Abstract: 【Purpose/significance】There are three paradoxes in multi-attribute academic evaluation. First, there are many multiattribute evaluation methods that are difficult to choose. Second, there is a dispute between subjective and objective evaluation meth⁃ ods. Third, the mainstream multi-attribute academic evaluation is linear weighted summary method. The existence of the problems makes the error problem in the selection of multi-attribute evaluation methods complicated and confusing.【 Method/process】This pa⁃ per analyzes the causes of the three paradoxes, and conducts an in-depth analysis of the error problem of multi-attribute evaluation methods, focusing on the subjective characteristics of the true value of academic evaluation, and proposes a new idea to reduce the se⁃ lection system error of multi-attribute evaluation methods.【 Result/conclusion】The research found that it is difficult to solve the sys⁃ tematic error problem in the selection of academic evaluation methods from the evaluation method itself, statistical method selection, and combined evaluation; linear subjective weighted summary is the best multi-attribute academic evaluation method; multi-attribute evaluation method innovation It is still necessary; the selection of academic evaluation methods is essentially a fusion of natural sci⁃ ences and humanities and social sciences.【 Innovation/limitation】There are still many areas to be studied for the systematic errors in the selection of multi-attribute academic evaluation methods, and further optimization and improvement are needed.