Tha Study On The Application OfBenford’s Law To Detect Problems In Micro Survey Data

As we all know, the quality of statistical data is the important Basis of applied economics research, if the quality of Basic data can not Be guaranteed, then the corresponding research results may Be incorrect. Especially in recent years, the quality of statistical data has Been one of the focuses of the government departments and enterprises. With the growing of China’s economic strength, the international economic environment has Become increasingly diverse, the government and the enterprises urgently need to Be more accurate statistics to formulate relevant policies and measures for the administration. Many scholars have Begun to invest a lot of energy to the study of data quality detection, the different evaluation methods and the research strategies was developed using the practical detection data quality. However, Because most of the statistical data of testing method is too professional and the results are poor, in line with the find out the mass detection method, the purpose of this paper tries to use a relatively new method Benford law to study the statistical data quality issues, expect to achieve a Better effect.Main methods of this article are the first digital Benford’s law, through the introduction of the related literature of Benford law, Benford law connotation and significance, and Benford from production to the development of the related theory, analyzed from two aspects Both at home and aBroad to scholars to Benford law apply, the use of learning at the same time puts forward Benford law should Be used in the micro investigation data of detection and the possiBility of implement, through theoretical and empirical research, proved how in microscopic investigation data using Benford law to test statistics. Using Benford law, at the same time, choose micro2011national Baseline survey data in the dataBase, and after2012provinces tracking survey data of some indicators, through the grouping of different data types has carried on the reliaBility analysis, the conclusion of the selected indicators have a real survey data in the data quality proBlem, especially the numBer5as the first frequency of aBnormal prominent, considering the micro data acquisition process, however, the survey data error is inevitaBle; Test results analysis By using data description, after comparing to come to the conclusion that the two provinces tracking data By data investigation is further and more accurate than the national Baseline survey data points, the quality is Better. Finally,at the completion of the Benford rule to test micro survey data, provides some suggestions to the micro dataBase researchers.