After the research objectives have been specified and the list of needed information has been prepared the researcher should determine whether such information is already available, either in company records or in outside sources. Certainly the researcher should not collect data from the field until the appropriate secondary sources of information have been reviewed.
If the needed information is not available from secondary sources, the researcher will have to collect data in the field and so it will be necessary to design a data collection project.
Problems in achieving the Scientific Method: Researchers encounter many possible sources of error when designing a data collection project. Some of these include using a poor or inappropriate research design (not using experimental designs where possible); not selecting the type of respondent who has valid and reliable information; asking unclear or ambiguous questions; using a large scale study instead of a small scale study or vice versa; using poor experimental designs, and still others.
Minimizing Potential Sources of Error: Researchers can do many things to minimize the possible sources of error associated with designing a data collection project. Many methods are suggested in reducing errors in marketing research projects. At this point five important issues are introduced that must be addressed when designing a data collection project. By giving each of these issues proper attention researcher can make good progress toward reducing the possible sources of error in their research projects. These issues are:
Should the research be exploratory or conclusive? To a great extent, this issue is likely to be determined by the research objectives and the situation leading to the request for research. Exploratory research is likely to be used when investigating a potential opportunity or problem, and conclusive research is likely to be used when the research findings are expected to results in specific decisions being made or specific actions being taken. Although there can be numerous exceptions to this generalization, the determination of whether to do exploratory research or conclusive research will be strongly guided by the objectives of the research.
Who should be interviewed and how? Next, the researcher must identify who has the needed information and how it can be obtained. For example, in a study to determine if children influence the brand (s) of ready-to-eat cereal purchased by their parents, the researcher must decide if only adults are to be studied, or if children are also to be included. The researcher must decide if data are to be collected through observing purchase behavior in supermarket or through interviewing and, if the latter, whether the interview is to be a personal one in stores near the cereal display or a telephone interview at the home where the cereal is consumed.
Should only a few cases be studied, or should a large sample be used? If the study is a conclusive one, if the findings are to be representative of the universe, the researcher may want to use a large scale statistical study. However, if the project is concerned with findings a best way rather than a typical way of doing something such as carrying out on the main floor of a fast food restaurant the research may well consists of the analysts of a few selected cases. The cases may be selected on the basis of the researcher’s judgment as to what are the most successful cases available and, possibly the least successful for comparison. Only a fewer cases would be chosen maybe only three or four.
How will experiments be incorporated in the research? If conclusive research is being undertaken, the research project may include one or more experiments of some kind. If an experiment is part of he research design. The researcher must decide such things as where and when measurements will take place and how the experimental variable will be applied. For example, in a test of advertising copy, respondents could first be interviewed to measure their awareness of and attitudes toward certain brands. Then they could be shown pilot versions of the proposed advertising copy. Following that, their attitudes could be measured once again to see if the proposed copy had any effect.