Rate My Professors gets 1 star for accuracy – The Appalachian

Every semester, college students eagerly review their course lists and study plans to create the perfect class schedule. A frequently used tool Rate My Professorsa website where students submit anonymous reviews of their professors, including their grades, class ratings, and additional comments. While this website may seem informative at first glance, it is plagued by several statistical issues that call into question the accuracy of these ratings.

Most departments will offer classes with two or three professors who may have drastically different teaching styles. However, students have no real way to gauge the best fit until syllabus week, when it is often too late to switch departments.

This is where Rate My Professors comes in. Since it integrates with a widely used scheduling app, Textbookstudents can view each professor’s overall rating as they browse through class options. While it’s tempting to draw conclusions about professors based solely on these online ratings, two statistical issues should make students think twice.

The first one is small sample sizes. Often professors, especially new ones, have a small number of ratings. This is a problem because a small sample size does not accurately capture the true range of possible outcomes. It’s like tossing a coin three times and concluding that it’s impossible to decide if it came up heads because they all came up heads.

Similarly, having just a few reviews from hundreds or thousands of people a professor has taught is not enough to create a representative sample of the opinions of all past students. This idea is law of large numbersIt is a statistical theorem stating that as the sample size increases, the sample mean will approach the population mean.

Simply put, a large enough sample size is needed for the average rating listed on Rate My Professors to approximate the true average rating of all students who have studied with that professor. A sample size of dozens of reviews may be large enough, but anything less than ten is certainly too small. However, even if the sample size is large enough, the sample must be randomly selected to draw accurate conclusions.

Statistical bias describes ways in which samples are not randomly selected. In the case of Rate My Professors, voluntary response prejudice That’s the main thing at play. This bias only includes a sample of people who choose to respond, and selects those with stronger opinions. Those who either admire or hate their professors will feel more compelled to leave a review, causing these extreme ratings to be overrepresented.

Meanwhile, students who don’t feel strongly about either way tend to be less motivated to comment, so their experiences are underrepresented. This can lead to particularly positive and negative comments being lumped together and perception being skewed to extremes.

However, there are still meaningful insights to be gleaned from the website. Students should look through the open-ended comments and tags section to see concrete facts about the professor’s teaching style, such as attendance policies or textbook requirements. Separating these comments from the opinions of disdainful former students is important for making educated class choices.

In the information age, it’s easy to be comfortable with information and anxious without it. The urge to tell the future is strong, and the urge to control it is even stronger. But the security blanket of others’ past experiences is more of a thinly veiled comfort than an accurate predictive model. Growing up as a college student means coming to terms with not knowing your future or your professors.