The perception of talent development environment in football: Scale development and psychometric testing

Abstract

Talent development environment (TDE) has been a key focus in sport as sport talents need proper development to transfer their potential to elite performance. The environment is increasingly recognised as a controllable factor in talent development. While several scales have been developed to measure TDE for general sport contexts, relatively few have been specifically designed for particular sport (e.g., football). This study developed a scale for measuring TDE in football and validated its psychometric properties. The participants (N = 551) were football talents (aged between 11 and 17 years) who were recruited from Chinese football specialised schools participating in the Chinese national youth football league. The participants included 377 males and 167 females, with seven participants not specifying their gender. Three phases were adopted: (a) item generation using a deductive approach based on a literature review and subsequent questionnaire design; (b) exploratory factor analysis (EFA; n = 270); and (c) confirmatory factor analysis (CFA; n = 200). Consequently, a six-dimensional scale was proposed (football coach, interaction with teammates, football training, football competition, academic support and family support) with 30 items. The results indicate that the scale has a sound overall model fit (χ²/df = 1.84, CFI = 0.93, IFI = 0.93, TLI = 0.92, RMSEA = 0.07, and SRMR = 0.06), as well as convergent (AVEs ranging from 0.51 to 0.75) and discriminant validity (no inter-dimensional correlations exceeding their respective squared root of AVEs). In the future, it is recommended that the scale be further validated and optimised using samples from different regions with different socio-cultural backgrounds.

Keywords

Academic support family interpersonal dynamics soccer teammates

Introduction

Talent development environment (TDE) is considered one of the most controllable components of the talent development process,¹ which influences a range of outcomes (e.g., athletic performance, well-being of athletes, and dropout rate) in sport.^2,3 Specifically, compared to other influential factors of talent development such as natural gift, chance, and personal traits (e.g., physical and mental characteristics), environment appears to be more easily manipulated for improving talent development practices.¹ Without environmental nurture, sport talents cannot automatically fulfil their potential and reach the elite level.⁴ In line with this, several psychometric instruments^4–6 have been developed for measuring the key environmental factors that are considered crucial for sport talent development. Although the environmental dimensions included in these measurements vary to some extent, they were all developed from a general sport perspective, typically by surveying athletes from multiple sport backgrounds. Such general instruments for assessing TDE are indeed valuable for understanding common environmental factors across different fields. However, they may not effectively capture specific features in a particular sport. For instance, previous TDE measurements have rarely touched upon team-related attributes such as inter-relationships and interaction effects within a team, which are critical to collective training and competition contexts, despite the fact that these factors are considered significant for talent development in team sports, such as football, given the nature of group dynamics.^7–9 Therefore, this study aims to fulfil this limitation and develop a TDE measurement model tailored to football (association football or soccer), which allows researchers and practitioners to refine football talent development practices. The following sections analyse the existing TDE scales and outline the theoretical foundation of the study, followed by three phases of scale development.

Key existing TDE measures in sport

The ecological systems theory emphasises that human development does not occur in isolation but is rather shaped through continuous and dynamic interactions with the multifaceted environments in which individuals are situated.¹⁰ Based on the ecological approach and the review of the literature on effective talent development,^10,11 talent development environment questionnaire (TDEQ) was developed with seven factors: long-term development focus, quality preparation, communication, understanding the athlete, support network, challenging and supportive environment, and long-term development fundamentals.⁵ TDEQ was subsequently further validated and applied across various sport settings.^12–14 However, several issues have emerged in empirical studies, such as a lack of internal reliability,¹⁴ poor model fit,¹⁵ and problematic items.¹⁶ Therefore, TDEQ required further optimisation.

To improve the measurement quality and simplify the content of the TDEQ, TDEQ was further refined.⁴ The optimised scale, named TDEQ-5, contains factors: long-term development, holistic quality preparation, support network, communication, and alignment of expectations. Then, TDEQ-5 has since been applied and validated across various sport contexts.^3,16–20 Compared to the TDEQ, the TDEQ-5 appears to have a more concise factor structure and exhibits greater reliability and validity.^4,5 Nonetheless, both TDEQ and TDEQ-5 lack consideration of specific cultural contexts. For example, various social and family cultures may develop different levels of expectations towards the development of football talents, hence constructing distinct developmental environments. For example, in countries where football holds a more significant position in society (e.g., UK, Brazil, and Argentina), both society and families may have stronger expectations for the development of football talents and thus invest more resources, whereas in cultures that prioritise the academic development of young people (e.g., China), parents may expect the environment to place greater (or at least equally) emphasis on supporting academic growth of football talents.²¹

Considering cultural differences between China and Western countries, a Chinese version of the Sports Talent Development Environment Questionnaire was recently developed to measure the environmental factors influencing the sustainable development of sport talents.⁶ This scale includes five factors: training environment, support network, family characteristics, communication, and team conditions. Traditional Chinese family culture is emphasised during the development of the scale, thus including a factor of family characteristics.⁶ In addition, the scale reflects consideration of the China's four-level sport talent training system.⁶ However, as participants were recruited from multiple sport programmes in China, this scale remains a general measure of TDE rather than one tailored to a specific sport.

In summary, previous TDE scales were developed using samples from multiple sport contexts to measure TDE broadly, while few TDE scales have been designed for a specific sport (e.g., football) using data from these single-sport samples. The next section illustrates the conceptual basis for developing the TDE scale for football.

A conceptual framework for TDE in football

Based on TDEQ⁵ and the athletic talent development environment model,²² a football TDE framework is conceptualised with six dimensions, including football coach, interaction with teammates, football competitions, academic support and family support.⁷ Compared to the earlier TDE frameworks, this framework specifies several contextual attributes (e.g., complex, open, and team-oriented) within corresponding dimensions. In addition, it also takes into account the significance of academic development and family support for young football talents, considering the academic and domestic challenges they face throughout their development.⁷ The dimensions were further operationalised to make them applicable for measurement development in this study. The six dimensions and their operational definitions are illustrated in detail as follows.

Firstly, football coach refers to “someone whose job is to teach football talents to improve at football.”⁷ Football coaches share many similarities with coaches in other sports, such as teaching motor skills, providing practice feedback, and motivating football talents.^23,24 However, in-game coaching ability and management skills as a team leader are thought to be more important in football than in many other individual sports.⁷ Therefore, the football coach is considered a key environmental dimension that needs to be explicitly addressed in the proposed model.

Secondly, interaction with teammates is defined as “a manner of behaviour in which football talents and their teammates act upon each other.”⁷ Previous research highlighted the importance of positive team attributes (e.g., motivational climate and teammate relationships) for the development of youth athletes.^25,26 However, the relationship between football talents and their teammates is even more significant, as football talents regularly practice and compete as a group, unlike many other individual sport talents. Therefore, interaction with teammates should be considered as a key dimension in the framework.

Thirdly, football training refers to football talents’ perceptions about daily football practice. Scientific training (e.g., load-appropriate, well-planned, and long- focused) helps athletes improve their sports performance and reduce the risk of injury.^5,27,28 However, considering the dynamic, complex, and open features of football, training should be more flexible and include rich content to accommodate the diverse and dynamic game contexts.²⁹ In line with this, the content related to training planning, flexibility and diversity should be incorporated into the items of this dimension.

Fourthly, football competition refers to football talents’ perceptions of matches in which their football team participates with the goal of winning. In football, competitions offer valuable opportunities to utilise physical abilities, understand game contexts, make strategic decisions, collaborate with teammates, and practice motor skills and tactics in specific situations.³⁰ Thus, regularly engaging in high-quality competitions is essential for football talent development. Accordingly, items that determine the quantity and quality of competition experience should be designed in this dimension.

Fifthly, academic support is regarded as essential for developing football talent, as sport talent should be nurtured holistically rather than in a solely sport-centric manner.²² This dimension is defined as actions taken to facilitate the academic development of football talents. Young athletes often encounter greater academic challenges than their non-athlete peers, as they must balance the demands of sport training with academic commitments.³¹ Hence, academic support is essential to help football talents develop as well-rounded individuals.

Finally, family support represents perceptions of football talents towards assistance provided by family members. Both financial and emotional support of parents plays a critical role in sustaining participation of youth athletes in sport.³² On the other hand, excessive control and pressure from parents increase the risk of causing young athletes to drop out of the sport development stage.³³ Consequently, family support is indispensable for football talents to fulfil their sporting potential and reach elite performance.

Overall, the six dimensions outlined above are hypothesised to be key dimensions of the TDE structure in football (Figure 1).⁷ Although this framework has been conceptually established,⁷ further quantitative validation is warranted to confirm its practical applicability. Therefore, this study aims to develop a measurement instrument theoretically grounded in the six-dimensional framework. The scale development process, as outlined in the following sections, adheres to the procedures in Churchill's research.³⁴

Figure 1.

The conceptual framework for talent development environment in football. Adapted from Xia et al.⁷

Phase 1: Item generation and questionnaire design

Generation of initial item pool

Based on the operational definitions of the six dimensions, a total of 61 items were initially generated: football coach (10 items), interaction with teammates (11 items), football training (10 items), football competition (10 items), academic support (nine items), and family support (11 items). Among them, 18 items were adapted from the previous measures, which include TDEQ,⁴ TDEQ-5,⁵ Student Athlete Relationship Instrument,³⁵ Student Social Support Scale,³⁶ Coaching Competency Scale,³⁷ and Coaching Competency Scale-II.³⁸ They were rephrased slightly to better align with this study. For example, the item “My training sessions are normally beneficial and challenging.”⁵ was modified to “The football training that I participate in is challenging.” The other 43 items were self-developed by using a deductive approach for item generation.^39,40 Specifically, these items were developed based on the six conceptual dimensions outlined in the framework,⁷ and the definition of each dimension served as a guide for the development of these items.^39,40 These items were initially developed based on a literature review combined with the authors’ experiences. Further content validity processes were conducted to test whether these items reflected the attributes of their dimensions.

Content validity

A panel of experts was formed to assess the content validity of the items to ensure their relevance to the corresponding factors and the clarity of individual item statements.⁴¹ Four experts from different fields (one expert in scale development, two scholars in football talent development, and one elite youth football coach) were invited to review the initial item pool.

Every expert was provided with a brief introduction to this study, including its purpose, along with a list of items and the definition of each dimension. Below each item, a space for comments was included for expert input. They were asked to examine whether the items fit their respective factors as conceptualised, and whether the item expressions were easy to understand. Several items deemed unsuitable by the panel were rephrased based on expert recommendations. For instance, “I can understand my training well” was replaced by “My training content is clear to understand”. In addition, the wording for many items (i.e., FC1, FC2, FC9, IWT1, IWT2, FT1, FT2, FT3, FCP1, FCP2, FCP4, AS1, AS2, AS4, FS1, FS5, and FS8) was refined to enhance clarity. For example, “The football competitions I played got a lot of attention” was refined as “The football competitions I played got a lot of social attention”. No items were removed in this phase.

In addition, a total of 30 postgraduate students with sport-related majors were recruited. They were all engaged in or had previously conducted research related to football. Twenty were from a university in the UK, and the other 10 were from a university in China. Both universities are famous for their sport majors in their respective countries, and all participants were able to read and understand English without barriers and had relevant expertise in sport. All 61 items were shuffled randomly, and the participants were asked to sort them into dimensions that best reflected their opinions. An item was retained if more than 60% of the participants assigned it to the same category.⁴² Consequently, six items were excluded, and 55 items were retained for data collection and analyses.

Questionnaire design

The questionnaire used for data collection included 55 items across six dimensions: football coach (10 items), interaction with teammates (eight items), football training (nine items), football competition (nine items), academic support (nine items), and family support (10 items). A 7-point Likert scale was applied to gather the degree of football talents’ perceptions, anchored from 1 (strongly disagree) to 7 (strongly agree) towards each question. Additionally, several demographic questions, such as age and gender, were included to understand the basic demographic profile of football talents.

Considering that the research samples were Chinese, the English-version needed to be translated into Chinese using a back-translation technique.⁴³ The English-version questionnaire was first translated into Chinese by a doctoral researcher who was proficient in both English and Chinese. Next, another bilingual who worked for the youth football department at the Chinese Football Association translated this Chinese version back to English. Finally, the first author reviewed both versions to identify any inconsistencies. If a discrepancy is found, the content is amended, and an iterative back-translation process is undertaken until no inconsistencies exist between the two versions.

Phase 2: Examining the proposed factor structure using exploratory factor analysis

Data collection and participants

Data were collected through football talents’ coaches, who were contacted via the first author's professional social network. Since all participants were minors, aged between 11 and 17, consent forms and questionnaires were distributed via an online questionnaire software, which generated a link for access via computer or mobile phone. The online settings required that the participants and their parents could only access the questionnaire after providing their consent to participate in the study. In this manner, all questionnaires were submitted with the consent of both participants and their parents. Participants were informed that they could withdraw from the study unconditionally at any time before submitting the questionnaire. The study was approved by the ethics committee of the authors’ affiliated university.

A convenience sampling strategy was adopted due to practical and financial constraints in accessing the entire target population (i.e., global football talents). Consequently, the findings may have limited generalisability beyond the sampled group, which is characterised by distinct social and cultural contexts. A total of 551 football talents from football specialised schools in 15 provinces of China completed the questionnaire. These participants were all in the talent development phase, as they had gone through talent identification tests and were identified as football talents by their schools. The samples were randomly split into two groups. The first split group (n = 300) was used for data analysis in this phase. The mean age of this group was 14.52 (SD = 1.82), with a range of 11 to 17 years old, which ensures that the participants are positioned at the developmental, rather than elite stage. Demographic information indicates that 71.00% of the sample (n = 210) were males, 28.67% (n = 86) were females, and four participants did not indicate their gender, representing 1.33% of the sample. All participants were declared to have received systematic football training (i.e., at least four football training sessions lasting more than 90 min per week).

Data analysis

Data analysis was processed by SPSS 29.0. Firstly, a preliminary analysis was conducted. Those questionnaires filled out carelessly were removed, such as unchanged responses (e.g., 7, 7, 7, …) and answered too quickly (e.g., an average response time of less than two seconds per question), as such rapid and uniform responses are unlikely to reflect a genuine understanding of the problem.^44,45 These exclusion criteria are widely used in survey research to reduce measurement error and enhance the reliability and validity of subsequent results analyses.^44,46 In addition, outliers were defined as cases with Mahalanobis distance greater than the chi-square critical value χ²(df) = 22.46 (p < 0.001).⁴⁵ Moreover, the univariate normality of each item was examined using skewness and kurtosis values, which should fall within −2.00 to 2.00 and −7.00 to 7.00, respectively.⁴⁷ The common method variance was evaluated using the Harman single-factor test.⁴⁸ If a single factor accounts for less than 50% of total variance, common method bias is unlikely to distort findings.⁴⁴ Next, the internal consistency of the measures was assessed. A factor is considered internally consistent if its Cronbach's alpha is over 0.70. In addition, item-to-total correlations of the items were calculated to inspect problematic items (<0.50).⁴⁹

Finally, the EFA was executed. Kaiser-Meyer-Olkin (KMO) and Bartlett's test were checked to assess data suitability for EFA. A value of KMO over 0.50 and a Bartlett's test p-value under 0.05 are considered adequate.⁴⁹ Considering the relatively small sample size (e.g., ≤300) and the exploratory stage of the factor structure, the principal axis factoring extraction method was thought to generate a more stable factor extraction result.⁵⁰ In addition, the six dimensions in this study were expected to be related (e.g., football coach and football training), and thus an oblique rotation was applied. Dimensions with eigenvalues greater than one were retained. The item purifying criteria were displayed in Table 1.^51–53

Table 1.

The item exclusion criteria and justification.

	Exclusion criteria	Excluded items (see Table 2)	Justifications for the criteria
1	An item loading on its primary dimension should be greater than 0.40.	FC2, FC8, FT1, FT2, FT3, FT6, FT7, FT9, FS2, FS4	Ensure the explanatory power and measurement validity of the items for the factors.
2	An item loading on an alternative dimension should be lower than 0.30.	FC1, FC5, FC9, FCP7, AS8	Avoid cross-loading that leads to unclear item attribution and confusion in factor structure.
3	There should be at least a 0.20 loading difference between their primary dimension and alternative dimensions.	FC6	Avoid cross-loading that leads to unclear item attribution and confusion in factor structure.
4	One dimension should comprise at least three items	N/A	Ensure the statistical identifiability and measurement stability of the factor structure.

Results: Preliminary analysis for EFA

The dataset had no missing values. Nonetheless, sixteen questionnaires that were filled out carelessly were removed. Accordingly, 284 samples were retained for further outlier detection, and 14 out of 284 samples were identified as outliers whose chi-square critical value was larger than 22.46 (p < 0.001) and thus eliminated, leaving 270 samples for later analysis. All items’ skewness (ranged from −2.00 to −0.20) and kurtosis (ranged from −1.39 to 1.53) values were within the acceptable range, indicating that all items were univariate and normally distributed. The common method variance was assessed using the Harman single-factor test. The single factor explained 43.07% of the variance, which is below the threshold 50%, indicating that common method bias is unlikely to distort findings.

The Cronbach's alpha values of the six dimensions ranged from 0.86 to 0.95, supporting the internal consistency of their measures. However, two items (“I receive extra lessons when I miss school for football training or competitions,” academic support, and “My parents help me make decisions about football,” family support) were found to have low item-to-total correlations (0.40 and 0.34, respectively). Thus, these two items were removed, and 53 items (see Table 2) were retained for further EFA.

Table 2.

The results of exploratory factor analysis (n = 270).

Item wording		Factor loading
Item wording		1	2	3	4	5	6	7	8
Football coach (α = 0.95)
FC1	My coach is good at helping me to understand my strengths and weaknesses in my sport.	0.28	0.09	−0.27	0.04	−0.35	0.11	0.44	−0.06
FC2	The feedback I get from my coaches almost always relates directly to my goals.	0.33	0.04	−0.18	0.23	−0.08	−0.12	0.13	−0.08
FC3	My coach demonstrates good football skills.	0.03	0.02	0.11	0.25	0.00	−0.00	0.72	0.05
FC4	My coach encourages me to have big dreams.	0.10	0.07	−0.20	0.09	−0.11	−0.08	0.47	−0.04
FC5	My coach manages the football team well.	0.25	0.06	−0.22	0.11	−0.40	0.09	0.43	−0.04
FC6	My coach makes sufficient arrangements before competitions.	0.08	−0.04	−0.03	0.29	−0.09	−0.26	0.48	−0.03
FC7	My coach emphasises good sportsmanship (e.g., respect for opponents and fair play, etc.).	−0.01	0.01	−0.06	0.02	0.13	−0.09	0.81	−0.04
FC8	My coach makes critical decisions during football competitions (e.g., changing tactics and substitutions, etc.).	0.04	−0.01	−0.26	0.23	−.032	−0.11	0.35	0.10
FC9	My coach helps me build up confidence.	0.19	0.04	−0.27	0.07	−0.38	0.03	0.42	−0.00
FC10	My coach can give me timely and effective feedback during football training.	0.24	0.05	−0.03	0.17	−0.19	−0.05	0.52	−0.08
Interaction with teammates (α = 0.95)
IWT1	My teammates and I praise each other when we play football well.	0.86	0.02	−0.02	−0.01	−0.01	−0.05	0.02	0.06
IWT2	My teammates and I encourage each other when we make mistakes in training or competition.	0.80	0.01	0.11	0.06	0.13	−0.07	0.01	−0.04
IWT3	My teammates and I enjoy listening to each other.	0.90	0.02	−0.01	−0.00	−0.10	0.10	−0.01	−0.10
IWT4	I share the same goals with my teammates.	0.48	0.03	−0.10	0.014	−0.08	−0.20	−0.05	0.05
IWT5	My teammates and I are willing to help each other to develop football skills.	0.79	0.05	−0.03	0.13	−0.07	−0.05	−0.03	0.04
IWT6	My teammates are friendly to me.	0.73	−0.03	−0.13	0.10	−0.04	0.00	−0.03	0.11
IWT7	My teammates and I are happy to help each other.	0.61	−0.01	−0.03	−0.01	0.26	−0.18	0.19	−0.01
IWT8	My teammates and I trust each other during training or competition.	0.89	0.04	0.04	0.01	0.03	0.06	0.07	−0.04
Football training (α = 0.90)
FT1	My training is specifically designed to help me develop effectively in the long term.	0.33	0.17	0.02	0.07	−0.05	−0.35	0.10	−0.26
FT2	My training programmes are developed specifically to my needs.	0.18	0.27	−0.02	0.06	−0.11	−0.09	0.08	−0.37
FT3	The football training that I participate in is challenging.	0.30	0.05	−0.27	−0.05	−0.24	−0.32	0.08	−0.08
FT4	The football training that I attend can be flexibly adjusted in time for emergencies (e.g., when faced with extreme weather and injuries).	0.15	0.08	−0.01	0.16	0.22	−0.52	0.15	0.13
FT5	My football team has a scientific training schedule (e.g., weekly, monthly and annual plans).	0.02	0.09	−0.07	0.13	0.06	−0.52	0.14	−0.20
FT6	My football skills are improved through football training.	0.36	0.01	−0.16	−0.10	−0.03	−0.37	0.15	0.07
FT7	My training content is clear to understand.	0.19	0.12	−0.23	0.17	−0.04	−0.36	0.07	0.23
FT8	My training plan incorporates a variety of physical preparations such as fitness, flexibility, agility, coordination, balance, strength training, etc.	0.15	0.09	−0.11	0.13	−0.12	−0.56	0.03	−0.03
FT9	My training load is appropriate.	0.23	0.12	0.06	0.13	0.09	−0.38	0.20	0.03
Football competition (α = 0.92)
FCP1	I can get good-quality competition experience at the level I require.	0.13	0.03	0.10	0.63	−0.25	−0.08	−0.01	−0.13
FCP2	The referees in the football competitions I play are professional.	0.02	0.06	−0.08	0.62	−0.10	0.12	0.14	0.12
FCP3	The schedule of the football competitions I participate in is properly arranged.	0.13	0.07	0.01	0.55	−0.03	−0.06	0.17	−0.16
FCP4	The football competitions I play in were fair.	0.11	−0.01	−0.03	0.61	0.13	0.15	0.20	0.12
FCP5	The ground conditions for the football competitions I participate in are normally good.	0.00	−0.01	0.02	0.75	0.12	−0.10	0.00	−0.09
FCP6	The football competitions I participate in are well organised.	0.02	0.01	−0.03	0.82	0.09	−0.05	−0.00	0.02
FCP7	The football competitions I play receive a lot of social attention.	−0.01	0.10	−0.04	0.44	0.01	−0.11	0.05	−0.35
FCP8	Our football team has many opportunities to compete.	0.15	0.03	−0.03	0.41	−0.20	−0.10	0.17	−0.20
FCP9	The football competitions I am involved in are competitive.	0.14	0.01	−0.13	0.53	−0.07	−0.24	−0.05	0.08
Academic support (α = 0.86)
AS1	I receive extra guidance on my assignments from my teacher.	0.00	0.79	0.02	−0.01	0.00	−0.01	0.06	0.02
AS2	My curriculum schedule can be adjusted according to my football training and competitions.	−0.11	0.58	−0.08	0.11	−0.05	0.01	−0.07	−0.29
AS3	My teachers allow me to have more flexibility in handing in my assignments.	0.06	0.71	−0.01	−0.05	0.16	0.10	0.02	−0.08
AS4	When I cannot take exams due to football training and competitions, make-up exams can be arranged for me.	−0.03	0.59	−0.08	0.12	0.08	0.02	−0.15	−0.13
AS5	There is someone to coordinate when training, competition and schoolwork conflict.	−0.05	0.56	−0.12	−0.04	−0.04	−0.10	0.05	−0.01
AS6	My classmates are willing to help me with my assignments.	0.08	0.72	0.03	0.03	−0.10	−0.11	−0.01	0.29
AS7	The football coach and teacher often communicate with each other about my academic performance.	0.08	0.68	0.08	−0.01	−0.03	0.01	0.08	−0.08
AS8	When I am absent from class, my classmates and teachers help me take class notes or share me their class notes.	0.05	0.60	−0.03	−0.00	−0.07	−0.08	0.02	0.32
Family support (α = 0.87)
FS1	My parents encourage me to take part in football.	−0.04	−0.06	−0.50	0.03	0.17	−0.26	0.13	−0.18
FS2	My family members always come to watch me play football.	0.06	0.08	−0.37	0.18	0.41	0.03	0.01	−0.03
FS3	My family gives me encouragement and comfort when I face setbacks in football.	0.04	0.04	−0.75	0.02	−0.21	−0.13	−0.06	0.04
FS4	My parents give me sufficient financial support for football (e.g., purchasing football equipment, transportation expenses, competition registration fees, nutrition products, etc.)	−0.00	0.10	−0.34	−0.08	0.27	−0.20	0.32	0.02
FS5	My parents praise me when I do a good job in football.	0.28	−0.03	−0.55	0.07	0.27	0.07	0.01	−0.04
FS6	My parents make suggestions when I am uncertain.	−0.01	0.06	−0.70	0.08	−0.15	−0.11	−0.03	0.12
FS7	My parents give me good advice.	0.01	0.07	−0.72	0.08	0.01	−0.02	0.06	−0.03
FS8	My family has a harmonious and supportive atmosphere.	0.01	0.07	−0.59	0.07	0.12	−0.02	0.20	−0.05
FS9	My parents don't put too much pressure on me.	0.08	0.25	−0.56	0.02	0.03	0.23	0.13	−0.04

Note: The retained items’ loadings were shown in bold, and the removed items were highlighted in grey.

Results: Exploratory factor analysis

The execution of EFA was supported by the value of KMO (0.94) and Bartlett's sphericity test (p < 0.001). The total variance explained showed that eight dimensions were initially identified, accounting for 69.32% of the total variance. The pattern matrix (see Table 2) shows how different items are loaded on different dimensions. After items purifying, six dimensions, comprising a total of 37 items, were retained in this phase (see Table 2 for details). Two dimensions (i.e., Dimension 5 and Dimension 8) were dropped because there were fewer than three items with factor loadings above 0.40.

Phase 3: Validating psychometric properties using confirmatory factor analysis

Data collection and participants

The same data collection procedures were followed in Phase 2. The second split group of data (n = 251) was used for this phase. The mean age of the sample was 14.25 (SD = 1.86), with an age range of 11 to 17. Regarding the demographic information, 66.53% of participants were male (n = 167), which outnumbered female participants (n = 81; 32.27%), while three participants (1.20%) did not mention their gender.

Measures

The measures in this phase were derived from Phase 2, which resulted in a six-dimensional scale with 37 items. The six dimensions in the measure were football coach (four items), interaction with teammates (eight items), football training (three items), football competition (eight items), academic support (seven items), and family support (seven items).

Data analysis

The preliminary analysis was processed by SPSS 29.0. The questionnaires filled out carelessly, outliers and item deviations, along with their skewness and kurtosis values, were identified using the same techniques employed in Phase 1. In addition, the common method variance was assessed by the Harman single-factor test and the common latent factor test (CLF).^48,54 The common method bias would not be a problem when (a) a single factor explains less than 50% of the total variance and (b) there are no significant differences in relevant model fit indexes between the models before and after the common latent factor is involved.

CFA was conducted using AMOS 29.0. Testing the goodness of fit for the model is a primary step in CFA. The goodness of fit was assessed using various fit indices, such as χ²/df, comparative fit index (CFI), Tucker-Lewis index (TLI), root mean square error of approximation (RMSEA), and standardised root mean square residual (SRMR). The cut-offs for acceptable model fit index values are as follows: χ²/df < 3.00, CFI ≥ 0.90, IFI ≥ 0.90, TLI ≥ 0.90, RMSEA ≤ 0.08, and SRMR ≤ 0.08.^47,54,55 Reliability was assessed using composite reliability (CR) value, which is considered acceptable when it exceeds 0.70.⁴⁸ Furthermore, construct validity was evaluated through convergent validity and discriminant validity. Convergent validity is considered satisfactory when the average variance extracted (AVE) is at least 0.50.⁵³ Another indicator of convergent validity is that all items should have factor loadings greater than 0.707 on their respective factors.⁵⁶ Discriminant validity was assessed by ensuring that correlation coefficients of each pair of factors are less than 0.85 and that no correlation coefficients exceed the square root of the respective AVEs.⁵⁴

Results: Preliminary analysis for CFA

The dataset had no missing values. Forty-one samples were removed from the second data group due to insincere responses. Ten samples were identified as outliers based on their Mahalanobis distances (p < .001). Thus, 200 samples were left for further analysis after eliminating outliers. Four items (FC4, FS1, FS3, and FS8) were further identified to deviate from normality and were removed, leaving 33 items for the subsequent CFA procedures.

In terms of common method variance. The results showed that the single factor explained 45.10% of the variance, which was acceptable.⁴⁸ Subsequently, the results of CLF indicated that the common method bias would not be a problem (i.e., CFI = 0.93, IFI = 0.93, TLI = 0.92, RMSEA = 0.07, SRMR = 0.06 in the original model, and CFI = 0.94, IFI = 0.94, TLI = 0.93, RMSEA = 0.06, SRMR = 0.05 in the model with CLF).⁵⁴

Results: Confirmatory factor analysis

Table 3 presents both the global and internal model fit results. The results indicated that the 33-item scale had an acceptable overall model fit to the data: χ²/df = 1.84, CFI = 0.93, IFI = 0.93, TLI = 0.92, RMSEA = 0.07 and SRMR = 0.06. The CR values for the six dimensions ranged from .86 to .96, reflecting good internal consistency.⁴⁷ Regarding convergent validity, the AVE values for the six dimensions ranged from 0.51 to 0.75, supporting convergent validity of the measures.⁵⁴ However, three items, AS2 (0.69), AS5 (0.68), and AS7 (0.62), in the academic support dimension were found to have low factor loadings below the 0.707 cut-off and were thus removed. Following their removal, the CR of the academic support dimension decreased from 0.87 to 0.77, while the AVE increased from 0.51 to 0.53; both values remain within acceptable ranges. Regarding discriminant validity, the correlation coefficients of each pair of dimensions were below 0.85, and no correlation coefficients exceeded the square root of AVEs, indicating that these six dimensions were distinct (see Table 4). Consequently, a six-dimensional scale with 30 items was concluded.

Table 3.

The confirmatory factor analysis result (n = 200).

Items		Factor loadings
Football coach
FC3	My coach demonstrates good football skills.	0.78
FC7	My coach emphasised good sportsmanship (e.g., respect for opponents and fair play, etc.).	0.90
FC10	My coach can give me timely and effective feedback during football training.	0.89
Interaction with teammates
IWT1	My teammates and I praise each other when we play football well.	0.90
IWT2	My teammates and I encourage each other when we make mistakes in training or competition.	0.87
IWT3	My teammates and I enjoy listening to each other.	0.88
IWT4	I share the same goals with my teammates.	0.73
IWT5	My teammates and I are willing to help each other to develop football skills.	0.88
IWT6	My teammates are friendly to me.	0.89
IWT7	My teammates and I are happy to help each other.	0.85
IWT8	My teammates and I trust each other during training or competition.	0.90
Football training
FT4	The football training that I attend can be flexibly adjusted in time for emergencies (e.g., when faced with extreme weather and injuries).	0.83
FT5	My football team has a scientific training schedule (e.g., weekly, monthly and annual plans).	0.74
FT8	My training plan incorporates a variety of physical preparations such as fitness, flexibility, agility, coordination, balance, strength training, etc.	0.87
Football competition
FCP1	I can get good-quality competition experience at the level I required.	0.87
FCP2	The referees in the football competitions I played in were professional.	0.78
FCP3	The schedule of the football competitions I participated in was properly arranged	0.81
FCP4	The football competitions I played in were fair.	0.83
FCP5	The ground conditions for the football competitions I participate in are normally good.	0.84
FCP6	The football competitions I participate in are well organised.	0.89
FCP8	Our football team has many opportunities to compete.	0.83
FCP9	The football competitions I am involved in are competitive.	0.78
Academic support
AS1	I receive extra guidance on my assignments from my teacher.	0.77
AS2	My curriculum schedule can be adjusted according to my football training and competitions.	0.69
AS3	My teachers allow me to have more flexibility in handing in my assignments.	0.71
AS4	When I cannot take exams due to football training and competitions, make-up exams can be arranged for me.	0.75
AS5	There is someone to coordinate when training, competition and schoolwork conflict.	0.68
AS6	My classmates are willing to help me with my assignments.	0.77
AS7	The football coach and teacher often communicate with each other about my academic performance.	0.62
Family support
FS5	My parents praise me when I do good job in football.	0.74
FS6	My parents make suggestions when I am uncertain.	0.75
FS7	My parents give me good advice.	0.76
FS9	My parents don’t put too much pressure on me.	0.88

Note: The low-loading items were highlighted in grey.

Table 4.

The result of CR, AVE, and discriminant validity test.

Factors	1	2	3	4	5	6	CR	AVE
1 Football coach	0 . 86						0.89	0.74
2 Interaction with teammate	0.83	0.87					0.96	0.75
3 Football training	0.80	0.69	0.81				0.86	0.66
4 Football competition	0.77	0.76	0.63	0.83			0.95	0.69
5 Academic support	0.36	0.29	0.43	0.43	0.71		0.87	0.51
6 Family support	0.49	0.43	0.55	0.60	0.22	0.79	0.86	0.62

Note. The square root of AVE values appears on the matrix diagonal.

Discussion

This study results in a six-dimensional, 30-item scale with sound psychometric attributes. Compared to previous TDE scales such as the TDEQ, TDEQ-5, and the Sport Talent Development Environment Questionnaire, this newly developed football-specific TDE scale incorporates distinct environmental factors that are particularly relevant to football talent development.^4–6 For example, the inclusion of teammate interaction captures the collective and team-based nature of football, offering a more context-specific assessment. Although the Sport Talent Development Environment Questionnaire⁶ includes a team condition factor, it primarily reflects general team characteristics in a more generic tone and does not fully capture the interpersonal dynamics unique to football. The use of more detailed and targeted items in this scale allows for better identification and resolution of context-specific issues in practice. Moreover, prior TDE scales have predominantly focused on the unidimensional development of athletes, primarily in terms of sporting performance, while overlooking other key developmental domains. In contrast, the present scale incorporates academic support as a critical environmental component, thereby providing a more holistic framework for understanding and fostering talent development in football. The rest of this section discusses these results according to the three scale development phases in this study.

In Phase 1, the excluded items were found to contain terms related to different dimensions, which may cause ambiguity. Specifically, for example, the item (“My teammates and I maintain positive communication during football competitions”) may simultaneously reflect football talents’ perception of both interaction with teammates and football competition. This violates the one-dimensional nature of an item that should be maintained in measurement.⁴⁰ The removal of these items was considered and negotiated to ensure that their exclusion would not compromise the content validity of the scale. For instance, there are other items (e.g., IWT1, IWT2, and IWT8) within the interaction with teammates dimension that can represent related content to the removed item (“My teammates and I maintain positive communication during football competitions”).

In Phase 2, 18 items were removed, leaving a six-dimensional scale with 37 items. The Cronbach's alpha values for the six dimensions ranged from 0.87 to 0.95, indicating satisfactory reliability. However, two items were removed due to low correlations with the other items in their factors. The low item-total correlation of the academic support item (“I receive extra lessons when I miss school for football training or competitions”) may be attributed to the fact that some schools have adjusted their schedules to accommodate students’ training and competition, thereby reducing the frequency of absences. Another reason for the low internal consistency of this item may stem from the phrase, “football training and competition”, which could lead football talents to perceive it as related to the dimensions of football training and football competition, rather than purely academic support. Similarly, the item (“My parents help me make decisions about football”) in family support may not align well with other items in that dimension due to variations in parenting style.

Specifically, different parenting styles may interfere to varying degrees with their children's decision-making.⁵⁷ For instance, parents with an autocratic parenting style exert a high level of control over their children's behaviour.⁵⁸ In contrast, those with a democratic parenting style are generally more permissive.⁵⁸ Consequently, parents with a democratic style may be less inclined to actively guide or support decisions related to football participation, compared to their autocratic counterparts. This difference in parental involvement may lead to divergent interpretations among football talents regarding the item.

In addition, the pattern matrix in Phase 2 revealed that 16 items were problematic and removed because of improper loading (e.g., low factor loading and cross-loading), and 37 items were left. These 37 items were not distributed in relatively equal numbers across all dimensions. Firstly, regarding the football coach dimension (Dimension 7), while FC1, FC5, FC8, and FC9 loaded properly on football coach, they also cross-loaded over 0.30 on Dimension 5. This suggests that they may also make sense on another potential dimension related to football coaches. Previous research has pointed out the multi-dimensional nature of coach-related structures (e.g., coach leadership and coach competency).⁵⁹ However, the various facets of the coach's role were not clearly specified when designing the items in the football coach dimension. This leads to the multi-dimensionality of many items in this dimension, leading to cross-loading issues. Next, three items (FT1, FT3, and FT6) in the football training dimension were removed because they were slightly lower than 0.40 for football training, but they also cross-loaded above 0.30 on Dimension 1 (i.e., interaction with teammates). This may be due to the collective nature of football. For example, football training is often organised as collective training, and football talents would not be able to enhance their skills through training without cooperation from teammates (e.g., FT6). Specifically, many complex contexts in football training, such as physical confrontation, tactical training, coordinated free kicks, and defensive coordination, arise through interactions among teammates. In these contexts, players’ perceptions of their interactions with teammates might significantly shape how they interpret and engage with training. This interdependence likely accounts for the high factor loadings observed for these items in the interaction with teammates dimension.

In Phase 3, three removed items, AS2, AS5, and AS7, were eliminated due to their low factor loadings, all of which belonged to academic support. In terms of AS2 (“My curriculum schedule can be adjusted according to my football training and competitions”) and AS5 (“There is someone to coordinate when training, competition, and schoolwork conflict”), their low factor loadings may be attributed to the diverse academic support policies implemented for football talents across different schools and regions in China. For example, a case study in Sichuan Province indicates that an academic support system has been established in local schools to develop and support sport talents.⁶⁰ This system precisely organises training and study schedules, ensuring that athletes’ training sessions do not conflict with their classroom learning.⁶⁰ In contrast, another case study in Beijing reveals that sport talents in schools was not provided such flexible academic timetable, which results in training activities taking up a large portion of their class time.⁶¹ Furthermore, regional differences in the scheduling of football training and competitions likely influence how football talents interpret these items. For example, football competitions usually take place on weekends or after school hours in many provinces in China, while these events are still scheduled during times that conflict with school hours in other provinces.⁶² Thus, AS2 and AS5 were omitted. Regarding AS7 (“The football coach and teacher often communicate with each other about my academic performance”). Its low factor loading may be attributed to the distinct responsibilities of football coaches and schoolteachers (e.g., football coaching for coaches and academic teaching for teachers). Monitoring the academic performance of football talents does not appear to be a primary responsibility of football coaches. Accordingly, football coaches and schoolteachers may have limited opportunities or interest in collaborating to discuss football talents’ academic development. Consequently, football talent may lack awareness or perceptions of such cooperation. In line with this, AS7 was also deemed unsuitable and removed from this dimension. The removal of these three items seems to have a limited impact on the construct validity. Although the related psychometric attributes of the academic dimension show slight fluctuations, the overall convergent and discriminant validity of its measures remain intact. Specifically, the content of AS2 and AS5, reflecting contradiction between academic and training or competition, can also be reflected by AS3 (“My teachers allow me to have more flexibility in handing in my assignments”) and AS4 (“When I cannot take exams due to football training and competitions, make-up exams can be arranged for me”). In contrast, AS7 primarily focuses on communication between coaches and teachers, rather than providing direct academic support. Therefore, the removal of these three items from the academic dimension in Phase 3 does not appear to compromise the construct validity of the scale.

Theoretical implications

Compared to earlier TDE scales (4–6), the Football Talent Development Environment Scale identified different TDE dimensions and validated their psychological properties. This provides empirical evidence for researching TDE from a novel framework. Moreover, several unique dimensions and items were utilised to capture and measure the TDE in football. For example, interaction with teammates is set to measure the team atmosphere of football talent during training and competition. This is consistent with considerable evidence that has demonstrated the significance of team attributes in football.^63–65 In summary, this study develops a tool for measuring TDE in the context of football and offers initial evidence for its reliability and validity.

Practical implications

The findings provide several practical implications for football talent development. Firstly, the TDE scale can serve as a tool for policymakers to recognise and monitor key environmental dimensions influencing outcome variables (e.g., turnover, motivation, athletic identity, and sport commitment) related to football talent development. This, in turn, allows for the formulation of more precise policies to promote the development of football talent. Secondly, this scale can help football coaches optimise their coaching and management strategies for the team. Thirdly, the scale can be used to monitor the effectiveness of intervention strategies for developing football talent. Measurements can be taken on a regular basis (e.g., monthly, quarterly) to monitor the changes in the perception of football talent to different environmental dimensions. It would help determine whether relevant interventions (e.g., educational support policies, communication with parents, and optimisation of football competitions) are effective.

Limitations and future research

This study has several limitations, which should be considered for further reference and application in future work. Firstly, all participants in this study are from China, which may limit the generalisability of the research results to a broader sociocultural background. For example, differences in academic education systems, parenting styles, and sport talent development systems between China and other countries may lead to variations in interpretations of the measures, particularly in the dimensions of academic support, family support and football training. Thus, it is recommended to validate the scale further with samples from other cultural backgrounds to enhance external validity. Secondly, although sound statistical criteria were followed in the process of scale development, it has been stressed that simply following quantitative interpretations may exclude qualitatively important content.^40,66 For example, FC4 (My coach manages the football team well) was excluded after the EFA because it loaded on two dimensions higher than .40, although it reflects the important qualities of a football coach. Hence, it is suggested that follow-up interviews with football talents or open-ended questionnaire surveys be conducted to further determine the implications of removing specific item content. Finally, while certain environmental features are identified to suit the football context better, the author contends that additional intricate TDE features for football can be further explored through qualitative research in the future, and thus to further optimise the current scale.

Conclusion

This study proposes a scale specifically tailored to the TDE in football, comprising six dimensions (football coach, interaction with teammates, football training, football competition, academic support, and family support) and 30 items that align with the originally conceptualised framework. Compared to existing TDE scales, the newly developed scale more effectively captures the unique contextual features inherent to football settings. The statistical analyses demonstrate that the scale possesses sound psychometric properties, indicating its reliability and validity for assessing TDE among young football talents.

Footnotes

Acknowledgements

The authors express their gratitude to Mr Yujie Huang from the Youth Football Department of the Chinese Football Association for his assistance in data collection.

Ethical considerations

The Ethics Review Committee at Loughborough University approved our study (approval: 2023-16323-17109) on December 19, 2023. Respondents gave written consent for review and signature before starting surveys.

Consent to participate

Before the study commenced, all participants and their parents received a written consent form outlining the study's purpose. It confirmed that participation was voluntary and that they could withdraw at any time by notifying the researchers.

Consent for publication

The athletes and their parents provided informed consent for publication. All data have been anonymised.

Data availability statement

The datasets produced or analysed in this study are not publicly accessible but can be obtained from the corresponding author upon a reasonable request.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the China Scholarship Council [grant number 202308060351].

ORCID iDs

Yang Xia

Minhyeok Tak

References

Gagné

. Transforming gifts into talents: the DMGT as a developmental theory. High Abil Stud 2004; 15: 119–147.

Wang

CKJ

Pyun

. Impacts of talent development environments on athlete burnout: a self-determination perspective. J Sports Sci 2016; 35: 1838–1845.

Thomas

Gastin

Abbott

, et al. Impact of the talent development environment on the wellbeing and burnout of Caribbean youth track and field athletes. Eur J Sport Sci 2020; 21: 1–14.

Wang

CKJ

Pyun

, et al. Further development of the talent development environment questionnaire for sport. J Sports Sci 2015; 33: 1831–1843.

Martindale

RJJ

Collins

Wang

JCK

, et al. Development of the talent development environment questionnaire for sport. J Sports Sci 2010; 28: 1209–1221.

Xiang

Dong

Tengku

FTK

, et al. Structural analysis of environmental factors of sports talent development. Curr Psychol 2023; 43: 6516–6532.

Xia

Pyun

Tak

. Nurturing football talent: a conceptual framework for talent development environment. Quest 2025; 77: 1–18.

Kleinert

Ohlert

Carron

, et al. Group dynamics in sports: an overview and recommendations on diagnostic and intervention. Sport Psychol 2012; 26: 412–434.

Cheng

Chang

AYC

Doya

. Information-theoretical analysis of team dynamics in football matches. Entropy 2025; 27: 224.

10.

Bronfenbrenner

. The ecology of human development: Experiments by nature and design. Cambridge: Harvard University Press, 1979.

11.

Martindale

Collins

Daubney

. Talent development: a guide for practice and research within sport. Quest 2005; 57: 353–375.

12.

Martindale

RJJ

Collins

Douglas

, et al. Examining the ecological validity of the talent development environment questionnaire. J Sports Sci 2013; 31: 41–47.

13.

Mills

Butt

Maynard

, et al. Examining the development environments of elite English football academies: the players’ perspective. Int J Sports Sci Coach 2014; 9: 1457–1472.

14.

John

WCK

Sproule

McNeill

, et al. Impact of the talent development environment on achievement goals and life aspirations in Singapore. J Appl Sport Psychol 2011; 23: 263–276.

15.

Ivarsson

Stenling

Fallby

, et al. The predictive ability of the talent development environment on youth elite football players’ well-being: a person-centered approach. Psychol Sport Exerc 2015; 16: 15–23.

16.

Alfermann

Lobinger

Nesges

, et al. German Version of the talent development environment questionnaire (TDEQ-5). Ger J Exerc Sport Res 2022; 53: 59–69.

17.

Brazo-Sayavera

Olivares

Andronikos

, et al. Spanish version of the talent development environment questionnaire for sport: cultural adaptation and initial validation. PLoS One 2017; 12: e0177721.

18.

Gangsø

Aspvik

Mehus

, et al. Talent development environments in football: comparing the top-five and bottom-five-ranked football academies in Norway. Int J Environ Res Public Health 2021; 18: 1–7.

19.

Gesbert

Crettaz Von

Hauw

Reviewing the role of the environment in the talent development of a professional soccer club. PLoS One 2021; 16: e0246823.

20.

Mitchell

Gledhill

Shand

, et al. Players’ perceptions of the talent development environment within the English premier league and football league. Int Sport Coach J 2021; 8: 1–7.

21.

Yin

Bin Abdullah

Hassan

MFBH

. The key factors of talent development environment for campus football athletes in China. Cogent Soc Sci 2024; 10: 2435190.

22.

Henriksen

Stambulova

Roessler

. Holistic approach to athletic talent development environments: a successful sailing milieu. Psychol Sport Exerc 2010; 11: 212–222.

23.

Hargreaves

Bate

. Skills and strategies for coaching soccer. Champaign: Human Kinetics, 2009.

24.

Smith

Quested

Appleton

, et al. Observing the coach-created motivational environment across training and competition in youth sport. J Sports Sci 2016; 35: 149–158.

25.

Hodge

Gucciardi

. Antisocial and prosocial behavior in sport: the role of motivational climate, basic psychological needs, and moral disengagement. J Sport Exerc Psychol 2015; 37: 257–273.

26.

Pacewicz

Smith

. Teammate relationships, loneliness, and the motivational and well-being experiences of adolescent athletes. J Clin Sport Psychol 2021; 16: 1–20.

27.

Gabbett

. Debunking the myths about training load, injury and performance: empirical evidence, hot topics and recommendations for practitioners. Br J Sports Med 2018; 54: 58–66.

28.

Graham

Cormack

Parfitt

, et al. Relationships between model estimates and actual match-performance indices in professional Australian footballers during an in-season macrocycle. Int J Sports Physiol Perform 2018; 13: 339–346.

29.

Sampaio

Maçãs

. Measuring tactical behaviour in football. Int J Sports Med 2012; 33: 395–401.

30.

Williams

Ward

, et al. Domain specificity, task specificity, and expert performance. Res Q Exerc Sport 2008; 79: 428–433.

31.

Larsen

Alfermann

Henriksen

, et al. Successful talent development in soccer: the characteristics of the environment. Sport Exerc Perform Psychol 2013; 2: 190–206.

32.

Turman

. Parental sport involvement: parental influence to encourage young athlete continued sport participation. J Fam Commun 2007; 7: 151–175.

33.

Fraser-Thomas

Côté

Deakin

. Understanding dropout and prolonged engagement in adolescent competitive sport. Psychol Sport Exerc 2008; 9: 645–662.

34.

Churchill

. A paradigm for developing better measures of marketing constructs. J Mark Res 1979; 16: 64–73.

35.

Donohue

Miller

Crammer

, et al. A standardized method of assessing sport specific problems in the relationships of athletes with their coaches, teammates, family, and peers. J Sport Behav 2007; 30: 375–397.

36.

Malecki

Elliott

. Adolescents’ ratings of perceived social support and its importance: validation of the student social support scale. Psychol Sch 1999; 36: 473–483.

37.

Myers

Feltz

Maier

, et al. Athletes’ evaluations of their head coach's coaching competency. Res Q Exerc Sport 2006; 77: 111–121.

38.

Myers

Chase

Beauchamp

, et al. Athletes’ perceptions of coaching competency scale II-high school teams. Educ Psychol Meas 2009; 70: 477–494.

39.

Hinkin

Tracey

Enz

. Scale construction: developing reliable and valid measurement instruments. J Hosp Tour Res 1997; 21: 100–120.

40.

Hinkin

. A brief tutorial on the development of measures for use in survey questionnaires. Organ Res Methods 1998; 1: 104–121.

41.

Worthington

Whittaker

. Scale development research: a content analysis and recommendations for best practices. Couns Psychol 2006; 34: 806–838.

42.

Ekinci

Riley

. Validating quality dimensions. Ann Tour Res 2001; 28: 202–223.

43.

Brislin

. Back-translation for cross-cultural research. J Cross Cult Psychol 1970; 1: 185–216.

44.

Huang

Curran

Keeney

, et al. Detecting and deterring insufficient effort responding to surveys. J Bus Psychol 2011; 27: 99–114.

45.

Tabachnick

Fidell

. Using multivariate statistics. 7th ed. New Nork: Pearson, 2021.

46.

Meade

Craig

. Identifying careless responses in survey data. Psychol Methods 2012; 17: 437–455.

47.

Hair

Black

Babin

, et al. Multivariate data analysis. 8th ed. Andover: Cengage Learning Emea, 2018.

48.

Podsakoff

MacKenzie

Lee

, et al. Common method biases in behavioral research: a critical review of the literature and recommended remedies. J Appl Psychol 2003; 88: 879–903.

49.

Field

. Discovering statistics using IBM SPSS Statistics. 5th ed. Los Angeles: Sage Publications, 2017.

50.

Hair

Black

Babin

, et al. Multivariate data analysis: A global perspective. 7th ed. Upper Saddle River: Pearson Education, 2010.

51.

Watkins

. Exploratory factor analysis: a guide to best practice. J Black Psychol 2018; 44: 219–246.

52.

Howard

. A review of exploratory factor analysis decisions and overview of current practices: what we are doing and how can we improve? Int J Hum Comput Interact 2016; 32: 51–62.

53.

Tucker

Viswanathan

Walford

. The SAGE handbook of measurement. London: Sage, 2010.

54.

Kline

. Principles and practice of structural equation modeling. 4th ed. New York: The Guilford Press, 2023.

55.

Bentler

. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Modeling 1999; 6: 1–55.

56.

Fornell

Larcker

. Evaluating structural equation models with unobservable variables and measurement error. J Mark Res 1981; 18: 39–50.

57.

Maccoby

Martin

. Socialization in the context of the family: parent-child interaction. In: Mussen PH (eds) Handbook of child psychology. New York: Wiley, 1983, pp.1–101.

58.

Aunola

Stattin

Nurmi

. Parenting styles and adolescents’ achievement strategies. J Adolesc 2000; 23: 205–222.

59.

Vella

Oades

Crowe

. The application of coach leadership models to coaching practice: current state and future directions. Int J Sports Sci Coach 2010; 5: 425–434.

60.

Wang

Luo

Yan

. Research on the training of youth sports reserve talents under the integration of education and sports: a case study of Yibin, Sichuan [in Chinese]. Sport World (Scholarly) 2016; 5: 123–128.

61.

Liu

Cang

. A study on the influencing factors of the conflict between learning and training among taekwondo athletes at Beijing No. 2 high school [in Chinese]. Broad Vis Sport 2017; 7: 226–227.

62.

Liang

. The factors and countermeasures that restrict the development of campus football in China from the perspective of football fever [in Chinese]. Sports Cult Guide 2017; 1: 150–153.

63.

Duarte

Araújo

Correia

, et al. Competing together: assessing the dynamics of team–team and player-team synchrony in professional association football. Hum Mov Sci 2013; 32: 555–566.

64.

Erikstad

Martin

Haugen

, et al. Group cohesion, needs satisfaction, and self-regulated learning: a one-year prospective study of elite youth soccer players’ perceptions of their club team. Psychol Sport Exerc 2018; 39: 171–178.

65.

Marcelino

Sampaio

Amichay

, et al. Collective movement analysis reveals coordination tactics of team players in football matches. Chaos Solitons Fractals 2020; 138: 1–10.

66.

Morgado

FFR

Meireles

JFF

Neves

, et al. Scale development: ten main limitations and recommendations to improve future research practices. Psicol Reflex Crít 2017; 30: 1–20.