logistic regression for association between Olympic silver medal and variables sex, birthplace (population, MHDI and region) and birthdate (birth quartile).

Table 5 .
Crude and adjusted logistic regression for association between Olympic bronze medal and variables sex, birthplace (population, MHDI and region) and birthdate (birth quartile)