Friday, March 30, 2012


          Research process starts once you are facing any unanswered question. These questions may cause confusion, anxiety, anger, or surprise you for something.. All such conditions of human behaviour are resolved by finding the answers to these questions, which invites researchers to follow formal procedure to investigate the case and bring out facts and underlying reasons:
For example, you might notice on any day during week an unusual turnover of customers in the store. This observation will create a research question in your mind.

“Why number of customers is high in store today”
Research question will set an objective for doing research.
          Immediately your though process will start working and display out a list of possible answers for your question. These answers are based on your existing level of knowledge.
They might be like this:
·          Due to national holiday today number of customers is high in store.
·          Due to first week of the month number of customers is high in store.
·          Due to sunny day today number of customers are high in store
·          Customer’s turnover is high because store has announced discounts on FM radio for some items.
·          Government has initiated mega road construction project which has caused inconvenience to majority of residents to reach to other stores. That’s why people prefer this store for their day to day shopping due to only convenient option left for them.
These possible answers are called Hypotheses.
          If nature of question is ordinary and you are not concerned with the accuracy of the answers of the research question then you may choose any one of the above possible hypothesis as best reason to your question.
          Up till this point every human being is applying research method every next moment in his life. Researchers or academicians or scientists move further to satisfy their intellectual zeal. They also need application of research method till further steps because their credibility is at stake. Any misinterpretation can cause damage to their reputation.
Research process consists of following steps:
1.         Clarity of the research question:
            First step is to rephrase the research question. Rephrasing means trimming the statement for research. We cannot start research unless we are not clear about the objective. And we also cannot do research on many directions. If we are facing problem of frequent electricity failure in our town, we can find number of possible reasons immediately. Each reason leads to a separate research. We may develop a research question that, “is this electricity failure due to heavy storm and rain outside”. Likewise another research question can be developed as, “what makes electricity supply so vulnerable to the climate?” and so on. Notice that  in each question research objective is clearly identified. Uni-directional questions are always preference of the researchers. Clearly defining research problem or research question is a pre-requisite of next step in research process. Uni-directional doesn't mean single answer. These may have more than one possible reason for that situation. That's why researchers avoid inserting question mark? with research questions.
2.         Preliminary investigation to list possible answers to the research question.
            Once research question is clearly defined, we move to second step in research process. In this step we explore further the situation and try to find immediate answers. These answers can be sought by using our wisdom, or by taking opinions from others, or by dipping into the available literature in libraries manuals, or internet. Formally we use the terms Literature Review and Interviews for this step. Literature review helps us to know previously conducted research on same or similar situation. By applying our wisdom, conducting interviews, and reviewing literature can help us to reach best possible answers to our research question.  The main purpose in this step is to identify all possible factors which may cause that situation which is under investigation. Without identifying these factors or variables we are unable to move to next step in research process.
3:         Development of relationship among variables:
            Third step in research process is to classify all factors or variables which are involved in the situation. These variables can be classified as dependent variable(objective variable), independent variables(factors or causes of the phenomenon), connecting variables(which intervene among relationship and their interdependence), and sensitivity variables(which define the strength of relationship between two connecting variables). For example, in our research question “what makes electricity supply so vulnerable to the climate?” following variables can be identified:
·          Dependent variable:                     Failure of electricity
·          Independent variables:                 Strom and rain
·          Connecting variables:                   Availability of electricity supply system, etc
·          Sensitivity variables:                    Quality of cables, Disaster management system,  etc
          Sometimes lurking or confounding variables are also important to identify. They are hidden factors and not apparently appeared. But their presence can affect the relationship between dependent and independent variables. For example, in above case the corruption on the part of technical staff may cause the situation. Then variable “corruption on the part of technical staff” is a confounding variable.
          Correctness and coverage of all possible variables depend on the rigor research puts into the second step. Step 3 provides a solid theoretical foundation to the research problem. This theoretical frame work will pave the way for further development in research process.
4.         Developing Hypothesis
            Hypothesis are the possible answers to the research questions on which researcher has reached after going through step no 2 and 3. These are wisely driven non-verified conclusions coming out of the mind of researcher as possible answers. For example one hypothesis can be framed as “Failure of electricity is due to poor quality of cables in the town”. Similarly researcher can frame other hypotheses for one research question.
          Notice that in our day to day life all human are going through the process explained in first four steps above. They take decision by selecting best possible reason out of the hypothesis they have framed based on preliminary investigation. They become professional scientist or researcher when they further investigate for the truth of the hypothesis so that they can interpret it according to current knowledge level prevailing in the society.
5.         Research Design
          This is most technical part assumed in research methodology. Selection of right design requires an overall experience of the researcher. Most of the time students are referred to their supervisors for specific advice. I also observe that majority of students are not comfortable when they reach to this step in research process, and they need external support to seek through guidelines to proceed with the research process. Research design is basically a program of investigation. There are more than one programs or design. Each design complete different objective.
          If you are doing research on any phenomenon which is quite new to the society, then you need Exploratory Research Design (ERD) technique. Which mostly apply surveys, interviews and primary data collection tools for getting right answer to research question? We should not confuse exploratory research design as a process involved in step 2. Here literature is not sufficiently available to help researcher on the subject. I remember that doing research on Knowledge management about ten years back was a daunting task, but now researchers are able to find some good material on that.
          If objective is to develop a profile of the variables or factors engaged in the research, then Descriptive Research Design (DRD) technique is the appropriate model. For example, If our research question is “what make Toyota market leader in car industry”, then we may require in depth analysis of the factors involved in the production process and their overall business strategies. We need DRD.
          If researcher is interested in studying  the relationship among variables then Relational Research Design(RRD) techniques  is appropriate. RRD helps researcher to develop a relation between variables and to understand the degree of causal relationship among different variables in the research. By causal relationship we mean functional relationship among variables. Causal relationship also indicates the presence of cause-effect among variables. For example, if hypothesis is that, “electricity failure is due to poor quality of cables”, then in order to study the cause and effect relationship between quality of cable and electricity failure, RRD is appropriate.
          Sometimes, researchers may adopt bench marking to reach their objectives. For example, in business research, one organization may try to understand the model set by her competitor in the market. In that situation The Case Study Research Design (CSRD) technique is advised. One should not confuse case study with the descriptive research design. Case study is the overall analysis of the organization or situation, whereas in descriptive research design one may be interested in some of few variables.
          The above mentioned research designs are also set hierarchy. That is researcher may have gone through ERD, DRD, RRD, and CSRD to complete his research.
          Above classification is based on the objective of the study. Besides above classification researchers also take care of some other factors in choosing right research design. If researcher can control the situation or have some capability to intervene the relationship then Experimental Research Design technique is appropriate. For example in Relational Research Design if researcher can control independent variable in some artificial environment, then he can generate outputs according to his desire. Later by comparing two outcomes (intervened and non- intervened) can be compared. Like teacher can compare two teaching methods, one group which is taught using internet and one which is taught using text book.
          Many times field study is appropriate because research has very less control on the situation then data is collected on the basis of available facts. For example opinion polls are done without researcher intervene.  Field study is normal executed on exploratory or descriptive research design.
          Likewise units of analysis are also important in selecting research design. Unit of analysis can be individual (group) or more than one group. For individual group one can opt for exploratory or descriptive research design. For more than one group relational research design can be appropriate. Likewise researchers also take care whether research is inter-temporal (on time scale), or cross-sectional (time is assumed constant). For example one may need to study the growth over a certain period of time. Or one may need to observe future affects of some treatment applied now on the cases. This is called longitudinal or inters temporal study. In other case researcher may need to compare two groups during same period of time. This called cross sectional study.
          In brief, it is very difficult to draw hard and fast rule about type of research design. Researcher may move from one research design to other to reach his objective. The researcher rigor in this case will set the value of research outcomes.
6.         Data collection
          Next step is the data collection. This is again a technical issue in research process. Mostly statisticians are involved from step 5 to step 7 during research process. They help researcher to determine appropriate sample size, and the way to collect date. Data or information is collected through interviews or through questionnaires. This data may be classified as quantitative or most of the time it is qualitative. Statisticians help researcher in operationalization of the variables in step 4 during the selection and defining of variables. Operationalization means converting variable into some measureable phenomenon. For example, if our variable is “Quality of capable ” then variable “quality” should be measured through cable specification or material or weight per meter etc. By operationalizing the variables researcher is able to tap qualitative features in some measureable quantitative scale. Likewise student’s intelligence can be tapped through his number of correct answers in any Quiz designed for the purpose. Here I would advise young researchers to learn basic difference between different measurement scales. I will cover further technicalities involved in some exclusive topic on data collection.
          I always advise my students to associate yourself with statistician right at from step 1. It is always helpful and save lot of energy, time and other resources of researcher. By statistician I mean an individual who has experience as applied statistician. Sometimes a person with good theoretical base, but with little research experience causes confusions in minds of students. Almost in every university there are valuable resources available as applied or research statisticians. They can advise and help students through statistical software like SPSS, SAS, R, Minitab, Microsoft Excel, Magastat etc.
7.         Data Analysis
          Data analysis is the output of the research design program. The data collected in step 6 is understood, and explained according to the objective of the research hypothesis. Statisticians prepare a reliable report based on collected data. Reliability of the report covers following areas:
a.       Has the data been collected from right source?
b.      Is data capable to tap desired information?
c.       Is the data capable to answer research question?
d.      Has researcher applied appropriate method to collect data?
e.       Is data complete and sufficient?
f.       Is the outcome unbiased?
g.      Is the hypothesis significant?
h.      Inference based on data.
i.        Answering research question
8.         Conclusion
          Outcomes of data analysis help researcher to write conclusion of the research question.

The above mentioned steps are written with care to avoid technical language.
(Author: Muhammad Azeem)