1.3.5: Education

The field of education has long depended on student performance data to make informed decisions about education policy and pedagogical techniques. One of the most important questions when assessing a pedagogical approach to teaching a subject is determining whether the students learned and retained the material taught during a class. A related question concerns whether changes in pedagogical techniques can result in increased rates of learning and retention. As education researchers have known for quite some time, these studies are very difficult to design and interpret because the concept of learning is very abstract, and hence their research has also considered how to measure the effectiveness of classroom pedagogy on learning in the first place.

Answering these types of questions requires many types of methods from statistics and data science. For example, one might be tempted to conclude that any student who gets a good grade, say a B or better in a course, has demonstrated that they have learned and retained the material taught in the course. But such a conclusion makes many assumptions about how the course grade relates to the work done in the course. One might consider an extreme example where the course grade was only based on attendance. In this case the student has not demonstrated that they have any knowledge of the course material, only that they have attended the course. An exam in a course will only demonstrate a student’s knowledge under certain conditions. For example, an instructor could review the actual exam the hour before the class and then give the exam. Or the instructor could do what is known as “teaching to the exam,” where class time and student work is focus on passing the exam and not on learning the material. The context under which the exam is given is also important. Students must be given the proper time and atmosphere to demonstrate their knowledge. Therefore, the design of the course assessment materials, as well as the content of the course is very important to assessing whether a student has learned the material or not. The field of statistics has developed methodology that aids with these problems, and it helps researchers get an idea of the true effects that things like teaching materials and different pedagogical approaches can have on student learning.

Prediction is another important, and sometimes controversial, aspect of education research. Those who work in admissions for colleges and universities are tasked with the problem of admitting students who are likely to be successful in their college career. It does neither the college nor the student any favor for them to be admitted to a program in which they are unlikely to succeed. This is the reason that colleges require transcripts and other information about potential students when they apply for admission. The problem is one of prediction, trying to determine what will happen in the future based on what is known at the present. Statistical methods can be used to make such predictions, and these predictions can be quite accurate if sufficient information is known about the present. This is the challenge for colleges and universities, and there have been questions from students and parents as to what type of information is relevant and should be used in the admission process. Current trends in this area include the use of artificial intelligence and data science methods to help answer these questions.