Robustness of bayesian factor analysis estimates

This paper presents the result of a study of the robustness of posterior estimators of the factor loading matrix, the factor scores, and the disturbance covariance matrix (the main model parameters) in a Bayesian factor analysis with respect to variations in the values of the parameters of their prior distributions (the hyperparameter). We adopt the ε - contamination model of Berger and Berliner(1986) to generate prior distributions whose hyper-paramters reflects small variations in the elements of the uncontaminated hyperparameters, and we use directional derivatives to examine the variation of the uncontaminated estimators with respect to changes in the values of the hyperparameters, in the directions of the main model parameters. Several matrix norms are used to measure the closeness of the resulting values. We illustrate the results with a numerical example.     

互联网视阈下大学生行为文化探析

互联网的迅猛发展和普及,对大学生行为文化的塑造带来机遇的同时,也面临着诸多挑战。一方面,大学生运用网络资源进行学习、交流、娱乐、休闲;另一方面,他们对网络产生的负面影响缺乏抵御能力,导致对大学生行为文化的塑造具有不稳定性。加强网络思想政治教育、网络行为规范教育以及大学生自我修养教育,是规避互联网负面影响,塑造大学生行为文化的重要途径。     

An analysis on basic income: Evidence using data from experimental projects throughout the world

Basic income was a hot issue in the recent Korean presidential election. Its pros and cons were used to scramble for the electorate's votes. However, their arguments largely relied on political rhetoric and lacked scientific facts. The problem is that politics without the backing of science is not only prone to a misguided conviction but also going to deal a blow to the future development of basic income. This study aims to narrow the gap between rhetoric and facts through evidence-based research on basic income. By employing methods of fact-finding and effect evaluation, we analyze using four aspects of assessment, namely, general information about basic income experiments; details on the experiments; development of the experiments; and the effect of the experiments. Basic income-related data are drawn from the World Bank and the Basic Income Laboratory at Stanford University. Significant results of the review are that some positive effects were identified at the micro level but any meaningful macro-level impact was impossible to confirm. To sum up, the outcomes are so erratic that it is difficult to detect any clear pattern in policy impacts. To approach basic income more scientifically, further research is urgently needed.     

Modeling the Effects of a Service Guarantee on Perceived Service Quality Using Alternating Conditional Expectations (ACE)*

This paper addresses the dearth of empirical research on the relationship between service guarantee and perceived service quality (PSQ). In particular, we examine the moderating effects of a service guarantee on PSQ. While a recent study provided empirical evidence that service quality is affected by service guarantee and employee variables such as employee motivation/vision and learning through service failure, the nature and form of the relationships between these variables remain unclear. Knowledge of these relationships can assist service managers to allocate resources more judiciously, avoid pitfalls, and establish more realistic expectations. Data was obtained from employees and customers of a multinational hotel chain that has implemented a service guarantee program in 89 of its hotels in America and Canada. As the employee variables could affect performance in a non‐linear fashion, we relaxed the assumption of model linearity by using the Alternating Conditional Expectations (ACE) algorithm to arrive at a better‐fitting, non‐linear regression model for PSQ. Our findings indicate the existence of significant non‐linear relationships between PSQ and its determinant variables. The ACE model also revealed that service guarantee interacts with the employee variables to affect PSQ in a non‐linear fashion. The non‐linear relationships present new insights into the management of service guarantees and PSQ. Explanations and managerial implications of our results are presented and discussed.