In December 2019, the American Enterprise Institute released a report, "American Family Diaries: Can Ethnographic Research Assist Public Policy?" ". The report includes the latest research using qualitative research methods such as ethnographic observation and in-depth interviews to investigate the causes of poverty and impediment to social mobility, and explores how qualitative research can be used to improve public policy development. Our reporter interviewed relevant scholars on the characteristics of qualitative and quantitative research methods and their application in social science and policy research.
Two revolutions in policy research
Aparna Mathur, a researcher in economic policy at the American Enterprise Institute and Jennifer M. Silva, an assistant professor of sociology and anthropology at Bucknell University, said in the United States, Policy research on poverty and social mobility mainly uses quantitative research methods that rely heavily on large-scale representative surveys and administrative data. Quantitative research focuses on a large and diverse set of information about people's beliefs and behaviors, and aims to generalize numerical relationships between variables to predict future behaviors.
According to Robert A. Moffitt, a professor of economics at Johns Hopkins University in the United States, there have been two revolutions in data collection and analysis in social science research, especially policy-oriented research, in the past 50 years. . The first revolution was the development of large-scale, population-representative household surveys, such as the current demographic survey and the National Health Interview Survey, which collected hundreds of key socio-economic variables. Estimate statistics and correlations at the total population level. Combined with household panel surveys (multiple visits to the same sample family at different points in time), these surveys have completely changed researchers' understanding of employment, income, family structure, and migration dynamics of low-income Americans.
The second revolution was the development of large-scale computing technology and large-capacity storage capabilities, which enabled high-speed download, storage, and analysis of data. The combination of individual-level household surveys and rapid analysis methods has created an unexpectedly massive accumulation of knowledge. Related to this is the rise of social experiments and randomized control experiments. The proposal for social experimentation originated from the research on negative income tax by Guy E. Orcutt and other econometricians in the 1960s and 1970s. Random control experiments require large-scale data collection, large-scale samples, and appropriate calculation methods. They are important tools for examining the impact of social policies on low-income people. They are generally considered to provide the best causal effect on a policy Evaluation.
Both quantitative and qualitative have their own advantages and disadvantages
Hu Shouping, founding director and chair professor of higher education at Florida State University's Higher Education Success Center, told our reporter that quantitative, qualitative, and quantitative-qualitative hybrid research methods are the basic methods of social science research. Quantitative research methods are regarded by the general public and many decision makers as more scientific and objective methods, while qualitative research methods have to spend time and energy to justify their science and value. Generally speaking, the quantitative research method is more suitable for more mature research fields, and it is more direct and powerful for the verification of research hypotheses. The qualitative research method is more valuable for the research fields that people know less, and it is conducive to generating research hypotheses. Quantitative research methods are conducive to verifying causality or to produce general and universal conclusions. Qualitative research methods are conducive to the understanding of unique individuals and unique phenomena. Quantitative and qualitative research methods are widely used in the field of humanities and social sciences, but they also vary depending on the academic tradition of the discipline and specific research issues, and are also affected by factors such as research purpose, time resources, and research funding. Generally speaking, small-scale research projects tend to use one of qualitative and quantitative methods, and large-scale and well-funded projects tend to use quantitative and qualitative hybrid research methods.
Hu Shouping said that the pros and cons of policy formulation depend on the rigor, scientificity, comprehensiveness and matching of the evidence used. Qualitative research methods emphasize the perspectives, feelings, and experiences of the parties, and also pay more attention to the vulnerable groups. They have a positive effect on the rationality of policy formulation, but because the parties are limited by their own interests, vision, and goals, it is necessary to be critical in some cases. Interpret and use qualitative research. Quantitative research results related to vulnerable groups may be obscured by "trends" in the data, while qualitative research can make up for this deficiency. In policy evaluation, quantitative methods can help establish or deny causality, and qualitative methods can help further understand why causality exists. In short, various research methods have their advantages and disadvantages, and should comprehensively consider factors such as specific research problems, goals, resources, time and environment.
Big data gives birth to "qualitative big data"
The rise and rapid popularity of big data has profoundly affected the scientific research process. Hu Shouping said that with the rapid expansion of data collection channels and capabilities and the improvement of computing power, big data has become a hot area for quantitative research, which will have more or less impact on qualitative research. The public may be more convinced of the science and practicability of big data and quantitative research methods, thus ignoring the rationality and contribution of qualitative research methods. Decision makers will also be more inclined to use big data and quantitative research findings as the basis for decision making and ignore qualitative research. result. Big data has created conditions for the broader application of quantitative research methods, and the two appear to be more "adapted", but the purpose, advantages, and applicable fields of qualitative research methods will not become obsolete or disappear. In contrast, qualitative research can reduce the negative consequences of vulnerable research based on quantitative research based on big data. Of course, due to limited resources, the popularity of big data may make it more difficult for traditional qualitative research projects to obtain resource support.
Lynn Jamieson, a professor of sociology at the University of Edinburgh, and Sarah Lewthwaite, a researcher at the National Research Methodology Centre of the Economic and Social Research Council of the United Kingdom, suggested that one shortcoming of qualitative research is that It is difficult for researchers or even teams to process large amounts of qualitative data quickly. Traditional, rigorous analysis methods require researchers to be "immersed" in the data to a certain extent, and even with software, it still takes a long time. This question aroused the interest of social scientists in big data and the preference of using quantitative methods to analyze quantitative data sets, which in turn gave birth to "qualitative big data" that combines the breadth of quantitative research with the depth of qualitative research.
According to Hu Shouping, "qualitative big data" refers to a qualitative database collected and constructed by applying qualitative research methods on a large scale and multiple angles. It is a new concept that coexists with quantitative big data. It has the advantage of observing and understanding the interactions of members in the system and the interactions between individuals on a large-scale, multi-angle, and all aspects, creating conditions for a more comprehensive understanding of the impact of individuals, organizations, systems, and policies.
According to Mather and Silva, the fundamental goals of quantitative research include universal applicability, efficiency, reproducibility, and transparency, which "coincides with" emerging big data-the purpose of big data analysis is to make the digital Speak "to show and predict human behavior patterns and trends in an unbiased manner. Although the size of quantitative data sets continues to expand and cover more areas of human experience, data analysis technology is also becoming faster and more complex, and social science researchers and public policy experts are increasingly interested in qualitative research.
Qualitative research is based on "discovery logic" and attempts to look at the world from the perspective of the investigator. The purpose is not to prove or falsify a theory about behavioral driving factors, but to start from the meaning system that people create and share in daily life. Propose a new theory. Mather and Silva emphasized that researchers "immersed" in the lives of the investigators in order to unveil the "world of meaning" that drives their behavior and guides their decisions, and to unearth the driving mechanism of the demographic model displayed by quantitative data. In the era of big data, researchers should seek to use all research methods in an optimal way to fill gaps in existing knowledge and provide better policy references, rather than sticking to one method.