Henry E. Brady and David Collier, editors, 2004. Rethinking Social Inquiry: Diverse Tools, Shared Standards. Lanham, MD: Rowman and Littlefield. Chapter 1.
Henry E. Brady, David Collier, and Jason Seawright.
Mainstream Quantitative Methods, Qualitative Methods, and Statistical Theory.
Social scientists should seek a shared framework allowing researchers using diverse analytic techniques to develop evidence that is convincing to analysts of differing methodological persuasions. A larger body of mutually accepted evidence can help contribute to finding better answers to the substantive questions that drive social research.
- Perspectives on "mainstream quantitative methods," an approach based on the use of regression analysis and related techniques for causal inference.
- To some, mainstream quantitative methods is superior to quantitative research.
- Qualitative methodologists doubt the quantitative approach is the only appropriate model for analysis.
- Those who follow "statistical theory"/"statistical rationale" are skeptical about applying assumptions behind regression analysis and related tools to real-world data and advocate alternative techniques to allow limited inferences based on fewer untested assumptions.
- Quantitative research / large-N research
- Superior tools for solving many problems of methodology and research design
- Can frame and generalize the findings of qualitative studies
- Suffers from procrustean quantification and a jumble of dissimilar cases
- Increasing the N may push scholars to compare cases that are not analytically equivalent and adding observations from a different spatial or temporal context or at a different level of analysis can extend the research beyond the setting for which the investigator can make valid inferences. Concern with context is a prerequisite for achieving descriptive and causal inference that is valid and rigorous.
- Increasing N may push scholars toward and untenable level of generality and a loss of contextual knowledge.
- Analysts must assume that they have the correct statistical model to begin with.
- Bounds the generality of research findings
- Sustained attention to conceptual issues
- Case-oriented scholars use flexible analytic frame that can be modified in light of the insight into cases they gain in the course of their research. This aspect makes of the case-oriented approach makes it well-suited for concept formation and theory development.
- Better suited for exploring the tipping points that play a critical role in shaping long-term processes of change and for providing more nuanced insight into findings derived from quantitative investigation.
- Handicapped by a lack of quantification and small numbers of observations
- It is difficult to make causal inferences from observational data, especially when research focuses on complex political processes.
- Forgoes large-N tools for measurement validation
- Loss of generality in research results
- Difficult to eliminate rival explanations
Research design involves fundamental trade-offs. Methodological advice needs to be framed in light of basic trade-offs among:
a) alternative goals of research
b) the types of observations researchers utilize
c) the diverse tools they employ for descriptive and causal inference
Alternative methodological tools are relevant and appropriate, depending on the goals and context of the research. Inference in quantitative research can sometimes be improved through the use of tools strongly identified with the qualitative tradition. Ideas drawn from qualitative methodology can improve quantitative practices by addressing weaknesses in quantitative approach
Scholars should develop shared standards. A basic goal of methodology should be to establish shared standards for managing these trade-offs. Shared standards can become the basis for combining the strengths of qualitative and quantitative tools.