Survey development and statistical analysis
This study was conducted in a period between March and November 2018. In this perspective study a design survey was used (provided as Additional file
1). Before each data collection, the study was announced a few days earlier in schools, sports clubs or international competitions. Coaches, teachers or parents of potential participants were contacted and introduced to the study in order to recruit athletes for the survey. The inclusion criteria were: the age between 15 and 18 and international competition level.
Three hundred and forty-eight athletes met the criteria and were surveyed, among which male and female participants were equally distributed. Also, the age distribution was balanced with half of the athletes of 15–16 years of age and the other half of 17–18 years of age. This international study included participation of young athletes from 4 countries: Serbia (39.4%), Germany (23.0%), Japan (20.1%) and Croatia (17.5%), all representing their countries at international competitions in 18 sports: kayak (27.9%), rowing (12.6%), canoeing (11.5%), basketball (8.6%), volleyball (8.6%), swimming (8.0%), athletics (4.0%), boxing (2.3%), soccer (2.3%), tennis (2.0%), karate (2.0%), handball (2.0%), water polo (1.4%), dance (1.4%), golf (1.4%), weightlifting (1.4%), archery (1.4%), and fencing (1.2%).
The survey consisted of 20 questions, divided into four main parts. The first part collected demographic and personal information on the study participants: age, sex, country, and the type of sport they are competing in. The second part obtained information regarding the usage, importance, source of information, safety and procurement of sports supplements. The third part tested the athlete’s knowledge about the proper use (timing, dosage and reason for use) of sports supplements. The last part investigated athletes’ beliefs and attitudes towards the use of sport supplements and possible Anti-Doping rules violations.
Athletes voluntarily completed the written survey on different occasions and places such as: international competitions, high schools or on individual basis at different sport clubs. The survey was previously reviewed by various certified coaches in different sports, physicians, university professors and researches specialized in food science and sport psychology.
The reliability analysis of the survey items revealed that all variables measured were reliable with reliability values of all the latent variables extracted above 0.7 (for Cronbach’s Alpha). The Composite Reliability (CR), which represents the overall reliability of a multi-dimensional construct reached values above 0.9, which is attributed as particularly significant. Data were normally distributed and negatively skewed with relatively flat peak. Average Variance Extracted (AVE) was estimated, and the significant values above 0.5 were obtained, meaning that the latent variables were bringing significant variation in the face of random measurement error.
All three conditions of convergent validity were satisfactorily met, i.e. regression weights/factor loadings were equal to or greater than 0.5, whereas squared multiple correlations (SMC) were equal to or greater than 0.7, while AVE values were equal to or greater than 0.5. All aforementioned conditions confirmed the convergent validity of the constructs. In order to test whether two constructs differ from each other, discriminant validity of the constructs was also checked and confirmed by showing that AVE was greater than SMC for each variable.
All surveyed athletes were previously informed about the study objectives and had a chance to clarify any possible misunderstanding of the survey questions with the team conducting the study. While filling out the survey a representative of the team conducting the study was present at the site.
This study was approved by the Ethics Committee of the Faculty of Medicine, University of Novi Sad, and all procedures were conducted in accordance with the Declaration of Helsinki.
Data were processed using Microsoft Excel (Microsoft Corporation, Redmond, Washington, USA) and analyzed using the statistical software Statistica 12 (Dell Software, Round Rock, Texas, USA). Descriptive data were calculated as frequencies. Data were evaluated by sex and age using chi-square (χ
2) analyses. Significance was determined at
p < 0.05. For the statistical analysis, two age categories were used: athletes 15–16 year olds (15-16Y) and athletes 17–18 year olds (17-18Y). The collected data about the proper use of sports supplements among different demographics were analyzed using the correspondence analysis. This analysis is a useful statistical technique for analyzing data collected in sport surveys by simple graphical presentation with a set of points with respect to two coordinate axes [
14]. Symmetric normalization model [
15‐
17] was suitable for exploring relationships between items of two nominal variables.