Qualitative and Mixed-methods Research in Economics

This article [gated] describes qualitative and mixed-methods in economics, their pros and cons. One interesting difference that the author, Martha A. Starr, makes between them is not the usual one in terms of numerical and verbal data, but in terms of close and open data respectively. It is close when the researcher knows the limits of the data, that happens in quantitative research. However, that is not true for qualitative data where the boundaries are harder to define - think about open-ended questions in semi-structured or unstructured interviews. 

The article describes seminal and fascinating qualitative-research in economics: 

Within the category of ‘case studies’, economists have a special tradition of ‘pin factory’ visits – traced back to Adam Smith’s observations on the 18 stages of making a pin, via which he laid out seminal ideas about the division of labor, productivity and income growth… Perhaps the most influential use of site visits was the year Ronald Coase spent traveling to US factories and businesses, talking to decision-makers and observing patterns of inter- and intra-business transactions; as discussed in Coase (1988), this material contributed centrally to the development of his understanding of horizontal and vertical integration, presented in his seminal ‘Nature of the Firm’ (1937). (p. 5)

Another interesting qualitative-method, not common in economics, is life histories that she discusses as well. Analytic narratives are missing in the article (here and here). Some Austrian-School economists have written fascinating cases which might be classified as analytic narratives. 

Some historical cases in economic growth are qualitative in nature and they can be very insightful (here and here). 

There are many good examples in professor Starr's paper. She explains instances in which qualitative methods are specially relevant: to explore a new phenomenon or to examine the validity and consistency of different samples (groups, communities, cities, etc.) used in field experiments. She also describes common criticisms to qualitative methods such as researcher bias, small n, and others. In general, they are overrated. She also gives advise on how to improve qualitative research. 

I find this paragraph quite elucidating: 

What is different in qualitative research projects is that, rather than setting up clear tests of hypotheses that result in zero/one judgments about empirical support, efforts to gauge the validity of theories or characterize causal relations tend to follow more of an ‘informal Bayesian logic’ (Bennett, 2004),* whereby researchers begin with a set of working hypotheses about the phenomenon of interest, then revise their ideas as they encounter new information. The end result then is typically not zero/one judgments about initial hypotheses, but rather sets of characterizations and explanations that have been revised and modified according to what was found in the field. While there is a good amount of discussion presently about practices that can be used to improve the rigor with which this is done, the expanding body of work in this area illustrates that, rather than being suitable for descriptive purposes only, qualitative and mixed-methods work often has good potential for identifying and characterizing causal processes, even if we are still in the process of figuring out best ways to realize it. (p 20-2)

Overall this reading was stimulating and inspiring.  

*Bennett, A. (2004) Testing theories and explaining cases. In National Science Foundation (2004). Workshop on Scientific Foundations of Qualitative Research, pp. 49–51.