|
主讲人 |
Tao Chen |
简介 |
<div style="font-size: 13px;"><span style="font-size: 10pt;">The abstract is: </span>We consider a method of moments approach for dealing with <span style="font-size: 10pt;">censoring at zero for data expressed in levels when researchers </span><span style="font-size: 10pt;">would like to take logarithms. A Box–Cox transformation is </span><span style="font-size: 10pt;">employed. We explore this approach in the context of linear </span><span style="font-size: 10pt;">regression where both dependent and independent variables are </span><span style="font-size: 10pt;">censored. We contrast this method to two others, (1) dropping </span><span style="font-size: 10pt;">records of data containing censored values and (2) assuming </span>normality for censored observations and the residuals in the model. <span style="font-size: 10pt;">Across the methods considered, where researchers are interested </span><span style="font-size: 10pt;">primarily in the slope parameter, estimation bias is consistently </span><span style="font-size: 10pt;">reduced using the method of moments approach.</span></div> |
