What Defines Early Specialization: A Systematic Review of Literature

Introduction: While practitioners and organizations advise against early specialization, the lack of a consistent and clear definition of early specialization reduces the impact of recommendations and policies in youth sport. An important first step in understanding the consequences of early specialization is establishing what early specialization is. Objectives: This PRISMA-guided systematic review aimed to determine the types, characteristics, and general content of early specialization papers within the literature, and examine how early specialization has been defined and measured in order to advance knowledge toward a clear and consistent definition of early specialization. Data sources: Four different electronic databases were searched (SPORTDiscus, Web of Science, Sports Medicine and Education Index, and Scopus). Both non data-driven and data-driven studies were included to ensure a comprehensive understanding of the literature. Eligibility Criteria: In order to be included in the review, the paper must: (a) Focus on specialization and explicitly use the term “specialization” (b) Focus on sport and athletes (c) Be papers from a peer-reviewed (d) Be in English. And finally, (e) be available in full text. Results: One thousand three hundred and seventy one articles were screened resulting in 129 articles included in the review after applying inclusion/exclusion criteria. Results indicated a clear discrepancy between key components of early specialization and the approaches used to classify early specializers. Conclusion: Future research should work toward developing a valid and reliable approach to classifying early specializers and establishing a consistent definition across studies.


INTRODUCTION
In 2016, the Canadian Lifestyle and Fitness Research Institute reported 77% of youth aged 5-19 participated in organized physical activity or sport. According to the Aspen Institute's Project Play (The Aspen Institute Project Play, 2019), 38% of children aged 6-12 participated in sport on a regular basis in 2018, based on United States government population statistics (Federal Interagency Forum on Child Family Statistics, 2020), equating to ∼9 million American children participating in sport regularly. Similarly, Australia reported 72.3% of children under the age of 15 participated in some type of sport related activity in 2019 (May, 2019), while in England 86.4% of children ages 5-15 were reported to participate in sport in 2018 (Lange, 2019). Due to the large number of youths participating in sport globally, researchers have attempted to better understand common sport pathways, and the benefits or consequences of sport participation.
One element of youth sport that has received more attention in recent years is early specialization, originally posited as athletes focusing on one sport that is practiced, trained and competed in year-round (Hill, 1993). Models of athlete development (e.g., Developmental Model of Sport Participation) (Côté and Fraser-Thomas, 2016) suggest early specialization excludes an important period of development where youth should be participating in a range of sports with the purpose of fun and enjoyment, in favor of dedication and skill acquisition in one sport. However, expertisecentered models of skill development (e.g., Deliberate Practice Framework) (Ericsson et al., 1993) suggest that individuals who begin focused practice early have an advantage over those who start later. Despite the prominence of the notion of deliberate practice in discussions of coaching and athlete development, a growing body of literature suggests early specialization is not a prerequisite of becoming an elite athlete (Soberlak and Cote, 2003;Buckley et al., 2017;Huxley et al., 2017;Black et al., 2019). Further, particular indicators of early specialization have been linked to a host of negative consequences. Researchers have found those who specialize early are at greater risk of injury, experience increased exhaustion, and are more likely to dropout than athletes who do not (Fraser-Thomas et al., 2008;Strachan et al., 2009;Bell et al., 2018c).
Over the past 20 years, at least seven major national and international sport and athletic associations, societies, federations, and organizations have released position statements advising against the practice of early specialization amongst youth athletes (e.g., American Orthopedic Society for Sports Medicine, American Academy of Pediatrics, International Society of Sport Psychology, National Association for Sports and Physical Education). Such strong consensus suggests there is clear and unambiguous evidence that early specialization is harmful and should be avoided in any context; however, further investigation indicates the evidence against early specialization may not be as robust as these position statements make it seem.
To begin, there are very few studies explicitly studying the consequences of early specialization; instead the literature is comprised heavily of review papers, commentaries, and editorials that reiterate previous findings. For example, a 2018 metaanalysis on specialization and overuse musculoskeletal injuries was comprised of only four studies (Bell et al., 2018c), suggesting an overall lack of research. More importantly, there is no standard definition of early specialization. Several researchers have emphasized the lack of a clear and consistent definition and have suggested that this inconsistency makes it unclear what exactly constitutes early specialization (Ferguson and Stern, 2014;Buckley et al., 2017). Some have defined early specialization as "year round intensive training in a single sport at the exclusion of other sports"  while others proposed "the time when the athlete defined one sport as being more important than other sports" (Moseid et al., 2019). Further complicating conceptualizations, some have suggested it is the type of participation (i.e., deliberate practice) that is a key marker of early specialization (Hendry and Hodges, 2018) while others designate early start age and early involvement in competitive sport as key parameters of early specialization . Without a consistent definition of early specialization, it is difficult to conclude early specialization is as harmful to youth as many organizations are claiming. More importantly perhaps-the lack of a clear definition of this phenomenon makes improving developmental training environments difficult given it is not clear what element of specialization (e.g., intensity, early start age, over-emphasis on winning) may be driving any negative consequences that do exist.
A recent systematic review of early specialization (DiSanti and Erickson, 2019) found that only 13 of 40 studies operationally defined "specialization." Among the few studies that provided an operational definition of specialization, the criteria used to distinguish early specializers varied considerably. Given these inconsistent criteria, athletes could be classified into different groups depending on the definitions used, raising concerning questions of reliability and validity of conclusions regarding early specialization. An important next step in determining the relationships between early specialization and developmental and performance outcomes, as well as identifying the mechanisms behind these effects, is to clearly define early specialization.
Practice and research in sport psychology is strongly influenced by policy decisions, and therefore, unlike previous reviews which have examined only data-driven studies (Fabricant et al., 2016;Bell et al., 2018c;Walters et al., 2018), this review will also include non-data driven articles. This will provide a more thorough understanding of the current state of literature (not just the state of the research) and overall understanding of the conceptions of "early specialization" in sport psychology and related fields of study. We believe this variation to the formula of systematic reviews makes this a novel approach to understanding a concept in its entirety.
It is important to note this review did not focus on scientific or measurement-related issues concerning definitions of early specialization (e.g., the implications of a yes/no dichotomy of specialization vs. a continuous measure). The necessary evidence for an empirically substantiated definition of early specialization has yet to be established and while these issues are clearly important in the study of early specialization, they were outside the scope of this review.
The aim of this review was not to come to a conclusion about the potential consequences or benefits of early specialization, as has been done in the past; the goal of this review was to gain a thorough understanding of the entire breadth of literature on the subject. As such, the objectives of this systematic review were: (a) to determine the types, characteristics, and general content of early specialization papers within the literature, and (b) to examine how early specialization has been defined and measured in the sport literature across all fields of study (e.g., biomechanics, psychology, talent development) and populations, in order to advance knowledge toward a clear and consistent definition of early specialization.

Research Protocol
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (Moher et al., 2009) was used as a guide for the exploration of literature. There is no protocol registered for this review.

Eligibility
In order to be included in the review, a priori criteria were established; specifically, the paper must: (a) Focus on specialization and explicitly use the term "specialization"; this meant that specialization had to be one of the key elements of the paper and not a footnote or added section. (b) Focus on sport and FIGURE 1 | PRISMA flow chart outlining flow of information through the different phases of the systematic review.
Frontiers in Sports and Active Living | www.frontiersin.org athletes; this ensured the focus was on sport specialization and not any other type of specialization (e.g., as it relates to medical expertise). (c) Be papers from a peer-reviewed journal rather than exclusively empirical studies; any review, commentary, editorial etc. was eligible for inclusion, in order to capture any and all definitions of early specialization and a more comprehensive picture of the current state of the literature 1 . (d) Be in English. And finally, (e) be available in full text.

Information Sources and Search Strategy
Beginning in June 2019, in consultation with a professional research librarian a rigorous search strategy was created. To identify relevant literature, thoroughly thought out search strings and key words were used within four electronic databases (i.e., SPORTDiscus, Web of Science, Sports Medicine and Education Index, and Scopus). Key words included "specialize" and "sport" as well as synonyms such as "year-round training" or "singlesport." For the keyword "early" synonyms included "youth, " "child, " and "adolescent." Various combinations of these key words were used for each of the four databases. In order to ensure studies captured all components, the connector "AND" was used, and to capture all variations, truncation was used. An example of a search string used in the Scopus database is (specialize * AND early AND sport * ). In order to get a thorough understanding of the research into early specialization, papers could be published any time before June 2019, with a final search date of August 2019.

Study Selection
The initial search resulted in 1,349 articles. An additional 22 were identified from reference lists of seminal papers, creating a total of 1,371. After duplicates were removed, 876 articles were screened. Information from all articles including title, year of publication, authors and abstract was compiled in an excel document for organization purposes. At this stage, the titles and abstracts for all articles were screened based on the above criteria, in order to determine inclusion or exclusion. If the first author was unsure, another author was consulted, and discussion continued until a decision was reached. This screening resulted in the exclusion of 725 articles, with 151 articles for full text review. Of the 151 articles read in-full, two were found to not be in English, 13 were deemed to have not focused on sport specialization, two were conference proceedings, one was not peer-reviewed and four were un-retrievable for a total of 22 studies being excluded in this step, resulting in a final total of 129 studies included in the systematic review. For a complete flow chart, see Figure 1.

Data Collection
The remaining 129 articles were then put into a new spreadsheet for data extraction. The definition used for early specialization and purpose of each study were transferred to this new file to allow for further analysis. To cover the objectives of the review, for empirical studies, additional information regarding methods 1 Book chapters were not included in the review as they do not undergo a rigorous peer-review process and they are not indexed the way journals are, therefore they would not show up in our four databases. used, sample size, country of study, age and sex of sample as well as sport studied were extracted from each paper.

Risk of Bias and Additional Analysis
Given that the objectives of this review were to determine the types, characteristics, and general content of early specialization papers within the literature, and to examine how early specialization has been defined and measured in the sport literature (i.e., not to summarize outcomes) a bias assessment was not performed. Additionally, because this was not a metaanalysis, no additional statistical analyses (e.g., meta-regression) were performed on the collected papers.

Paper Types, Characteristics, and Content
To achieve the first objective of the review, studies were first categorized based on article type (i.e., non-data-driven editorials/commentaries/reviews, systematic reviews/meta analyses, and data driven studies), to gain a better understanding of the overall composition of the literature. Of the 129 papers included in the study, 43.4% (n = 56) were non-data driven papers (i.e., editorials, reviews and commentaries) and 3.8% (n = 5) were systematic reviews. The data-driven studies (n = 68; 52.7%) were further divided into those that explicitly included specialization in the purpose (i.e., specialization specific; n = 48; 37.2%) and those that did not include specialization in the purpose but met all criteria to be included in the review (i.e., specialization general; n = 20; 15.5%). Table 1 provides an overview of the characteristics of non-data driven papers which included 36 reviews, eight commentaries, seven editorials, and five position/consensus statements. Among this category, the areas of focus were injury, talent development and policy (for full breakdown see Table 1). The five position statements were from five different organizations, which were all either against or relatively neutral toward early specialization, indicating a negatively skewed perception of early specialization. Table 2 presents the data from the systematic reviews. The number of studies included in each review ranged from 3 to 40. Injury was the main focus of these reviews (n = 3) while the remaining two were multidisciplinary in nature.
The characteristics of the data-driven studies are provided in Table 3. Within the 48 studies in the specialization specific category (i.e., explicitly included specialization in the purpose), there were a variety of outcomes studied. Injury studies (n = 14) were the most prominent and were often epidemiological examinations of rates or risk of injury in early specializers. Specialization characteristics such as age of specialization or prevalence were also heavily studied (n = 10). Talent development studies (n = 9) focused on the training activities of elite athletes, often comparing them to their less successful peers. Psychological outcomes (e.g., burnout, mental toughness) and physical outcomes (e.g., landing error, anterior y balance performance), were less heavily studied (i.e., n = 5 and n = 4, respectively). The least studied areas in relation to early specialization were later physical activity (n = 3), and skill   transfer to other sports (n = 1). There were also single studies that considered how specialization affected (a) ability to learn basketball skills in non-basketball players (Santos et al., 2017), and (b) health related quality of life (Patel and Jayanthi, 2018). The average sample size of studies in this category was 499.7 with a range from 1 to 3090. The studies were comprised of retrospective (n = 16), cross-sectional (n = 15), case control (n = 8), descriptive epidemiological (n = 4), longitudinal (n = 1), prospective (n = 1), case study (n = 1), case report (n = 1), and a single cohort studies. Studies came from a total of 11 different countries and there was a large variety of individual and team sports examined. Finally, specialization general studies (i.e., did not explicitly include specialization in the purpose) were largely comprised of talent development studies (n = 16). These studies generally focused on the developmental activities of athletes who became elite or differences between elite and non-elite athletes, which meant that while early specialization was a focus in the article, the actual purpose of the paper was not necessarily to advance understanding of early specialization. The average sample size was 314 with a range from 12 to 1,558. Studies were retrospective (n = 17) or cross-sectional (n = 3) with one having a combined longitudinal/retrospective design. Participants were generally either males only or mixed samples of males and females, with only one study examining females only. Lastly, data was collected in nine different countries.

Definitions and Measures
As the second objective of the review was to examine how early specialization has been defined and measured, this section focuses on the conceptual and operational definition of early specialization as well as the approaches used to determine early specializers, across all types of papers. In their 2019 scoping review, DiSanti and Erickson (2019) established that year-round intense training in a single sport at the exclusion of other sports was the most commonly used definition in empirical studies. In line with this review of empirical studies, the four key components of this definition were used as a starting point for our analysis (i.e., year-round, intense training, single sport, and exclusion of other sports). Deliberate practice was also added as a definition component, as the previously mentioned Developmental Model of Sport Participation (Côté and Fraser-Thomas, 2016) suggests deliberate practice is also a key indicator of early specialization. Finally, as this review focuses on early specialization definitions were also coded depending on whether they included any information regarding an age threshold.
Definitions were extracted from all 129 articles and coded for each of the six individual components (i.e., year-round, intense training, single sport, exclusion of other sports, deliberate practice, and age threshold) of early specialization, which are presented in Table 4. Just over 20% (i.e., 20.9%, n = 27) of the articles included the initial four-component definition of early specialization. The most frequent individual component of early specialization was single sport participation (i.e., 73.6%, n = 95), while the least frequent individual component was    high amounts/volume of deliberate practice at 9.3% (n = 12). Additionally, 44.2% (n = 57) included year-round training, 41.9% (n = 54) used exclusion of other sports and 31.8% (n = 41) considered intense training to be a key facet of early specialization. A particularly interesting finding was the lack of distinction between early specialization vs. sport specialization; only 30.2% (n = 39) of the papers included some mention of early or young age as part of the definition for early specialization. Finally, 17.1% (n = 22) of the 129 papers discussed and focused on early specialization yet had no explicit definition of early specialization. While definitions lay the foundation for understanding components of early specialization, it follows that studies in turn must classify athletes according to these definitions. Further analysis was conducted on the measures used in the 48 data-driven specialization specific studies in order to better understand how researchers classified athletes as early specializers. A key step to measuring early specialization is determining what is meant by early, yet only 25 studies (52.1%) included a measure of age in their screening tool. Of those, 56% (n = 14) used "before the age of 12" as the cut-off for early specialization. To determine specialization status in the empirical studies, 18 different approaches or strategies were employed. It should be noted that while different indicators of early specialization were used, some of the constructs overlap. Fifteen (31.3%) of the 48 studies used the "Sport Specialization Scale" by Jayanthi et al. (2015), 11 (22.9%) used a single item question (e.g., "Did you specialize before high school, yes or no?") while 10 (20.8%) collected a full developmental history of the athlete (e.g., hours in each sport, practice history, and number of sports at different ages). For a complete list of the different approaches used, see Table 5.

DISCUSSION
Early specialization is currently a "hot button" topic in athlete development research in particular and sport science more generally. Our review suggests much of the discussion in this area is driven by non-data driven, commentaries, Limiting sport participation to a single sport with the specific goal of guiding the child athlete to top achievement Yes Yes Baker et al. (2005) A shift from activities that are play like in nature to more structured and effortful training activities. In addition, the number of sport-specific training hours dramatically increases from initial involvement in the sampling years to committed involvement in the investment years Baker et al. (2009) Four specific parameters: early start age in sport; early involvement in one sport (as opposed to participating in several sports); early involvement in focused, high intensity training; and early involvement in competitive sport

Yes Yes Yes
Baker (2003) To limit their childhood sport participation to a single sport, with a deliberate focus on training and development in that sport

Yes Yes
Beese et al. (2015) Year-round participation in a single sport to the exclusion of other sports and activities    (2015) Focused involvement in one sport and a large number of hours of deliberate practice with the goal of improving sport skills and performance outcomes during childhood

Yes
Güllich (2017) Early concentration in one sport with reinforced sport specific DP/training that is subsequently expanded through all age periods Yes Yes Yes Güllich and Emrich (2014) None Güllich and Emrich (2013) Represent the poles of a continuum differing in exclusivity and intensity of early, sport-specific practice/training, involvement in different sports and non-organized sport activities.

Yes
Yes Yes Hall et al. (2015) With intense year-round training in a single sport at the exclusion of other sports or activities   (2017) Year-round intense training in a single sport with the exclusion of other sports Yes Yes Yes Yes Jayanthi et al. (2013) Intense, year-round training in a single sport with the exclusion of other sports Yes Yes Yes Yes Jayanthi et al. (2018) Intensive, year-round training in a single sport at the exclusion of other sports Yes Yes Yes Yes Kaleth and Mikesky (2010) This practice typically involves children (ages 6 to 12) who commit almost exclusively to a single sport, train and compete year-round, and have high internal-and often external-expectations Year-round sport-specific training, participation on multiple teams of the same sport and focused participation in a single sport

Yes Yes
Myer et al. (2015) Intensive year-round training in a single sport at the exclusion of other sports Yes Yes Yes Yes Myer et al. (2016) Including year-round sport-specific training, participation on multiple teams of the same sport, and focused participation in a single sport   Wiersma (2000) Year-round training in a single sport at the exclusion of other sport or non-sport activities Yes Yes Yes Wilhelm et al. (2017) Intense, year-round training in a single sport with the exclusion of other sports Yes Yes Yes Yes Williams (2018) Participation in a single sport and reporting more than 8 months per year training for that sport Yes Wilson (2006) Year-round training in a single sport at the exclusion of other activities Yes Yes Yes Zibung and Conzelmann (2013) None Yes Yes editorials, and reviews, which undermines the extent to which recommendations about early specialization can be seen as evidence-based. Only 37% of the literature in this review included data-driven studies that were explicitly designed to advance our understanding of early specialization specifically, with 43% of the papers comprised of editorials, commentaries, or reviews. Common rhetoric around this issue assumes early specialization leads to injury, yet only 14 studies have actually examined this relationship with certain indicators of early specialization and of those only five measured early specialization. Similarly, despite broad recommendations that early specialization increases risk of burnout from sport, only three studies explicitly examined this relationship. Given the findings contained in this review, we believe there is insufficient evidence to provide the foundation for the strong and "conclusive" position statements around this topic. Importantly, there is also insufficient evidence to conclude there are no risks to early specialization. Despite messages to the contrary, the benefits and risks of early specialization remains an open topic for sport researchers. The work summarized in this review raises important concerns about the state of the evidence against early specialization and how future research could be improved to resolve outstanding issues. The first issue relates to the conflating of "early specialization" and "sport specialization." Most researchers would agree that the considerable training required to become an elite athlete necessitates eventual specialization at some point . Researchers however advised against the practice of early specialization, suggesting this leads to negative outcomes such as increased injury rates (Hall et al., 2015) without associated benefits (Baker, 2003). In the current review, only half of the studies identified measured an aspect of "early." This distinction between "early specialization" and "specialization" is important. "Specialization" in a single sport may be associated with injury or other negative outcomes due to the link between specialization and overtraining (Ferguson and Stern, 2014) not the age at which it is occurring. Further, in order to properly study the effects of early specialization, it is important to clearly operationalize "early." Of the few studies in this review that measured early only about half used the same criteria (i.e., before age 12).
Another issue relates to the validity of the scales or tools used to determine specialization. The most commonly used scale is Jayanthi et al.'s (2015) Sport Specialization Scale, which uses three criteria [1. single sport training, 2. exclusion of other sports, and 3. year-round training (>8 months)] to rank athletes as low (having only one of the criteria), moderate (two of three) or high on specialization (all three). Over 30% of the datadriven specialization specific studies in this review used this scale, despite concerns about the validity of this scale (Smith et al., 2017). With this scale, for example, a recreational athlete who participates once a week for 2 h in basketball, but quit soccer at age seven, would be regarded as more specialized than a competitive basketball player who participates for 6 h a week but only ever participated in basketball, despite the fact that  Furthermore, 20% of studies in this review used only a single item to measure specialization, raising further concerns about whether a single item is nuanced enough to adequately capture this multi-faceted concept. As noted in the results, 18 different approaches have been used to determine specialization status often inconsistently categorizing athletes. For instance, one study compared a self-classification method (i.e., are you a single sport or multi-sport athlete) to the 3-point "Sport Specialization Scale, " resulting in only 38% agreement on the athletes' categorization and differing results on the relationships between specialization status and injury history (Bell et al., 2016). There were also inconsistencies between the definitions of early specialization and the markers researchers used to measure it. Over half of the studies mentioned "intense" training in their definition of early specialization yet only three studies included any measure of intensity. These were unique case reports that collected a thorough background on one athlete. This misalignment between definition and method further highlights concerns with validity that mar this area of research.
These issues highlight the precarious foundation of the early specialization evidence base. Ferguson and Stern (2014) noted "All position statements are slightly different, but there is not one single position statement that supports early specialization" (p. 380)-but it is unclear why researchers have been so quick to conclude against early specialization given the lack of a consistent definition or method of classifying athletes. Also concerning is that researchers are recommending multisport participation in lieu of early specialization (Coté et al., 2009) without understanding the harmful mechanism behind early specialization.
Around 73% of the papers in this review agreed that singlesport participation was a key component of early specialization, yet this component of specialization alone was not found to be associated with injury history (Bell et al., 2016). The harmful mechanisms behind early specialization are undoubtedly more complex than just single-sport participation and advocating for multi-sport participation without fully understanding what aspect of early specialization is harmful may be short-sighted.

Future Directions
There are several important next steps for research in this area. First and most important, there needs to be a clear and consistent definition of early specialization that can be utilized across disciplines, organizations and researchers. The field will be unable to understand the potential consequences or benefits of early specialization without first establishing a clear understanding of the components and the requirements of this concept. Although it may be difficult to come to a consensus on a definition for early specialization, a Delphi-type approach (i.e., using experts' answers to questionnaires) could be a useful way to reach convergence. Experts could reflect on which previously used facets of early specialization are essential, which are less important, and which are missing. This could help the field establish a definition of early specialization that most agree with. Second, a valid and reliable scale that captures and categorizes early specializers is needed. Any future scales should include some measure of age in order to distinguish "early specialization" from "sport specialization." Additionally, researchers may consider adding measures of intensity to the classification of early specializers to separate those who participate more recreationally from those at risk of overtraining. As noted by a previous systematic review (DiSanti and Erickson, 2019), 92.5% of studies used a dichotomy (i.e., specializer or not) to classify athletes. This likely over-simplifies a highly nuanced topic and future research should consider establishing a continuum of early specialization. Finally, there is a need for more research overall on this topic. Suggestions and statements need to be evidence-based and in order for that to happen there needs to be more evidence.

Limitations
While this review provides the first comprehensive look at all papers related to early specialization in sport, it is not without limitations. First the inclusion of non-data driven studies, while important for understanding the composition of the literature, made it impossible to synthesize all papers in the review uniformly. Additionally, the range of approaches used to classify athletes also made it impractical to perform a meta-analysis. Second, the inclusion criteria that studies "explicitly use the term 'specialization"' might have eliminated studies that focused on the same area but used other words to describe this pattern of participation. Finally, while the search strategy was created in consultation with a profession research librarian, the search string used could have limited the number of studies found through each of the four search engines (e.g., using the connector "AND" could have excluded studies that did not include all the required search terms but were still relevant to the review).

CONCLUSION
This review has shown that there are troubling inconsistencies in the definitions of early specialization and the approaches used to classify athletes. Although this review does not directly establish a clear and consistent definition of early specialization, it is an essential first step. While practitioners and organizations advise against early specialization, this review raises significant questions around the validity and reliability of the evidence underpinning these claims. Once a consistent definition of early specialization is established and researchers have created a valid and reliable measure to capture it, the work to determine negative consequences and benefits of early specialization can begin.
Until then research and any recommendations around early specialization should be viewed with caution. To understand the mechanisms behind early specialization and why it is potentially harmful or beneficial, the field must first establish what early specialization is and how best to measure it.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.