Berkeley Blog: Second Semester Subjects - Statistics
Degrees in liberal arts are many times the butt of jokes about how easy they are, or how they do not prepare students to hold any difficult or demanding job, or any legitimate job at all. These jokes suggest that liberal arts degrees prepare kids to be nothing more than really amazing drive-thru employees, and that science degrees are far more important and superior. Yes, science is imperative, but the reason I even know about science is because of liberal arts. The fact that you understand what I'm typing, that we use a common language, and the fact that I can type at all, are thanks to language arts. Mass media, marketing, journalism, and so on, inform me about science. The entertainment industry, graphic designers, writers, directors, and so on, allow the public to fantasize, dream, and think about science in a variety of ways. Public policy, history, law, and political art degrees help to design better systems to promote science and literacy. Science literacy does not come from science or scientists but from liberal arts teachers.
Hey Matt! What does this all have to do with statistics! Well, I'll get to that. But first what is statistics?
75% of statistics is spent learning about and correcting for errors. Imagine trying to test the entire population of America. 213 million people would have to participate and give consent each time we want to know something about human behavior. This would be impossible. Even our census, which is a simple survey done every four years, costs millions of dollars and is far from perfect. If we have trouble with a simple survey, imagine the problems we'd have with more complex research. So, in these cases we need to test a smaller, more contained group of people. But since we are not testing everyone, errors can arise from assumptions we make about a smaller sample versus the larger population. Statistics account for these errors in a variety of ways: random sampling and random assignment, single and double-blind studies, and making sure the data and the mechanics of measurement are reliable, among others. Mathematically speaking, ideas such as standard deviations, standard errors, means, medians, modes, distributions, probability, and much more, introduce ways to correct for errors when running statistical tests. Both of these things, proper sampling and proper error correction, account for the majority of what stats classes are all about. The rest of the class, maybe 10-15%, is actually doing research.
There seems to be two ways to teach a stats class. The first way is to simply teach students how to identify which formulas go with each type of test. This is designed to allow students to focus more on the research than the mathematics. The second way is to provide students with a much more in depth, mathematical heavy view, requiring students understand the genesis of formulas. Learning the logic behind formulas may give students an advantage later in life but requires much more time and energy. I've been part of both techniques and many students probably do not need such an in depth view and would do fine with the simpler way to teach the class. At Berkeley however, all psychology students are required to take the more in depth, mathematical statistics course. In fact, Berkeley is the only public university in California that requires an upper level stats course for psychology majors. And I think Berkeley is doing it right. Amid the sea of liberal art majors, Berkeley sticks to its roots in research. Berkeley has nearly 30 Nobel prize winners and is also one of the best in colleges for engineering, physics, computer science, and much more "hard" sciences. Therefore, the psychology degree at Berkeley has a much more balanced approach than other universities, focusing on science, math, and social issues together.
While I'll still be the butt of many liberal arts degree jokes, people tend to forget that to fully comprehend, communicate, test, and perform any type of science, you must first learn a lot of liberal arts like history, writing, language, communication, psychology, and of course math -- another liberal art.
Hey Matt! What does this all have to do with statistics! Well, I'll get to that. But first what is statistics?
75% of statistics is spent learning about and correcting for errors. Imagine trying to test the entire population of America. 213 million people would have to participate and give consent each time we want to know something about human behavior. This would be impossible. Even our census, which is a simple survey done every four years, costs millions of dollars and is far from perfect. If we have trouble with a simple survey, imagine the problems we'd have with more complex research. So, in these cases we need to test a smaller, more contained group of people. But since we are not testing everyone, errors can arise from assumptions we make about a smaller sample versus the larger population. Statistics account for these errors in a variety of ways: random sampling and random assignment, single and double-blind studies, and making sure the data and the mechanics of measurement are reliable, among others. Mathematically speaking, ideas such as standard deviations, standard errors, means, medians, modes, distributions, probability, and much more, introduce ways to correct for errors when running statistical tests. Both of these things, proper sampling and proper error correction, account for the majority of what stats classes are all about. The rest of the class, maybe 10-15%, is actually doing research.
There seems to be two ways to teach a stats class. The first way is to simply teach students how to identify which formulas go with each type of test. This is designed to allow students to focus more on the research than the mathematics. The second way is to provide students with a much more in depth, mathematical heavy view, requiring students understand the genesis of formulas. Learning the logic behind formulas may give students an advantage later in life but requires much more time and energy. I've been part of both techniques and many students probably do not need such an in depth view and would do fine with the simpler way to teach the class. At Berkeley however, all psychology students are required to take the more in depth, mathematical statistics course. In fact, Berkeley is the only public university in California that requires an upper level stats course for psychology majors. And I think Berkeley is doing it right. Amid the sea of liberal art majors, Berkeley sticks to its roots in research. Berkeley has nearly 30 Nobel prize winners and is also one of the best in colleges for engineering, physics, computer science, and much more "hard" sciences. Therefore, the psychology degree at Berkeley has a much more balanced approach than other universities, focusing on science, math, and social issues together.
While I'll still be the butt of many liberal arts degree jokes, people tend to forget that to fully comprehend, communicate, test, and perform any type of science, you must first learn a lot of liberal arts like history, writing, language, communication, psychology, and of course math -- another liberal art.
Comments
Post a Comment