Last winter 21 Northwestern students worked together, and in 11 weeks they built five functioning journalism tools. About a third of them were journalism undergraduates, the rest were computer science students. Each project was designed so that those who used it could create, consume or distribute news and information. Like the students, the faculty members were an interdisciplinary team. And the pedagogical methods we used were as experimental as the projects the students worked on, including these:
NewsTube: Aggregates and displays relevant supplementary news and information for a user who is viewing a video from the C-SPAN archive. This supplementary information “pops up” alongside the video.
News Here: Knows what’s happening where you are. It automatically delivers the most recent news to any mobile browser for the on-the-go user looking to find out what’s going on around the corner.
HomeSpot: Highlights the best neighborhood for a person looking to move based on cost and the quality of life aspects that the user indicates matter most.
Living Stories: Conveys to online readers the emotional impact of an ongoing news story by highlighting the voices and reactions of the people involved through quotes layered over strong images.
Tweebune: Finds relevant news articles for a Twitter user based on people, places and things mentioned in tweets.
JourNet (currently not available to the public): Allows journalists to obtain statistical information for their stories through a platform that integrates both the writing and the research processes.
Meshing Technology with Journalism
This collaboration actually began three and a half years ago when I interviewed at Medill, Northwestern University’s journalism school, with a dean who was focused on the importance of evolving technology but unclear what its inclusion into a journalism curriculum would mean. We talked with a pair of Northwestern computer scientists about the work coming out of their lab—projects like auto-generated, animated movie reviews and a system that finds and presents related materials to viewers watching broadcast news reports. After we left, the dean asked: “What if we could bring tools like this to newsrooms?”
The projects we’d heard about were much less important than the learning opportunities they presented for journalists and computer scientists—from faculty teaching together to having students for these two disciplines collaborate to build digital tools. Thinking back on my own experience at Medill a decade earlier, I realized that the actual tools were not what mattered. At that time, I was working on the capstone project for my master’s in journalism. Led by Medill professor Richard Gordon, this was the journalism school’s first joint effort with computer scientists. Unfortunately, my project suffered from a communication barrier; it was too difficult for the students to understand one another and the learning goals were not clear enough for either side to continue at that time. Still, for me the rewards of this experience were many and lasting.
When I returned to teach at Medill in the fall of 2008, I knew that developing a closer relationship with the computer science field was critical. My first class, co-taught with Richard, was a capstone class that included two programmers who had received Knight Foundation grants toward pursuit of a master’s degree in journalism. They brought to the class an agile software development philosophy that emphasized building working pieces of software every week. The class project—a website prototype called NewsMixer (no longer active) that tested new ways for users to comment on news stories—meant they partnered with other journalism students to introduce Q. & A.’s into comments, create a space for digital Letters to the Editor, and force more succinct and thoughtful comments by limiting the number of characters available to commenters.
The programmers’ philosophy was a good fit with journalism and, therefore, with how the class functioned. We held few meetings, and the meeting “scrums” we did have were stand-up, focused affairs that included targeted weekly deliverables with team-wide reviews. Our way of working pushed us toward usability and testing, another key area of emphasis for my brand of human-centered media design. The project was so successful it was a runner-up in the Knight-Batten awards and our team of Brian Boyar and Ryan Mark went on to found the Chicago Tribune’s News Application team.
But Medill’s foundation-funded model of scholarships for seasoned programmers had its limitations. It depended on us finding the money and experienced software developers willing to make a career switch. I set out to look for other options; I devoted a quarter of a semester to exploring another direction for collaboration. Instead of bringing trained computer scientists to the university, I decided to look for some at Northwestern who would collaborate with Medill. I watched a computer science practicum without any formal instruction; students put into practice skills they learned in other classes. I could see how a combination of team-based learning, experimentation and presentation created a cycle of exploration and explanation.
Yet my own experiences convinced me that even in a practicum arrangement, we’d need to provide more structure if we were going to collaborate across disciplines. Most important was for us to create a setting in which students forged a shared language. The journalists had to know enough about computer science to know what was easy, hard or even possible. The computer science students needed to understand the goals of journalists, the ethics of the practice, and key tenets like narrative structure, news judgment, and AP style. Just creating interdisciplinary teams did not guarantee the kind of shared experience or exchange of information this effort would require.
Our 2011 winter class had six teams with three to five students on each one. The classes met for a total of three hours a week for about 10 weeks. Half of the class was spent with the students presenting their editorial and programmatic progress. We pushed each team to start showing a demo as quickly as possible.
The students were required to work together outside of class and had to show proof of their work during in-person project time, sharing with us photographs or anecdotes from their time together. They were encouraged to practice “paired programming,” which happens when two students work on one laptop writing code together. At the end of the quarter, they were required to document their progress, identify areas of future improvement, and present a working demo to the public.
Every team managed to complete its working demo.
Yet our learning goals go much deeper than proving that software can be built—and content brought into it—during 10 weeks of working in teams. The journalism students teach the computer science students about media, narrative and editorial judgment. The computer science students introduce logical thinking, programming languages, and software tools and libraries like entity detectors, natural language processors, and Web scrapers. Success comes when students forge a common language and an understanding as they experience what it means to conduct a true experiment with an uncertain outcome. They must function as a team with members being equally able to explain what they did, their challenges, and what they need to do next.
Assessing Our Progress
To test the impact of the class, and ones like it, I conduct a pre- and post-class survey. The surveys from this class produced interesting results, including these:
While the overall knowledge of computer science did not change markedly during the course for the class as a whole, at the start of the course seven students reported that they knew nothing about journalism and by the end of the quarter all of the students reported that they knew a little bit (or more).
Before the class began, three-quarters of the students reported they were “very interested” in journalism but by the end of the quarter that interest had shifted to “somewhat interested.” But the students who were “very interested” in journalism had risen more than 10 percent.
Some of the student comments that emerged in interviews, e-mails and student surveys include:
Thanks... for all your guidance that quarter. One of the best classes I've taken.
I found the weekly feedback each Tuesday incredibly useful in moving forward with the projects.
Good class, more work than expected but got a working project at the end and my team was really happy with what we accomplished. The professor does a great job at motivating you.
The success of the class is almost completely dependent on your group. It's likely that things will come together in the end, but things can get rocky if you don't have great group mechanics. Otherwise, it's great to see a project through from start to finish.
Good class! At first I was hesitant and annoyed with how the majority of the work is for the computer scientists but in the end it all worked out and realized that the structure of the class enables students to learn from other students and groups, and is a good approach for a joint major course.
When asked to list what they’d learned during the quarter, students reported:
How to collaborate (related to programming).
How essential active communication is to being productive from week to week.
How to set reasonable deadlines for myself.
How to deal with difficult group members.
Familiarity with computer science terms. Now have a context to talk about or at least understand basic operations.
Students also shared ideas about how to improve the class:
More research, less iterative development. Iterative development forces us to test theories quickly and not spend enough time formulating theories. The best balance is difficult to master.
Give journalism students more opportunities to learn things related to computer science (and, maybe, vice versa).
I would love it if you would add a question to the peer evaluation (after you ask “What was accomplished?”) like the following: List the tasks assigned to each member.
The professors can provide some great feedback, but more often their criticisms seem inconsistent from week to week. Professor Gilbert, though, is consistently a great source of feedback and helpful, reasonable critique.
Each group really needs to have an adequate number of people with the technical skills for implementation—particularly so on projects with heavy emphasis on functionality.
Over the last three years, this collaboration with computer science has been formalized in Northwestern’s Center for Innovation in Technology, Media and Journalism, launched with Northwestern’s engineering school. With a Knight Foundation grant, we pay a staff person to implement technologies such as those developed in classes for use by those working in news media.
Improving the students’ collaborative learning experience is something we think about constantly. Without knowing their skills and experience when they enroll each semester, our ability to create balanced, interdisciplinary teams remains challenging. Because this class is a practicum, and therefore lacks conventional lectures and lessons, it is important that students arrive with a certain level of skills and the practice in deploying them. It is not an environment in which to begin learning them.
Properly evaluating each student can be almost as difficult as arranging the teams. The commitment each brings to the team project is almost as critical as their skills. Even with our peer evaluations, students still feel they are being unfairly graded as individuals, as can often happen in working as a team member on a project. Involving professional news organizations can add other layers of complication. While clients offer more than a touch of reality, doing so can limit the students’ most creative thinking. But when the team chemistry works well—and the project excites them, the combination of different types of students can result in terrific collaborations and innovative outcomes. On occasion, these classes shape career paths as students enter with one direction in mind and leave heading in another.
Jeremy Gilbert is an assistant professor at Northwestern University's Medill School of Journalism, teaching media product design and digital innovation. He has directed award-winning, student-based digital projects and helped revamp the interactive curriculum. In addition to teaching, he is researching the future of mobile journalism and designs mobile, tablet and Web-based media applications.