PEER'S POST

We collect data daily and compile it monthly to look for behavior trends and ways to update and improve their support plans to decrease further any maladaptive behaviors they may have. For Susan, cluster sampling would not represent the population well for her desired results. If she randomly selects and utilizes all groups and units, this will not account for her concern with event turnout. To fully account for all students and achieve her desired result, I believe utilizing stratified sampling would be the most advantageous. This way, she could select equal amounts from each grade and each gender.

ORIGINAL REFERENCE STORY:

You will take on the role of Susan, a high school principal. She is planning on having a big movie night for the high school. She has received a lot of feedback on which movie to show and sees differences in movie preferences by gender and also by grade level. She knows if the wrong movie is shown, it could reduce event turnout by 50%. She would like to maximize the number of students who attend and would like to select a PG-rated movie based on the overall student population's movie preferences. Each student is assigned a classroom with other students in their grade. She has a spreadsheet that lists each student's name, classroom, and grade. Susan knows a simple random sample would provide a good representation of the population of students at their high school but wonders if a different method would be better.

1.) Can you identify bias in the peer's reply post sample and explain how you think they could remove the bias from their sample?