Statistically Significant? Data Analysis Classes Can Answer This


“It wasn’t about the survey, it was about the process. The kids learning the different techniques about the sampling. It was about the reflection, realizing what worked and what did not work, learning from it”       –Griffin

The Data Analysis classes, taught by Ms. Griffin and Mr. Lyons, create the school-wide surveys you occasionally take on Schoology. These surveys cover a broad range of topics.  Ms. Griffin shared some examples of topics, including ethnicity within Bolles, boarders vs. non-boarders, grades, and favorite subjects.

Data Analysis students choose the topics of their surveys as well as administer them to the student body. “This was totally their baby. I gave them parameters and said fly and they did,” said Griffin. They chose their topics in their smaller groups within their class and then create their survey questions. The whole process is student-led.

According to Lyons and Griffin, the key to making a well-written survey is asking questions in a neutral manner, not asking leading questions. If a question sways you to a specific opinion/answer, then it is a leading question. For example, “Would you vote for Mr. Incredible even though he is currently being investigated for unauthorized use of superhero powers?”

Another quality sought for in a survey is repetition. Mr. Lyons said, “Repeats are purposeful and they are there to measure consistency in a survey.”

Another task while creating a survey is getting equal representation of opinions. In order to find this equal representation of student opinions from around the school, different methods of sampling are used such as cluster sampling, proportional sampling, and stratified sampling.

A well-distributed survey represents a range of opinions. If one opinion dominates, then that survey has not been properly administered. If people choose not to respond to a survey, then then that survey was also not properly administered. Lyons said, “People have told us ‘no’, they won’t do it. You just have to deal with it.”  If students face these situations, it can be a learning experience. Ms. Griffin said, “They could see how statistics could be used to vastly misrepresent a topic.”

Once the students receive their answers, they analyze the results. Many times, the results  differ from pre-survey predictions. For example, in a survey about favorite subjects around the school, the students predicted English or science and were surprised that math was the favorite.

To visually show the data, they create bar graphs and determine whether their data was statistically significant.

   Once the graphs are created, groups begin the hardest stage. They must give a presentation analyzing the results and the effectiveness of their survey.

Taking surveys can be sweet! Lyons told a story from his own statistics class from when he was in high school. “We had a correlation activity with M&M’s concerning the weights of bags of M&M’s. I made up a joke and sent a letter to M&M Mars telling them that my bag of M&M’s weighed the least and I was ridiculed by my friends. They ended up sending me a coupon for a free bag of M&M’s- I consider that a win!”