Empirical Accounting Research I .pptVIP

* Discussion question 2 A Chinese study finds that the B-share firms have worse performance than A-share firms Conclusion: foreign investors are no better monitors of listed companies than domestic investors in China What are the problems in this conclusion? How to overcome these problems? * Step 4. Sample Selection （样本选择） Random sampling （随机抽样）: To select a sample from the population, so that, all possible samples of size N have the same possibility of being selected Filtering: eliminate certain group s from the population过滤：从总体中排除某些特定的样本 e.g., excluding banks due to different meanings of certain financial ratios Directed sampling: restrict to subjects which are likely to show the effects直接取样：只分析可能出现结果的样本 e.g.,只分析fortune 500中发生了merger的公司 * Sample size样本容量问题 As large as possible越大越好 Relationship between error deviation from population values and sample size: decreasing function 样本越大，误差越小 The larger the better?越大越好? Cost and benefit容量越大，数据收集成本越高 Problem of too large samples 太大样本的问题: 标准方差可能很小，因此很容易拒绝H0 假说 * Population vs. sample: when all firms for certain years in the database are used in the study, is it still a sample? Two views: 样本与总体：对于某些研究，使用的是全样本：比如研究对象是所有上市公司 This is not sample, but population. So, empirical research uses statistical language, not exactly statistical meaning, in describing results. The firms in the database listed firms are representative of all firms. The study period is representative of all periods. So, there is still statistical meaning. * How people evaluate an empirical study? （如何评价实证研究？） Is it interesting relevant ? To policy decisions, to practice, to other research, etc. Is it new creativity ? What has been known already ?what does this study add? Is it in-depth? Is it rigorous reliable ? Methodology * How to measure the rigorousness of research?如何衡量实证研究的严格性 Internal Validity内部有效性: How valid is the inference that X causes Y 从X推断Y，其因果关系有多强 Primary issue: control of “extraneous variables” 外生变量的控制：学历和成功的关系 学历越高的