As a student newspaper, “The Patriot Post” strives to produce the most accurate, thorough and interesting news, sports, features, opinion and entertainment stories for our audience. It’s clearly written in our mission statements, both online and in print. As student journalists, the Patriot Post staff works hard to put out every issue, every online article and every photo gallery that reaches the public eye. We try not only to report important occurrences and discuss relevant issues, but to appeal to students from all across campus, regardless of grade level, grade point average or social circle.
Part of appealing to the student body is incorporating their opinions into our work. Here is where the student polls come in. Usually, “The Patriot Post” includes statistics of some sort, whether it be in the large centerspread or in one of our previous one-page spreads from the newsprint version of the paper. At the bottom of the percentages reads, in small print, “(X number) of students polled.” While what follows may be targeted at my small circle of friends who accuse our polls of being unrepresentative of the student body, let me tell you why our polls are statistically sound.
In the statistics classes offered at Heritage, the first chapter the students learn is that of sampling procedures. We learn how to record data without bias so that it is represents the population of interest to its best ability. What “The Patriot Post” does is send out a survey to all students of the Upper School student body. Nobody is required to fill out this survey; only those who choose to fill it out actually have their responses recorded and have an influence on the next paper. While this causes what is known as response bias, any alternative proves to be difficult as well. Essentially, each survey can be thought of as an attempted census (a method of collecting data in which the entire population is surveyed as opposed to a selected sample).
Another common misconception is that more responses is the equivalent of a more accurate depiction of the population. As my statistics teacher put it, when cooking a large vat of soup, only a small spoonful is necessary to understand how the entire thing tastes. With over 1,950 students in the American Heritage Upper School, it’s even difficult to have at least a 50 percent response rate (975 responses). The survey from most recent issue of “The Patriot Post” had 280 responses on the topic of sexual harassment. Looking at some of last year’s numbers, since the other two issues this year did not include student polls, the May 2018 issue survey had 474 responses on the topics summer activities and shoes; and the March/April 2018 issue had 249 responses on the topics political activism and candy.
In statistics, there is something known as the “ten percent condition.” It states that, when sampling from a finite population, observations cannot necessarily be considered independent of each other because knowing the outcome of one trial helps to predict the outcome of future trials. However, by journalistic standards, a survey should have at least ten percent of the population sampled. As always, there are exceptions to statistics, and we don’t “sample” but rather census as part of a journalistic publication, the journalistic standard presides. Ten percent of the population is about 195 students, and we surpass that standard.
Here’s what you can do as students and recipients of our survey to help make our surveys more journalistically sound: respond to them. The more people that answer, the less response bias we have, the less nonresponse bias we have (bias that occurs when people don’t respond) and the closer we are to achieving a census. Our surveys are sent in our best interest, and the only required questions are generally multiple choice. Sometimes your response may require a bit more thinking, but that’s another one of our goals — to provoke thought, start a conversation and have an impact.