Weil P A, Schleiter M K
Ann Intern Med. 1981 May;94(5):691-703. doi: 10.7326/0003-4819-94-5-691.
With data collected from a national sample of residents in internal medicine (Questionnaire III of the National Study of Internal Medicine Manpower), two models are empirically tested: the first predicting interest in primary care medicine in contrast to subspecialty medicine and the second predicting interest in academic medicine in contrast to clinical practice. Correlation and multiple regression (path analysis) show that background and personal variables specific to each person such as religion, personality type, and desire for control over working conditions have greater predictive power than variables describing the environment such as the type of medical school or residency program the trainee attends. The single most important predictor in terms of unique variance explained in both models is the desire for control over working conditions. The variables positively related to interest in primary care are negatively related to interest in academic careers.
通过从全国内科住院医师样本中收集的数据（内科人力全国研究问卷三），对两个模型进行了实证检验：第一个模型预测对初级保健医学的兴趣与专科医学的兴趣对比，第二个模型预测对学术医学的兴趣与临床实践的兴趣对比。相关性和多元回归（路径分析）表明，每个人特有的背景和个人变量，如宗教信仰、性格类型以及对工作条件控制的欲望，比描述环境的变量，如受训者就读的医学院类型或住院医师培训项目，具有更大的预测能力。在两个模型中，就解释的独特方差而言，最重要的单一预测因素是对工作条件控制的欲望。与初级保健兴趣呈正相关的变量与学术职业兴趣呈负相关。
