Athletes' career transition out of sport: a systematic review

The purpose of this study was to provide a systematic review of studies on athletes' career transition out of sport from 1968 until the end of 2010. A total of 126 studies were evaluated and reported in three sections: sample characteristics, research designs and correlates of athletes' career transition adjustment. Samples ranged from 1 to 1617. Investigators examined a wide range of competitive levels, both genders and various sports. Researchers have used qualitative (44%), quantitative (44%), and mixed-model (12%) designs. Variables correlated to athletes' career transition adjustment were categorized into factors related to career transition (e.g., self-identity) and available resources during the career transition process (e.g., social support). The discussion focuses on the current status of the study area, limitations in its knowledge, suggested practical implications (e.g., providing proactive support) and future research directions (e.g., examining athletes' retirement decision-making process).

Research on career transitions in sport has been growing gradually over the past three decades, and as a result investigators have found various predictors (e.g., athletic identity, voluntary control over the decision to retire) of the quality of the career transition process for athletes (e.g., Kerr & Dacyshyn, 2000;Lally, 2007). In the early stages, researchers focused on the consequences of athletes' career transition out of sport, but more recently they have distinguished between specific types of transitions, such as young athletes' disengagement/withdrawal from sport and within-sport career transitions. Since the 1990s, researchers have developed appropriate models describing athletes' career transitions (e.g., Taylor & Ogilvie, 1994), and several review papers have also been published on the phenomenon providing guidance for future research directions (e.g., Baillie & Danish, 1992;Wylleman, Alfermann, & Lavallee, 2004).

General findings
A total of 139 studies met the inclusion criteria, and among them 13 were inaccessible and four were excluded because of a lack of information (e.g., sample characteristics, methods or findings). The remaining 122 study papers included 57 published journal articles, 23 published conference proceedings, four published book chapters and 38 dissertations. Among the 122 papers, four contained two independent samples (Alfermann, 1995;Fernandez, Stephan, & Fouqereau, 2006;Herman, 2002;Zaichkowsky, Lipton, & Tucci, 1997). A total of 122 papers were reviewed, and in a final bibliography table (Tables 1 and 2), 126 studies were listed. Among the 126 studies, 10 were published before 1990, 48 in the 1990s and 68 between 2000 and 2010.

Sample characteristics
The total number of participants was 13,511 and the range of sample sizes was between 1 and 1617. Across the samples, 1909 were current athletes, 51 were athletes' entourage (e.g., families and coaches), 219 were non-athletes and 11,332 were retired athletes. The number of studies with fewer than 50 participants was 71, and 36 studies were conducted with samples between 51 and 200. Eleven studies had samples between 201 and 500, seven studies examined over 500 participants and one was not identified because of a lack of information. Most studies (121) investigated athletes' career transition experiences or compared experiences between athlete and nonathlete groups, and five included both athletes and their entourages (Gilmore, 2008;Kane, 1991;Redmond, Gordon, & Chambers, 2007;Stambulova, 1994;Zaichkowsky et al., 2000). These five studies aimed to discover how athletes' families, coaches and administrators influence or are influenced by athletes' sport career termination. Across the studies, 56 contained both genders, 38 contained male athletes only, 24 contained female athletes only and gender was unspecified in eight studies. The studies included a wide range of competitive levels, including student (32), club (7), professional (27) and elite/Olympic-level athletes (50). Mixed-level athletes were examined in six studies, two studies were conducted with disabled athletes and in two studies the level was not identified.
Researchers have examined team sports (36), individual sports (26) or a combination of both (59), and five did not report the type of sport. In 53 studies the athletes were aged between 16 and 26, in 21 studies athletes were aged between 27 and 40 and in five studies the athletes were over 40. A wide range of age groups (aged between 15 and 84) were examined in 17 studies, two studies were done with athletes aged under 16 and 28 studies did not report the age of participants. The majority of studies were conducted in Western countries (60 in North America, 45 in Europe and 10 in Australia). Three studies had been conducted in Asia and South America, and two studies had been done in the Middle East. One study existed with African athletes, and two studies did not identify where data originated.

Correlates and consequences of career transition
Slightly under half (55) of the studies investigated the psychological, emotional, social and physical consequences of athletes' retirement from sport. The other studies examined variables that influence the quality of athletes' career transition out of sport. Among the 13,511 participants, 11,332 (84%) of them had experienced termination from their sport and 1768 (16%) of them reported that their career transition experiences had accompanied adjustment difficulties or problems. In addition, the majority of studies (86) reported that some of their participants expressed career transition difficulties or negative emotions, including feelings of loss, identity crisis and distress, when they ended their career and adjusted to postsport life (e.g., Baillie, 1992;Kerr & Dacyshyn, 2000;McKenna & Thomas, 2007). Four studies reported that the career transition process for athletes was neither a positive nor negative event for them (e.g., Alfermann, 1995;Johns, Lindner, & Wolko, 1990;Schwendener-Holt, 1994;Torregrosa, Sanchez, & Cruz, 2003).
We identified 63 correlates related to the quality of athletes' career transitions. These variables were reduced to 19 during the analysis and categorized into two themes based on two existing models (e.g., Gordon, 1995;Taylor & Ogilvie, 1994): (a) factors related to the quality of career transition; and (b) available resources during the career transition.
Factors related to the quality of career transition Table 2 shows the correlates and their directions associated with the quality of career transition. We identified 15 variables associated with the quality of athletes' career transitions, including athletic identity, demographical issues, voluntariness of retirement decision, injuries/health problems, career/personal development, sport career achievement, educational status, financial status, self-perception, control of life, disengagement/drop-out, time passed after retirement, relationship with coach, life changes and balance of life. These variables are presented below in the order of the number of studies that examined each of them.

Athletic identity
Athletic identity refers to self-identity in the sport domain (Brewer, Van Raalte, & Petitpas, 2000). A total of 35 independent studies demonstrated correlations between athletic identity and athletes' career transitions. Among them, 34 studies indicated that both a strong athletic identity and high tendency towards identity foreclosure were negatively associated with the quality of athletes' career transitions. These studies have revealed that the retired athletes experienced a loss of identity when they had a strong athletic identity at the time of their sport career termination (e.g., Kerr & Dacyshyn, 2000;Lally, 2007), and they needed a longer period of time to adjust to post-sport life (e.g., Grove et al., 1997;Warriner & Lavallee, 2008).

Demographical issues
Although 24 studies examined demographic differences between athletes at the time of career termination, such as gender, age, social status, type of sport, race, marital status, competitive levels and cultural or national factors, the overall association was indeterminate.
Four studies examined differences in adjustment to post-sport life between male and female athletes, and only one study (Stambulova, 2001) reported that female former athletes adapted to post-sport life more quickly than male former athletes. Studies conducted by Cecić Erpič, Wylleman, and Zupančič (2004) and Leung, Carre, and Fu (2005) indicated positive relationships between retired athletes' ages and experienced transition difficulties, but Chow (2001) and Gilmore (2008) reported that those who terminated their sport career at a younger age expressed higher career transition difficulties. Three studies reported that athletes' social status was positively related to the quality of their career transitions (Conzelmann & Nagel, 2003;Houlston, 1982;Koukouris, 1994). Three studies examined relationships between type of sport and transition difficulties, but no differences were found (Fernandez et al., 2006;Gilmore, 2008;Tate, 1993). Two studies, which demonstrated transitional differences between Caucasian student athletes and African American student athletes, showed opposite directions (Caucasian American athletes experienced more transition difficulties [Lewis, 1997]; African American athletes experienced more transition difficulties [Perna, Zaichkowsky, & Bocknek, 1996]). Two studies assessed marital status (Fernandez et al., 2006;McKnight, 1996), and both showed that married athletes experienced a higher degree of perceived support from their partners and less difficulty in the transitional process. Two studies showed that competitive level is positively related to financial and occupational adjustment (Gilmore, 2008;Leung et al., 2005).
In cross-national comparisons among European countries, three studies reported some differences, such as age of career termination, athletic career satisfactions and use of coping strategies. Researchers, however, concluded that athletes' degree of athletic identity and pre-retirement planning had more influence on the quality of their career transitions than their nationalities or cultural differences (Alfermann, Stambulova, & Zemaityte, 2004;Schmidt & Hackfort, 2001;Stambulova, Stephan, & Jäphag, 2007). Huang (2002) examined cultural differences between Chinese and German elite-level athletes' career transition experiences. The study reported differences in several areas of post-sport life, including involvement in sport after retirement (more Germans involved in sport) and current occupations (more Chinese working in sport fields), but no significant differences were found in age of career termination and perspective of sport career termination.

Voluntariness of retirement decision
Voluntariness of retirement decision can be explained as the degree of control athletes have over their decision to retire. The variable has been examined in 21 studies, with 18 indicating a positive association with the quality of career transition and three finding no relationship. Studies (7) with athletes who experienced forced retirement found that participants experienced high levels of negative emotions, such as fear of a social death or dying (Blinde & Stratta, 1992;Fortunato & Marchant, 1999;Zaichkowsky et al., 2000), a sense of betrayal and social exclusion (McKenna & Thomas, 2007) and a loss of identity (Butt & Molnar, 2009;Lotysz & Short, 2004;Lynch, 2006).

Injuries/health problems
Ten out of 11 studies found that injuries and health issues were sources of transitional difficulties for retired athletes (e.g., Gilmore, 2008;Kadlcik & Flemr, 2008;Muscat, 2010). Injuries and health problems were negative factors for retired athletes wishing to move towards post-sport life, and the participants who had physical problems needed longer periods of time to adjust after entering post-sport life (e.g., Werthner & Orlick, 1986). In addition, retired athletes' physical condition was one of the most immediate concerns for the quality of their post-sport lives, and some athletes expressed difficulties in dealing with post-sport life because of their physical pains (Gilmore, 2008;Hughes, 1990). In contrast, Perna, Ahlgren, and Zaichkowsky (1999) did not find any relationship between former collegiate athletes' injuries and their long-term life satisfaction.

Career/personal development
All nine studies that examined vocational and life skills development indicated positive associations with the quality of athletes' career transitions. Former professional athletes showed difficulties in dealing with non-sporting situations, and they experienced delayed identity shifts because of a lack of non-sporting life experiences during their sport careers (Kane, 1991;Muscat, 2010). Educational involvement and career planning were also positively associated with post-sport life adjustment among college athletes (Lantz, 1995), and athletes attributed their limitation of life choices after sport career termination to a lack of personal development (Chow, 2001;Stronach & Adair, 2010;Swain, 1991).

Sport career achievement
The number of studies measuring sporting goal achievement was nine, and all of them reported a positive correlation with the quality of career transition. Retired athletes who had succeeded in their sport showed stable levels of self-identity, selfesteem and global self-concept, and fewer occupational difficulties (Cecić Erpič et al., 2004). In contrast, those who had not achieved their expected sporting goals expressed a high degree of psychosocial difficulties, such as loneliness, missing people related to sport and difficulty in organizing their post-sport lives, including taking a longer period to adjust to post-sport life and having negative evaluations of their adaptation process (Chow, 2001;Koukouris, 1994). Additionally, Sinclair and Orlick (1993) revealed that those who achieved more in their sport showed a higher satisfaction in post-sport life.

Educational status
All eight studies that examined educational status reported positive correlations with the quality of athletes' career transitions. Athletes reported that their educational progress was negatively influenced by their sport career, and low educational attainment was related to vocational difficulties during the career transition process (Marthinus, 2007;Stronach & Adair, 2010). In addition, athletes' educational and college graduation status influenced both short-term and long-term adjustment after retirement (e.g., Williams, 1991).

Financial status
Financial status has been reported in eight studies, and all indicated positive associations with the quality of career transition for former athletes. The studies found that some former athletes experienced financial problems, which caused transition difficulties and limited their post-sport life choices (e.g., Lotysz & Short, 2004;Menkenhorst & Van Den Berg, 1997).

Self-perception
Self-perception incorporated athletes' perceived body image, self-confidence and selfworth. Eight studies discovered a positive correlation between self-perception and the quality of athletes' career transitions. Five studies revealed that athletes' body image can be influenced by their retirement and their negative perceptions of their body can be a source of distress during the career transition process (Kerr & Dacyshyn, 2000;Lavallee & Robinson, 2007;Sparkes, 1998, Stephan, Bilard, Ninot, & Delignières, 2003a, 2003bStephan, Torregrosa, & Sanchez, 2007). In addition, findings indicated that athletes' feelings of loss of control over their bodies can be sources of identity crisis during the career transition process (Sparkes, 1998). Three studies discussed athletes' perceived self-worth and self-confidence (Missler, 1996;Newell, 2005;Stephan et al., 2003a). Missler's (1996) study suggested that gaining self-worth without sport performance was positively related to former golfers' quality of career transition. Stephan et al. (2003a) revealed that former Olympians showed a significantly lower mean physical self-worth and self-perceived physical strength than active athletes. Newell (2005) indicated that retired athletes who had high self-confidence also showed positive perspective on their post-sport career in terms of self-belief in their abilities to achieve new career goals.

Control of life
Control of life refers to athletes' perceived autonomy and power over their decisions while competing and during the career transition process. Seven studies examined control of life, and all of them indicated that athletes who had less control over their life expressed more negative emotions during the transition process than those who had more control (e.g., Kane, 1991;Kerr & Dacyshyn, 2000). College athletes who showed more control over their lives had greater self-esteem, higher life satisfaction and were more positive about the future (Webb, Nasco, Riley, & Headrick, 1998). Some athletes attributed their forced retirement and negative emotional experiences during the career transition to unbalanced power in their sporting system, which they could not control (Blinde & Stratta, 1992;Parker, 1994). Koukouris (1991Koukouris ( , 1994Koukouris ( , 2001 applied the term 'disengagement' to examine Greek adolescent and young adult athletes' drop-out experiences, and four other studies (Alfermann, 1995;Alfermann & Gross, 1997;Butt & Molnar, 2009;Johns et al., 1990) also explored athletes' drop-out experiences. Drop-out refers to the premature sport career termination among young athletes before they reach their full potential (Alfermann, 1995). Findings indicated that athletes who dropped out mostly terminated their sport careers voluntarily. The decision might have been premature and related to lack of control over their situation and decisions (e.g., anticipation of deselection, conflicts with coaches [Butt & Molnar, 2009;Koukouris, 1994]). Feelings of failure, therefore, can be sources of career transition difficulties (Alfermann, 1995;Butt & Molnar, 2009). For this reason, researchers have distinguished 'drop-out' from the term 'retirement' (e.g., Alfermann, 1995;Wylleman et al., 2004).

Disengagement/drop-out
Among seven studies examining athletes' disengagement/drop-out, six indicated a negative association between disengagement/drop-out experiences and the quality of career transition, and one showed no association with the quality of career transition. Alfermann's (1995) and Alfermann and Gross's (1997) studies reported that some athletes who dropped out from their sport experienced identity problems and negative emotions, and they also used more passive coping strategies than those who experienced retirement from sport. Butt and Molnar (2009) revealed that athletes who dropped out expressed a loss of social networks similar to retired athletes, but they also reported feelings of rejection from former friends who were still in sport teams. Transition difficulties were found among disengaged athletes in Koukouris's (2001) study as well. The study highlighted that athletes' disengagement from sport could lead to social, psychological and medical difficulties during their post-sport life adjustment. In contrast, Johns et al. (1990) showed that young gymnasts who dropped out from the sport perceived their sport career termination in both positive and negative terms, including changes in social relationships and lifestyles.

Time passed after retirement
Seven studies reported positive associations between time passed after retirement and former athletes' perception of the quality of their adjustment to their post-sport life. Five of those seven studies used a longitudinal design, and the results revealed that retired athletes perceived fewer transition difficulties over time (Douglas & Carless, 2009;Lally, 2007;McKenna & Thomas, 2007;Stephan et al., 2003b;Wippert & Wippert, 2008). Stephan et al. (2003b) found that post-sport life adjustment difficulties reduced as time passed, and in the other studies the participants started to experience a balance in their lives after 18 months of their sport career ending (Douglas & Carless, 2009;McKenna & Thomas, 2007). Lally (2007) reported that athletes appeared to have new roles and identities a year after sport career termination. Wippert and Wippert (2008) also revealed that retired athletes' selfperceived stress levels significantly declined three months following the end of their career, compared to after 10 days of retirement.

Relationship with coach
All six studies that investigated coachÁathlete relationships revealed that conflict between both could be a source of athletes' career transition difficulties. Retired athletes who did not have a good relationship with their coach expressed more difficulties during the career transition process (Chow, 2001;Kerr & Dacyshyn, 2000;Muscat, 2010) and the participants in the studies blamed their coaches for an unsatisfying retirement (Lavallee & Robinson, 2007). The studies also indicated that unbalanced power between coaches and athletes and an unpleasant coachÁathlete relationship were associated with injury or disengagement from sport (Johns et al., 1990;Werthner & Orlick, 1986).

Life changes
Life changes refer to changes in lifestyles and daily routines (Stephan et al., 2003a). Five studies indicated that changes in former athletes' lives had negative associations with the quality of career transition among retired athletes. Former athletes reported feelings of anxiety associated with their new routines and feelings of being lost resulting from no more competition and training (Kerr & Dacyshyn, 2000;Schwenk, Gorenflo, Dopp, & Hipple, 2007). In addition, accepting a new lifestyle was one of the transitional difficulties for athletes (Stephan et al., 2003b).

Balance of life while competing
A total of three studies reported positive correlations between the balance of sporting and non-sporting lives prior to retirement and the quality of career transition. Lawrence (2003, 2004) found that student athletes perceived balancing academic and athletic activities during their sport participation as a significant predictor for post-sport life adjustment. Kerr and Dacyshyn (2000) revealed that retired gymnasts who had found the right balance between sporting and non-sporting lives during their athletic careers reported higher life satisfaction after their retirement compared with those who did not have a balanced life while competing.

Available resources during the career transition
We examined four correlates in the available resources category with respect to the quality of athletes' career transitions, including coping strategies, pre-retirement planning, psychosocial support and support programme involvement.

Coping strategies
A total of 32 studies examined the frequency of coping strategy used during the career transition process and found no clear evidence that certain strategies are more effective than others, except for searching for new careers or interests. Additionally, studies have shown that finding new careers or interests was closely related to identity shift processing because of role changes following the end of a sport career (e.g., Kerr & Dacyshyn, 2000;Lally, 2007). Six studies reported that athletes employed keeping themselves busy as a way of coping. Furthermore, results from two studies revealed that keeping busy during the career transition period was one of the beneficial coping strategies for athletes to reduce career transition difficulties (Kadlcik & Flemr, 2008;Lotysz & Short, 2004).
Some researchers (e.g., Grove et al., 1997) attempted to investigate associations between certain variables (e.g., athletic identity, nationality) and the use of coping strategies. Grove et al. (1997) reported that higher athletic identity was positively associated with the use of denial, venting emotions and searching for social support. Alfermann and Gross (1997) discovered that athletes who dropped out from their sports used more passive ways of coping than retired athletes who did not drop out from their sport. Stambulova et al. (2007) found that French former athletes used a denial strategy more than Swedish former athletes. In addition, six studies noted maladaptive coping strategies among retired athletes, such as alcohol dependence, increased smoking, committing suicide or drug use (Douglas & Carless, 2009;Koukouris, 1991;Mihovilovic, 1968;Schwenk et al., 2007;Sinclair & Orlick, 1993;Wippert & Wippert, 2008).

Pre-retirement planning
Pre-retirement planning included vocational, psychological and financial considerations following the end of a sport career. Twenty-nine studies described preretirement planning as a variable, and 28 of them reported that it was positively associated with the quality of athletes' career transitions. Planning for post-sport life included psychological preparation before sport career end, and having a clear goal outside of sport gave retired athletes a feeling of comfort (e.g., Warriner & Lavallee, 2008;Young, Pearce, Kane, & Pain, 2006). Pre-retirement planning was also related to former athletes' vocational adjustment to post-sport life (Coakley, 2006), and financial planning was one of the influential factors in the quality of athletes' postsport life adjustment (Fortunato & Marchant, 1999).

Psychosocial support
Psychosocial support incorporated all kinds of support from non-sporting (e.g., spouses, families, friends and significant others) and sporting sources (e.g., coaches, trainers and teammates). Investigators have reported various kinds of psychosocial support, including emotional, esteem, information, network and tangible support. Among 29 studies, 27 reported that support from others had a positive influence on the quality of career transition. The results showed that being supported by close others eased transition difficulties during the post-sport adjustment period (Alfermann, 1995;Werthner & Orlick, 1986;Young et al., 2006). Participants in two studies (McKnight, 1996;Schmid & Seiler, 2003) discussed more than one kind of psychosocial support (i.e., emotional, information and tangible) and seven studies did not specify the kind of support that participants had received from close others.
Among 29 studies, 10 reported the value of emotion and esteem support for athletes' career transitions. Emotion and esteem support included encouragement and help for emotional challenges through storytelling, account making and mentoring. Studies revealed that greater amounts of account making were related to decreases in athletic identity, and completion of account-making procedures was associated with greater overall success in coping with retirement (Barners, 2002;. Studies that examined mentoring effectiveness found that higher mentoring scores were positively associated with intimacy levels among athletes and athletes' demands of mentoring during the career transition process (Lavallee, Nesti, Borkoles, Cockerill, & Edge, 2000;Perna et al., 1996).
Information support has been examined by four studies and included supportive transition (i.e., providing information prior to a transition [Fernandez et al., 2006;Wippert & Wippert, 2008]) and information from organizations, former teammates and coaches during the career transition (McKnight, 1996;Stephan et al., 2003b). Fernandez et al. (2006) and Wippert and Wippert (2008) discovered that athletes who had been given pre-retirement information from coaches or trainers before being cut from the team expressed fewer transition difficulties or negative emotions during the career transition process compared to those who did not receive the information. Results from McKnight's (1996) study showed that the recognition of guidance on post-sport life or preparation within organizational policies was positively associated with athletes' career transitions, and participants from Stephan et al. (2003b) perceived teammates' and coaches' support during career transition as helpful for post-sport adjustment.
Six studies reported that social networks played an important role in athletes' career transitions. Studies revealed that a loss of social networks after a sport career termination was one career transition difficulty (e.g., Kane, 1991), and athletes who experienced less difficulty in career transition had a stronger social support network (Schwendener-Holt, 1994).
Tangible support was discussed in two studies (Leung et al., 2005;Schmid & Seiler, 2003). The study by Leung et al. (2005) showed that tangible support from national organizations mediates the quality of athletes' career transitions, and participants who had such opportunities when seeking a career after retiring from sport (e.g., funding) experienced a relatively healthier transition than athletes who did not receive organizational support. Schmid and Seiler (2003) revealed that tangible support from athletes' close others was positively related to successful transitions.

Support programme involvement
Eight studies indicated positive associations between athletes' support programme involvement and their life skills development and the quality of career transition. Albion (2007) and Redmond et al. (2007) evaluated the effectiveness of the Athlete Career and Education (ACE) programme in Australia and found that the programme helped athletes become less inclined towards identity foreclosure and increased motivation to make career decisions. Gilmore (2008) examined Scottish former athletes in a performance lifestyle programme and found that they perceived job seeking and interview skills development as useful institutional support. Four studies examined college student athletes' support programme involvement and revealed that the programme helped them to develop their life skills, including leadership skills, decision-making, career planning and communication skills (Goddard, 2004;Selden, 1997;Stankovich, 1998;Torregrosa, Mateos, Sanchez, & Cruz, 2007). Lavallee (2005) revealed that providing Life Development Interventions (LDI; Danish, Petitpas, & Hale, 1992) had a positive influence on athletes' career transition experiences because it helped them to develop appropriate coping skills.

Discussion
The present study aimed to provide a systematic review of athletes' career transition out of sport. A total of 126 studies met the inclusion criteria, and results demonstrated that the study area has been growing steadily. Investigators have used both qualitative and quantitative methods, and the number of studies that employed longitudinal designs has increased since 1990. A wide range of competitive levels, both genders and various types of sports have been examined. The current review identified 19 variables related to the quality of athletes' career transitions, and the correlates were categorized into two themes: factors related to career transition (e.g., athletic identity, voluntariness of decision and degree of individual development) and available resources during the career transition (e.g., coping strategies, psychosocial support and pre-retirement planning).  reported 80 independent sample studies in their review, and the present review found 126 studies, showing that the study area has grown significantly over the past 10 years. The findings from the current review also revealed that studies conducted via qualitative methods (44% compared with 29% reported by , with female athletes (19% compared with 11% in 2000) and with participants from non-Western regions (nine studies compared with no studies in 2000) have increased gradually over the past decade. These results can be interpreted as a diversification of the methodologies used and a broadening of sample populations in the study area.
Early reviews (e.g., Baillie & Danish, 1992;Crook & Robertson, 1991) pointed to the importance of individual development, self-identity, voluntariness of retirement and social support in the career transition process. The current review supports the significance of these factors. Indeed, to review existing studies from the early stages up to 2010 made it possible to examine changes in research trends. As Wylleman et al. (2004) highlighted, in the early stages of investigation researchers focused on causes and consequences of athletes' retirement and then moved on to identify predictors for the quality of athletes' career transitions. Since then, researchers have also refined well-known variables (e.g., athletic identity) and have examined a broader range of correlates. For example, a number of previous studies (e.g., Sparkes, 1998) discussed athletes' identity shift after leaving their sport, but a more recent longitudinal study (Lally, 2007) revealed that athletes experience decreases in athletic identity in the latter stages of their sport career, not just after actually retiring. These findings from longitudinal studies (Lally, 2007;Stephan et al., 2003a) have documented detailed changes in athletes' perceptions and attitudes through the career transition process, including the latter stages of their sport careers and post retirement. They have also made recommendations for how to support athletes not just in terms of the kind of support they need, but also when this support should be provided.
Although previous literature has highlighted the importance of social support in athletes' career transition experiences (e.g., Baillie & Danish, 1992), past reviewers have typically not stratified research according to specific kinds of support. Social support is conceptualized in many different ways across psychology. Recently, in sport and performance psychology, Murphy (2009) introduced five types of social support (emotional, esteem, information, network and tangible), which help interpret existing data. Some researchers (e.g., Curtona & Russell, 1990) have claimed that certain forms of social support (e.g., tangible, emotional) may have more beneficial influence on specific kinds of stress (e.g., losses, transitions) than others. Therefore, identifying types of social support and examining the forms of support athletes received during the career transition out of sport may help practitioners design appropriate social support-based interventions. This review has several limitations. Only English language studies were included. During the search process four foreign language studies (two Chinese [Liu & Li, 2007;Wang, 2008] and two Korean [Chung, 2010;Hong, 2010]) were excluded. The exclusion of these non-English studies might influence sample characteristics (e.g., location of study) and lead to the omission of potential correlates, such as cultural or sport system-related issues. Finally, the review could not present all examined correlates individually, due to limited space. In line with previous and accepted guidelines (Goodger et al., 2007;Sallis et al., 2000), correlates with fewer than three individual studies were either grouped into similar predictors during the analysis process or dropped from the summary table. Providing categories with conceptually similar variables might lead to more robust results in terms of suggesting directions for examining similar correlates.
Based on previous findings, several future research directions can be identified. There has been a growing body of interest in the development of conceptual models in athletes' career transitions, but these models have not been systematically tested. Only one doctoral dissertation, for example, examined Taylor and Ogilvie's (1994) model (Coakley, 2006), and one study developed theories through the grounded theory method (Torregrosa, Boixadó s, Valiente, & Cruz, 2004). Testing available models and developing sport-specific theoretical frameworks might help researchers to clarify the conceptualization of athletes' career transition out of sport.
In terms of research design, the majority of investigators have employed retrospective data collection methods, and many of them have noted memory and recall bias as a limitation (e.g., Kerr & Dacyshyn, 2000;Lavallee & Robinson, 2007). Employing prospective longitudinal research designs to study athletes' career transitions might allow researchers to examine the dynamic processes over time. Therefore, more prospective longitudinal studies are needed in the study area.
Nearly half of the studies used a qualitative research method, and most researchers employed individual interviews for data collection and thematic analysis of transcripts. Investigators analysed data in different ways, including interpretative phenomenological analysis (e. g., Warriner & Lavallee, 2008) and narrative analysis (e.g., Gearing, 1999), yet there are more ways to explore athletes' career transition phenomena from a qualitative perspective. Since various types of research designs provide different methodological advantages and limitations, we recommend future researchers build upon this research by employing diverse methods, including focus groups, case studies and action research methods.
Several investigators (e.g., Cecić Erpič, 2000;Lavallee & Wylleman, 1999) have developed measurement tools to investigate athletes' career transitions, but these available questionnaires have not been validated with a range of samples. Further validation research of the questionnaires may assess their utility in athlete populations.
The results from the present review showed that there are many different variables that influence the quality of athletes' career transition adjustment. Since findings revealed that coachÁathlete relationships (e.g., Chow, 2001;Lavallee & Robinson, 2007) and athletes' physical health (e.g., Gilmore, 2008) were positively associated with the quality of athletes' career transitions, employing relevant issues from other study areas (e.g., burnout, coachÁathlete relationships and injury) may help to explain the multidimensional aspect of the career transition process.
Regarding sample characteristics, although several studies have attempted to discover demographic differences, such as gender, age, types of sport and marital status, no consistent evidence has been found except for a positive association between marital, social, educational and financial status and the quality of career transition (e.g., Lotysz & Short, 2004;Marthinus, 2007). Research findings across studies have indicated that athletes who are of different ages during their career transitions showed different needs post sport career, because of diversities in individual development stages and life plans. For example, studies with student athletes who terminated their sport career in their teens or early 20s showed that they often chose to become students rather than find employment (e.g., Lally, 2007;Lavallee & Robinson, 2007). In contrast, the majority of professional or elite-level athletes who retired in their late 20s or 30s made a transition into the world of work (e.g., Lavallee, Gordon, & Grove, 1996;Marthinus, 2007). Some studies (e.g., Allison & Meyer, 1988) did not report specific details of sample characteristics (e.g., age, type of sport), and only four studies examined athletes from more than one country in order to identify cultural or social differences in the quality of their career transitions (i.e., Alfermann et al., 2004;Schmidt & Hackfort, 2001;Stambulova et al., 2007;Wheeler et al., 1999). Examining differences in demographic variables might help practitioners to provide appropriate support for each athlete, whether the differences are based on individual factors (e.g., age, gender) or environmental factors (e.g., culture, sport context). Future research needs to focus on examining sport-specific or life span and age-related issues in the quality of career transitions.
As Stambulova, Alfermann, Statler, and Côte (2009) noted, athletes' career transition studies have been conducted mainly in European countries, North America and Australia. Findings from the current review confirmed this trend and revealed that eight out of nine studies from non-Western countries (i.e., Africa, Asia, the Middle East and South America) were conducted between 2001 and 2007. We agree with , who have called for more cultural studies in the career transition area, in terms of investigating the influence of sport systems and environmental contexts on the quality of athletes' career transitions. Moreover, as several researchers have suggested (Berry, Poortinga, Segall, & Dasen, 2002;Si & Lee, 2007), investigating cultural similarities and differences could assist in testing the generality and validity of existing knowledge and theories and lead to practical implications, such as providing suitable and appropriate support in applied work.
Few studies mentioned the importance of the decision-making process for postsport life adjustment (e.g., Cecić Erpič, 2007b;Kirby, 1986;McPherson, 1980). Kerr and Dacyshyn (2000) indicated that the sport career termination decision-making process is multifaceted and complex. Athletes' sport career termination decisionmaking processes usually occur over an extended period of time (Kerr & Dacyshyn, 2000) and are considered different from retirement itself (Kirby, 1986). Recently, investigators have tried to specify aspects of the decision-making process, both athletic and non-athletic elements, that lead to sport career termination (Cecić Erpič et al., 2004), and the (anti-)pull or (anti-)push factors related to sport career retirement decision making (Fernandez et al., 2006). According to Fernandez et al. (2006), a better understanding of the retirement decision-making process is useful for applied work, when practitioners assist athletes who are planning for post-sport life.
Numerous authors (e.g., Taylor & Ogilvie, 1994) have discussed intervention strategies to support athletes' career transitions. However, only one published study (Lavallee, 2005), three published conference proceedings (Albion, 2007;Redmond et al., 2007;Torregrosa et al., 2007) and four dissertations (Gilmore, 2008;Goddard, 2004;Selden, 1997;Stankovich, 1998) have tested the effectiveness of a specific intervention strategy or programme for supporting athletes' career transition out of sport. In order to evaluate the effectiveness of interventions, further research is required into whether specific psychological interventions can assist practitioners in supporting athletes in transitions. For example, although researchers have frequently examined coping strategies employed by athletes during the career transition process, no study has been conducted on how sport psychologists can support athletes in developing effective coping strategies.
None of the previous reviews divided the voluntariness of the retirement decision and disengagement/drop-out factors into different categories. However, the current review supports the suggestion by Wylleman et al. (2004) to separate the two factors, in terms of differences in the processes between athletes' voluntariness of retirement decisions and disengagement/drop-out issues. The processes involved in disengagement and dropping out from sport participation are non-normative and occur without individuals' intention, but they are still different from forced retirement. Disengagement and drop-out processes are not as uncontrollable as forced sport career termination because athletes often have a choice to continue or stop their sports career (e.g., Piffaretti, Schnyder, Mahler, Barbat, & Keller, 2003). In addition, findings have suggested that disengagement/drop-out can be related to burnout (Smith, 1986). Research examining disengagement/drop-out and forced retirement might extend knowledge on outcomes of specific transitions.
Several practical implications for sport psychologists, advisors working with athletes in transition and sport organizations emerged from the present review. The findings provided positive associations between the quality of athletes' career transitions and athletes' programme involvement. In addition, results from longitudinal studies (e.g., Douglas & Carless, 2009;Lally, 2007) indicated that athletes showed changes in their degree of athletic identity and required certain periods of time to adjust to their post-sport lives. Therefore, to assist athletes' career transitions, practitioners need to provide athletes with both proactive (e.g., career planning, education in transferable skills) and reactive support (e.g., coping with emotions, supporting the identity reformation process) programmes to help them prepare for their career transition out of sport and to adjust to post-sport lives.
Findings indicated that coachÁathlete relationships influenced the quality of athletes' career transitions, in terms of coaches' influences on athletes' retirement decision making and athletes' expectations for coaches' support (e.g., Chow, 2001). Practitioners may need to examine coachÁathlete relationships and consider using coaches' support to assist athletes' career transitions.
Studies (e.g., Alfermann, 1995) revealed that dropping out from sport without reaching their full potential could lead young athletes to feel like failures and require special attention. Practitioners may need to employ different kinds of assistance to support these young athletes as they withdraw from their sport, such as helping them to reduce feelings of failure or self-disappointment through building positive selfimage and self-confidence.
Researchers (e.g., Albion, 2007;Lavallee, 2005) revealed that many athletes did not participate or use career transition intervention programmes even if they had opportunities to receive such services. Previous literature (e.g., Petitpas, Danish, McKelvain, & Murphy, 1992) also highlighted athletes' resistance to such programmes while actively competing and attributed this attitude to their perception that involvement would be a distraction from their sport performance. For active athletes with negative perceptions regarding support programme involvement, practitioners need to provide education related to the importance of pre-transition planning and life skills development.

Conclusion
Research in the area of career transitions in sport has increased gradually over time, as reflected in the growing number of studies and reviews. The investigations to date have contributed to a better understanding of the athletic career transition process, but further research is needed. This systematic review reported the current status of athletes' career transition study and also highlighted limitations in its knowledge. In addition, the present review suggested several future directions, including diversification in research design (e.g., employing longitudinal designs), sample characteristics (e.g., examining demographical differences), correlates (e.g., discovering roles of coachÁathlete relationships in career transitions) and practical implications (e.g., providing proactive intervention).