Is the traditional social survey out of date in the era of big data?

In the era of big data, the popularization and application of modern network information technology and smart devices bring challenges to the traditional social investigation methods. Some scholars have suggested that compared with the massive information obtained by data mining technology, the information obtained by traditional social surveys is only "small data". This has caused controversy in academic circles: Do traditional social surveys still need in the era of big data? In the era when big data technologies and methods are widely used, how do traditional social survey methods show their unique value? Recently, a reporter from China Social Sciences News interviewed relevant scholars.

Big data technology promotes data collection and analysis.

"The mobile Internet enables the attitudes and behaviors of social actors to be rapidly informatized and recorded by Internet devices, providing researchers with a lot of information that could not be collected by previous information collection methods. At the same time, it has greatly improved the ability of human beings to record and collect relevant information, and greatly reduced the cost of obtaining certain information. " Li Ding, an associate professor at the School of Social and Population Studies at Renmin University of China, said.

Big data technology has changed the way of obtaining, processing and understanding data. According to Du Haifeng, Executive Dean of School of Public Policy and Management, Xi Jiaotong University, the way of data acquisition has changed from collecting questionnaires or interviews to the comprehensive application of network, multimedia and other technical means, and more importantly, the change of objects. Traditional methods need to scientifically sample from the matrix, and the data acquisition object of big data may be the matrix directly; The data processing mode has changed from the traditional attribute data analysis method to the comprehensive integrated analysis based on structure and intelligent information processing. The way of understanding data has developed from the traditional statistical causality to the analysis of the law of "correlation", especially the relationship between different information.

According to Kuiyu Tang, a professor of sociology at Harbin Institute of Technology, big data technology not only has advantages in collecting data, collating data and analyzing data, but also provides effective evidence for the overall analysis of social problems. These changes are laying a new methodological foundation for sociology to return to "social facts" as a whole, and at the same time, they undoubtedly bring challenges to traditional questionnaire and in-depth interview survey methods.

Social investigation method has special advantages.

Since big data technology has such outstanding advantages in the field of information acquisition and analysis, does it mean that traditional social surveys will be replaced? The scholars interviewed do not agree with this view.

On the one hand, compared with traditional information collection methods, big data technology still has its limitations; On the other hand, traditional information collection methods still have unique value. Kuiyu Tang said, taking sampling survey as an example, in some cases, sampling survey is more suitable for those "lost" data and representative samples. Faced with the analysis of complex and interpersonal social problems, big data methods are not nuanced enough.

"A very important feature of big data is' low value density'. The content of data may not be of concern to specific researchers, so it may not always meet the needs of research on specific issues. " Du Haifeng pointed out that the traditional social survey is not only a necessary supplement for big data to obtain information, but also a more necessary basic material for special research.

The information obtained by big data technology is equivalent to census and non-probabilistic samples. However, big data is not borderless. If the boundary cannot be recognized or agreed, even though the data is very large, it cannot be used for scientific research. As Ding Li analyzed, social actors perceived and recorded through the Internet and smart devices cannot cover all actors. If you don't know the coverage of big data and what kind of group it represents, even if the sample size is large, the knowledge and laws obtained may be misleading.

In addition, the boundary of big data lies in the meaning of variables. "Although different companies and research units collect a large number of data samples according to their own needs, the variable information of each sample is very small. If these different types of database information cannot be added in series, that is, the effective information of each research object, then the research value is very limited. " Li Ding said.

Li Ding believes that the information obtained by traditional social surveys is very dense, and its purpose is more direct, its design is more standardized and its efficiency is very high. "Without using the traditional social survey method, even the most powerful Internet companies in the world today may not be able to obtain a data set with the same representativeness, reliability, validity, information density and the same variables as the comprehensive social survey in China from the existing Internet trace data."

Realize the complementary advantages of the two methods.

As Li Ding said, on the one hand, in the context of the era of big data, extracting valuable information and knowledge from big data may gain new knowledge about actors and new laws of social operation; On the other hand, researchers should recognize the limitations of big data and the advantages of traditional research methods and avoid blind worship. Traditional survey methods still have cost advantages and scientific advantages in obtaining some high-density and statistically representative data.

Regarding the academic controversy about the two methods being either one or the other, Kuiyu Tang believes that when analyzing social facts and social problems of different types and complexity, traditional social surveys or big data methods should be appropriately selected and used.

Future social science research may realize the complementary advantages of big data and traditional social survey methods. The interviewed scholars put forward some ideas. Li Ding believes that traditional qualitative research methods and sampling survey methods can supplement the deficiency of big data and help us understand its social significance. Big data can also provide important information supplement for traditional investigation and research. Qualitative research can have a more comprehensive understanding and grasp of the research object if it can obtain the trace data, social communication data and action trajectory data of the interviewee on the Internet on the basis of existing interviews and observations.