Materials

Have you used our book? We want to hear from you!

Please e-mail Sara Stoudt ()

Teaching Materials

Original Class (that inspired the book)

Taught by Deborah Nolan and Sara Stoudt at UC Berkeley

Syllabus

First Year Seminar: Storytelling with Data

Taught by Sara Stoudt at Bucknell University

Day-to-day overview Fall ’25: coming soon

Day-to-day overview Fall ’24

Day-to-day overview Fall ’22

More on Reading to Write

Nolan, Deborah, and Sara Stoudt. “Reading to write.” Significance 17.6 (2020).

More on Storyboarding

paper coming soon

Nolan, Deborah, and Sara Stoudt. “Storyboarding as Part of the Process of Statistical Investigation” USCOTS (2021)

More on Presentations

Stoudt, Sara. “Can TV make you a better stats communicator?.” Significance (2023).

More on Revision

Nolan, Deborah, and Sara Stoudt. “Revision Beyond the Copyedit: Encouraing Students to Reorganize, Restructure, and Refocus in Writing Assignments” ICOTS 11 (2022)

More on Captions

Nolan, Deborah, and Sara Stoudt. “Captions: The Unsung Heroes of Data Communication” IASE (2021)

More on Portfolios

Nolan, Deborah, and Sara Stoudt. “The promise of portfolios: Training modern data scientists.” Harvard Data Science Review 3.3 (2021).

Errata - p.206 Chapter 8

A rapid test for HIV was developed for its sensitivity: the new test was shown to have sensitivity of 83%, compared to 60% for a commonly used test. This increased detection rate matters because those with false negative results (HIV-positive individuals whose disease goes undetected by a test) are likely not to return for further testing. The specificity of the rapid test is 97%. For example in a population of 692,000 with 12,000 HIV-positive, we expect \((0.83*12,000 =) 9,960\) HIV-positive people to test positive and \(((1-0.97)*(692,000 - 12,000) =) 20,400\) HIV-negative people to test positive. That is, the test has a precision of only \((9,960/(20,400+9,960) =) 32.8\%\). However the rapid test can be immediately administered a second time for anyone with a positive result. With this two-step approach the precision increases to \(((12,000/69,2000*0.83^2)/((12,000/69,2000*0.83^2) + (680,000/692,000*0.03^2)) =) 93.1\%\).