Students in Ms Herringshaw's statistics class are deciding on a method to survey the 1,600 students in their four year high school for a project. Of the following survey methods, which is most likely to introduce bias in the data?
1) Randomly selecting 400 students from each grade level
2) Surveying all students in the first grade only
3) Surveying all students in the last grade only
4) Surveying all students in the math class only