Oral interviews

Background

One of the learning goals of this class is to:

effectively communicate statistical ideas and results, both verbally and in writing

Whether you choose to stay in academia or pursue a career in industry, the ability to communicate clearly is of paramount importance. As a data scientist, the burden of proof is on you to convince your audience that what you are saying is true. If your audience (who may very well be less knowledgeable about statistics than you are) cannot understand your results or their interpretations, then the technical merit of your project is irrelevant.

We are well aware that AI tools can likely complete much or all of this project to a satisfactory degree. However, we are not interested in that—we are interested in your ability to discuss your findings in reasonable depth. While you are allowed to use AI tools to prepare your project, we don’t recommend them, and we will conduct an oral interviews so that we can assess your proficiency in an AI-free environment.

Procedure

Important

Pick a slot on my calendar!!!

You will prepare a Quarto document that contains (at least) the following:

  1. one data graphic or other visual display from your project: How the graphic was created is not so important – the questions are going to be about what the data graphic means. So make sure that you understand how the graphic you chose relates to the project.
  2. one other technical thing that you contributed to the project: For this, it is important that you are accountable for the work. Whether you wrote the code from scratch, or prompted AI for it, you are accountable for it, and so you need to be prepared to answer questions about what it’s doing, why you chose to write it that way, etc.

Do NOT make slides! You may simply use your draft of the final paper. You will hook up one of your laptops to the computer in McConnell 213 and project it for us to see.

Be prepared to discuss your project and your findings orally with us. We will ask you questions that force you to think on-the-fly. The questions will be straightforward—we’re not trying to test your knowledge of esoterica.

Some of our favorite questions are:

  • What does that number mean (in the context of the problem)?
  • What would happen if you changed this to that?
  • Why did you choose to do this instead of that?

Advice

  • DON’T over-prepare. If you actually did the project and learned the material in the class, you shouldn’t have any trouble answering our questions.
  • DON’T try to filibuster us by delivering a long, set of prepared remarks about your data set in an attempt to run out the clock. This is not a speech.
  • DON’T try to read from a set of notes, a draft of your paper, or other printed materials. Don’t try to read from the Quarto document on the screen, either. We know you can read—we want to see you think on your feet.
  • DON’T try to read from other laptops, phones, tablets, or other devices. Be fully present. You will be surprised by how much you remember!
  • DON’T try to shift accountability for part (2) to one of your teammates or AI. “I didn’t work on that part” or “AI wrote the code for me” are not satisfactory answers (if even true!).

Grading and rubic

Due to their complexity, oral interviews are coarsely graded:

Oral interviews rubric
100 90 80 50
Excellent. I’m totally convinced that you know your stuff. Good. I’m convinced that you did your work earnestly and you know how your piece fits into the big picture. Maybe there were some details that were’t quite right, but nobody’s perfect. Keep up the good work. Not there yet. There are some big gaps in your understanding of what you’re doing and/or how it fits into the overall project. Your grasp of the material does not yet inpsire confidence. We have a problem. You don’t appear to have done any substantive work on this project and/or your grasp of what you’ve done is weak.