Referee's records (players and coaches sanctions)
Referees’ records obtained from the 275844 games played in 1401 league competitions in the 2014-2015 and 2018-2019 seasons in the Valencia Community (Spain). A total of 53031 sanctions were recorded by the referees of these matches, being most of them minor (93.334 %). The most frequent sanctions were produced because of accumulation of cards (41.8 %), violent or dangerous play (6.5 %) and contemptuous and inconsiderate attitudes towards referees (5.5 %). Assaults (3.3 %), insults and serious offenses (2.7 %) and very serious offenses (0.004 %) received the severe sanctions by referees. Furthermore, we found significantly (p<.05) more minor than serious sanctions, in 11-a-side soccer than in 8-a-side soccer, in the men's category compared to the women's category and in the U18 and U15 men's categories compared to the rest of the lower age categories. This suggests that the strategies adopted for the improvement of sportsmanship should especially target players in the U18 and U15 male categories.
Steps to reproduce
We uploaded the data to the Version 3.6.0 R program to run the necessary analyses. First, the team calculated the descriptive statistics of the variables of interest using the R base functions and the ggplot2 library. For example, for the variables of player sanctions in the different seasons, we calculated the measures of central tendency (mean), dispersion (minimum, maximum and standard deviation) and shape (skewness and kurtosis). We also graphically represented the most common reasons for sanctions. We then performed a multifactor analysis of variance (ANOVA). This type of statistical test, widely used in sports science, can be used to examine the interactions between at least two ordinal independent variables with respect to the continuous dependent variable (Malek et al. 2019). In this case it allowed us to determine the influence and interactions between the type of soccer, gender, playing category and severity of the sanction (independent variables) and the sanctions (dependent variable). The factorial ANOVA had two goals: to determine the main effects of the different variables and the existence of a significant interaction (p<0.05). We obtained significant main effects for all the variables. Next, we performed multiple comparison procedures using Tukey's HSD post hoc test (Malek et al. 2019), which compares all the groups or levels of a categorical variable. Through this test we compared differences between means instead of comparing pairs of values. Tukey's HSD involves the calculation of the absolute value of the difference between pairs of means divided by the standard error of the mean (SE), determined by a one-way ANOVA test. The SE is, in turn, the square root of the variance divided by the sample size.