Voting
The United States presidential election is just a few days away. Many news reports focus on polls to indicate which candidate is in the lead to win the race. What is the relationship between election polls and statistics? To understand this, we first need some background in statistics.
Some people consider statistics as a science related to developing methods of collecting, analyzing, interpreting and presenting data. I consider statistics as a highly versatile field since many scientific and non-scientific disciplines use statistics for their data analysis. To sum up in much simpler terms, statistics can be considered like a quality control process, especially when very large numbers are involved. This type of methodical data quality control is crucial during democratic elections since millions of people cast their votes to decide the next leader. Furthermore, the data can be used to analyze the outcome of past elections for future strategic planning. For the general public, the main focus is on the outcome on election day; however, behind the scenes, statistical methods are applied much earlier in the election process. For example, data analysis methods are applied to know the population size, age groups, income levels, political party preferences across the country. This type of analysis gives political leaders an insight into which slice of society is favorable to which belief systems, thus helping them strategize their campaigns.
Counting
Any data collection, big or small, is prone to errors. The statistical methods of confidence intervals or the margin of error are used to minimize the impact of errors on crucial outcomes such as elections. A Confidence Interval (CI) means the level of confidence one has about the outcome in the entire population by looking at the outcome in a much smaller sample. In a dynamic process such as elections, the population outcome cannot be predicted with 100% confidence. Mistakes made during predicting election outcomes or counting votes cannot be ignored; therefore, CI is always associated with the margin of error. The margin of error is defined as the number of points by which your prediction will differ from the actual result. For example, suppose I report a poll with candidate A having 60% support and candidate B having 35% with a margin of error of +/- 4%, based on a 95% confidence interval. This result means that 95% of the time, candidate A has support between 56 – 64 %, while candidate B has support between 31 – 39%.
Results graphs
The beauty of statistical analysis of elections is that it not only tries to make a prediction of the current outcome but also gives valuable information to political parties for future elections. To the world, the USA seems like a polarized nation with a parallel set of values among st its citizens. In my opinion, the common fabric that is binding us all together is voting and the data interpretation using statistics. Whatever might be the outcome this November, researching for this article has given me a new appreciation for numbers and the statistical methods. Age might be just a number, but when it comes to voting, it is more than just numbers!
