考虑M eta分析中偏倚对总体效应估计的影响。在模型中引入表示偏倚的参数，构造基于偏正态分布的模型，将偏倚与异质性同时纳入模型，并进行估计，由估计量分析偏倚对异质性和总体效应的影响。在基于偏正态分布的模型下，对总体效应量采用加权最小二乘估计，并进一步修正为无偏估计。与以往基于正态分布的模型相比，该模型将偏倚进行了量化，并从总体效应的估计中予以剔除，从而使合并效应量剔除了偏倚的影响，而且较基于正态分布的模型更符合实际情况。%The bias influences on the overall effect are considered in Meta analysis .The bias parameter was introduced to structure the model based on the skewed normal distribution .At the same time ,bias and heterogeneity were taken into account and estimated .The influence of bias to the heterogeneity and the overall effect was explained by the estimators .Based on the model of the skewed normal distribution , the overall effect was estimated by using the weighted least squares which was further corrected to be the unbiased estimator .Compared with the model based on normal distribution ,the bias is quantified and removed from the overall effect estimator ,so that the combined effect is gotten rid of the bias influence , and more in line with the actual situation .
展开▼
