Share article
Disclosure: To support our site, Class Central may be compensated by some course providers.

A Passion for Stats and the Beauty of Bayes: A Duke Professor on Teaching Online

Written by Charlie Chung
5 minute read
0 Comments

Statistics and data analysis is a hot field, and “Data Scientist” has been called the sexiest job of the 21st Century. Dr. Mine Cetinkaya-Rundel is an Assistant Professor of the Practice in the department of Statistical Science at Duke University. When Mine started teaching Data Analysis and Statistical Inference on the Coursera platform, she had over 80,000 registrations for her first session, and there have now accrued over 220,000 registrations. Mine attributes the surge in interest in statistics to the dramatic increase in the availability of data:

“There’s data everywhere and it is now really easy and cheap to record that data, and people are realizing more and more that you can make sense of this data.” 

Mine is clearly passionate about statistics and educating people about it. She grew up in Turkey, where there was an emphasis on math and statistics in high school. When she came to the U.S., she saw that most students aren’t exposed to very much stats in high school: “the way we’re teaching stats in the U.S. is only at the college level, and then it is kind of dry.

But stats is very important in our society, and not just for computer science. An example Mine points to is Bayes Theorem, which can be challenging to explain (see, for example, the counter-intuitive ‘Monty Hall Problem’ that caused a stir in Parade Magazine). This concept is important in decision-making, for example in understanding the chances you have a disease when you have a positive test result (this is where understanding p-values also comes in handy). Mine has been working on an open-source statistics textbook project, and was in the process of recording her own videos when he opportunity to produce a MOOC came up. As it turned out, the impact of creating the MOOC on her teaching was much larger than she would have thought.

 

Class Content Gets an Internet-driven Overhaul

When Mine started to design her online course and thought about how she would arrange her content into modules, it forced her to re-think her instructional approach completely:

“The process of creating this online course makes you really follow the principles of sound instructional design in a way that you did not have to before” 

As a result, Mine had an explicit list of learning objectives that all of the content mapped to. It also provides solid grounding for all of the assessments “to tie every single assessment item into one of the learning objectives”. And the results have been positive. Because the outline of the course matches the content and assignments so clearly, learners understand exactly where they are and what they need to learn every step along the way.

The Many Faces of the Statistically Curious

Among the 220,000 enrollees in the course are high school students and people with PhD’s. Mine was surprised that people with advanced degrees were in her course, but found that many wanted to learn more about statistics or brush up on their stats knowledge. And also a surprising fact: Mine has been advising two high school students in Missouri who contacted her through her course! Mine relays the story:

“They emailed me saying: we took your course, we learned some statistics, we learned some R, we’re trying to do an experiment. Can you help us a little bit?” 

Mine agreed and has been providing some quick feedback through periodic emails. The two students conducted an experiment to test the impact of accents on people’s perceptions by showing recorded videos to participants and collecting their reactions. What a fantastic example of finding a mini-thesis advisor through a MOOC when you are not even in college, let alone the college the professors is at!

We Can’t Just be ‘Sheldons’

Besides covering the fundamentals of statistics, Mine’s course includes a project where students can select their own data sets and perform some real-world analysis of their own choosing. This is important, because as passionate Mine is about statistics, she recognizes this is just part of the requirement for being a good data analyst:

“Doing statistics as three parts: one part is mathematics, one part is computation, and the other part is actually being able to communicate your results” 

Thus, we can’t be effective by being like Sheldon Cooper in Big Bang Theory who is brilliant, but lacks the communication skills to interact well with others. We need to be able to form an argument, apply analysis to back it up, and then craft a compelling story. For this reason Duke set up an unusual Specialization on Coursera, Reasoning, Data Analysis, and Writing. In addition to Mine’s course, it also includes the popular Think Again: How to Reason and Argue, and English Composition I: Achieving Expertise, as well as a Capstone Project.

Most people will probably understand the value of this combination of subjects, but will the credential be meaningful to organizations or employers? Mine is not sure herself, but does point to a key value of pursuing a credential that is psychological: “it seems to be one of the things that make people stick to an area of study”. At some point, maybe someone will analyze the economic benefit of earning this credential–and maybe that person will have developed their analysis skills in Mine’s course!

You can still sign up for the Data Analysis and Statistical Inference, which is ongoing through May 10.