Statistical methods play an important role in social science research。The two major branches of statistical analysis methods，Descriptive statistics and inference statistics have also changed in the application of social science research。Recent，Kevin Murphy, an honorary professor at the Business School of the Business School of Ireland, wrote on the official website of the London School of Political Economics，As the amount of data continues to increase，Quantitative social research has become increasingly dependent on complex inference statistical methods，But at the same time, we should not ignore relatively simple descriptive statistics。

　　bet365 live casino games

Description statistics refers to the use of watchmaking and classification、Graphics and computing general data to describe various activities of data characteristics，It is a common method in social science research。British scholar John Grandt's "Nature and Political Observation of the Journal Bet365 lotto review of Death" can be regarded as descriptive statistics.。in the book，Grandt presented a large amount of data in some tables for the first time，It is easier to understand relevant information。

Inference through statistical data can be traced back to the 5th century BC。In the Berlone Pennie War, Greece，The number of layers of the soldiers of the soldiers of the soldiers of the soldiers，Calculate the length required to calculate the ladder。So inferring statistics are using sampling surveys，Inference overall from partially，and make a statistical method for decision -making on uncertain things。

Descriptive statistics and inferred statistics each have thousands of autumn，each has its own reasonable scope of applicable。Deputy Dean of the School of Journalism and Communication of Shandong University、Ni Wan, Executive Director of the Construction Project of Big Data and Accurate Communication Laboratory，The application of descriptive statistics has a certain subversive effect on social science research，It can help researchers clearly understand the overall context and appearance of research on the basis of big data，But this is not enough to support or even replace inference statistical research。Descriptive statistics and inference Bet365 app download statistics will not reduce their role because of the general application of big data，Both have their own strengths。

　　Avoid too dependence on inference statistics

Ni Wan said，For social science research，Big Data Generation mainly depends on the data of literature、Dataization and perspective of daily behavior、Opinions、Dataization of attitude。Where，Data of the document information by researchers design data collection index system，Use related software to generate，Meet the goal of descriptive analysis of researchers; data of daily behavior is generated by the system monitoring natural flow of natural flow，It is easier to confirm and structured; and related views、Opinions、Attitude data is relatively difficult to collect difficulties，Mainly non -structured data or semi -structured data，And these data are difficult to structure。

Murphy thinks，When researchers start to incorporate big data into their research，The tendency of filtering and analyzing data with complex inference statistical methods becomes more and more serious。Increasing statistical complexity can help researchers explore the problems that traditional data analysis tools cannot be revealed，But it also brought a lot of challenges。bet365 best casino games First，When statistical analysis of certain research becomes more and more complicated，Researchers or users of research results, the more difficult it is to understand the content of the study。Next，There is clear evidence to indicate，The more complicated the statistical analysis，The possibility of major deviation in data analysis and interpretation is greater。Last，Many methods commonly used by researchers in social and behavioral science researchers depend on "zero fake significant inspection" for statistical inference，And this method has become increasingly used in the scientific community。Realty testing of zero fake is a statistical inference method commonly used in social science research。What is a zero fake? For example, the government has launched a policy on alleviating employment difficulties，Zero fake setting is to assume that this policy is invalid，and a significant test is used to overthrow the zero fake。Assume that the government’s policy is invalid，The employment situation obtained through the survey of graduates should not be optimistic，But the final fact is that the employment rate has increased，Then the possibility of invalid policy is very low。But，Bet365 app download This method has been questioned by researchers，Because statistics are significant, there is certain uncertainty，Can't simply determine that there is a real effect based on the significant results。

Murphy said，Key decisions in statistical analysis often depend on whether the special parameters in a complex statistical model are statistically significant。But statistical significant testing is easy to understand in simple statistics，When the analysis becomes complicated，The key elements of statistical testing have become more and more complicated。

Ni Wan proposed，Too dependence on inference statistics is easy to fall into the state of blindness，Descriptive statistics can be combined with inferred statistics under the support of big data，Avoid studying a leaf barrier。

　　Actively use descriptive statistics

Ni Wan thinks，In the era of big data，Descriptive statistical research on literature data is not only convenient, but also more comprehensive analysis of research objects，Behavior data will also be more accurate，It has a basic role in overall social science research。For descriptive statistics，The value of the large sample volume is prominent in determining the data distribution and related relationship，Regardless of literature bet365 Play online games data、Behavioral data or attitude data，The analysis of the overall trend is more accurate。

Murphy said，The defect of complex inferring statistics is，Researchers are difficult to concise、Explain to users of the research results accurately，To help them formulate intervention measures or policies。In this sense，Description statistics reflect its advantage as a communication tool。Social Science researchers should fully use comprehensive charts and other tools，A good communication with policy makers and the public。other，Accurate、Comprehensive、Correct description is the basis for all empirical analysis。more attention to descriptive statistics，The possibility of a useful impact on the study。