Avoid Survey Bias

One of the greatest challenges facing survey researchers is their own agenda. It’s not uncommon for the researcher’s goals to play a role in the value of the survey. An example is when a researcher decides to use a survey in order to justify an opinion they already have about a particular course of action. For example, the company is planning on initiating campaign X, so they use their survey to show why campaign X is going to be effective, rather than to see what the population feel about the idea behind campaign X.

When researchers or businesses use surveys to justify positions, they run the risk of introducing several potential problems: Interpretation – If the results of the survey are either not statistically significant but leaning in their favor or not very strongly in favor of the idea being researched (but still statistically significant), the researcher may decide to interpret the results as proof that their theory is correct, ignoring the fact they have no proof of that statement. Wording Biases – It’s not uncommon for those that want to run the survey to also write the survey, and if you write the survey yourself you also run the risk of introducing biases with your wording that cause you to receive the results you expect. It’s very hard for language to be 100% objective, and writing the survey yourself introduces the potential to alter your results to meet your needs.

Ignoring Other Information – When you create a survey for a single result, you also run the risk of ignoring other valuable information that may be apparent in the survey, because you find yourself too focused on proving what you believed to be right as right. You want every survey you run to provide you with as much valuable information as possible, and if you are too strongly interested in one result you may not see all of the other data that’s there. Research needs to be objective.

While the researcher may have an expectation about the results of the survey (and indeed, a great deal of research is completed where the researcher has a pretty good guess what the outcome will be), they cannot be using the research to justify a decision they’ve already made in their minds. If they do, they are likely to alter the results of the survey and yield data that isn’t indicative to the feelings of the population.