D) One hundred random students from the sixth grade
First off, the survey should be focused on the sixth graders, since they are the demographic that will be voting for 6th grade treasurer.
Next, the group should be random, because random groups better represent the entire class than an alphabetic list. Maybe those first 100 kids in the list like Noah because he has a last name that begins with A or something. It's not likely, but it's something that could corrupt the data collection.
