Critical evaluation of two differing approaches in research

Critical Evaluation 10

CRITICAL EVALUATION

Statistical versus Theoretical Generalizability

Definitions

Generalisation is a reasoning act that encompasses the drawing of broad inferences from specific observations by researchers. Statistical generalizability or generalisation is a form of inductive simplification that uses the statistical attributes of a sample to make statistical claims regarding the entire population. Under statistical generalisation, the researcher draws conclusions about a subset or member of a population with reference to the available information about the population (Polit & Beck 2010). Therefore, statistical generalisations make use of reference classes. On the other hand, theoretical generalizability also known as theoretical generalisation focuses at the formation of a theory through the systematic testing of hypotheses and conjectures and using intentionally selected cases to reformulate them. From the definition, it is proper to refer to theoretical generalizability as theoretical or theory-carried sampling or generalisation.

Under the approach, the researcher halts collecting data on new research cases when the already formulated theory does not encounter any contradictions (Smaling 2003). There are two forms of theoretical generalisation: abductive generalisation and subsumptive generalisation. The former aims at forming hypothesis and theories whereas the latter targets to generalise research conclusions and results by placing them in the context of a general expression. The researcher also stops collecting data on new cases of research even the focus shifts to ‘negative cases’ that provide contradicting theory. The researcher arrives at the ‘theoretical saturation’ point after studying about five ‘negative cases’ and having been unable to identify contradicting theory to the existing one. Having reached that point, the researcher stops looking for other cases and regards the theory as a suitable vehicle or carrier for generalisation to other cases that have not undergone the research process.

In statistical generalisation, a researcher cannot generalise one case to the other case in the event that both cases are representative samples of the population that encompasses both cases. In the event that the population in question is heterogeneous, the case-to-case generalisation depends on chance-capitalisation. This turns out to be a stroke of luck or a guessing game. It is proper to use replication logic rather than a sampling logic in the event that the statistical generalisation has a total connection with the already mentioned statistical representativeness of the cases.

Comparison

From a broad perspective, it is proper to state that both generalisations recline on inductive reasoning as the basis. The objective of both generalisations is to draw conclusions about a class of situations, cases or subjects that have not undergone any research process while basing their argument on a limited number of cases, situations and persons that have undergone the research process (Smaling 2003). As a result, both forms of generalizability yield generalisations from the results of the research to a scope or population that belongs to the theory. Therefore, they suffice to be forms of inductive reasoning since they yield universal or general propositions. One of the major limitations of both forms of inductive reasoning is the fact that they do not take into account the differences and similarities existing between cases, subjects, situations among other areas of research.

As a result, the use of explicit analogical reasoning is imperative in the quest to reinforce inductive generalisations (Tsang & Williams 2012). The use of analogical reasoning is necessary despite the fact that both statistical and theoretical generalisations assume that the researcher had prior knowledge concerning the differences and similarities between cases, situations, subjects and the other areas of interest to the research. It is apparent that having prior knowledge about the similarities and differences enables the researcher to make firmly based generalisations on the research topic.

Business Examples

Below is an example of a statistical generalisation:

I have been to one of Apple’s distribution centres in London on several occasions to buy an iPad and I have found a queue all the time; thus, Apple distribution centres always have a queue. The generalisation makes other people believe that Apple distribution centres always have a queue that makes it difficult for a person to purchase a product within the shortest time possible. The generalisation exhibits biasness since it is evident that there are other Apple distribution centres in other parts of the town or country that do not have queues every time.

The following is an example of a theoretical generalisation:

Due to high demand of Apple products, Apple distribution centres always have a queue thereby making it difficult for a customer to purchase Apple’s products. The researcher that uses the generalisation reclines on the baseline theory that Apple products have a high demand. As a result, an individual that visits the stores with the intention of purchasing the company’s products will witness long queues.

Circumstances of using the alternative approaches to generalisation

Researchers apply statistical generalisation after having identified the underlying population that they would use to generalise results. In this case, the population comprises of the totality of people or elements that have similar, defined characteristics that are relevant to the research topic (Smaling 2003). As a result, the researcher should be able to select sample participants or elements from the population before deciding to use theoretical generalisation. Therefore, the occurrence of random sampling from the reference population is mandatory for the use of statistical generalisation in research.

The use theory-carried or theoretical generalisation necessitates the prior existence of conjectures and hypotheses that the researcher would reformulate in the generalisation process (Smaling 2003). Following the existence of the hypotheses, the researcher would then intentionally select the desired hypothesis that best suits the prevailing conditions. The existing hypotheses form the theory that provides the backbone of the generalised data. However, generalising the desired theory or hypothesis occurs only if the researcher does not find any contradicting theory upon conducting further research on ‘negative cases’. Most importantly, the researcher generalises the selected theory upon having studied five ‘negative cases’ and having been unable to identify contradicting theory. In such a case, the point of research attained by the researcher is known as the ‘saturation point’. It is only at this point that the researcher stops searching for other cases.

Parsimony versus Rigour in Research Endeavour

Introduction

In research endeavour, parsimony refers to the ability to provide explanations using relatively few statements and terms. Therefore, the use of parsimony in research endeavour requires the researcher to use simple hypotheses in explaining the underlying phenomena provided that the simplicity does not have a negative impact on the understanding of the research topic. The fact that empirical data comprises of both a replicable, structured part and a non-replicable, idiosyncratic part, the use of the principle of parsimony is beneficial to researchers that use prediction models (Vandekerckhove et al. 2014). On the other hand, rigour in research targets to enhance the credibility and reliability of the research process and data. The use of rigour requires the researcher to allow the participants to guide the process of inquiry. The researcher should also compare meanings derived from participants about the phenomenon with the already generated theoretical construction of the phenomenon (Chiovitti & Piran 2003).

The theory should also entail the use of the actual words of the participants. A rigorous research endeavour should also articulate the views and opinions of the researcher towards the phenomenon under study. The research should also specify the criteria that govern the thinking of the researcher. Moreover, using rigour in research also requires the researcher to specify the process for selection and reasons behind the selection of the participants used in the study. Furthermore, using rigour also requires the researcher to define the scope of the research. Finally, a researcher that uses rigour in the research endeavour should also provide an in-depth description of the relevant literature associated with each segments of the research topic as well as the literature that emerges in the theory (Chiovitti & Piran 2003). Even though there are different methods of applying rigour in research, the researcher should avail adequate information about the process of generating theory for the research.

Business Examples

In applying parsimony in research, the motivation of businesses is to ensure that the research as simple and economical as possible. Businesses apply simplicity in defining problems and explaining the solutions to the problems as opposed to complex measures of attaining the objectives. In order to achieve the economy objective, researchers use few variables as opposed to using more variables that leads to greater variance. Brief and concise explanations of the factors influencing a phenomenon in the company is an evidence of the use of parsimony in research as it enhances the understanding of the factors on the part of the other stakeholders of the company. In the quest to enhance the understanding of the phenomena under study, a researcher that employs the principle of parsimony undertakes a study of the associated literature. The researcher also conducts interviews with the concerned stakeholders.

In the case of Apple Inc., a possible research area would involve determining the motivating factors behind the continued preference of the company’s products to those of its competitors. The application of parsimony on the research topic would necessitate making reference to few existing literature about the factors that motivate ICT consumers to increase their levels of customer loyalty towards the products of a company. The researcher would then select few variables such as the innovativeness of the applications, the quality of the hardware, the reputation of the company, the value of the company’s products with respect to their price and the ability of products to meet customer needs and expectations. A parsimonious research endeavour would then entail conducting interviews with consumers and product developers to determine if it is possible for the company to produce products based on the specifications and expectations of consumers. At the end of the interview and related review of literature, the researcher would conclude the research endeavour and provide recommendations to the company.

However, in the case of a rigorous research, the researcher would have to undertake the research process through the standard process of a rigorous research. In the case of the Apple Company, a research on the factors determining the factors requires the company to allow consumers and other stakeholders to guide the process of inquiry. The researcher would then evaluate the meanings stated by the participants regarding the participants in relation to the existing theoretical construction. For instance, Apple’s consumers would provide their meanings of the determinants of success for the company’s products. The research would also make use of the actual words uttered by the participants during the interviews. The firm would then integrate the views of the participants into the existing theory to yield relevant conclusions and recommendations. The firm would also indicate that the selection of the aims and objectives was based on the fact that customer satisfaction and customer loyalty are crucial to its success.

Circumstances of using the alternative approaches to generalisation

Applying the principle of parsimony in a research endeavour requires the researcher to identify the simplest hypothesis for the study (Epstein 1984). By so doing, it would be easy for other stakeholders to understand the theories underlying the research. However, the use of the simplest hypothesis necessitates ascertaining that the hypothesis does not contradict the available information in any way. The acceptance of the propositions requires the existence of empirical support. As a result, the parsimony principle advocates for the rejection of propositions such as mysticism, alien abductions, and extrasensory perceptions among others.

The necessity for validity and reliability in research suffices to be the basis for the use of rigour in a research endeavour (Golafshani 2003). The approach is also necessary in the case of research endeavours that focus on assessing the utility and trustworthiness of the research strategies rather than the mere focus on the implemented strategies. In essence, the use of rigour in research is applicable in research endeavours that necessitate the use of the eight research methods mentioned above.

Reference List

Chiovitti, R.F. and Piran, N., 2003. Rigour and grounded theory research. Journal of advanced nursing, 44(4), pp.427-435.

Epstein, R., 1984. The principle of parsimony and some applications in psychology. Journal of Mind Behavior, 5, pp.119-130.

Golafshani, N., 2003. Understanding reliability and validity in qualitative research. The qualitative report, 8(4), pp.597-606.

Polit, D.F. and Beck, C.T., 2010. Generalization in quantitative and qualitative research: myths and strategies. International journal of nursing studies, 47(11), pp.1451-1458.

Seale, C. and Silverman, D., 1997. Ensuring rigour in qualitative research. European journal of public health, 7(4), pp.379-384.

Smaling, A., 2003. Inductive, analogical, and communicative generalization. International Journal of Qualitative Methods, 2(1), pp.52-67.

Tsang, E.W. and Williams, J.N., 2012. Generalization and induction: misconceptions, clarifications, and a classification of induction. Mis Quarterly, 36(3), pp.729-748.

Vandekerckhove, J., Matzke, D. and Wagenmakers, E.J., 2014. Model comparison and the principle of parsimony.