Case Study
The primary purpose of doing research is to broaden our understanding about different issues that inform our understanding of specific topics. With this aim in mind, researchers have used different methodologies for achieving this goal. Among these methodologies is a case study. Case studies normally involve an in-depth analysis of different scientific phenomenon.
However, case studies are not as common today as during the onset of the modern science period. Indeed, early researchers such as Austrian-born anthropologist Bronislaw Malinowski heavily relied on case study information to come up with their findings (Moses and Knutsen 2007).
Since then, quantitative research methods have commanded a hegemonic position in modern-day research. However, recently, case study research has realized a new lease of life because researchers are slowly starting to embrace it again as an effort among researchers to align epistemology to ontology.
Aside from its rejuvenation, there have been many views held by naturalists and constructivists regarding case studies but more importantly, these views highlight the advantages and disadvantages of the research methodology. Some of them are discussed below.
Advantages
In-depth Understanding
One way that this paper highlights the advantages of case studies is comparing it with other methodological approaches. When case studies are analyzed viz-a-viz other methodologies, there is a clear consensus that it is superior to other research approaches because it is critical (for example, when analyzing individual perceptions or hegemonic discourses) (Moses and Knutsen 2007).
The ‘critical’ nature of case studies hails from the fact that case studies invest heavily in in-depth interviews and discourse analyses (Della-Porta and Keating 2008, p. 227).
Moreover, case studies tend to be more concerned with the pathways defining the cause and effect relationship between different mechanisms as opposed to assessing the strength of all the variables involved. Positivist methodologies also agree that case studies have a stronger insight into the depth of an analysis but this should be understood to mean the empirical completeness and natural wholeness of an analysis.
Validity
The validity of case studies is guaranteed through its internal mechanisms. For example, the construct of case studies contributes to the validity of its internal mechanisms through the use of a range of indicators for analyzing the theoretical intrigues of a research analysis.
In retrospect, case studies have contributed immensely to the growth and development of theories (although Large-N studies determine the empirical scope of these theories with a lot of emphasis on causal factors as opposed to descriptive factors, like case studies do) (Della-Porta and Keating 2008).
Disadvantages
Lack of Generalization
Since case studies are often dynamic in the number of participants, it is often difficult to generalize the findings if only a few subjects are used. For example, there are many case studies that have been done involving one participant or a group of participants, thereby making it difficult to generalize the findings among populations that do not fit the criteria of the subjects used (Della-Porta and Keating 2008).
Reliability
Throughout the history of case studies, there have been several concerns expressed by researchers regarding the lack of rigor that case studies tend to have (Moses and Knutsen 2007). In this regard, some researchers pursuing the case study approach have been sloppy and their findings dismissed as a result. Their sloppiness has often led to the introduction of bias or errors in the data generation process, thereby compromising the reliability of case study findings.
Naturalist View on Case Studies
The naturalist view of case studies often advocates for social diffusion and learning through natural generalization; however, even though researchers often do not go beyond the case analyzed, other researchers can still use the findings according to their suitability (in their studies).
Naturalists usually rank case studies to be among the lowest preferences for scientists doing research because of their affinity to historic information (Moses and Knutsen 2007). The main reason for having a low accreditation for case studies is that historic information has often been linked with fiction. In other words, here, the relationship between history and science is tenuous because science does not rely on myths or unfounded information (but history does).
The general perception of case studies among naturalists is therefore founded on the principle that cases are unitary observations that constitute part of a larger understanding of the same complex. Naturalists also believe that case studies constitute a great part of theoretical developments and even though they have a strong empirical dimension, they are still not very useful in themselves (Moses and Knutsen 2007).
However, the strength of case studies in empirical analysis stems from the fact that case studies have a strong criterion for establishing their sources and a strong utilization of data for the realization of historical objectives. In addition, naturalists often believe that in case studies, there is only one objective reality that is independent of human influences (Moses and Knutsen 2007).
This ontological assumption has also been shared by positivists. This analysis is however unsurprising because naturalists often try to deduce the natural elements of an analysis by analyzing different issues closely. Empathy and participatory explanations are therefore highly condoned by the naturalist view.
Constructivist View on Case Studies
As opposed to naturalists, constructivists often believe that empirical reality and theories are usually mutually constitutive because in their understanding, bridging is the narrowing of the divide between concrete understandings and their abstract meanings (Moses and Knutsen 2007).
Usually, interpretive techniques are used to understand this divide. Constructivists also prefer to use multiple theories to understand a given phenomenon because using interpretive techniques to analyze a given issue may lose the real meaning of the analysis in the first place (if such conclusions were mirrored against quantitative techniques).
Experimental Method
The experimental method has been part of modern day science for a long time. Respectable scientists and researchers such as Aristotle and Galileo have used experiments to come up with different findings (Moses and Knutsen 2007). Since the 1990s, many researchers have greatly relied on experiments to undertake different studies. The use of experiments in past decades has even been more dramatic. Especially, the use of experimentation in modern day science is starting to be more profound on the issue of theory testing.
Disadvantages
Applicability
Even though science has greatly relied on experimental research methods to build its body of knowledge, some fields still do not have a very mutually fulfilling relationship with this research methodology. For example, experimental methods are rarely used in political science because politics is more of an observational discipline as opposed to an experimental discipline (Moses and Knutsen 2007).
However, the recent advancement in technology is quickly changing this paradigm to make political science an experimental discipline. Nonetheless, it should be understood that the advancement in technology only encourages a specific trend as opposed to mirroring the situation as it is.
Intervention
From the nature of conducting experiments, the biggest weakness for experimentation is the ability of the researcher to influence the experiment. In other words, the intervention of the researcher is critical to the structure of experiments, thereby creating a possibility of data bias (Della-Porta and Keating 2008, p. 200). In non-experimental research, the data generation process is usually independent of the researcher.
Therefore, the role of the researcher is to observe the intricacies of the study as opposed to intervening. However, there is a thin line between interventions and observations because even though observations are perceived to be non-intrusive, the Hawthorne experiments of the 1920s showed that observations can also be intrusive (Moses and Knutsen 2007).
In detail, these experiments were aimed at investigating ways to improve worker productivity by establishing two worker groups which were distinguished by the presence of a change stimuli and the lack of it.
In both groups, it was established that productivity improved. However, it was interesting for the researchers to witness productivity even among workers who operated without a change stimulus. Nonetheless, it was later observed that the workers knew they were being watched and therefore, they improved their productivity (Moses and Knutsen 2007).
The main lesson learned from these experiments was that even observations can be intrusive. This analysis therefore exposes another weakness of the experimental research method because data generation can be biased since subjects may produce the ‘expected’ results as opposed to the ‘real’ results.
Imaginative
Another criticism leveled against experimental research (usually by constructivists) is the fact that experimentation can be fanciful. In other words, experimentation can be perceived to be a pigment of the researchers’ imagination and therefore, they may be designed to be fanciful (according to the researcher’s thoughts). To some people, this weakness is the epitome of anti-science (Moses and Knutsen 2007).
To explain this weakness, it is crucial to highlight the fact that most experiments started in the researchers’ minds as “what ifs?” Since such ideas were rooted in the imagery of experimentation, these researchers went out of their way to find proof of their imageries. Most of the researchers who used such steps therefore found bits of empirical evidence to support their ideas and therefore, their conclusions were convincing.
Some of these researchers include Darwin (in Darwin’s discovery of evolution), Lyell (in the determination of the age of the earth), Wegener (in tectonic plate theory) and Hubble (in the big bang theory). Since most of such experiments started as intuition, experimentation has often been perceived by some critics as ‘unscientific’. Other researchers who have conducted ‘though-out’ experiments include Plato, Hobbes and Rousseau (Della-Porta and Keating 2008).
Advantages
Reliability
Experiments work on a three-tiered philosophy which is constituted of interventions, controls and random selections. Interventions are normally understood within the confines of manipulations to create variations. The aspect of ‘control’ is usually understood to be the limitation of variations so that two groups (divided by controlling factors) can be established.
One group is normally known as the experimental group while the other is known as the control group. The ‘control’ aspect normally establishes the pre-testing and post-testing of findings for both groups (Moses and Knutsen 2007). The third foundation for experimentation is the random allocation of research objects among both groups of research so that it is simpler to control any external factors in the experiment. The three-tiered philosophy improves the reliability of experimental data.
Relevance
When trying to understand the strengths and weaknesses of experimental studies (viz-a-viz other research approaches), we see that, in experiments, the researcher can control factors of influence in the experiment while other methodologies are independent of this fact. This ability improves the relevance of information obtained.
Constructivist View on Experimentation
Constructivists do not usually prefer to use experimentation (even though they still rely on the research methods for their studies) (Moses and Knutsen 2007). The relationship between experimentation and constructivism is that experimentation could be perceived to be thinking which is immersed in context while context forms an important part of constructivism.
Comprehensively, constructivists think of experiments as distorting the context that gives meaning (Moses and Knutsen 2007). However, this understanding is only true for ‘real’ and not thought-out experiments. The aspect of control is also criticized for de-contextualizing the meaning and for naturalist, this analysis amounts to denaturalization.
Naturalist View on Experimentation
Naturalists often perceive experimental research to be at the center of ground breaking natural science because experimental research usually observes phenomena in their natural elements. For example, a ground-breaking experiment of 1939 was designed to test a reigning orthodoxy about the main influences of behavior where rats were studied to investigate how human behavior is categorized (Moses and Knutsen 2007).
In the experiment, it was concluded that behavior can be categorized into two contexts: fixation seizures and goal-oriented behaviors. To naturalists, this type of experimentation is not extraneous. The relevance of experimentation in politics is seen through the study of electoral behavior, measuring public policy impacts, and in investigating political psychology.
Small-N Designs
From Moses and Knutsen’s understanding of small-N designs, they tend to be pre-experimental and qualitative. One unique feature of small-N designs is that there is a careful analysis of the observations present in a given research. The incorporation of several data collection tools such as observations and interviews is usually a predominant feature of this research design.
J.S Mill is often perceived to be the main source of knowledge for information on small-N designs and he explains that five logical designs are considered realistic approaches for small-N designs. These designs are: the method of difference, the method of agreement, the indirect method of difference, the concomitant variation and the method of residues (Moses and Knutsen 2007). The method of residues is however considered to be inappropriate for doing social science research because it eliminates the effects of all causes except one.
Some of the advantages and disadvantages of small-N designs which are to be discussed in later sections of this paper will encompass the characteristics of the five designs discussed above. However, it is important to point out that the ‘method of difference approach’ seeks to control most variables by identifying similar cases (Moses and Knutsen 2007). This approach has been largely used in the comparison and analysis of the governance structure of different countries such as intra-state analysis (especially, federal state comparisons).
The ‘method of selection’ is different from the method of difference because it selects cases to be analyzed as a strategy for reducing variables. For example, this method has been used to investigate why there are some social and political revolutions which have seen a widespread participation of peasants while there are others that do not involve the peasant population (Moses and Knutsen 2007).
When we analyze the method of agreement, its preliminary analysis dictates that different cases may be chosen while only one similarity may explain the difference in outcomes. The indirect method of difference is however the most reliable method used in small-N designs because it is easy to apply and it equally has a high validity rate. Nonetheless, the concomitant variation is more sophisticated because it measures quantitative variations.
Notably, Durkheim has emphasized its application in small-N methodologies (Della-Porta and Keating 2008, p. 170). Concomitant variations have been known to produce non-dichotomous data and they have also been used to conduct not less than 5000 interviews in the US, UK, Mexico, Italy and Germany (Moses and Knutsen 2007).
These interviews were used to create the body of knowledge found in the 1960 book written by Almond and Verba – Establishing the civic culture: Political Attitudes and democracy in Five Nations. Based on the application of small-N designs, several advantages and disadvantages have been associated with it.
Advantages
Control
Unlike other research methods such as the Large-N design, small-N designs have a better control of extraneous factors because it ensures these factors remain constant. This strength is founded on the premise that small-N designs tend to prefer using fewer variables hence establishing more control over the study (Della-Porta and Keating 2008).
Validity
When compared to the other research methods analyzed in this paper, small-N designs tend to have a higher validity. This validity stems from its ability to better record the performance of the subjects involved in the study. Moreover, the performance of such subjects is also better assessed using the small-N design.
A closely related advantage to this understanding is the ability of the small-N design to be particular about the variables being investigated in the study (thereby providing more accurate and specific findings regarding the research topic) (Della-Porta and Keating 2008). Comparatively, the Large-N technique uses pooled data to examine research variables. Through the high validity preferred to small-N designs, the data generation process may equally not be subjected to time-consuming statistical testing.
Replication
Among the greatest advantages pointed out by naturalists (about the small-N technique) is its ability to be replicated. This advantage shows that the small_N design can be used in different contexts.
Authoritative
Another advantage pointed out by naturalists is the use of authoritative data as a reliable data source for the small-N technique. To naturalists, this advantage improves the reliability of the data technique (Moses and Knutsen 2007).
Disadvantages
Selection Bias
The small-N research methodology has a weakness in selection bias where non-observable variances are often made to be the cause of variation. Ordinarily, a random sampling technique would be perceived to have no selection bias, but this is often not the case with the small-N research methodology because it uses a non-random sampling technique (Moses and Knutsen 2007).
Over-determination
Another weakness of the small-N sampling technique is its tendency to be over-determined. Here, there are several variables that can be hardly controlled throughout the research process. The numerous variables that are uncontrollable make it difficult to establish which variables are causal.
In this regard, it is difficult to determine if self-generalization is even possible but to reduce the number of variables to a manageable estimate (preferably one), it is crucial to provide explanations. Naturalists would prefer to increase the number of cases to reduce variables because according to their understanding, this is the best way to make an experiment controllable (Moses and Knutsen 2007).
Constructivist View on Small-N Design
Regarding the nature of data used, naturalists and constructivists often have different ideologies about the small-N technique. For example, since the small-N technique uses de-contextualized data, constructivists often explain that de-contextualized data may have the capability of distorting the intended meaning of the research (and in some cases, it may make it more difficult to point out the meaning altogether) (Moses and Knutsen 2007).
Furthermore, considering the small-N technique uses authoritative data, constructivists point out that this strategy may introduce bias to the research process. Their argument also stems from the fact that such data is almost always interpreted anyway and therefore bias is still unavoidable. Regarding the argument that studies done using the small-N technique can be replicated; constructivists point out that its ability to replicate is often unnecessary (Moses and Knutsen 2007).
Naturalist View on Small-N Design
Naturalists differ from constructivists in the way they view the research designs for small-N techniques. Naturalists view cases analyzed within the Small-N framework as anonymous and independent as opposed to complex and systematic. Their perception of concepts is also operationalized and predetermined (Moses and Knutsen 2007).
This view is contrary to the perception held by constructivists because the latter perceives cases to be constructed through the course of the study. Naturalists assumes that the number of cases that can be handled by the small-N design are many, although this is often contrary to the constructivist view that only a few cases can be handled using this research design.
Naturalists also believe that case selection should be variable and the diversity created in the research should be controlled through parameterization. The number of variables is also preferably reduced under the naturalist approach although constructivists prefer them to be increased for thick description (Moses and Knutsen 2007).
Large-N Designs
The main difference between Large-N and Small-N designs is often understood through the number observations in a research. Large-N designs are normally used when there is a strong lack of uniformity within specific variables of a research (Della-Porta and Keating 2008, p. 210). Large-N designs are also used to draw inferences by explaining what is not ordinarily seen by the naked eye.
The explanation by J.S Mill regarding the development of inferences closely resembles the naturalist view of science because he explained that nature has a range of regularities which can be easily inferred (Della-Porta and Keating 2008). For example, J.S Mill would ordinarily try to draw a link between two things that would occur at the same time. The main lesson learned here is that if there is the presence of co-variation, a researcher should explore it to its depth.
Advantages
Investigating Complex Relations
Large-N designs are often beneficial when investigating complex relationships (or studies that involve many variables). Its efficiency in measuring variables is often the main point of departure with small-N samples. Its efficiency in doing so is undisputed because different techniques used in large-N research designs have a high efficacy.
For example, Bayenesian techniques have been used in military expeditions with a high degree of success. For instance, the identification of a damaged U.S submarine in the 60s was done using the Bayenesian technique (Moses and Knutsen 2007). This technique worked by investigating which possible areas that the submarine accident could have occurred and therefore triangulating such a zone for more search.
Breadth
Large-N methodologies differ with other research methodologies because they have a broader understanding of different variables as opposed to the understanding of their depth. Abstractly, the broader understanding of the variables in large-N methodologies can be mirrored in the understanding of co-efficient variables that are present in published research.
In this analysis, the use of regression analysis surfaces because it is normally used in the triangulation of research co-efficients. The ability to undertake regression analysis helps people to be more critical about the research findings and by extension; such skills will assist them to understand different encounters in the world of quantitative research (Della-Porta and Keating 2008).
External Validity
Large-N studies are also advantageous in the sense that their external validity is more secure because of the use of statistical controls. The external validity of the large-N methodology can equally be extrapolated to include counterfactual depictions where variables (other than the ones) investigated are used to understand a research problem (Moses and Knutsen 2007).
Disadvantages
Weak Internal Validity
Among the greatest disadvantages of large-N designs is their weak internal validity. Indeed, as opposed to experimental research designs, large-N research designs do not have a high level of control (Della-Porta and Keating 2008).
Time
Another great weakness for large-N studies as explained by naturalists is their failure to account for time because people are ordinarily dynamic (not static) and therefore accounting for time is very important. Large-N studies fail in this regard (Moses and Knutsen 2007).
Naturalist View on Large-N Designs
The concept of naturalism indicates that knowledge is often generated through direct experiences from sensory perceptions and albeit it is crucial to be logical and reasonable in the creation of knowledge, these two concepts are not sufficient to the true generation of knowledge. In the analysis of Large-N designs, three main contemporary issues emerge: falsification, prediction, and the formulation of frameworks to analyze the knowledge created (Moses and Knutsen 2007).
Through the understanding of the naturalism concept, we can point out a few key features of the Large-N design. First, among the most notable explanations of naturalism is its explanation that within the world we live in, observable patterns exist.
Another strong characteristic exhibited by the naturalism concept is that statements based on these observable patterns can be easily tested (Moses and Knutsen 2007). A similar characteristic of the naturalism concept is that, facts are often very different from values and the aim of science is normally generalized (nomothetic) as opposed to individualistic (ideographic).
Constructivist Approach to Large-N Designs
Contrary to the concept of naturalism, the concept of constructivism argues that the patterns exhibited in Large-N designs (such as regression analysis, graphs and the likes) are not usually rooted in nature because they are ephemeral, and contingent upon human agency (Moses and Knutsen 2007).
Regarding the apparent differences in observer perceptions for social research in Large-N designs, the constructivism concept states that age and gender differentials, which are also entangled within social characteristics such as era, culture and language, are responsible for differences in observer perceptions. However, some of these social factors are ordinarily important in explaining our everyday lives including the progress made with science.
References
Della-Porta, D & Keating, M 2008, Approaches and Methodologies in the Social Sciences, Cambridge University Press, Cambridge.
Moses, J & Knutsen, T 2007, Ways of Knowing. Competitive Methodologies in Social and Political Research, Palgrave Macmillan, London.