Studies On Methods Of Evaluating Agreement Among Multiple Investigators (Methods Of Measurement)

Large-scale and joint scientific researches and new methods keep emerging in the sphere of sociology, behavior science and biomedicine. In some cases a project involves a number of investigators who use different methods of measurement. Their findings usually stem from the comprehensive analysis by different investigators using different methods. Therefore, the accuracy and reliability of the results among investigators (or methods) is very important, which are the basis for later study. Different methods are based on different theories. To date, there has been no good agreement evaluation method intended for the continuous variables among more than two investigators (or methods), despite many agreement evaluation methods. In practice, few scientists could clearly understand agreement evaluation methods, nor could they choose a proper method in agreement evaluation. The conclusions they arrive at may be incorrect because of improper methods used. Furthermore, many scientists tend to draw a conclusion of agreement or otherwise without due attention to the reason for disagreement and to methods of dealing with the disagreement. Sometimes the reason for disagreement is important to later studies. Based on current studies of agreement evaluation, this study focuses on: (1) systematic summarization of various agreement evaluation methods; (2) systematic analysis of the advantages and disadvantages, and conditions of application of agreement evaluation methods ; (3) improvement of the method of limit of agreement by suggesting a new agreement evaluation method for more than two investigators (or methods); (4) raising a new agreement evaluation method for more than two investigators (or methods) using methods of survival analysis; (5) investigation of the bias of agreement evaluation of all kinds of data and offering new methods that can recognize and solve the bias; (6) new applications of the methods of conducting comprehensive analysis of the same measurement variables tested by different instrument; (7) suggesting an evaluation method of multiple investigators in...