Exploring the Relationship between Adult Women’s Participation in Recreational Soccer and Self-Efficacy in Rural Kenya

Background Increases in non-communicable diseases in Kenya are attributed in part to a decrease in physical activity. Although, sports programs for girls and young adult women promote health, reduce risk behaviors and build self-efficacy, similar programs for adult women in Kenya are largely unavailable. The Nikumbuke Soccer League, a recreational soccer program for adult women in southeastern Kenya, is a notable exception. Methods Data for this analysis were taken from a 2018–2019 study of women’s experiences in an adult women’s soccer league coordinated by the Nikumbuke Project in ten communities in southeastern rural Kenya. 702 project members, 229 of whom play soccer, participated in semi-structured interviews containing questions on demographics, self-efficacy, and health. Frequencies were run for all variables with separate bivariate analyses examining socio-demographic attributes and two outcome variables of interest, general self-efficacy (GSE) and participation in soccer. Linear and logistic regressions were run to explore the relationship between women’s self-efficacy and participation in soccer, controlling for socio-demographic covariates. controlling point increase on the GSE scale. Results regression point GSE scale is suggesting a bi-directional relationship between

3 Background Physical activity is a major determinant of health that is associated with reductions in premature mortality and the global burden of non-communicable disease (NCD). 1 Despite the known benefits of physical activity, physical inactivity continues to account for an estimated 6-10% of the global burden of disease attributed to coronary heart disease, type 2 diabetes, and breast and colon cancer, as well as 9% of all premature mortality. 2 Physical inactivity increases with age, accounting for an estimated 3.2 million deaths each year. 3 Strategies to encourage participation in physical activity, particularly leisure activity targeting adults, are an integral part of the NCD-risk reduction programs in many countries. 4 Physical activity behaviours occur within an ecosystem of individual, household, community, and societal attributes that shape individual behaviors. 5 Research suggests that a number of attributes are consistent predictors of initiation and adherence to physical activity by adult women, including age, education, marital status, self-efficacy, and health status. 6-10 A substantial body of research supports the link between self-efficacy, defined by Bandura as "people's beliefs about their capabilities to exercise control over their own level of functioning and over events that affect their lives," 11 [p. 118] and certain health behaviors. 12-16 High levels of perceived self-efficacy have been positively associated with smoking cessation, recovery from posttraumatic stress, HIV preventive behaviours, and treatment adherence for a variety of communicable and non-communicable diseases. In Kenya, a number of recreational programs targeting girls and young adult women have been shown to promote health, lower risk behaviours, and build selfefficacy. 17,18 Studies investigating self-efficacy in exercise groups show a consistent positive relationship between efficacy and exercise participation. 19 Perceived self-efficacy has been shown to be a significant predictor of exercise adherence and compliance, as well as a consequence of physical activity participation. 9 Self-efficacy beliefs about one's ability to carry out a task may influence the adoption of physical activity while outcome expectations due to those beliefs may influence adherence. 20 Women in Sub-Saharan Africa face a number of socio-cultural, environmental, and behavioural obstacles in terms of access and engagement in leisure-time physical activity. In countries in Sub-Saharan Africa, physical inactivity has been linked to overweight/obesity, hypertension, and diabetes among women. 21-25 A 2011 study of physical activity in 22 African countries found that approximately 76% of females and 83.8% of males met the World Health Organization recommendations for at least 150 minutes of moderate-intensity physical activity per week.
Importantly, the study found that the majority of physical activity stemmed from work or transportation and physical activity during leisure was rare. 26 Reasons for high rates of physical inactivity among women in this setting include a lack of material resources, cultural norms discouraging the wearing of tight-fitting clothes when participating in sports, and the gendered nature of domestic and household chores that afford women limited time for leisure activities of any kind. 21,27 Women in Nigeria were reported to experience structural, interpersonal and intrapersonal constraints to participating in recreational sports activities, with lack of time and norms reinforcing women's responsibilities as mothers and homemakers presenting the biggest obstacles. 28 While opportunities exist in the region for adults to participate in sports like soccer, rugby, and boxing, these are considered traditionally 'male' activities; sports remains a gendered domain that preclude adult women from participating in organized leisure sports. 22 Several studies, for the most part conducted in Nigeria or South Africa, have variously identified self-efficacy, social support, marital status, high parity, and norms regarding women's roles as significant predictors of adult women's participation in physical activities. 22,29−31 The burden of non-communicable diseases in Kenya is on the rise, with NCDs accounting for 27% of all deaths and 50% of all hospital admissions in 2015. 32 Among the risk factors to which this increase in NCDs is attributed is the corresponding decrease in physical activity in the population; reversing this trend has been identified as a central objective in the country's national strategy for combatting NCDs. 33

Research examining the relationship between self-efficacy and physical inactivity in Sub-Saharan
Africa has largely focused on patient populations with health conditions such as hypertension and type-2 diabetes. 30,31,34 A few studies link higher levels of self-efficacy to physical activity in university students in low-, middle-, and high-income countries 35 and healthy women ages 18-64 years study in Nigeria. 29 Despite a growing burden of non-communicable diseases in East Africa and the adoption of healthy lifestyles as a risk-reduction strategy in countries like Kenya, almost no research has been conducted to date on adult women's engagement in organized recreational sports nor on the psycho-social attributes that might encourage their participation in and adherence to such activities. This study contributes to filling that gap by examining the relationship between women's self-efficacy and their participation in a recreational soccer league in rural southeastern Kenya. We hypothesized that women's participation in the soccer league would be positively associated with self-efficacy that this relationship would be bi-directional.

Data and sample
Data for this analysis were taken from a 2018-2019 study focused on women's experiences in a small, but established, adult women's soccer league coordinated by the Nikumbuke Project in Lunga Lunga, Kenya on the border with Tanzania. 36 The Nikumbuke Project is a local, internationally supported non-governmental organization that offers basic health and literacy programs for its female members, provides subsidies for their daughters' school fees, and runs a vocational training program in tailoring for young women who did not complete secondary school. In 2014, at the request of the membership, the project launched a modest four-team adult women's soccer league with assistance from a US-based non-profit established to raise funds for the program and provide guidance in the sport. At the time of this study, the Nikumbuke Project included 702 members, 229 of whom played in its recreational soccer program, the Nikumbuke Soccer League. The teams in the League represent the ten communities in which the Nikumbuke Project members resided at the time of the study. These communities included Lunga Lunga, a market town and the seat of the Kwale County subconstituency, seven settled rural villages -Godo, Mgombezi, Lunga Lunga, Kidomaya, Juakali, Perani, Mpakani and Jirani, a pastoralist community of Maasai, and Umoja, a mixed pastoralist/settled village.
Teams, which reflect a diversity of religions and distinct tribal groups, practice weekly, compete monthly in inter-village competitions during the non-rainy season and participate in a League-wide "World Cup" held each May/June.
All 702 members of the Nikumbuke Project were invited to take part in the study, which included semi-structured interviews containing questions relating to demographics, self-efficacy, health, gender-related norms, and participation in the Nikumbuke Soccer League. In order to participate in the study women had to be at least 18 years old and provide verbal informed consent. Interviews were conducted by a team of 18 female members of the Nikumbuke Project who were elected by their groups to assist in the study and who were trained in research ethics, survey methods, and survey techniques. Trainees received a small stipend for successfully completing the program and for each survey instrument they administered. Women who agreed to participate in the interviews were also given a small stipend in appreciation for their time. Interviews lasted between 35-65 minutes and took place in a private setting either in the Nikumbuke meeting hall in each community or in respondents' homes.

Measures
The primary outcome variables of interest in this analysis were self-efficacy, measured using the General Self-Efficacy (GSE) scale. 37,38 and participation in the recreational soccer league. The 10item GSE scale measures perceived coping ability across a wide range of demanding scenarios such as 'I can always manage to solve difficult problems if I try hard enough' and 'I can usually handle whatever comes my way.' Responses to these items range from '1 = not true at all' to '4 = exactly true;' thus, summed scores on the GSE range from 10-40. In this analysis, the scale had an internal consistency of alpha = .82, well within typical ranges reported elsewhere 39 . For the purposes of bivariate analysis, a dichotomous variable representing self-efficacy was also created from the summed scores on the GSE to reflect 'at or below the GSE mean score' and 'above the mean GSE score'. Socio-demographic variables included age, education, marital status, number of children, work outside the home, self-identification as head of household, and monthly household income.

Analysis Strategy
Eighteen of the 702 members in the Nikumbuke Project were excluded from the analytic sample as they participated in the study as field staff responsible for administration of the interviews andtherefore, had a different knowledge of the instrument and study than other members of the Nikumbuke Project at the time of survey implementation. The resultant analytic sample included 684 women.
Data analyses were conducted using StataCorp statistical software, Release 16. 40 Frequencies were run for all socio-demographic variables as well as two outcome variables of interest, i.e. GSE and participation in the recreational soccer program. An item non-response analysis of all variables in this study indicated minimal missing data (3% or less on all variables except age, which had 9.5% missing). Single-imputation was used to approximate responses for missing data. 41 Chi-square and Fischer's exact tests were run to explore bivariate relationships between high self-efficacy (measured as GSE scores above the sample mean), participation in recreational soccer, and the demographic covariates in this study. A two-sample t-test was conducted to compare the mean GSE scores of women who participated in the soccer league and those who did not. Linear and logistic regressions were used to explore the relationship between women's self-efficacy and participation in the recreational soccer program while controlling for a number of socio-demographic covariates. Two regression models were run. Models 1 used linear regression examine the association between women's participation in the soccer program, and GSE scores after controlling for socio-demographic covariates. To consider the question of a bi-directional relationship between GSE and soccer participation, a second model using logistic regression was run to explore the association between GSE scores and women's participation in the league.

Descriptive statistics
Descrptive statistics appear in Table 1. The majority of the 684 women included in this analysis had joined the Nikumbuke Project within the previous four years. The average age of respondents was just over 40 years, with women as young as 18 years and as old as 80 years participating in the surveys.
Over three-quarters of the women were married and living with their partners, although close to 42% of respondents considered themselves the head of their households. About 84% of women had three or more children, with close to 20% having seven or more children. Just over 40% of the women had no formal education, and an additional 31% had not completed primary school. The majority of households had monthly incomes under 5,000 KES (about US $50), which falls well below the Kenyan national poverty line. 42 Table 1 Distribution of respondents by demographic characteristics according to self-efficacy and participation in soccer league (n = 684) Self-efficacy Participation in recreational soccer 25 46 *p < 0.05, **p < 0.01, ***p < 0.001, † fisher's exact p-value i at or below the mean, ii above the sample mean Table 2 Model 1: Linear regression model reporting standardized beta coefficients (95% confidence intervals) for the association between self-efficacy and participation in a recreational soccer league (n = 684) Self-efficacy (mean 33

Multivariate analyses
A linear regression analysis was performed to investigate whether participation in the soccer program was associated with the self-efficacy of respondents (

Discussion
This study is the first of its kind in Kenya to examine the factors that may be associated with adult women's participation in sports, particularly the potential role played by self-efficacy. Its findings suggest a bi-directional relationship between self-efficacy and women's participation in a recreational sports program. While it is beyond the scope of this cross-sectional study to determine the presence of a causal or temporal relationship between the two phenomena, it does provide strong evidence Kenya engage in physical activity to be 231 minutes, well above the WHO recommendation of 150 minutes daily, but, of this total, only 5.5 minutes were expended during recreational activity. 49 In addition, slightly more than half of women surveyed did not engage in any strenuous physical activity of any kind. In the Nikumbuke soccer program, women engage in weekly three-hour practices and monthly inter-village matches during the soccer 'season' that provide team members with both a venue and an occasion to engage in leisure sport.
In addition to the weakness of cross-sectional data in determining the causal and temporal relationship between self-efficacy and sports participation, as well as the extent to which the relationship may be bi-directional, this study had several other limitations. Almost 3% of respondents reported that they did not know their age (not a surprising result in an environment with limited birth registries) and were coded as missing.

Consent for publication
Consent was provided orally by respondents as signed documentation would have provided the only link between respondents and their participation in the study. No personal identifiers were collected as part of the interview in order to reduce the likelihood that an individual would linked to specific responses.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.