Stabile Seminar

The empirical spine of investigative reporting

Syllabus     Weekly schedule     Tip sheets / handouts

Spring 2014, Tuesdays 9-noon

Instructor: Sarah Cohen, The New York Times
Office hours: At Columbia, 8:30-9am Tuesday; by phone most evenings 5-7pm. / 212.556.8027 (work) / 202.213.6980 (cell – no texts)

Teaching assistant: Rob Barry, Wall Street Journal

Course website:
Sign up for the Piazza discussion site at

Course description and objectives

This is a class in computer-assisted reporting, a branch of the data journalism movement. In previous incarnations of the class, Maurice Tamman of Reuters dubbed it the “empirical spine” of journalism in the tradition of Philip Meyer’s Precision Journalism. Whatever the name, the idea is the same: original reporting and analysis in the service of stories, particularly in the investigative and enterprise genres.

Using records in electronic form gives you a powerful way to document patterns no one else has seen or sought. It changes your interviews from fact-gathering exercises to explanations and illumination. It can help you confirm or disprove a tip, isolate anecdotes and mash up sources for exclusive stories. Best yet, a computer doesn’t complain when you ask new questions and only lies when you make a mistake or misunderstand.

By the end of the semester, you should be able to:


You should prepare for and participate in every class. Classroom sessions will usually be broken into three sections.

First we’ll discuss the assigned story or stories. Most weeks, students will assign the stories and lead the discussions. Next we’ll go over questions on the week's assignments. It’s best, though, to post your questions on the Piazza site as they arise. Other students can benefit from your questions and we can give you help and resources before it’s too late. The last portion will be a lab working on your story memo.

Most of the work for your story memo can be done in class. Out of class, you'll be required to review video lectures and do tutorials on skills. There will be a deliverable almost every week that shows mastery of the material and your ability to turn that skill into stories. It will usually be a data set and a 3-paragraph memo. This structure, while unusual, allows us to see when some topics are more difficult than we'd imagined and to give you one-on-one feedback during class. There will be some weeks in which some of you are already proficient in the skill. If so, it will take you less time but I will also post resources that you can use to get to the next level.

There are no excused absences except for emergencies, religious holidays and incapacitating illness. No accommodation will be made for any other absences or failure to complete your work. No late work will be accepted. Covering an event, attending a wedding and going on a job interview are scheduling choices that reflect your priorities. They may be rational choices, but they’re not emergencies. You will have one no-questions-asked missed class and one no-questions-asked missed deadline for the weekly assignment.

Requirements and evaluation

In-class work and weekly assignments: 60 percent. This will also reflect your overall engagement and participation in class and your work in the labs.

Story presentation and discussion: 15 percent. Prepare a presentation on a story that included a social science, computer-assisted reporting or other empirical element. It should include what data was used, how it was obtained, how it was analyzed and what it added to the story. The best of these will often include comments from the author or the person who did the analysis. You should prepare a 15-minute presentation and plan to lead a discussion for an additional 15 minutes. Do not use either of the instructors' stories for this assignment.

Final investigative story memo: 25 percent. Write an investigative story memo proposing a records or data-driven story idea. It should be well enough researched to know what records you will use, how you will get them and what they might show. You are encouraged to tie this project in with your master’s project or another story you are working on outside of class. The memo should include a coherent analysis of a dataset that you find on your own with at least 2,000 records or a dataset you build yourself using public records as source documents. Its nut graf should tell a reader why it is important and include a rough outline of the story that could emerge. Expect to do some shoe-leather and documents reporting and be prepared to make newsworthy comparisons. This will be assessed at each stage, not just for the final product.

Books and materials

All students:

There is a wide range of backgrounds in this class, so you should review these materials to see which ones will be helpful for you:

Over the course of the semester, you may decide you want to go further and buy some low-cost software packages that will make your job easier. Consider buying software while you're a student. It will get much more expensive once you graduate.