Virginia Mathematics Teacher Spring 2017

Polling Data and the 2016 Presidential Election Carrie Case and Jean Mistele

To understand the polling data that was reported prior to the 2016 Presidential elections, it is useful to be statistically literate and to be politically literate. In particular, a statically literate person who understands confidence intervals can make sense of the polling data. Likewise, a politically literate person who understands the election process used in the United States, Electoral College, is better able to interpret the statistical data situated within the political context. In this article we address these two forms of literacy and when they are used in concert, these literacy skills allow people to understand and make sense of the polling data and the results from the Presidential elections last fall. In the following paragraphs we first address statistical literacy followed by political literacy. We close with the way these two forms of literacy serve an engaged citizen. Statistical literacy is a required skill to make sense of our data rich world. Katherine Wallman, a former Chief of Statistical Policy in the United States Office of Management and Budget during the Clinton Administration offers her perspective on statistical literacy: Statistical literacy is the ability to understand and critically evaluate statistical results that permeate our daily lives-coupled with the ability to appreciate the contributions that statistical thinking can make in public and private, professional and personal decisions (Wallman, 1993, p.1). Gal (2002) describes statistical literacy as having two interconnected parts: knowledge and the ability to communicate that knowledge. Shaughnessy (2007) also describes statistical literacy as two interrelated components, the first as a learner of statistics and the second as a consumer of statistics. All three perspectives share the notion that a person needs to know statistics on a level that allows that person to use that knowledge to

critically assess social, economic, or political issues in our world. Next, we discuss the statistical knowledge needed to better understand and make sense of the reported polling data. Statistically, we need to understand how to interpret the polling results so we can make sense of them. Specifically, understanding confidence intervals associated with the polling numbers sheds light on the projected presidential race outcomes. The chart below shows the results as reported by Jennifer Agiesta in the CNN/ORC obtained from the Opinion Research Corporation that polled voters on October 25, 2016. Clinton 49% Trump 44% With a margin of error, E , of 3.5%. In this discussion, we contend that valid polling organizations understand statistics and would implement sound sampling techniques to ensure the people polled were selected at random and that each person was asked the same question; questions such as, “Are you a likely voter?” and if the person answered “yes,” then the pollster may ask, “Which candidate will most likely receive your vote?” In addition, we maintain that polling organizations would use large samples. A large sample would include 1000 to 3500 people from the entire population of likely voters, which is approximately 120 million to 130 million people. The polling data shows the results in proportions— the percentage of people favoring each candidate. Specifically, the polling organizations attempt to calculate the range of the true proportion of voters that will vote for a particular candidate. This range is called a confidence interval. This confidence interval estimates the proportion, p, of voters who indicate they will vote for that candidate, which includes a margin of error, E. The margin of error is the amount of variation calculated from all of the responses in the sample. This is the percentage above the proportion p and the percentage below

Virginia Mathematics Teacher vol. 43, no. 2

42