青少年在發展階段中,由於期望被以成人般看待,因而模仿其心目中所謂的成人行為。亦即抽菸、喝酒或吸食藥物等行為。而青少年時期會抽菸、喝酒或吸食藥物多是因為好奇或同儕壓力影響。故本文欲探討同儕效果對於青少年時期抽菸、喝酒或吸食藥物的影響。 本文利用中央研究院社會學研究所之計畫「青少年成長歷程研究」,從2000年至2006年共7年的追蹤資料。將資料分為國中時期、高中時期與高中、大學時期共三個時期。利用Probit Model探討國中時期好友與同儕對於青少年抽菸、喝酒或吸食藥物的影響。利用Pooled Probit Model 與 Conditional Fixed Effects Logit Model分別探討高中時期與高中、大學時期的好友影響。 研究結果顯示,國中時期同儕與好友抽菸、喝酒或吸食藥物行為對於青少年有顯著的正相關,男性抽菸、喝酒或吸食藥物的機率顯著高於女性。成績與抽菸、喝酒或吸食藥物為顯著的負相關,成績愈高者抽菸、喝酒或吸食藥物的機率愈低。而在高中時期與高中、大學時期時加入了在學變數後發現到,不在學者抽菸、喝酒或吸食藥物的機率顯著大於在學者。而隨著追蹤期間的拉長,家庭因素中的父親存歿狀況與父親教育程度對青少年的影響愈趨顯著。父親已逝世者其抽菸、喝酒或吸食藥物的機率顯著大於父親存活者。而父親教育程度愈高者抽菸、喝酒或吸食藥物的機率也愈高。
The data source of this study is Taiwan Youth Project from Academic Sinica. Our purpose is to use TYP data collected year by year since 2000 to 2006. We try to divided the data into three parts, one is when the sample is junior, the other is senior and another is senior to college. We use Probit Model in junior, Pooled Probit Model and Conditional Fixed Effect Logit Model in senior and senior to college to find out how peer effect influenced the youth smoking, drinking and drug using. Our result is that peer effect has a positive effect on youth smoking, drinking and drug using. The male is significant than female. The higher in grades is less in smoking, drinking and drug using. On the period of senior and senior to college, we add the variable of in school or not. We find that the youth in school is less significant in smoking, drinking and drug using than not in school. When the period is getting longer, we find that the variables of father alive or not and father education is more significant. Father alive or not is significant negative in youth smoking, drinking or drug using. Father education is significant positive in youth smoking, drinking or drug using.