Abstract
Yourdon and Constantine (1979), De Marco (1979), and Gane and Sarson (1979) introduced the data flow diagram (DFD) more than a quarter of a century ago, as a systems planning tool that is particularly useful in the fields of software engineering and information systems development. But the DFD is not restricted to those fields. Empirical research projects are systems too (which consist of interconnected sources, data, collection processes, files, analysis processes, knowledge, and users), and those systems are similar to information systems. This article reports how the DFD can also be useful in planning empirical research projects. This finding should be advantageous to research planners, individual researchers, research advisors, research supervisors, or research managers. And it should be especially advantageous to research planners in information and communication technology (ICT) because they know DFDs already, so they can get the planning advantages with little or no extra learning effort. This finding was obtained from two research projects. The first was planned without the aid of a DFD and failed. It was then replanned with a DFD and redone in a second project, which succeeded.
TopIntroduction
Yourdon and Constantine (1979), De Marco (1979), and Gane and Sarson (1979) introduced the data flow diagram (DFD) more than a quarter of a century ago, as a systems planning tool that is particularly useful in the fields of software engineering and information systems development. But the DFD is not restricted to those fields. Empirical research projects are systems too (which consist of interconnected sources, data, collection processes, files, analysis processes, knowledge, and users), and those systems are similar to information systems. This article reports how the DFD can also be useful in planning empirical research projects. This finding should be advantageous to research planners, individual researchers, research advisors, research supervisors, or research managers. And it should be especially advantageous to research planners in information and communication technology (ICT) because they know DFDs already, so they can get the planning advantages with little or no extra learning effort. This finding was obtained from two research projects. The first was planned without the aid of a DFD and failed. It was then replanned with a DFD and redone in a second project, which succeeded. The DFD that turned failure into success is Figure 1.
Figure 1. DFD of research on DFD’s in research
The second project had the exploratory aim of demonstrating that DFDs can be useful in planning empirical research projects (the more ambitious aim should calls for further research). Figure 1 shows that the aim was achieved by means of DFD examples that speak for themselves: so no elaborate data collection was necessary and neither was any sophisticated data analysis. First, textbooks of research methods were surveyed to develop a structural model of empirical research projects: that model identifies seven major types of components as key planning issues. Then DFDs were drawn from the research proposals of three recent research projects to demonstrate that those DFDs explicitly identify the major decision components. This means that DFDs can be useful in focusing attention on the key planning issues. Third, faulty DFDs were drawn, some from initial proposals of old research projects, and others by conducting thought experiments that distorted the structural model in various ways (Brown, 1992): these examples demonstrate that DFDs readily expose planning errors. This means that DFDs can be useful in identifying planning errors. Therefore, by focusing researchers’ attention on key planning issues and by enabling them to identify planning errors, DFDs can be useful in planning empirical research projects.
TopBackground
DFDs were introduced in 1979 by Yourdon and Constantine, De Marco, and Gane and Sarson, and thereafter were recommended by many other authors (e.g., Awad, 1985; Budgen, 1994; Burch, 1992; Buxton & McDermid, 1991; Coleman & Baker, 1997; Conger, 1994; Fairley, 1985; Hawryszkiewycz, 1991; Jeffrey & Lawrence, 1984; Kendall & Kendall, 2002; McDermid, 1990; Millet, 1999; Powers, Cheney, & Crow, 1990; Satzinger, Jackson, & Burd, 2004; Schach, 1993; Sommerville, 1992; Stevens, 1981; Weinberg, 1980; Whitten & Bentley, 1998; Wieringa, 1998). DFDs have mostly been used to plan software packages and information systems by outlining their major components in a top-level DFD, and then decomposing major processes into more detailed lower-level DFDs. But only top DFDs are considered here. Yourdon and Constantine noted that the major processes in the top DFDs are usually afferent or efferent (this important insight has largely been ignored by current textbooks). In the case of an information system, afferent means data collection, and efferent means information extraction: so the top DFD of an information system can be drawn as a series of rows shaped like Figure 2 (Mende, 2007).
Figure 2. Structural model of an information system
Therefore, Figure 2 serves as a structural model of an information system.
Key Terms in this Chapter
Empirical Research Project: The entities and processes that answer a group of related research questions or confirm a group of related research hypotheses using empirical methods of data collection and data analysis.
Data Collection Process: A combination of human activities and computer processes that get data from sources into files. It gets the file data using empirical methods such as questionnaire, interview, observation, or experiment.
Research Hypothesis: A tentative answer to a research question. A previous researcher may have suggested a potential answer that could be verified empirically, or the researcher may be able to deduce consequences of a previous researcher’s model or theory.
Data Source: An entity or group of entities from which data can be collected. The entities may be people, objects, or processes.
Research Question: An unsolved subproblem of a research problem, for example, a classification scheme that is needed but not yet available, a classification scheme that is not entirely comprehensive, an unknown cause-effect relationship, or an unknown technique.
Data Flow Diagram: A diagram of the data flow from sources through processes and files to users. A source or user is represented by a square; a data file is represented by rectangles with missing righthand edges; a process is represented by a circle or rounded rectangle; and a data flow is represented by an arrow.
Data Analysis Process: A combination of human activities and computer processes that answer a research question or confirm a research hypotheses. It answers the question from data files, using empirical methods such as correlation, t-test, content analysis, or Mill’s method of agreement.
Knowledge: Information about a class of entities, events, or relationships, which is likely to remain true for a relatively long period of time.
Research Problem: A phenomenon that people cope with by applying knowledge. It usually divides into several subproblems, for example, a classification of the different kinds of entities and events involved; cause-effect relationships between attributes of entities and events; and techniques for manipulating relationships to achieve desired outcomes.