Improving Validity of market research

As discussed in our previous issue, Market Research involves problem definition, developing a research methodology or an approach to the problem defined, research design, collection of data from participants, data extraction for analysis and finally reporting.

A good research practice is to collect secondary data and analyze it completely before analyzing primary research data. This is because primary data addresses the issue at hand, unlike secondary data which is used to analyze all the other ‘related’ issues to the problem under concern. To be more precise, secondary data is mostly related to background information and can only add to or support the information that is later generated from primary research.

Having realized the importance of conducting primary research, it is now critical to understand the nature of primary research to use to achieve an objective. It is therefore necessary to define the approach towards problem. The approach adopted may be formulation of research questions, hypothesizing and developing theoretical frameworks or models. This is then tested by developing appropriate research methodology which could be either ‘analytical and quantitative’ or ‘exploratory and qualitative’. A quantitative analysis is not as error prone as qualitative research due to its dependence on numerical values and statistical analysis. The only bias issue with quantitative analysis could be when the sample chosen is erroneous and is not reflective of the population characteristics. The method of obtaining data may be through surveys, experimentation and observation.

A qualitative analysis is extremely beneficial in understanding group dynamics and the trends that are emerging in evolving markets. It is important to gain an insight into the changing needs of people by actually speaking to groups. This could be achieved through focus groups, case studies, in-depth interviews and word associations. However bias could be generated in every phase of discourse starting from the way focus groups are conducted to the way information is analyzed.

How can this bias and ‘generalisability’ issue be avoided? Studies indicate that best practice in conducting such a research is by applying intellectual vigor right from the point of sampling to applying techniques to analyze non-numerical data. In practice, it is hard to achieve the ideal sample, because research may sometimes involve hard-to-reach population. But one must try to have a well-selected and diversified sample to add to validity of data. It has been found that mutual trust between researcher and participant enables a constructive conversation which is helpful in eliciting more accurate data. Social theories can be used to provide explanatory framework and claim that the outcome of research must be applicable to a range of other social groups. If there is a deviation from the social theory and the secondary research, then the validity of data elicited is questionable.

Studies have indicated that the best way to avoid the researcher induced bias is by having holistic description and engaging in corroboration through triangulation. According to Denzin, triangulation is of four types data, investigator, methodological and theory triangulation. It is a proven method of obtaining data from 2 or 3 different sources to test the accuracy and validity of data. The overall idea is to overcome intrinsic biases associated with single theory, single method and single user.