Research design is of critical importance in social research, despite its relative neglect in many methods resources. This book discusses the nature of design, offers a flexible approach to new designs, and looks at a range of standard design models. Stephen Gorard‘s book is illustrated with case studies of real work and concludes with suggested readings and topics for discussion in seminars and workshops, making it an essential read for all social science researchers, concludes Barbara J. Cooke.
This review originally appeared on LSE Review of Books.
Research Design: Creating Robust Approaches for the Social Sciences. Stephen Gorard. SAGE Publications. February 2013.
It is rare to find a clearly written social science research book that is entirely generalizable across the social sciences and even more rare to find one that is genuinely enjoyable to read. However, with his new book entitled Research Design: Creating Robust Approaches for the Social Sciences, Stephen Gorard has managed to achieve the seemingly impossible. The book not only provides an essential introduction to research design in the socials sciences for new researchers, but also acts as a good reminder of best practice design for seasoned researchers. Best of all, throughout the entire book, Gorard provides clear instructions, examples, and rationale for the various types and components of research design. Supplementing his fourteen chapters are suggested readings and exercises with accompanying notes. I found these exercises and notes particularly well thought out and helpful for kinesthetic learners or anyone that particularly benefits from practical exercises.
Perhaps the most important feature of Gorard’s book is the narrative of best practice social science research that runs right through the book. From the beginning he asserts that sloppy social science research has the potential to have as equally detrimental effects as medical science research, but because these side effects are not as immediate as drug trials per se, we are less stringent about properly evaluating public policy. Another problem he notes is that many public policies and interventions are implemented without ever piloting them to evaluate their effects. Additionally, many policies are put into place based on shoddy evaluations and faulty results. Gorard asserts that the field of social science should demand rigorous research designs to avoid wasting taxpayer money on ineffective or harmful policies and/or interventions.
With this aim in mind, Gorard sets out to define best practice social science research design. One of the main strengths of the book is that he makes this definition accessible to all relevant fields within the social sciences by clearly defining his terminology and sticking to the bare bones of research design and staying clear of discussion of methodology. This adds to the abovementioned clarity of Research Design.
Gorard breaks the book into five parts, guiding readers through the research design process. He starts with the essential distinction between research design and research methods. Though this might seem like a simple distinction, he argues that new social science researchers often muddle the concepts and it is essential that the design come first as it will inform the methods. Gorard goes on to discuss research questions and actual research design that is based on the objectives of the research as outlined by the research questions. Throughout these discussions he continually refers back to the issue of warranting research claims and how research design affects our ability to substantiate any claims made by researchers. Gorard repeatedly asserts that current practice repeatedly makes unwarranted claims that can be traced back to poor research design.
Chapter 4 “Warranting Research Claims” is designated entirely to the discussion of research claims. As this discussion is the heart of the book, I feel is deserves further discussion in this review. Gorard defines a warrant as “the logical argument, with supporting evidence if needed, that shows why a specific research conclusion should be preferred over all others” (p. 41). He argues that the warranted research claim put forth by the research document should be the most simple and logical conclusion of the research and evidence in question and that too often social science researchers do not sufficiently explain why their conclusion is the most simple and logical in comparison to the plethora of other conclusions that could be drawn. Additionally, Gorard accuses social science researchers of not fully understanding the statistical analyses that they cling to and rely on to support their research claims, particularly statistical significance testing which he further argues against in Chapter 12. Though many of Gorard’s assertions and claims could be seen as controversial, this chapter serves as an excellent reminder of the impact that research design has on warranted research claims.
Another chapter that serves as a good warning for new researchers and as a good reminder for seasoned researchers is Chapter 11 “Challenges for Validity”. Here, Gorard rakes through all of the possible challenges to a study’s validity and can consequently make the reader cringe and consider new career prospects. These are essential considerations though and are valid threats to research claims. Gorard does make helpful suggestions and some of these threats may have never before been brought to the researcher’s attention (e.g. routine SPSS errors in calculation). I found one of Gorard’s assertions particularly significant: the simpler the design the less error the researcher has to account for. Social science researchers can often be tempted with complicated research design, but this complexity exposes them to error and biases (e.g., measurement error) that can invalidate their results and conclusions.
Gorard’s tough and unapologetic criticisms of current practice may be a hard pill to swallow for some, but I believe that Research Design is an essential read for all social science researchers. My one criticism would be his apparent oversight of the challenges particular fields face in gaining research access and maintaining control over an idyllic research design (e.g., prisons research). That being said, he does assert that the book lays out best practice and that alterations to a best practice design can be made, but only as long as the researcher acknowledges how these alterations affect the design and subsequent claims. Gorard wants the field to hold higher standards for research and to subsequently stop implementing policies and/or interventions that are either untested or poorly tested—this is particularly relevant for certain fields, like criminology, that increasingly value evidence-based interventions. However, this seemingly thorny task is entirely feasible and Gorard beautifully lays out a framework for best practice research design.
Barbara Cooke is a PhD candidate at the University of Cambridge Institute of Criminology. Previously, she studied at the University of California, Santa Barbara and Royal Holloway, University of London. She tweets as @bcooke121. Read more reviews by Barbara.