NFORMACIÓN GENERAL Título del conjunto de datos:Dataset used for the study entitled: Gender, Match Outcome, and Attacking Performance in Elite Rugby Union... Información del autor: [Completar la información para cada autor] Nombre:DIEGO HERNÁN VILLAREJO GARCÍA Institución: Correo electrónico:dvillarejo@um.es ORCID:https://orcid.org/0000-0001-6149-9253 Fecha de recopilación de datos (fecha única o rango de fechas): [2025-04-21] Fecha de depósito: [2025-05-23] Idioma:Español/inglés INFORMACIÓN METODOLÓGICA This study followed a retrospective, observational, and analytical design (Thomas et al., 2019). A 2x2 between-subjects factorial design was employed to investigate the influence of Gender (Male vs. Female) and Match Outcome (Winner vs. Loser) on key attacking performance variables. Participants and Sample The study population consisted of the national teams participating in the men's and women's Six Nations tournaments (England, France, Ireland, Italy, Scotland, and Wales). The primary unit of analysis was the offensive performance of each team in a specific match ('team-game observation'). The study sample consisted of 270 such units of analysis, obtained from a total of 135 matches. Specifically, data from 75 men's tournament matches (generating 150 units of analysis) and 60 women's tournament matches (generating 120 units of analysis) were included. Men's matches spanned the 2021, 2022, 2023, 2024, and 2025 tournament editions. Women's matches corresponded to the 2022, 2023, 2024, and 2025 tournament editions. Data Collection Match performance data were collected and processed using the Hawk-Eye technology system (Jayalath, 2021; Uzor, et., at., 2024). Hawk-Eye is an advanced object tracking system based on triangulation (Hartley & Zisserman, 2003; Morgulev et al., 2018). The information provided by Hawk-Eye is derived from the analysis of the spatial position of the ball and players throughout the match using a system of calibrated cameras (Fernandez et al., 2018). This system generates a wide range of game statistics with high precision (Morgulev, et., al., 2018). The data collected by this system are publicly available on the official Six Nations tournament website: https://www.sixnationsrugby.com/en. Additionally, with the aim of evaluating the reliability of the performance metrics provided by the Hawk-Eye system for the key attacking variables, a concurrent manual verification was conducted on a subsample of matches. This subsample consisted of 30 matches, selected using stratified sampling by gender (15 male and 15 female) and final outcome (winner/loser), representing approximately 22.22% of the total sample of 135 matches. Manual observation of these 30 matches was performed by one of the study authors with 20 years of experience in notational analysis. Standardized operational definitions were used for each event, and the coding process was conducted using Longomatch 2.0 video analysis software (Gil, 2018). The variables selected for this manual verification were the 7 key discrete metrics: Tries Scored, Linebreaks, Defenders Beaten, Total Passes, Offloads, Handling Errors, and Turnovers Conceded. These variables were selected because they represent discrete events whose automatic recording may be more prone to variability compared to continuous metrics (e.g., possession percentages). To evaluate the reliability of the Hawk-Eye system compared to human observation, the Intraclass Correlation Coefficient (ICC) was calculated for each of the 7 variables in the 30 matches. An ICC(2,1) model for absolute agreement of a single observer was employed (Koo & Li, 2016). The results of this reliability analysis indicated moderate to good consistency between the Hawk-Eye data and manual observation for the evaluated variables. ICC values obtained ranged between .72 and .89. The minimum reliability value observed was .72 for the variable Defenders Beaten. Variables Thirteen original variables related to attacking performance were analyzed. These variables, as provided by the Hawk-Eye system, were: Possession (%): The percentage of total match time during which a team had control of the ball or participated in attacking phases; Territoriality (%): The percentage of total match time that a team spent within the opponent's half of the field; Attacking_Minutes: The total time (in minutes) that a team spent performing offensive actions with ball possession; Tries_Scored: The total number of tries scored by the team during the match; Carries: The number of instances in which a player individually carried the ball, advancing with it; Metres_Carried: The total accumulated distance (in meters) advanced by a team's players while carrying the ball; Metres Gained: The total accumulated distance (in meters) that a team managed to advance up the field during their attacking phases, measured from the initial to the final position of each phase; Linebreaks: The number of times a player with the ball successfully broke through the opponent's defensive line, penetrating into their advantage zone or open space; Defenders Beaten: The number of occasions in which a ball carrier successfully evaded an attempted tackle by an opposing defender; Total_Passes: The total number of passes made by the team's players during the match; Offloads: Passes made by a player while being tackled or in contact with a defender; Handling Errors: The number of times a player lost control of the ball due to a technical error, such as a knock-on or a forward pass; Turnovers_Conceded: The total number of times a team lost possession of the ball to the opponent, including handling errors, penalties resulting in change of possession, being tackled into touch, etc. These variables represent standard metrics used in rugby union performance analysis to quantify offensive activity, territorial progression, and efficacy in creating scoring opportunities, as well as errors committed during possession. Ethical Considerations As the study is based on official and publicly available performance data from elite competitions, individual informed consent from players or teams was not required (American Psychological Association, 2017). Access and processing of the data were carried out for academic research purposes and adhered to the regulations for the use of publicly available sports data. Data Availability Statistical Analysis Descriptive statistics (mean and standard deviation) were calculated for the 13 original attacking variables. Subsequently, a Principal Component Analysis (PCA) was conducted on the dataset of these variables to reduce dimensionality and identify underlying dimensions of offensive performance. The suitability of the data for PCA was assessed using the Kaiser-Meyer-Olkin (KMO) index and Bartlett's Test of Sphericity. Three principal components were extracted based on the criterion of initial eigenvalues greater than 1 and considering the percentage of total variance explained (75.9%). Varimax rotation was applied to facilitate the interpretation of the components. The scores of these three components were calculated and stored for use as dependent variables in subsequent analyses. To evaluate the differences in the principal component scores based on the between-subjects factors of Sex (Male/Female) and Match Outcome (Winner/Loser), three independent univariate analyses were performed: 2x2 factorial ANOVAs for each of the component scores (PC1, PC2, PC3). Prior to interpreting the results of each ANOVA, the assumptions of residual normality were assessed (using the Kolmogorov-Smirnov Test and visual inspection of Q-Q plots) and homogeneity of variances between the four groups (using Levene's Test). Where the assumption of homogeneity of variances was violated for main effects, specifically for PC2: Efficacy and PC3: Errors, Welch's ANOVA was employed for these effects. In such cases, the Welch's F-statistic, adjusted degrees of freedom, p-value, and the original partial eta-squared (ηp2) are reported. For interaction effects, and for all effects in components where homoscedasticity was met (e.g., PC1), standard ANOVA results are presented. Given that three independent ANOVAs were conducted, a Bonferroni correction was applied to the alpha significance level (α = 0.05) to control the Type I error rate for the family of the three tests. The adjusted statistical significance threshold was p < 0.0167. The statistical significance of the main effects (Sex, Outcome) and the interaction (Sex * Outcome) in each ANOVA was determined by comparing the obtained p-values with the adjusted significance threshold. Effect sizes were reported using partial eta-squared (ηp2). Estimated marginal means plots were used to visualize the effects. To explore the nature of observed effects and determine significant differences between groups following the identification of significant main effects or an interaction (according to the adjusted threshold of p < 0.0167), inferential post hoc comparisons were conducted. Pairwise comparisons on the estimated marginal means were performed, and the Bonferroni correction was applied to the resulting p-values to control the family-wise Type I error rate associated with the multiple comparisons within each significant ANOVA. All statistical analyses for this study were performed using the statistical software Jamovi (version 2.6.26) and the R statistical programming environment (version 4.4.1). ARCHIVOS Nombre del archivo: Datos reposistorio UMU.pdf Formato del archivo:pdf PALABRAS CLAVE Rugby Union; Performance Analysis; Gender; Attacking Performance; Match Outcome. INFORMACIÓN DE PATROCINIO E IDENTIFICADORES DE SUBVENCIONES Sin financiación externa a la Universidad PUBLICACIONES RELACIONADAS Publicación relacionada: Conjunto de datos relacionado: LICENCIAS Y PRIVACIDAD Licencias:Creative Commons Privacidad: MÁS INFORMACIÓN [Incluir cualquier otra información sobre el conjunto de datos que no esté reflejada en esta plantilla y que se considere relevante.]