本研究乃利用 1990 年國民健康調查資料，應用群聚分析法，以十二項健康行為 組合所呈現整體型態的差異，對 2782 名 18 歲以上國人進行分類，結果共得出七組健康行 為群體。根據對香菸、酒、及檳榔三項成癮物質之使用，此七組群體又可區分為三大類型： （１）少癮型：包括循規保健組、多運動及零食組，以及無早餐多咖啡組．（２）雙癮型： 包括多菸多檳榔組及多菸多酒組，以及（３）多癮型：包括正負混合組以及自我摧殘組。進 一步之信度及效度檢驗支持比 -- 分類之適切性。大體而言，相對於少癮型，雙癮型及多癮 型以男性佔絕大多數，平均年齡較低，教育程度以小學國中居多，且職業則以從事半技術之 藍領工人所佔比例最高。而進一步分析則發現不同類型之各組間亦存有相當大的差異。此一 群體間之差異的現象可以 Max Weber 的地位群體（ status group ）及次文化理論加以詮 釋，即健康行為型態乃個人與社會環境及次文化不斷交互作用下的產物。因此本研究建議： 當前之健康促進研究與實務，不該將行為視為獨立的、個別的危險因子，而應強調全貌性整 合性的行為概念化，如此才更具有理論與政策意義。

　　　　　Using data from a 1990 national household health survey, a typology of 2782 adults 18 years of age and above was developed by a cluster analysis of 12 health-related behaviors. The analysis generated seven distinct health behavioral groups, which were further classified into three major types based on the use of three addctive substances: (1) limited-addiction type, including three groups: Health-promotive, Exercise and snack, and Coffee but no breakfast, (2) double-addiction type, including two groups: Smoking and drinking, and Smoking and betal-nuts chewing, and (3) multiple-ad-diction type, including two groups: Mixed and Self-destructive. Reliability and validity of this classification were examined and the results supported this seven-group solution. The health behavioral profile of these seven types was presented. Overall, in comparison with the limited-addiction group, both double-addition group and multiple-addiction group were characterized by younger male populations, with a lower educational attainment, and were employed in blue collar, semi-manual occupations. Futher analyses indicated that considerable variation in sociodemographic characteristics might exist among different types of the same groups. The findings of this study can be well understood within the framework of Max Weber's "status group" and "subculture" theory, i.e., health behavioral pattern was a product of continuous interactions between individuals and their social environment and subculture. It is suggested that rather than treating various behaviors as discrete and independent entities as it is in current health promotion research and practice, a holistic conceptualization of health behavior should be adopted for greater theoretical and policy implications.