PURPOSE-MOTIVATION

The authors are motivated by published recommendations of other researchers espousing the use of various research methods. The authors’ purpose is to articulate the usefulness of multiple research methods in order to compensate for the inherent weaknesses within each method and to provide a more complete understanding of research phenomena by examination from different perspectives. The choice of a specific research method is guided by the instant research question, the current state of knowledge regarding the phenomenon and the feasibility of conducting a particular method. The use of different methods at various stages of a research stream is viewed as a circular, iterative process (see Figure 1 on pg. 53), wherein the results of subsequent studies reinforce, augment and triangulate the findings from prior studies, or raise new questions to be addressed in subsequent studies. Two variations of the multiple method approach are suggested: 1) using more than one method when conducting a single study and 2) using differing methods to examine a phenomenon by conducting a coordinated series of studies.

SUMMARY

Empirical research in management accounting is typically either basic research, conducted to describe, explain and predict management accounting phenomena or applied research, conducted to evaluate a particular design or implementation of an accounting system with the goal of policy recommendation or design improvement.

The authors have defined the following three types of empirical research method (p. 34):

>Field research-occurs in natural settings that are not created for the sole or primary purpose of conducting research. Three primary types of field research are case studies, field studies and field experiments.

>Experiment – involves the manipulation of the independent variables and the observation of their effects on the dependent variables. True experiments require random assignment of sampling units to treatments. Quasi-experiments do not require random assignment to treatment groups. Laboratory experiments are differentiated from field experiments by the nature of the setting, with a laboratory experiment conducted in an artificially created setting designed for research control while a field experiment is conducted in a natural setting designed for other purposes.

>Survey – employs a standardized approach in order to collect information from sampling units to make inferences about the population. Typically conducted through the mail, telephone or personal interview.

The authors compare the three research methods using five dimensions, three derived from Runkel and McGrath (1972): actors, behaviors and contexts; and two additional dimensions: researcher and research application. The comparison illustrates the strengths and weaknesses inherent in each of the research methods. The comparison is summarized in Table 1 (p.36-37), reproduced below.

TABLE 1: Comparison of Empirical Research Methods

The four validity criteria are defined (p. 38-39):

Ø Internal validity-assessment of whether the changes in the dependent variable are caused by changes in the independent variable. Statistical validity is a necessary condition for internal validity.

Ø External validity-whether the observed causal relations in one study can be generalized to or across groups, settings and times. Construct validity and internal validity are necessary conditions for external validity.

Ø Statistical conclusion validity-whether the independent and dependent variables covary, given the statistical evidence. A necessary condition for internal validity.

Ø Construct validity-whether a particular operational definition of a construct is a valid measure of the construct. Assessed by measuring convergence of various measures and divergence of various measures.

The researcher may inadvertently introduce two behaviors in participants: demand effects, in which participants behave in a certain way because they believe it is required of them or expectancy effects, in which participants behave in certain ways because they believe it is what the researcher expects or wants them to do.

Some of the considerations related to Actors, or participants, are concerned with the usefulness of students as appropriate surrogates for other groups. Key considerations include the nature of the task and the ability of the surrogate to perform the task. The authors (p. 45) also point out two reasons why the representativeness of the sampling units may not be a factor (Lynch, 1982): 1) test of theory are stronger if there is small error variance, which can occur when a homogeneous sample is used and 2) if a theory is universal it should apply to any and all people who satisfy its boundary conditions.

The six methods of recording data are (p. 48):

Ø Subjective reports – an actor is aware of being observed and recorded.

Ø Trace – record of an actor’s behavior when the actor is not aware the record will be used for research purposes

Ø Visible observer – actors know that a researcher is observing and recording their behavior

Ø Hidden observer – actors are not aware that a researcher is observing and recording their behavior

Ø Public records – written by someone other than the researcher, actors are aware that their behavior is being recorded

Ø Archival records – written by someone other than the researcher, actors are not aware that their behavior is being observed and recorded or may be aware that the behavior is being observed and recorded but not have an expectation that the record will be used for research purposes.

Research context may vary significantly and is an important concept in its effect on the validity of research results. Two elements of context are defined (p. 49-50):

Ø Experimental realism – increases when the research context engages or involves the participant fully in the research. The goal is to increase the probability that the research context captures the intended essence of the theoretical variables

Ø Mundane realism – whether the research context attempts to include superficially relevant features of the "natural world".

The authors illustrate the usefulness of the multiple method approach by an examination of selective literature on budget setting, an important area of management accounting research. They summarize this illustration as follows (p. 57-58):

"The catalyst for the investigation of budget setting was non-accounting research on participative decision making…initial accounting studies were primarily done in the field and documented the nature of budget setting. Surveys followed that expanded the sample of firms and the situations observed. Laboratory studies were then used to test for causal relations among key variables identified by the other methods. Finally, field studies were used to reexamine some of the original issues, but with insights gained form both the laboratory and survey studies."

This demonstrates the circular nature of research and the way in which each research method is informed by and supports the other methods. The use of multiple methods contributed to the extension of the research stream.

The authors believe that significant changes in organizations will result in the near future and advocate the use of multiple research methods in the examination of the impact of these organizational changes on management accounting.

	Web maaw.info