- Donald R. Cooper
A professor at Florida Atlantic University, teaching research methods, statistics, and organizational behavior. Cooper is also the principal of Cooper Research Group, a customer satisfaction and market research consultancy.
Continued from: Theoretical framework and hypothesis development
Research design is a roadmap for researchers. It is step by step approach. Research design is prepared keeping in view some basic questions like “What would be scope of my study?” or “What data do I need to collect?” or something like “What methods will I use to collect the data and how will I justify them?”
A leading author, Donald R. Cooper, had defined research design as follows: “The research design constitutes the blue-print for collection, measurement and analysis of data. It aids the scientist in allocation of his limited resources by posing crucial choices: Is the blueprint to include experiments, interviews, observations, the analysis of records, simulation, or some combination of these? Are the method of data collection and research situation to be highly structured? Is an intensive study of a small sample more effective than a less intensive study of a large sample? Should the analysis be primarily qualitative or quantitative?"
Essential of Research Design
The above definition covers the essential of the Research design. It would be an activity and time based plan depending upon the research question. It would guide what types of information is to be collected and from what source. It would be a framework for specifying the relationship among the study's variables, Finally, the desisgn outlines procedures for every research activity.
This has been briefly described in the sketch on the left side.
Types of investigation
Basically, there are three types of studies: (i) descriptive, (ii) correlational and (iii) causal.
Descriptive study just describes the situation. A bank manager may like to know how much loans were advanced and how much were recovered within the given time and how much were never recovered. Similarly, a cane manager of a sugar mill may like to study how much cane is available within a radius of 5 km, 10 km, 15 km and so on. In both case, actual figures are just being described or presented with percentages.
Correlational studies identify factors leading to a certain situation. For example, smoking, drinking and chewing tobacco maybe associated with cancer. While there is a link, one cannot establish a relationship between cancer and its causes: number of cigarette smoked, drinks taken and quantity of tobacco chewed.
Causal studies, on the other hand, bring up a definite cause and effect relationship. It isolates the variable that changes the dependent variable. For example, smoking causes weak lungs and not vice versa. Also to establish cause and effect relationship, it must be demonstrated that (i) both X & Y are co-varying, (ii) X (the independent) was preceding Y, a time lag, (iii) no other variable was possibly causing the change in Y and (d) when X is eliminated, Y is eliminated.
PURPOSE OF STUDY
The purpose ranges from pure research to problem oriented research. In the former, the researcher just wants to satisfy his or her curiosity. For example, a researcher is curious to know if there is a difference between work ethics of Japanese and American? Later such studies would be used for correlating the worth ethics with industrial advancement and innovation. Eventually, these would guide the policy maker in human resource planning and development.
Problem-oriented research are undertaken with a specific problem in mind. Education department may be considering increase the number of text books to increase knowledge of students. This maybe because of some previous finding that "the understanding of a subject showed improvement when number of books were increased."
The following table shows various studies and their purposes:
PURPOSE OF STUDY
Interviwing individuals, Basic Research, Extensive Survey
Study of secondary data, Analysis, Correlation
Studying Relationship: Cause and Effect
Work ethics - Asian and American. Quality of Life Index, Investment Opportunities in a far off area
Cane Zone of a sugar mills, Aging of Loan defaults, LIFO Followers, Destruction of a city
Advertising and Sales, Price and Demand, Whistle Blowers-more men than women
Nature, Non-contrived, Field Studies
Artificial setting, contrived, Field experiment, Lab Experiment
Degree of Interference
A researcher may decided to keep his or her interference as minimum as possible. Events would be studied as they occur; data would be collected though a questionnaire and no changes made in the input or comments. But this would NOT be applicable for a causal studies where excessive interference or manipulation is called for to establish a cause and effect relationship. In order to find out effect of salary on the output, the researcher may ask the company to increase pay of employees of one department, lower the pay of employees of another department and leave the third department as it is. There maybe a change in output or may not be depending on the motivation of the employees.
Setting may be natural or artificial. In natural setting, field studies are conducted. For example, a bank manager may try to find out relationship between returns and bank deposits. For this purpose, various saving schemes of the banks would be studied to find out whether more returns attracted more deposits or motives of the depositors were other than returns like safety and liquidity. Similarly, researcher may observe wild life in their natural habitat though hidden cameras or field glasses.
Artificial environment are staged in lab or field experiments. There would be lot of manipulation and interferences. Lights maybe increased or decreased to observe their effect on productivity. Deliberate distractions maybe created for study of the loss of concentration.
A researcher may show inspirational films to students and note improvement in their attitude towards achievement. In a field experiment, sugar water may be placed at different distances from a bee nest to observe how the difference in distance affects the dance that bees do on returning home to communicate where the food is.
UNIT OF ANALYSIS
Unit of analysis may be individuals, pairs or groups. When we are interested in knowing as to how many managers are interested in learning computers, we would be using individuals as a unit of analysis. In study of pairs or dyads, pilot will be studied along with co-pilot, trainer with trainee, doctor with nurse and driver with conductor etc. On still higher level, we may study clusters of a community, branches of a bank or residents of various localities which would fall under "groups" as unit of analysis.
It may be cross sectional or longitudinal. In cross-sectional, data is collect only once. It is not necessarily gathered at the same time. It may take a longer period for example to find out from bank employees scattered all over the country their favored month for taking leave. Or one can ask men between the ages of 35-44 for choice of color of a shirt in a survey lasting for two weeks.
In Longitudinalstudies, data would be collected at several points of time say temperature of a patient in the morning and evening for three consecutive days etc. Similarly, one may like to note employee behavior before and after raise of pay or promotion. There are other examples like sales before and after advertisement, annual confidential reports of employees for the last five years etc.
What next : Experimental Design