科研项目评分程序与方法

The grading method for scientific research program and its procedure

定量评分不正确一直是科研项目 (课题 )评价中存在的大问题。本方法采用先定性排序 ,后定量评分的程序 ,增加了评分者对诸课题的重要程度的认识 ,从而能较容易地把握评分尺度 ,减少评分的错误。第一步用符号检验法去掉错误数据 ,按各课题的重要程度排出前后次序 ;第二步定量评分 ,新创了用“波动系数平均值”来确定评分值的控制线 ,在掉落在上下控制线之外的不合理数据 ,最后到了较为满意的评分结果。该方法基本上解决了课题评分不正确的问题 ,据验证 ,正确率从原来的 70 %左右可提高到95%。它不但适用课题评分 ,同样也适用其它技术经济评估。

The incorrectness of fixed quantity grading is always a big problem in grading scientific research programs. The above mentioned method uses the procedure of ranking all the programs according to the degrees of their importance first and then grading with the fixed quantity, making the markers be more aware of every program's importance, it's easier for them to hold yardstick`and reduce the mistakes in grading. The first step is to use the sign checking method to remove the incorrect data, then rank all the programs according to the degrees of their importance; the second step is to grade the programs th the fixed quantity, it for the first time use 'the average value of fluctuation coefficient' to determine the control line of grading data, and then remove the unreasonable data which fall out of control line so as to achieve more satisfactory grading result.This method basically solves the problem of incorrect grading for programs. It has been proved that the percentage of correctness has been raised from originally 70% to 95%. This method is not only applied to program grading but also can be used to in other technological economy grading.