Opinion Polling

Opinion Polling and Random Sampling

Obtaining a true random probability sample continues to be a challenge for researchers. Opinion polls have owed their accuracy to the appropriate use of probability sampling, traditionally conducted by landline phone through random digit dialing. However, due to advances in technology, the Internet and mobile devices now overshadow print material and landline methods of communication, requiring the nature of polling to evolve as well. The struggle today lies in attempting to obtain the most representative samples possible, which may not be feasible without requiring traditional methods to adapt to societal changes.

According to survey researcher Scott Keeter, the two major challenges are avoiding nonresponse and insuring adequate population coverage, given the decreased use in landline phones and increased use of cell phones. Through dual-frame probability sampling, whereby the sample selected incorporates both landline and cell phones, pollsters have found a solution to adapt traditional telephone polls to actually enhance coverage. However, increasing trends of nonresponse to polls and surveys remain troublesome.

In addition, widespread Internet access has resulted in another challenge: addressing the gains in popularity of non-probability-based online polls, e-mail polls, and social media polls. Although supporters of such polling argue that the proliferation of the new technologies call for new ways to [Page 1150]survey public opinion and gain quick access to specific content users, the fact remains that social media users are not representative of the population and that such polls do not follow the scientific reasoning that underpins probability sampling and make it a useful tool for making projections about a broader population.

In 2014, The New York Times, a long-term advocate of probability sampling, decided to reevaluate its polling standards in light of the aforementioned issues. Although this action was initially met with much criticism from survey researchers, there is now general agreement within the polling community that traditional methods may be hindering the ability to gather public opinion effectively, given technological advances. The American Association for Public Opinion Research has taken the lead in studying ways in which modern survey methodology can be improved.

Mark L. Berenson, Nicole B. Koppel, and Laura L. Chapdelaine

See also Respondents; Sampling, Methodological Issues in; Sampling, Nonprobability; Sampling, Probability; Sampling, Random; Sampling Frames; Survey: Sampling Issues

Further Readings

American Association for Public Opinion Research. (2014). Furthering our understanding of today’s survey methods. Retrieved from http://bit.ly/VPFdAL

Cohn, N. (2014. July 27). Explaining online panels and the 2014 midterms. Retrieved from http://nyti.ms/1rtIOSB

...