A Study on Causes of Errors of Enterprise R&D Statistics & Accounting Data

: The level of the R&D statistics quality is directly related to the accuracy of the S&T policy research and making, so to explore the causes of the R&D statistical error is great significance for improving the R&D statistics quality. On the basis of the results of previous studies and seek expert advice, the paper summarizes the possible causes of the R&D statistical error. Through field research to get first-hand data of Shaanxi province, the paper identifies the key causes of the R&D statistical error based on principal component analysis. We found that: the key causes of the R&D statistical error are followed by inconsistent of R&D statistical and accounting port, lack of training to the R&D fill staff, not allowed to grasp the connotation and extension of the R&D indicators and more difficult to obtain data. Therefore, the R&D statistical work should: on the one hand to strengthen among the various departments exchange and communication to clarify the port of R&D statistical and accounting; the other hand should be carried out theoretical studies to solve technical problems in the accounting and define the connotation and extension of the R&D indicators; Moreover, it should be do R&D statistics training to improve professional quality of R&D fill staff. Key words: Principal component analysis; R&D statistics; Statistical error