본 연구는 프로 경륜선수를 대상으로 직업 전환에 영향을 미치는 요인들 간의 인과관계를 설명하는 구조모형을 개발하기 위해 프로 경륜선수 448명을 분석대상으로 하였다. 통계분석 방법으로는 기술통계분석, 신뢰도분석, 탐색적 요인분석, 확인적 요인분석, 상관관계 분석, 구조방정식 모형분석, 잠재평균분석, 다중집단분석 등을 SAS 9.1과 SPSS 18.0, AMOS 18.0을 사용하였으며, 분석한 결과는 다음과 같다. 첫째, 신뢰성과 타당성이 인정된 최종 4개의 측정모형을 바탕으로 설정한 구조모형의 적합도 지수들은 모두 기준치를 만족시켰으며, 인과효과를 분석한 결과, 진로자기효능감에는 자아정체성(γ = .730)의 영향력이 가장 크고, 사회정체성(γ = .017)은 매우 작은 것으로 나타났으며, 직업 전환에는 진로자기효능감(β = .552)의 영향력이 가장 크고, 다음은 자아정체성(γ = .591), 사회정체성(γγ = -.002) 순으로 나타났다. 또한, 최종 구조모형의 교차타당성이 인정되었다. 둘째, 집단 간의 차이를 알아보기 위하여 잠재평균분석을 실시한 결과, 고연령 집단이 저연령 집단에 비해 직업 전환이 높은 것으로 나타났으며(p < .001), 비 선수출신 집단이 선수출신 집단에 비해 직업 전환이 높은 것으로 나타났다(p < .001). 또한 고경력 집단이 저경력 집단에 비해 직업 전환이 높은 것으로 나타났으며(p < .001), 은퇴 전 계획이 있는 집단이 없는 집단보다 자아정체성, 진로자기효능감, 직업 전환이 높은 것으로 나타났다(p < .001). 셋째, 최종모형의 적용성 검토를 위해 다중집단분석을 실시한 결과, 경력별에 따라 자아정체성에서 진로자기효능감과 사회정체성에서 직업 전환에 이르는 경로에서 고경력 집단이 저경력 집단에 비해 보다 강력한 영향을 미치고 있는 것으로 나타났으며, 은퇴 전 계획 유무별 집단에서는 자아정체성에서 진로자기효능감과 진로자기효능감에서 직업 전환에 이르는 경로에서 은퇴 전 계획이 있는 집단이 은퇴 전 계획이 없는 집단에 비해 보다 강력한 영향을 미치고 있는 것으로 나타났다.

This study, which targeted professional cycle racers, aimed to develop a structural model explaining causal between factors having an effect on career transitions. It was conducted by 448 cycle racers in order to design a structural model. Methods of statistical analysis are: descriptive statistics, reliability statistics, exploratory factor analysis, confirmatory factor analysis, correlational relationship analysis, structural equation modeling analysis, latent mean analysis, multi-group analysis, etc. This study used SAS 9.1 and SPSS 18.0 AMOS 18.0 statistical analysis software program. the analytical results of this study are as follows. 1. As a result of an analysis about measurement model, career exploration behavior was deleted in the process of model modification by modification index and career maturity failed to reach inspection standard of reliability and validity, and was removed. Finally, the result of confirmatory factor analysis of 4 measurement models corresponding with reliability and validity met the requirement of most standards and thus this measurement model has been confirmed as valid. 2. Set goodness-of-fit-indexes of structural model based on 4 final measurement models which have been proven in reliability and validity came up to all the standards. As an analytical result of causal effect, career self-efficiency was the most powerful in self-identity(γ = .730) and was very small in social identity(γ = .017). In career transitions, career self-efficacy(β = .552), the most influential, followed by self-identity(γ = .591), and social identity(γ = -.002) were present in sequence. Also, final structural model was accepted as cross-validation. 3. For the search of differences between groups, a latent mean analysis was conducted. As a result, First, it showed that possibilities of older age athletes group were higher than those of younger age athletes group in career transition(ρ < .001). Secondary, chances of non-people of athletes group were higher than those of people of athletes group in career transition(ρ < .001). Thirdly, athletes group with much experience showed higher possibilites than athletes group with lack of experience in career transitions(ρ < .001). Fourth, athletes group with pre-retirement planning showed higher chances than athletes group without the plan in self-identity and career self-efficacy, and career transitions(ρ < .001). 4. The outcome of multi group analysis conducted for the review of application about final model revealed that by experience, in view of self-identity, career self-efficacy and in social-identity, the process of career transitions powerfully affected those who have much career compared with those who have less career. By groups of checking whether pre-retirement planning is or not, in terms of self-identity, career self-efficacy, and in career self-efficacy, the process of career transitions are much more influential for groups with pre-retirement planning than groups without one.

국문초록
 Ⅰ. 서 론
 Ⅱ. 연구 방법
  1. 연구 대상
  2. 검사도구
  3. 자료처리
 Ⅲ. 결 과
  1. 측정모형 분석
  2. 구조모형 개발
  3. 교차타당성 분석
  4. 잠재평균분석
  5. 다중집단분석
 Ⅳ. 논 의
 참고문헌
 ABSTRACT

• 박광진(창원대학교) | Park Kwang-Jin (Changwon University)
• 이강헌(창원대학교) | Lee Kang-Hun (Changwon University)