Development and evaluation of a practical tool for the quantification of readiness to train in female youth soccer players

  • Francisco Songane

Student thesis: Doctoral Thesis


Participation in female soccer is growing rapidly yet girls’ soccer talent programmes operate under a largely part-time workforce with far less funding and infrastructure when compared to males. Girls participating in such programmes do so on top of school education, physical education, and extracurricular sport which can lead to unstructured and occasionally high training loads. Assessing players’ psychophysiological state of recovery, termed ‘readiness’, prior to training maybe useful to help balance high training loads with sufficient recovery. However, assessment of readiness is particularly challenging, as validated measures are often time-consuming and non-specific to girls’ soccer. Therefore, there is a need for an iteratively validated readiness measure that can quickly collect such data to subsequently inform practice. The aims of this programme of work were to: 1) develop a short practical Athlete Readiness to Train questionnaire (ART-Q); 2) iteratively test the ART-Q’s validity in girls’ soccer, and 3) implement the ART-Q into routine monitoring practices to help inform session planning.
A thorough literature review revealed the following as relevant measures of readiness: mood, health, fatigue, sleep, soreness, and nutrition. These measures therefore constituted the ART-Q’s initial pool of measurement items scored on 5-point Likert scales. The ART-Q was subsequently developed via programming in C#, in-house, and downloaded as a tablet application to facilitate data collection. Thirty-three players from a Regional Talent Club (RTC) initially completed the ART-Q during a 12-week competitive period (706 observations) to assess the ART-Q’s factor structure via a Principal Component Analysis (PCA). There was good internal consistency (Cronbach’s alpha, ± 90% confidence limits; α = 0.85; ± 0.01) and the PCA extracted one principal component (eigenvalue = 3.49) consisting of all six items explaining 58% of the variance – suggesting readiness is a single-factor construct. This contrasts with the consensus in the literature surrounding the multifactorial nature of readiness, which prompted an assessment of the questionnaire’s face validity.
Fifteen RTC players were subsequently recruited to three focus groups to understand their perspective on athlete readiness and ART-Q user-experience. An inductive thematic analysis revealed the players recognised the multifactorial nature of readiness; however, they felt 1) their experiences at school were an important aspect missing in the ART-Q, 2) there was a need for the differentiation between food intake and hydration within the ‘nutrition’ item, 3) the ‘health’ item needed to be re-worded to reflect only instances of illness, not physical fitness too, and 4) a more sensitive (7-point Likert scale) could improve their accuracy.
Informed modifications were then applied, which prompted a reassessment of the questionnaire’s validity via another PCA on data collected during a 13-week period from 35 RTC players (507 observations). There was good internal consistency (α = 0.82; ± 0.01) and the PCA extracted three components (eigenvalue threshold at 0.75). Soreness loaded highest (0.97) in the first component (eigenvalue = 3.92; 49% of the variance), hydration loaded highest (0.92) in the second component (eigenvalue = 0.89; 11% of the variance), and school loaded highest (1.00) in the third component (eigenvalue = 0.76; 10% of the variance). These findings suggest communicating data on these three items to coaches, as these would explain most of the variance. However, practitioners should consider collecting all eight items regularly to record the ‘unaccounted’ variance.
A decision support system was subsequently developed and assessed to enhance the ART-Q’s implementation to practice. The ART-Q’s minimal important change, defined as a change from an individual’s 28-day average (± SD) item-score greater than one arbitrary unit, was calculated to inform session planning. This threshold was applied alongside visual Excel-generated sessional reports signposting players of concern. 35 players recorded 829 ART-Q observations across an 8-week period, of which 44 were highlighted, with 16 true-positives – a 36% efficacy. This system prompted useful player-coach conversations that highlighted potential well-being/welfare issues crucial for player safeguarding practices. Future use of the ART-Q should consider whether biological maturation may impact its factor structure.
Date of Award1 Feb 2023
Original languageEnglish
Awarding Institution
  • Teesside University
SupervisorMatthew Wright (Supervisor), Iain Spears (Supervisor) & Jonathan Taylor (Supervisor)

Cite this