Polling and Research – New Challenges, Proven Benefit?
In the wake of the US presidential election, much has been written about the accuracy and usefulness of modern polling. There is no shortage of commentary suggesting that modern day polls are inaccurate, useless, and given more attention and credit than they deserve. Specifically, this commentary has been primarily directed towards the publicly distributed horserace numbers that have come to dominate daily election coverage. While modern polling and research goes well beyond publicly available vote intention surveys, their criticisms do touch on valid concerns.
These sentiments largely come from two concerns. First and foremost, they assert, although possibly not in these words, that modern polling suffers from a selection bias in which the makeup of the individuals that participate in a poll differs from the makeup of the broader population. Since it is unlikely that the participant pool will be perfectly representative of the broader population, pollsters and market researchers apply various weightings to the polling data in an attempt to ensure its representativeness. While they primarily do so based on demographic data and historical voting patterns, there are obviously other, less tangible, factors at play that may not be so easily accounted for. In short, there may be factors that relate to an individual’s willingness and ability to participate in the research that may also relate to their preferences.
Secondly, pollsters face the constant challenge of equating an individual’s opinions and statements given in a moment of time with the actions they will take in the future. There is obviously the potential for individuals to alter their opinions between these two points in time, but there is also the possibility of some individuals providing alternative responses if they perceive that there is a stigma associated with the opinions they hold. In the case of public office holders and vote intention, this is known as the Bradley effect, named after the Los Angeles Mayor who lost the 1982 California gubernatorial race despite leading in the polls, and has been pointed to as an explanation for the poor performance of recent electoral polls.
Given the above issues, it is almost no wonder why individuals are skeptical about the merits of modern polling and public opinion research. While any type of research ultimately stands to benefit from skepticism and constructive criticism, many critics of modern polling make the mistake of treating as a crystal ball as opposed to a tool.
For those that use polling to inform their decision making, polling is a valuable tool and strategy for obtaining external information in general. It is a means of validating or testing internal strategies, tactics, and observations. Sophisticated organizations will use polling to assess the quality and accuracy of data retrieved through other means, and of course, using that data to assess the quality and accuracy of the polling itself. Should they differ, the organization has a starting point to re-examine their own data, their collection methods, their understandings and possibly other conclusions that have been drawn from the polling.
Modern engagement campaigns provide organizations with data from a variety of sources, each with their own potential for bias and inaccuracy. To discard one of those sources, and one of the longest-used sources at that, would be foolish. While polling may no longer meet the standards of prediction and forecasting that the media and the public have come to expect from it, it is still a vital tool for organizations.
Contact us to learn more about how Candidate Cloud can help improve make your organization's research efforts.