Research Paradigms and Concepts of Ontology and Epistemology

Chapter 2: Research Methodology

This chapter of our thesis will illustrate the methodology that was undertaken in order to conduct our research. This chapter is further divided into two sections. In 1st section of the chapter we will discuss theoretical methodology. It includes research philosophy, research approach and research strategy. In 2nd section of the chapter we will discuss practical methodology. It includes choice of the subject, preconception, perspective, literature, data collection and sampling of methods. Here we will not only describe the methods that are opted during our research process but also try to explain and justify why they are preferred over the alternatives methods that are available.

2.1: Theoretical methodology

2.1.1 Research philosophy

During a business research it is imperative to think about different research paradigms and theme of ontology and epistemology. These research paradigms represent a parameter that controls the research carry out from research design to the conclusion and recommendations of the research. That’s why it is of great significance to understand these features in order to move in harmonious manner and actions leading towards unambiguous investigation and making sure that researcher biasnesses are minimized (Flower, 2009, p 1).

2.1.1.1 Epistemological Choice

“Epistemological issue concerns with the question of what or should be regarded as acceptable knowledge in a discipline”.(Bryman & Bell, 2007, p. 16). Eriksson and Kovalainen describe epistemology as “what knowledge is and what are the sources and limits of knowledge” (Eriksson and Kovalainen, 2008). It is important that the given piece of knowledge is studied in the relative manner. Epistemology is further divided into positivism, interpretivism and realism.

Positivism:

Positivism is the resultant of research in natural science where a hypothesis is tested that is derived from a prevailing theory. Positivism emphasizes that true and reliable knowledge is that which stands upon logic, practical experience and affirmative authentication. “The purpose of theory is to generate hypotheses that can be tested and that will be thereby allow explanations of laws to be assessed”(Bryman & Bell, 2007, p. 16). Positivism is generally linked with quantitative research where one selects a theory and piece of knowledge, then collects data and interprets it and hence proves the hypothesis to be true or not. “The researcher seeking to adopt a decided positivist stance exercises choice of the study, the research objective to pursue and the data to be collected” (Saunders et al., 2009, p. 114). It is believed that the positivist researcher will use structured methodology with the aim of producing replicability.

Interpretivism

Interpretivism is a stance contrary to positivism and is also known as anti-positivism. Philosophers of social science believe that subject matter of a research in a social science that is undertaken upon some individuals and personals or institutions are different from that of the natural science (Bryman & Bell, 2007, p. 17). According to Saunders et al., (2009) phenomenology and symbolic interactionism are the two intellectual traditions that interpretivism comes from. “Phenomenology refers to way in which we as humans make sense of the world around us. In symbolic interactionism we are in continual process of interpreting the social world around us in that we interpret the actions of others with whom we interact and this interpretation leads adjustment of our own meaning and actions” (Saunders et al., 2009, p. 116).

Realism:

Realism is another branch of epistemology that is similar in nature to positivism. According to this philosophic position reality does exist independently of the human realization. There are two main types of realism. Bryman & Bell, 2007 named them empirical realism and critical realism. Whereas Saunders et al., 2009 name them as direct realism and critical realism. Main difference among critical and direct realism is that critical realism claims that our knowledge is developed in two phases. Firts there is an object and substance we see and experience trough our senses and in second phase process goes on after the senses receive the sensations. Whereas direct realism claims first phase is enough and what we see is what we get (Saunders et al., 2009, p. 117).

In epistemological consideration of our research we take the positivist position. The reason behind selection of positivist stance is what we studied in the literature we develop a hypothesis on the basis of prevailing theories and in our empirical findings we will test this hypothesis for acceptation or rejection. What we are going to study is the usefulness of banks internal evaluation model (CAMELS) with respect to external credit rating models such as PACRA and Moody’s. Another reason behind selection of this stance is the use of secondary data and positivism is usually linked with researches that are quantitative in nature such as statistical tools and figures.

2.1.1.2 Ontology:

This pillar of the research paradigm deals with the nature of social entities and realities. This raises the question of “whether social entities can and should be considered objective entities that have reality external to the social factors, or whether they can and should be considered social construction build up from the perception and actions of the social actors” (Bryman & Bell, 2007, p. 17). Above we discussed two aspects of ontology, are known as objectivism and constructivism respectively.

In ontological consideration, our research thesis is objectivist in nature. For empirical finding of our thesis we will use statistical tools such as financial ratios. The result provided by these ratios will be rated on the scale of 1 to 5 based on the issued ranking system of CAMELS rating model. Then we will compare the result provided by CAMELS rating model with the issued ratings of PACRA rating agency of the same period of time. So it is obvious from our choice of ontological consideration that the research will be free from every sort of biasness of the authors.

2.1.2 Research Approach:

At this stage of a research one has to decide on the basis of chosen theory that’s is either learnt or contained in the literature that he/she will follow a deductive approach or inductive approach. In deductive approach one develops a hypothesis from the theory and devises research strategy to test the hypothesis to accept or reject it. Whereas in inductive approach one has to collect data, analyze it and develop theory based on the result of analyzed data (Saunders et al., 2009, p. 124).

2.1.2.1 Deductive approach:

This approach represents characteristics of the correlation among theory and research. On the basis of gained knowledge that may be inferred from the theoretical reflection about the research field, one assumes a hypothesis that will be observed in empirical analysis (Bryman & Bell, 2007, p. 14).

According to Robson (2002, cited in Saunders et al., 2009, p. 124) deductive research goes through 5 steps. First develop a hypothesis from a theory, second express the hypothesis in operational terms, third is hypothesis testing, fourth analyzes the particular result of the inquiry and in the fifth and final stage verifies or modifies the theory on base of your findings. Collection of quantitative data is an important characteristic of deductive approach although it may use qualitative data some times. In deductive approach highly structured methodology is implemented to assist replication to guarantee reliability. Generalization is another important characteristic of deductive approach (Saunders et al., 2009, p. 124-125).

2.1.2.1 Inductive approach:

In inductive approach theory is developed on the basis of research observations. Researchers in inductive approach are predominantly concerned with the events that are actually taking place and deal with the small sample as against the large number in deductive approach. Researchers most likely use qualitative data in this approach. Structure of the research in inductive approach is more flexible as less generalized as compared to highly structured methodology and generalized characteristics of deductive approach (Saunders et al., 2009, p. 126-127).

In our research we will use deductive approach and not the inductive approach. The primary reason behind selection of this approach is correlation among the theory and hypothesis. On the basis of studied literature and constructed theoretical frame work we have derived a hypothesis. Our thesis findings will lead us towards the acceptation or rejection of hypothesis which is superiority of bank’s internal evaluation models with respect to external rating agencies models. As we know that deductive approach is highly structured methodology so our thesis will follow the predefined structure..

2.1.3 Research strategy:

According to Bryman & Bell, 2007 “by research strategy, we simply mean a general orientation to the conduct of business research”. Two separate clusters of research strategy are qualitative and quantitative research. Both quantitative and qualitative researches are different from each other not only on the basis of quantification and measurement of the result but also on the basis of epistemological and ontological foundations (Bryman & Bell, 2007, p. 28).

2.1.3.1 Quantitative research:

Quantitative research is mostly used in deductive approach where the aim of the study is to test a hypothesis for proving a theory. In quantitative research analysis of the result is mostly in numbers and quantify. Another characteristic of quantitative research is size of the sample is very large. Quantitative research is subject to a very low level of biasness in the interpretation from the researchers as statistical tools are used for analysis of the results. Quantitative research is more generalizable. Quantitative research is more positivist in nature when it comes to epistemological orientation. While in ontological orientation quantitative research is objectivist in nature (Bryman & Bell, 2007, p. 28). According to Bryman & Bell, 2007 quantitative research have 11 steps. Steps mentioned by them in their book Business research methodsare as follow 1st Theory, 2nd Hypothesis, 3rd Research design, 4th Devise measures of concepts, 5th select research site, 6th select research subject / respondent, 7th administer research instruments / collect data, 8th Process data, 9th Analyze data, 10th finding and conclusion and 11th is write up findings and conclusions.

Fig 2.2 Based on the idea from Bryman & Bell, 2007 the process of quantitative research.

2.1.3.2 Qualitative research

Qualitative research is the contrast of the quantitative research in almost every aspect. Qualitative method is mostly used in inductive research where emphasis is on the generation of a theory that is based upon research observation. Analysis of the qualitative research is based upon the words and statements and mostly in textual form. Qualitative research is more interpretivist in nature when it comes to epistemological orientation. While in ontological orientation qualitative research is constructivist in nature. Size of the sample in qualitative research is small. Researcher in the qualitative research requires special skills for the interpretations of the results. Contrary to quantitative research, qualitative research is less generalizable and very low level of replicability. (Bryman & Bell, 2007, p. 28). Below is differentiation chart of quantitative and qualitative research based on different research aspects.

Self made chart based on the studied books and articles.

In our research thesis we will use quantitative research strategy instead of qualitative research. Selection of quantitative research strategy is based upon number of arguments other than presentation of result in numerics and figures. As in epistemology consideration our research is positivist in nature and in ontological consideration we has opted the stance of objectivist, this is another reason for our selection of quantitative research. Our research is basically concerned with numbers and use of statistical tools to measure and analyze these numbers and figures. We have selected deductive approach which is usually related with quantitative research strategy and we argued about its selection in previous section. So keeping in mind our research question and objective we selected this strategy that will be appropriate and will guide us though out our research process.

2.1.4 Nature of Research

The relationship between events or factors is often described, examine, and explain by the researchers. The research can either be descriptive, explanatory, exploratory or predictive nature. Nature of the research totally depends on research question. In descriptive research the researcher tries to identify or describe the events. For example in descriptive research if the question is “What is the present or past state of events?” for this the researcher selects the representative sampling of the people. Whereas explanatory research is for events to be explain by the researcher and look for fundamental reasons. Explanatory research is also referred to interpretive research. For example the researcher conducts explanatory research if they ask “Why have these events happened in the manner they did?”  Or “What are the implications of these events occurring as they have? (Rubin et al, 2010, 198)”. Exploratory research is a type of research when researcher is unable to find any or very few prior studies about the research question. Exploratory study aspire researcher to look for the ideas, Patterns, or hypothesis, instead of testing or verifying a theory. In exploratory research typical techniques are used here researcher comprises observation, case studies, and previous studies. In this type of research the researcher uses both qualitative and quantitative data. Exploratory research totally focuses on achieving the imminent and acquaintance with the subject area to examine at later stage (Collis &Hussey, 2009, p. 6). Predictive research advances one step further than explanatory research. Aim of the study is to create justification for what is happening in particular circumstances. Predictive researchis concerned with the anticipation of a possibility of some occurrence. Aim of predictive research is to generalize the investigation by forecasting some phenomena on the foundation of hypothesized, general relationships. Hence we can say that solution provided by a predictive research in a particular situation can be applicable to some other problems of similar nature, provided that the solution is valid (Collis &Hussey, 2009, p. 6).

Nature of our research is exploratory in nature as we didn’t find any previous research work on the same research problem. There are some articles and thesis that reflects the usage of camels rating model framework on any particular or limited number of banks. As in our research we will try to investigate the superiority of bank’s internal rating models with respect to external credit rating agencies, we found no research work in this particular field. Our research will open a new window of research in the field of banking rating systems.

2.1.5 Time horizon of research:

Time horizon of research plays a very significant role for both authors of the research and readers of the research. From researchers point of view it is important to understand that what type and quality of data they are collecting for the research and how easily it can be collected. Some time it is very difficult and time consuming to collect the data spread over long period of time. From readers point of view time horizon of the research is of great importance as it shows credibility and quality of the research observations. According to Sunders et al. 2009 time horizon of the research can be divided into two types; cross-sectional research and longitudinal research (Saunders et al.2009, p. 155).

2.1.5.1 Longitudinal study

Longitudinal study is such type of a research in which same sample of population is observed over a longer period of time. Longitudinal research is a type of observational research in which the subjects are observed without manipulations and hence can be argued that it has less potential to detect cause and effects relationships of variables as do by the experimental studies. In longitudinal research, researchers have the opportunity to observe changes and improvement that took place over the period of time (Lindborg & Ohlsson, 2009, p. 15).

2.1.5.2 Cross-sectional study

Cross-sectional research is a study of sample observations or of a population in which a researcher makes her/his study and get result for a short period of time or on a single occasion. From the population researcher takes sample and within that sample he/she distribute the variables, and sometime the variables which are predict and designated on the bases of reasonable information which he/she gets from other sources. Descriptive and exploratory studies are frequently cross-sectional. For example a single survey of a country to describe the population of the specific country at a given time. Cross-sectional study is also very close to explanatory studies. A researcher conducts a survey on national base to examine the nation problem at a given time (Hulley et al, 2007, p.109). 

Keeping in mind above arguments about longitudinal and cross-sectional studies, we can easily say that our research is a cross-sectional study. As in our thesis we will manipulate one year annual financial reports for the year ended on 31st Dec 2010, and all 17 banks belongs from Pakistan. It would be of great significance to use annual reports of more than only year but the problem is at the same time we need to collect published credit ratings of banks for the same year and that was very difficult to collets as most banks do not keep their old records of ratings. As we discussed above that cross-sectional studies are mostly exploratory or descriptive in nature and our research is also exploratory in nature.

2.2 Practical Methodology:

Selection of the research topic:

Both authors of the thesis come from Islamic Republic of Pakistan where banking industry is on the boom despite of the general economic turmoil. As far as our educational background is concerned, one of the authors has done of BBA (Hons) with major in Finance and second author has done MBA with major in Finance, in which we studied several courses about banking and finance. This was the point where our interest was developed in the field of banking. Now being the students of Master in Finance, keeping in view our future professional careers in the field of banking and finance, we choose this research topic. As our research is focusing on CAMELS rating model that takes into account 6 important component of banking industry, it will give us better understanding and knowledge about performance of banking industry particularly in Pakistan. Another reason behind selection of this research topic is our personal curiosity about the ratings of the banks particularly in Pakistan.

Preconception:

Preconception about the research topic is quite important as it develops interest of the researcher and involves him neutrally in the topic. Both authors of the thesis worked as an internee in conventional bank in Pakistan after completion of their studies and have gained some practical experience of banking. In our practical experiences we observed that not all of above but some of these factors that are addressed in ratings models are of great importance for better performance of the banks.

Perspective:

Perspective provides point of view about the chosen field of study from several angles with different aspects, providing a pool of options from which the most suitable one is to be selected while keeping in mind your research question and purpose. Our thesis is based upon the superiority of CAMELS rating model results with respect to external rating agencies; we will be looking at it from internal point of view of the banks and their regulatory authorities. But it readers can also be external users such as rating agencies and will be useful for them as well.

Data Collection Method:

For any type of research study data collection is an important aspect. Data is the source from where researchers can get relevant information to answer the research questions. To gather applicable information researchers use primary and secondary data as a sources. Primary data is collected or perceived straight from the first time experience. Or we can say that data collected for the first time particularly for this research problem. The sources contained in primary data are questionnaires, observations, social surveys, experiments and interviews. On the other hand secondary data is published and the data collected by someone else in the past. We use the published and collected data by someone else to solve our problems but the problem might be different from others. The sources from which we can collect secondary data are articles, books, journals, and web-based data (Ghauri & Gronhaug, 2005, p. 91-102).

Theoretical framework of our study that is based upon secondary data, we read all relevant literature to our study that gave us full knowledge and beneficial understanding of our research questions. On the basis of this further study should be conducted. Findings of our research are totally dependent upon secondary. Findings of our research are complex in nature but our secondary data will help us to achieve this objective. The electronic search engine is the main source in our study we used Umeå University electronic library, Google search, electronic books; we also use printed materials like books.

Literature studied:

Access to the relevant and authentic literature for a researcher is of great importance. It provides basis for the researcher to build upon the theoretical frame work of a chosen field of study and research design (Brayman & Bell, 2007, p. 94). In the beginning of our research we studied some articles regarding the financial crises caused by the collapse of leading financial institutions of that time and banking industry of Pakistan afterwards we collected some articles and books through university archives database and internet search about research methods to construct research methodology chapter of our thesis and to guide us throughout our thesis. Articles are mostly collected from journals of Banking and Finance, journal of international banking regulation and the review of financial studies accessed through university provided logins on its database. It is a difficult task to gather relevant articles as one comes across so many articles that look relevant but they are not in real. We also studied some online books that are available via different websites such as www.book.google.com. We also gathered some material such as brochures and working papers about the procedures and methods used for ratings of financial institutions form rating agencies websites including S&P’s, PACRA, and regulatory authorities such as Basel and SBoP. The collected materials provide us better understanding and capabilities to work in the field rating systems. 

Sampling:

Sample is a subset of whole population that is selected to represent the population in any specific research to perform statistical inferences and to make judgments about the whole population on the basis of selected sample. For researchers it is very important to select appropriate sample from population to make inferences. There are two primary reasons why researchers select a sample for their research instead of studying the whole population, 1st it is very costly and 2nd is the time limitation for a research. According to Bryman & Bell, sample is a fragment of population chosen for examination or research. Talking in a broader sense there are basically two types of sampling approaches: Probability and non probability approach (Bryman & Bell, 2007, p. 182).

Probability sampling is an approach in which each and every unit of population has equal chance of being selected in the sample and their probability of selection is greater than “0”. This is the most suitable approach that eliminates bias in sample selection and reduces sampling error. Simple random sampling, systematic or interval random sampling, stratified sampling and cluster sampling are some types of probability sampling (Brayman & Bell, 2007, p. 182). Whereas on the other side non probability sampling is an approach in which probability of selection of elements is not known or some elements of the population have no chance to be selected as a sample. Referral (Snow ball) sample, quota sample, criteria sample, homogeneous sample, critical sample and matched sample are the types of non-probability sampling.

Referral sampling method is also knows as snow ball sampling. This type of sampling procedure is used when it is difficult to find subjects or samples for the research. In this type of a research sample we cannot identify our sample in advance and important consideration is to find initial subjects and informants. From initial subjects we can ask to identify further subjects for our research sample and the process continues to go on like a snow ball (Swisher, 2010, p.1). In Quota samplingpopulation is first divided into subgroups based on some criteria and then subjects are selected as a sample from each group. Quota sampling can be further divided in two sub groups known as proportional and non proportional quota sampling. In proportional quota sampling, sample of every sub group of population is represented by its proportional weight age of the population. Where as in non proportional quota sampling a limit of minimum numbers of sample subjects are set as a standard and are not necessary to represent its proportion in the population. It only has to justify that the chosen sample is able enough to represent a small group in a population. Heterogeneity sampling is a procedure when we select sample for our research that will represent every group of thought or variables of the population, on the other hand Homogeneous sampling is a procedure of sampling when we want to study about some specific factors or variables and their effect on the outcome that represent the population (Trochin, & William, 2006).  Criteria sampling is a procedure to select sample from the population that is based upon some specified and thoughtful criteria. Criteria set by the researchers fulfilled by any subject that represents the population should be included in the sample and any subject fail to meet the criteria is not included in sample that will represent the population (Swisher, 2010, p.4).

To make a sample for our research we have selected criteria sampling method. As we are working on the CAMELS rating model and its replicability to the results of the external credit rating agencies published ratings, so for that we have to work upon bank’s annual financial reports. Our criteria for the banks to be included as a sample is the availability of their audited annual financial reports for the year ended 31st Dec 2010. Total population of number of banks operating in Pakistan is 38. Out of 38 banks 5 banks are Islamic commercial banks and operating on Sharia standards that is to great extent different from commercial banks system. CAMELS rating model is designed for conventional banks and its applicability on Islamic banks is argued by several authors but contradicted by others. So for this particular reason we did not include Islamic banks in our sample. There are 6 foreign banks operating with in Pakistan as subsidiaries of other multinational banks, and these banks prepare their annual financial reports and submit it in their Head offices that further add up in their consolidate annul reports. So there were some complications in collection of their annual reports and its interpretation. Because of this reason we did not include these foreign banks as our sample for our thesis. Further on searching for the annual audited financial reports of the banks, we find out that till date 10 banks did not published or announced their annual financial reports. As these banks did not meet with the predefined criteria of availability of their annual financial reports, they are also excluded and are not part of our sample. At the end we are left behind with 17 banks and their audited annual financial reports, so our sample includes 17 commercial banks from Pakistan.