Summary
Contents
Subject index
It charts the new and evolving terrain of social research methodology, covering qualitative, quantitative, and mixed methods in one volume.
Combining Different Types of Data for Quantitative Analysis
Combining Different Types of Data for Quantitative Analysis
A man [sic.] with a watch knows what time it is.
A man [sic.] with two watches is never sure.
Segal's Law
Introduction
Data do not occur naturally, nor do data ever speak for themselves, nor does there exist an obvious interpretation for a datum. Instead, data are manufactured and interpreted to fit a particular research purpose or line of argumentation. Empirical detection and interpretation of presences or absences, patterns, order, structure, or change, regardless of whether inductively or deductively derived, are the outcome of theoretical models and assumptions underlying analysis, of which data are an integral part. Already in 1964, Coombs wrote: ‘knowledge is the result of theory – we buy information with assumptions – “facts” are inferences, and so also are data and measurements and scales’ (1964: 5). If data production is part of the constitutive process of research, then from where do they come and of what are they made? And if data are indeed thus produced, what are the advantages of using more than one dataset for a particular research purpose?
This chapter is about what and how data are detected and used, and, as a consequence, how certain limitations thus arising may be overcome by using more than one dataset. Of particular interest are different types of data and how they are selected and combined in modern research designs. For this objective, it is necessary, first, to conceptualize data and their integral position within the research process, second, to understand the process of data production, and, third, to explain the possibilities and limits of using more than one dataset for a research project. This chapter will not deal with data analysis issues specifically but will nevertheless cover reasons for which more than one dataset could be used in quantitative research. In addition, while many of these issues could be applicable to qualitative or mixed-methods analysis, the explicit focus here is on quantitatively oriented research. Finally, there exists an excellent literature on validity and reliability, which connects in many ways to the use of more than one dataset for a particular research purpose. In this text, however, such issues are not covered in detail. The utility of using multiple datasets transcends quality issues relating to classical validity concerns but tends to be under-theorized. This chapter addresses this omission.
Data and the Research Process
Prompted by various introductory texts and lectures on research methods and methodology, most people understand empirical research as a tripartite process: the conceptualization of a research question, the collection of data, and the analysis of these data, from which the research results emanate.
The Conventional View of the Research Process
The conventional model about the research process connects the four principal research components, i.e. research question, data collection, data analysis, and the research results1, in a specific way. Figure 35.1 illustrates this conventional view of the research process.
There are three fundamental problems with this research model: chronology, fragmentation, and apparent inevitability:
- Chronology: This conventional model implies a chronological ordering of the different parts of the research process such that researchers appear to have settled on a research question before they collect or select appropriate data, and only then would they consider how these data are to be analyzed. Thus, the model strongly implies a deductive approach to research, while inductive research, including data exploration and visualization, are either ignored or spurned2. In practice, researchers often either formulate or at least adjust their research questions according to the characteristics of the available data. This is particularly the case with secondary analysis of existing data, where researchers often create proxies from variables that may be related to, but do not fully connect with, a construct under investigation, or they adjust their research questions or models to create a more adequate fit between the constructs embedded in the research question and the data available. Moreover, few researchers are unclear about what analytic techniques they will use, at least in general terms, before they have collected their data, often selecting the analytic strategies and methods according to their analytic competences and habits. Quite often, specialists in multidimensional scaling, correspondence analysis, latent class analysis, etc. tend to stick with the technique with which they are familiar.
- Fragmentation: Due to the conventional tripartite division of the research process, researchers tend to focus on the details relating to the components of the research process – research question, data collection, and data analysis, while neglecting the intricate relations between them. However, the quality of the research process and its results are at least as dependent on the interconnectedness between the components as they are on the components themselves. Due in part to this fragmented research design, many research results are unconvincing or incommensurable with other research findings, despite the availability of appropriate data and the application of sophisticated analytic techniques. This connects to some extent to John Tukey's suggestion that there exists an error source far more treacherous than the Type I or Type II error: the greatest threat to validity, the ‘Type III error,’ is asking the wrong questions of the data (cited in Raiffa, 1968).
Figure 35.1 The conventional view of the research process
...
- Loading...
Get a 30 day FREE TRIAL
-
Watch videos from a variety of sources bringing classroom topics to life -
Read modern, diverse business cases
-
Explore hundreds of books and reference titles
Read next
More like this
Sage Recommends
We found other relevant content for you on other Sage platforms.
Have you created a personal profile? Login or create a profile so that you can save clips, playlists and searches