The methodology section of the APA’s Executive Summary covers twelve different methodological approaches to research design. The various approaches are specified, including those based on principal factors (such as sampling, statistical inference, or the use of logistic regression), those determined by the practices (such as propensity matching and principal components analysis), and those grounded in principles of research management (such as the use of systematic reviews). A further twelve methods are recommended as being based on empirical study and practice, but not explicitly stated as such. One other suggested approach is to apply methods that have been shown to be effective research tools in other fields.
The main body of the article briefly describes the various methods listed in the paragraph above. All but one of these methods are used in virtually all fields of research. One exception is that a very few researchers opt to combine two or more methods; for instance, when applying a mathematical method to identify the population size of an animal through counting its individuals. The APA does not endorse the combining of these methods, but recommends using methods that make it easy to identify a significant threshold population. For this threshold population, a mathematical technique is preferable over an empirical method.
Methods can differ depending on the research problem at hand. In addition, there are a number of different theoretical foundations that underpin the various methods of design. Therefore, when deciding on the appropriate methodological approach, it is important to consider these various theoretical bases and understand their implications. The following discussion highlights the key features of each major methodological approach.
Principal factors refer to those aspects of the research process that drives the data, results, and recommendations. Examples include the extent to which any individual or set of participants have power to alter the outcome of the research process. Similarly, the magnitude of the effects of any variable. For instance, the inclusion or exclusion of a particular factor can significantly change the estimates of the effect of another factor. Likewise, changes in the mean level of the effect of a variable can alter the estimate of the standard deviation.
Procedure-based approaches involve obtaining enough information to allow for a confident estimation of the effect of a statistical parameter. This is usually coupled with the analysis of the data sources used. Additionally, it is assumed that the procedures followed are adequate for the purposes of good research design. Typically, this is coupled with the assumption that the participants have sufficient knowledge to effectively control the variables studied.
Based on the principles of ecological theory, mixed methods provide a way to combine the theoretical models with real data. The mixed methods section should describe how the model was constructed and how the results are applicable to the research problem. The mixed approach consists of both the theoretical models and real data sources. The description of the theoretical models should focus on the strengths of the methodology and the assumptions and the resulting models should be compatible with the real problems.
The third section of the overview explains the basic procedures used in the research problem. This includes the selection of a particular method of analyzing the data and the estimation of the effect of that method on the outcome of the study. It is assumed that the participants in the research have knowledge about the theoretical concepts and that they will execute the chosen method. This makes the choice of the particular method a part of the methodological approach.
The fourth section briefly discusses the methods considered as reliable empirical techniques and their applicability to specific research questions. This includes the criteria for selecting appropriate methods as well as the relative merits of various methods. It is also assumed that the participant’s knowledge and experience form part of the empirical evidence base. Methodological approaches should respect the boundary of acceptable methodology. If a procedure deviates from this boundary, it should be rejected.