Statistical Data Quality Assessment Methods

Statistical data is the information carrier. With the development of marketeconomy, the needs of information from all aspects of society are growing. Thegovernments and enterprises ask for higher statistical data quality in order to take thedominant position in the fierce market competition. The quality of statistical datawhich is directly related to the economic development, plays an important role in allkinds of decision-making. By reviewing the recent statistical data published by theNational Bureau of Statistics of China, and comparing the information accuracy withpublic demands or international standards, you’ll find there is still a long way to go.Especially there are a lot of problems with our GDP. Scholars have found that the sumof the provinces released GDP statistical data is much greater than national GDPstatistical data released by NBS. It shows that the statistical data released by NBS iscredible to some extent. however, on the basis of the development of the society andeconomy, there are still some problems with the quality of statistical data in ourcountry, such as false statistical data, statistical data lag, statistical system fails tomeet the requirements of the current economic management system, etc. As a result, areasonable and appropriate assessment and diagnosis method is required to maximizethe statistical data quality.As we all know, GDP is one of the best indicators to measure a country’seconomic situation, not only can it reflect a country’s economic performance, but alsonational strength and wealth. The accuracy of the GDP statistical data would affect theeconomic entities’ judgment of the market development, therefore the quality of GDPstatistical data has become the focus of widespread social concern. In this context, inorder to improve the GDP statistical data quality and strengthen the efficiency of theeconomic operation, government started to improve the original GDP assessmentapproach. NBS developed “regional quarter GDP statistical data assessment approach”in2009which audited the regional quarter GDP statistical data according to thedevelopments and changes of main macroeconomic indicators of the relevant departments and industries. If GDP doesn’t match the rules, NBS will adjust theregional GDP and the growth rate with current prices, and then announce to thepublic.GDP statistical data assessment approach developed by NBS just based on thesimple descriptive statistics. We tried to work on modeling the approach from theangle of match rules, which means that, firstly a reasonable model would be chosen toexplain the relationship between one of the microeconomic indicators and GDP, thenstepwise local influence analysis would be introduced to detect the abnormalobservations. The paper is divided into five parts, the first part briefly presents thesignificance of the topic, a literature review of past research and methods, as well asinnovation of the paper; the second part gives a specific description from the thoughtto theory of the stepwise local influence analysis which is mainly used in part four;the third part discusses the match degree between the match indicators in order to findthe right model fitting the relationship between them; the fourth part examines theGDP quality and finds the outliers in nearly30years of China; and the last part comesto the conclusion and suggestion.