2. Materials and Methods
2.1. Study Design
This study used a cross-sectional survey design. This design is helpful to gather quantitative data, such as scores on instruments that produce specific numbers that can be statistically analyzed, as it can yield results to assess the relationship between variables and measure prevalence for the variables of interest [
52]. The independent variable in the analysis was the program quality components, which consisted of four domains: current practice in youth volleyball, perceived experience in youth volleyball, perceived challenges in youth volleyball, and coach–athlete relationship in volleyball. The dependent variables included quality of talent identification and development system, youth volleyball sport participation, integrated system for youth volleyball, and national and regional volleyball competitions.
2.2. Population, Sample Size, and Sampling Techniques
Ethiopia consists of regional states and special city administrations, and these are divided into zones, depending on geographic proximity. Hence, zones are the second level of the administration. Each zone consists of districts of Ethiopia that are known as Woredas, and these are the third level of the administration of Ethiopia—next to zones and regional states. The target population of this study encompasses all under 17 youth volleyball players in Ethiopia (between the age of 15 and 17)—(n = 13,200).
Each Regional state consisted of 10 youth volleyball development projects (both male and female youth players) found in 10 different youth volleyball development facilities. Each youth volleyball development project facility had a total of 120 youth players containing 60 (girls) and 60 (boys); therefore, 1200 youth volleyball players were involved in each region’s U-17 youth volleyball training development projects.
Sample selection followed a multistage sampling procedure, dividing the population into smaller and smaller groupings to create a manageable sample. First, one Regional State, that is, Southern Nations Nationalities, Peoples’ Regional (SNNPR) State was selected using a simple random sampling method. Following that, the 10 zones in the selected Regional State were purposefully considered as the 10 youth volleyball sites as they were located in these zones. Finally, 20% of the sample youth volleyball players were selected from each volleyball project. The youth players sample (n = 22, girls = 11, and boys = 11) in each study site was approached, and a total of 220 participants were involved in the study. To identify the youth player sample participants in each project site, the principal author obtained the list of youth players’ names from each youth volleyball development facility head. Then, sample participants (n = 22, girls = 11, and boys = 11) were selected using a simple random sampling method with two categories of girl and boy groups. Then, each questionnaire was administered to each sample participant. Eligible participants were youth volleyball players (n = 220, girls = 110, and boys = 110), each of whom had an active engagement in the youth volleyball project in each sample youth and sports development facility during the 2019 project implementation season as youth players.
The project studied is the training arm of the sports institutions as it runs training, learning, and capacity development services for youth volleyball players. The average age of the youth participants was 16.18 years (SD = 0.69). Youth volleyball projects are located in three geographical divisions called zones. These include: western zone (Konta, Segen, & Dawaro, SNNPR, Ethiopia), central zone (Wolita, Kanbata, Hadiya & Gedio, SNNPR, Ethiopia), and eastern zone (Sidama, Gamogofa, & South Omo, SNNPR, Ethiopia) (2019).
Table 1 presents the summary of the demographic characteristics of the study participants (2019).
The study sample involved 107 males (49.8%) and 108 females (50.2%) with a mean age of 15.81 years (SD +0.76) for females and a mean age of 16.54 years (SD +0.62) for males. Attrition rates were low with 69.3% attending for four years (69.3%) and 30.7% attending three years. Youth volleyball players involved in this study were distributed across the three divisions, with the western zone accounting for 40.5%, the central zone 29.8%, and the eastern zone 29.8%. All had 3 to 4 years of playing experience.
2.3. Study Variables
Again, this comprehensive assessment included a program quality measure of the process where the object of measurement is the program [
28]. Hence, the program quality measures emphasized perceived experiences and social processes or interactions between people within the program [
32]. The developmental outcome measures were designed to provide a framework of essential measures of sport-based development outcomes [
53]. The specific indicators of each domain/component are drawn from youth sport program practices and research [
54], and while not exhaustive, they represent concrete ways that characterize a youth sports program and its impacts on development.
Table 1 presents the categorization of variables included in the study.
The majority of items used in the indicators of program quality helped to assess promotive interactions between and among youth players, and their coaches, and the extent to which young volleyball players are engaged in the program-representing participant perceptions of experiences in youth programs. However, the program quality indicators also addressed the perceived challenges and the coach–athlete relationships within the program. The other domains/components of developmental outcomes focus on four major elements of developmental outcomes: Perceived quality of talent identification and development system, participation in youth volleyball, integrated system for youth volleyball, and national and regional volleyball competitions (
Table 2).
2.4. Data Collection Instrument
In this study, the authors used a questionnaire to collect data from youth volleyball players. In order to clarify the factors and developmental outcome measures,
Table 3 illustrates the definition of each measure and its sources.
The validity and reliability of the items were checked by expert review and a pilot test with participants from similar youth volleyball projects. Item-to-total correlations were assessed, as well as additional confirmatory factor analysis, which suggested a model of good-fit and constructs validity of the scale. For reliability, the internal consistency coefficients ranged from 0.81 to 0.87, which satisfies the criteria outlined in sports sciences research [
18]. Each item in the questionnaire was measured using a five-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree). Furthermore, demographic information was collected to compare responses between subgroups. Level of training experience, youth volleyball development facility locations and low achievers), and gender were the primary areas of interest.
2.5. Data Collection Procedure
The principal author collected the relevant data from each study site beginning from 10 November to 30 December 2019. To do so, the principal author first obtained permission from the respective youth and sports authorities at three levels: The Southern Nation and Nationalities Peoples’ Regional (SNNPR) State, Zone, and Woreda.
Moreover, the permission included each youth volleyball development facility head to ensure their cooperation in the study. Besides cooperation, their permission also acknowledges that they understand the purpose and ethicality of the study. Additionally, the authors asked for parental/guardian consent for the youth players’ voluntary participation in the research study. Moreover, respondents were given a clear description of the purpose, scope, and intended outcomes of the research. The type of information required for the research was clearly stated, as in the policy for anonymity and confidentiality. Ethical clearance was sought for the study from Bahir Dar University Sport Academy Ethical Review Committee (S/A/D 6974/11) to ensure that the study did not involve questions that were offensive or personal in nature and there were no identifiable risks to the respondents’ health. The survey questionnaire data were completed by the youth player samples participants (n = 220, girls = 120, and boys = 120) from the 10 project sites. A total of 215 participants returned the questionnaires. The final sample included 215 participants, reflecting a response rate of 97.7%.
2.6. Data Analysis
Data were analyzed using IBM SPSS Statistics, Version 23.0, and the collected data were analyzed at four levels. First, descriptive statistics were used for the subscales of the program quality and reported developmental outcome measures to understand the perceived state of these measures in the studied context (Research question 1). Then, t-tests and one-way ANOVAs were calculated on the dependent and independent variables to examine the pattern of significant differences between sample youth volleyball players classified by their gender and project site zone (Research question 2). Significant differences between groups were determined by
p-values ≤ 0.05. Effect sizes are important because whilst the independent t-test or ANOVA tells us whether differences between group means are “real,” it does not tell us the “size” of the difference. Effect sizes were computed to overcome this limitation and measure the magnitude of differences [
57]. The most popular effect size measure is Cohen’s d [
58]. In this study, the authors used an online calculator to compute the different effect sizes using Cohen’s d and eta squared [
59]. The results are interpreted in the discussion section.
Then, two sets of analyses were used to examine the relationship between the quality of youth volleyball outcomes and the process correlates, including participation, experience, challenges, and coach–athlete relationship (Research question 3). The first set of bivariate correlations were calculated between the total scores of the eight self-report measures as a preliminary step to examine the relationship between the process and outcome variables of interest. The second set used multiple regressions to assess the relative influence of the four program quality variables in predicting the quality of youth volleyball outcomes.
2.7. Preliminary Analysis
The authors conducted preliminary analyses to ensure no violation of the assumptions of normality, linearity, multicollinearity, and homoscedasticity. A histogram was used to test the normality distribution of residuals. The result indicated that the majority of the scores lie around the center of the distribution; in addition, the coefficient of the skewness data value is between −3 and 3 and the kurtosis value is not far from zero. Thus, it fulfills the assumption of multiple regressions.
A normal probability plot or normal quantile plots of the residuals are an indicator for best tests for normally distributed errors [
57]. In this study context, the points on a plot fall close to the diagonal line. This result showed that the distribution is linear. Hence, it fulfills the assumption of multiple regressions.
In examining the correlation matrix of independent variables, none of the pair of correlation coefficients exceeded 0.86 [
60]. Similarly, the results revealed that no tolerance value found below 0.1 and all-variable inflation factors (VIF) values are below 10 (the VIF current practice of youth volleyball, perceived experience in youth volleyball, perceived challenges in youth volleyball, and coach–athlete relationship in youth volleyball is found below 10). Based on this, multicollinearity was not a problem in this study context [
57].
3. Results
The results showed that the nature of the variables under examination, the existing group variations and which variables are significantly related to the quality of youth volleyball, and to what degree the program quality indicators relate to one another. The results are presented in different headings, namely, demographic statistics of participants, variations across gender and project site zones, relationship between program quality and developmental outcomes with one another, and the predictive capacity of the program quality indicators.
Overall, the ratings of program quality and developmental outcome measures were compared across gender. According to the results of this study, program quality and developmental outcomes did not vary significantly across gender, except for perceived experience, which was rated higher by female volleyball players than their male counter parts.
Table 4 presents the summary of the independent
t-test result.
As shown in
Table 4, the female volleyball players’ group perceived significantly higher (M = 2.68, SD = 0.258) compared to the male volleyball players’ group (M = 2.58, SD = 0.318) in terms of perceived experience. The mean difference is 0.35 SD, which indicates a medium effect size [
58].
In this study, the authors used descriptive statistics to measure the extent of positive participation, relatedness, and quality youth volleyball developmental outcomes, and provide summaries of the sample and the measures used.
Table 5 presents the descriptive statistics of the variables studied.
As shown in
Table 5, the mean score values for the total sample ranges between 2.53 to 3.39 with the standard deviation ranging between 0.293 to 0.818. In terms of the score distributions, for each project site division, the values ranged between mean values of 2.38 to 3.68 and the standard deviations ranged between 0.248 to 0.948. It was clear from the results of
Table 5 that the ratings across the different project site zones varied considerably. However, it was unclear whether those differences were statistically significant differences or not. For this, the authors used one-way ANOVA.
One-way ANOVAs were performed to compare the effects of project site location difference on the program quality and developmental outcome variables. More specifically, there were statistically significant differences between project site zones at the
p < 0.05 level for the current practice in youth volleyball (F (2, 212) = 21.41,
p = 0.000,
η2 = 0.16), perceived experience in youth volleyball (F (2, 212) = 15.79,
p = 0.000,
η2 = 0.12), perceived challenges in youth volleyball (F (2, 212) = 5.84,
p = 0.003,
η2 = 0.05), and coach–athlete relationship in youth volleyball (F (2, 212) = 8.22,
p = 0.000,
η2 = 0.07).
Table 6 presents the summary of the one-way ANOVAs for the process quality variables studied.
According to the results in
Table 6, there was a statistically significant difference between the three groups in terms of the four program quality domains related to the youth volleyball program quality measures as demonstrated by one-way ANOVAs, current practice (F (2, 212) = 21.41,
p < 0.001), perceived experiences, (F (2, 212) = 15.79,
p < 0.001), perceived challenges (F (2, 212) = 5.84,
p = 0.003), and coach–athlete relationship (F (2, 212) = 8.22,
p < 0.001). The results of the ANOVAs for the three groups (
p > 0.131) in terms of the four developmental outcome measures were non-significant.
In order to identify which group is different from the other groups, the authors ran post hoc tests. Depending on the results of the Levin test of homogeneity of variance, the authors carried out a Tukey post-hoc comparison test when the Levin test results were not significant. Contrary to this, the authors carried out the Games Howell post-hoc test instead of Tukey’s when the Levin test results were significant.
Table 7 presents the summary of the post-hoc tests.
As shown in
Table 7, in three of four variables of program quality, youth volleyball projects found in the central zone were rated significantly higher than those in the western zone: Perceived experiences (
M = 2.68 versus 2.56), Perceived challenge (
M = 3.50 versus 3.17), and coach–athlete relationship (
M = 3.43 versus 2.94). Also, the mean score of players’ ratings of youth volleyball projects in the central zone was significantly higher than those projects in the eastern zone. Likewise, the mean score of players’ ratings of youth volleyball projects in the eastern zone was significantly higher than the western zone in terms of the perceived challenge (
M = 3.48 versus 3.17). However, the mean score of players’ ratings of the projects in the central zone was significantly lower than the western and eastern zones only in terms of the perceived challenge (
M = 2.38 versus 2.64 and 2.60). Except for the current practice, a consistent perception difference was present between youth project players in the central zone and those youth project players in the western zone. This means the youth volleyball projects in the central zone were rated significantly higher than those in the western zone. taken together, these results suggest that project site location differences really do matter in program quality and attaining developmental outcomes.
For establishing relationships between constructs, and to answer the second basic research question and determine individual associations between these variables, Pearson correlation matrices were conducted for the total sample (
n = 215). Correlations were computed among four program quality variables and another four developmental outcome variables on the data for sample (
n = 215) participants.
Table 8 presents a summary of total intercorrelations for the sample.
As shown in
Table 8, the results indicated that 21 out of 28 correlations were statistically significant (16 of them had significant positive correlations and five of them had significant negative correlations). In terms of positive correlation results, 16 out of 21 correlations were statistically positively significant and were greater or equal to
r(215) = 0.18,
p < 0.05, two-tailed. The correlations of the perceived experience variable with the other process quality and developmental outcome variables were not statistically significant, with the exception of current practice rating,
r(215) = 0.18,
p < 0.01, two-tailed.
In terms of negative correlations, 5 out of 7 correlations were statistically negatively significant and were greater or equal to r(215) = −0.14, p < 0.05. Accordingly, the correlations of study participants’ ratings of perceived challenge with the other program quality and developmental outcome variables were statistically negatively significant and were greater or equal to r(215) ≥ −0.14, p < 0.05. The correlations of the perceived challenge rating with the other developmental outcome variables such as perceived experience and coach–athlete relationship were not significant, r(215) = −0.03, p > 0.05.
The third research question of this study was to investigate the predictive capacity of the four-program quality indicators on the predictions of the four developmental outcomes. For this, the authors conducted four multiple regression models. This approach was used to determine the unique variance in the dependent variable (i.e., quality of youth volleyball) that each of the independent variables explains.
The contribution of each independent variable to the variance of the dependent variable was calculated and the coefficient of determination, which is the proportion of variance in the dependent variable, that is, each developmental outcome indicator that can be explained by the independent variables was calculated. Predictors included: Current practice, player’s experience, perceived challenges, and coach–athlete relationship in youth volleyball.
Table 9 presents the results of the regression models.
As shown in
Table 9, in the first regression model, the four selected independent variables together predicted the quality of talent identification and development outcome, explaining 29.4% of the variance. In this model, the beta values indicated that the current practice and coach–athlete relationship in youth volleyball were found significant predictors at a comparably equal magnitude, that is, current practice (
β = 0.28,
p < 0.004), and coach–athlete relationship (
β = 0.282,
p < 0.003).
Again, in the second regression model, the four program quality indicators, together predicted the youth volleyball participation outcome, accounting for 24.7% of the total variance (R2 = 0.25, F(4, 210) = 17.18, p < 0.000). Furthermore, the effect of current practice and coach–athlete relationship to the volleyball sport participation significantly predicted youth volleyball sport participation (β = 0.25, p < 0.013), and (β = 0.26, p < 0.008).
In the third regression model, the four program quality indicators, together predicted the integrated system in youth volleyball outcome, accounting for 31.8% of the variance explained (R2 = 0.32, F(4, 210) = 24.47, p < 0.000). Furthermore, the effect of current practice and coach–athlete relationship on the integrated system in youth volleyball were found significant (β = 0.23, p < 0.016), and (β = 0.33, p < 0.001).
In the fourth regression model, the four selected independent variables collectively explained 18.9% of the total variances in national and regional sport competition outcome (R2 = 0.18, F(4, 210) = 12.21, p < 0.000), as P is less than 0.05 and F value is large, the model is significant. Furthermore, it was found that the direct effect of the variables on the national and regional volleyball sport competition was determined using a beta coefficient. Perceived challenges in youth volleyball significantly negatively predicted (β = −0.17, p < 0.008), the national and regional competition. Also, coach–athlete relationship significantly positively predicted the national and regional sport competition (β = 0.31, p < 0.003).
In general, these results of regression analyses indicate that differences in youth sports program quality, particularly current practices, perceived challenges, and coach–athlete relationships, may help to explain variations in the sport development outcomes. The persistent positive influence of coach–athlete relationship variable, which was examined as a predictor of the four developmental outcomes, further signals the relevance of the coach–athlete relationship variable in youth sport research.