本研究旨在利用重複捕取實驗進行台灣地區馬拉松運動人口數估計,並瞭解全程及半程馬拉松運動人口數之估計情形,採系統取樣方式來進行人口抽樣,以2015年台北國道馬拉松賽、高美溼地馬拉松賽、北頭洋走標路跑賽、八卦山台地馬拉松賽、三重區全國馬拉松賽及葫蘆墩全國馬拉賽之全馬與半馬跑者為研究對象,共計發放1,986份問卷,回收有效問卷共1,949份,有效回收率為98.13%,所得資料經描述性統計及Program CARE-2族群估計軟體進行資料分析,結果如下:1.台灣地區馬拉松運動之跑者主要特性為男性、年齡在41歲以上、職業為工業、已婚、教育程度為大專校院、主要參加半程馬拉松組、慢跑鞋品牌為MIZUNO、為參加賽會之個人花費為3,001~5,000元、本次同行人數為4人以上、多透過網路通訊得知賽會,2014年平均每人參與4.26場馬拉松賽。2.研究結果顯示,在估計馬拉松運動人口數時,個體行為反應及抽樣時間點之因素影響較不明顯,在個體異質性模式下,以Chao (1987)下界估計式估算結果為32,761人,然以Lee and Chao (1994)樣本涵蓋率之估計結果較為可信,大約為33,188-46,210人。最後根據本研究結果,提出實務之建議,藉以提供相關單位作為未來規劃之參考依據。
The purpose of this study was to use capture-recapture approach estimating the population size for marathon sports in Taiwan. It was moreover, meant to comprehend the estimating situation of the population size for full marathon and half marathon. The researcher used systematic sampling and taking the participants of 2015 Taipei international express marathon, Taichung Gaomei wetland marathon, Beitouyang marathon, NanTou Mt.Bagua Terrace Marathon, Sanchong National Marathon and Huludun national marathon. There are total 1,986 copies of questionnaire were issued, 1,949 valid samples and the effective response rate is 98.13%. The collected data were analyzed by descriptive statistics and program CARE-2, the results were illustrated as follows. First, the majority of marathon participants in Taiwan are male. Their age ranges from over 41. Most of them engaged in the industry. The marital status of the participants is mostly married. They are mostly university and college graduates. The main participants take part in half marathon. Running shoes of the main brand is MIZUNO. About Personal expenses NT$3,001-$5,001 for the tournament. The accompany numbers have more than 4. The majority of participants are informed of the tournament through the internet. The average per capita to participate in 4.26 field in 2014. Second, The result indicated that model M_b and M_t is less obvious impact on the estimated population of the marathon. For model M_h, The lower bounds at the results for the 32,761 by Chao (1987). Based on the sample coverage approach, the estimated population size under model M_h is approximately the 33,188-46,210 by Lee and Chao (1994). Finally, based on the above results, this study provided some suggestion for some research directions for the organizers planning in the future.