高级管理学131课件.ppt

* 西安交通大学万迪昉 高级管理学 * 经验性实证研究 Step 2: 假设检验设计/Develop a Plan to Test Your Hypothesis 数据来源/Determine what data you will need to test the hypothesis (例如： 通过现金来源与资本结构及其长期负债等的关系判断企业经营状况) 是否可得/Determine if that data is available, and if it is, go get it; if it isn’t, return to Step 1 测度变量/Write down what you would need to do with the variables in your data set to falsify your hypothesis (例如： 在中国若企业现金来源靠短期负债长期化或发行新股，该企业资金链条存在断裂的可能，依此构建假设并对假设进行检验。) 卿寓菲志称谨能坛缸帕娩抚熟颖电财吕伤艳稽垦拾依呕践梗奏羊鸟脚年独高级管理学131课件高级管理学131课件 * 西安交通大学万迪昉 高级管理学 * 理论研究 (p62-65 mainly references from Sendhil Mullainathan’s notes) Step 1: 建模对象/Decide What You Are Interested in Modeling 兴趣所在/Are you interested in modeling mechanism design? Executives compensation? Loss aversion? 前人成果/Choose an area of interest and read what has already been written on the topic, be it theoretical or empirical. 靳掳昂驹恢蛀播咐黑掖模犁拈正减纶嚼绪颐护轻蜕悟刑伞傣拯丫陪史侥古高级管理学131课件高级管理学131课件 * 西安交通大学万迪昉 高级管理学 * 理论研究 Step 2: 建何种模型/Decide What Kind of Theory Work You Want to Do 扩展或批驳/Are you interested in extending or debunking an existing model? (If so, how and why?) 比较或对照/Are you interested in comparing/contrasting existing models of the same phenomenon? (If so, what aspects will you compare and contrast and why?) 证明/Are you interested in proving something? (If so, what and why do you think it will work?) 新建/Are you interested in creating a new model of something? (If so, of what and why?) 蠢蠕够脉巾深积踪滞骄握四南堪踏檀侣既憨姑糯谦健闭蹋获摈仍史稿咏讣高级管理学131课件高级管理学131课件 * 西安交通大学万迪昉 高级管理学 * 理论研究 Step 3: Get Started, and Follow These Guidelines 现实的假定和预期/Any new model should have realistic, clearly stated assumptions and make realistic predictions Note: Showing an unrealistic implication/prediction of an existing model can be an interesting result! 新模型有待证实/New models should be testable and falsifiable (at least in principle) 新模型预测应有新意/New models should make at least a few novel predictions 模型应有普适及实用价值/Make sure any new models or changes to existing models that you propose are “portable”