For conducting a truly objective research, it is imperative that the researcher is free from any inherent bias — a task easier said than done.
Bias is defined as inclination or prejudice for or against one person or group, especially in a way considered to be unfair. It is an accumulation of a person’s entire life experience, and may be embedded so deep in his subconscious that he is himself unaware of it.
Apart from preventing an objective assessment of survey results, biases may lead us to falter at the beginning itself.
For example, while preparing a questionnaire on assessing a bank’s customer service, the researcher can frame a question as follows:
How awesome do you think our bank’s customer service agents are?
As we can see, this is a leading question, and is pressurising the respondents to answer in a particular way. Additionally, it contains the ambiguous word “awesome”.
Sometimes, the same question can be reworded as following:
How frustrated do you get while talking to our customer service executives?
This shows a negative bias towards the customer service department and is thus not a neutral question.
Having such questions in a survey is bound to give skewed results, thus defeating the entire purpose of research.
Such mistakes can prove very costly for the companies relying on research results.
For example, in the late 1970s, American car companies like GM, Ford and Chrysler did extensive surveys on consumer car preferences. However, the survey questions were totally biased in favour of the companies. Some questions listed various models produced by these companies and asked consumers to rank them.
Since many models got high marks in the limited choices given to respondents, the companies assumed that they were on the right track.
But within a few years, smaller and more efficient cars by Japanese car companies began to be the first choice of American consumers, leaving US car executives dumbfounded.
Biases can also creep in during sample selection.
If a company limits its survey only to its current customers, the results may not be trustworthy as the sample consists of people who view the company’s products favourably. It would make sense to include some ex-customers also to know what made them stop buying the product.
Similarly, if a research agency conducts door-to-door surveys during the weekdays, this means it is excluding the big chunk of working professionals from their sample. The same problem arises with internet surveys as well, because people with access to the net may not be the true representatives of the general population, especially in a country like India.
Thus, we see how important it is to be unbiased while conducting research to get accurate results.