Automated media

In the old days, the first thing an aspiring journalists learned about the news business was that the business model was about “delivering eyeballs to advertisers.” Journalism historians tell us that advertising revenue facilitated the turn from partisan news media and pamphleteering to independent news outlets staffed by people who became professionalized as journalists with norms, ethics and standards of practice. The growth of advertising-financed news paralleled the growth of the industrial-era consumer economy and made news a viable business. The staff-and-line structure of industrial organizations made it easy and logical to separate the editorial and business functions of the newspaper, freeing journalists to focus on reporting without necessarily having to think about the business consequences – the so-called “wall between church and state.” The digital turn in journalism has, of course, destroyed that wall, but it’s done much more – it’s changed the business logic behind paying for content. Mark Andrejevic and his colleagues at Monash University are helpful here:

Since the mass media has few technological mechanisms for targeting specific groups of people, advertisers developed very rough proxies – concentrating ads for household products, for example, during daytime hours to reach homemakers (hence the term “soap opera”); or placing toy ads alongside Saturday morning cartoons.

The ads followed the content, and, in some cases, its timing and geography. These ads were available to large groups of people, and thus available for public scrutiny – and often became the topic of concern about stereotyping and predatory marketing tactics.

Going Dark: Holding platforms to account over targeted online advertising, July 5, 2021

As Andrejevic notes in his 2019 book, Automated Media, digital publishing platforms — including, but not limited to social media — give advertisers a way of targeting and surveilling consumers directly, just as they give governments and other powerful actors ways of shaping public opinion without debating openly in the public square. Instead, databases, machine learning and AI technologies foster opportunities to create media environments in which people can construct their own shared meanings and notions of epistemic authority. We see this in the power of hashtag activism and the infodemics of baseless conspiracy theories that endanger both public health and the viability of democratic regimes around the world. What Andrejevic gives us in this book is a way of understanding how the logic of the automated systems themselves — the machine-to-machine communication — helps to undermine the Enlightenment-era assumptions about the ability of reason and evidence win out in the marketplace of ideas.

Here’s how Andrejevic puts it:

Whereas automated machinery offloaded the social labor of production onto mechanized infrastructures, automated media seek to offload culture itself onto artificial intelligence and data-driven forms of social sorting and decision-making. The result is what might be described as an ongoing process of social de-skilling accompanied by the dis-embedding of key decision-making processes from the forms of social life and social interaction upon which they rely. The attempt to abstract core elements of human culture from the realm of social interaction (by offloading them onto automated information systems) makes it easier to misrecognize and ignore the underlying forms of social interdependence and recognition that enable the formation of shared or common interests and understandings. This social de-skilling is the result of what the chapter describes as the “cascading” logic of automation that characterizes the contemporary information environment: automated information collection generates so much information that only automated systems can meaningfully organize it. Once sense-making becomes automated, the next logical step is toward automated response, which, in turn, promises to surpass the capacities of human subjects. If automated machinery displaced human labor, automation targets the figure of the subject. 

Mark Andrejevic, Automated Media. Introduction

Andrejevic discusses the implications of this in some detail in this New Books interview:

I’m going to be spending some time with Andrejevic’s work. I think it has a lot to teach me as I grapple with the questions I raised in my February, 2021 blog post that asked what journalism educators need to know in the era of the fourth industrial revolution. I suspect I will want to put it in conversation with Ramesh Srinavasan’s “Beyond the Valley.

Here’s my 2021 blog post, if you’re curious.

Toward a more perfect union: the case for culturally responsive computational journalism

The slides below are from a presentation I gave today as this semester’s Faculty Senate Colloquium lecturer at The College of New Jersey. To be chosen by one’s peers to deliver such a research talk is a singular honor. I am particularly grateful to my English department colleague, the distinguished scholar and pundit Cassandra Jackson, whose introduction made me sound like someone I’d like to meet.

Here is the presentation abstract:

I moved from industry into academia 25 years ago because I had come to an understanding that the “hollowing-out” and flattening, of corporate, political and cultural hierarchies would make the role of professional communicators more central to the effective functioning of businesses and communities. As the expansion of the Internet and online technologies upended the news and communication industries, I became increasingly engaged with understanding how professional communicators could adapt to these seismic changes. This ultimately led to my current research in the development of culturally responsive models for teaching and practicing computational journalism. In this talk, I will draw upon that research to articulate a vision for a culturally responsive journalism. I will argue that culturally responsive computational journalism is essential to realizing the constructive potential of the seismic changes that computer science has visited upon the news industry. Properly crafted and implemented, culturally responsive journalism could:

1. Create an inclusive epistemology of journalism that moves beyond naive empiricism and the current propagandistic journalism of assertion
2. Democratize access to media technologies by broadening participation in the development and deployment of civic media
3. Deepen and broaden critical user engagement with the news
4. Deepen and broaden civic engagement
Computing technology and networks afford almost everyone the opportunity to be a publisher, but they also reward those who are computationally fluent with superior access to the public square. For this reason, I envision a future in which broad application and refinement the pedagogical models being developed here and elsewhere can actually empower citizens and strengthen democracy.

Here are links to sources for the presentation:

“Newspaper Newsroom Workforce Continues to Drop.”  Pew Research Journalism Project. March 20, 2014

Broadband technology fact sheet.” Pew Research Internet Project.

Computer and Internet Use 1984-2012 US Census

Closing the Digital Divide: Latinos and Technology Adoption Pew Research Hispanic Trends Project

The State of Digital Divides. Pew Internet Research Project. Nov. 5,2013

The Digital Divide is Still Leaving Americans Behind.” Jessica Goodman,  Mashable,  August 20, 2013

Yahoo Latest Tech Icon to Reveal Lack of Diversity.” Jessica Guynn, USA Today, August 15, 2014

Interactive Journalism Institute for Middle Schoolers

CABECT research website

CABECT in a nutshell (flyer describing the project, with some preliminary data)

As we change journalism education, we need to study journalism learners

After years of exhortation and industry convulsions, journalism education is changing. The argument for infusing digital  media education – even programming — into the journalism curriculum is over. The questions are mostly logistical – what type, in what sequence, how much and to what ends? Driven largely by business needs, college newspapers are becoming sites of experimentation with new business and management models. Professional news organizations are expanding their relationships with journalism schools beyond their traditional roles as providers of internships and first employers. In some cases, they are collaborating on beat coverage and special investigations. In at least one instance, the local professional news outlets have physically moved on campus.

At the graduate level, Medill’s Innovation program helped spawn Narrative Science, a company that programs robots to generate stories. We faculty at small programs, who have thinking through what these changes mean for institutions like ours, finally have our own journal, Teaching Journalism and Mass Communications. The 2013 edition of Georgia Tech’s groundbreaking Computation + Journalism Symposium will likely drive the conversation even further.

All signs of progress, but something important is being lost amid the frenzy.

As former President George W. Bush famously put it, “Rarely is the question asked, ‘Is our children learning?'” Mindy McAdams speaks for many of us who have spent years looking for ways to infuse digital skills into the journalism curriculum:

“We can offer a course that focuses on Web technologies — HTML, CSS, JavaScript, etc. But there is no data journalism in that class. And a lot of the students are going to hate typing those little brackets and so on. They’ll be so happy when that course is done and they never have to do that again.

“Moreover, they won’t practice what they learned, and very soon, they will forget all of it.

“We can offer a course about scraping and doing stuff with large data sets. We can teach students how to find stories in data. Students who like this, who learn how to do it and want to continue doing it, are probably among those most likely to get a journalism job. Like the Web technologies course, though, this is a class that many students will either avoid like the plague or take and then count the minutes until it’s over.”

Please, please read the whole post. She points to a real challenge that we haven’t yet cracked: how to engage students who think that journalism is about writing, not math or technology. Students who have convinced themselves that writing is something they are inherently “good” at, while math and tech are something they are inherently “bad” at. Students who don’t see why they need to understand html when they can just use a wysiwyg platform to build a website.

And my colleague and friend Michelle Johnson adds another layer: too often, the students who are least successful in adapting to journalism’s digital evolution are students of color, apparently another manifestation of the racial achievement gap. She writes:

“[F]or the past 20 years, I’ve read literally hundreds of applications for journalism training programs and scholarships, as well as for admission to journalism school. And sadly, I’m seeing some troubling signs.
“This isn’t just hand-wringing about a decline in writing skills among young people with short attention spans who communicate via texting abbreviations — I’ve noticed that among all the students.
“Simply put, I’m seeing that many of the students of color lack experience with the tools and technologies that will be fundamental to journalism innovation going forward. And this comes at a time when funding for training programs for students of color has shrunk, along with the bottom lines of the news industry and professional associations.”

These are exactly the concerns that keep me awake at night, even as I champion interactive journalism as a way of bringing members of under-represented groups into computing fields. (I’d also add working-class students to Michelle’s list, by the way.)

I would submit that amid our frenzy to learn and then incorporate all the skills that our graduates need into our curricula, we need a better understanding of what students absorb, and what affects their sense of self-efficacy as they confront the unexpected skills and content we are asking them to learn. That’s part of what I’m hoping to better understand with the new research project that I’ve embarked upon with Dr. S. Monisha Pulimood, of TCNJ’s Computer Science Department. The formal title is TUES: Collaborating Across Boundaries to Engage Undergraduates in Computational Thinking.(NSF Award #1141170). As we state in our abstract:

“To adequately prepare a workforce for the changing economic and global landscape, the project is developing a model that enables students with diverse perspectives and disciplinary backgrounds to learn how to collaborate and integrate concepts from their respective fields to develop technology-based solutions for complex real-world problems.”

It’s a tall order that we’ve set ourselves, and we are grateful to have Diane Bates, our independent evaluator, on board to help us assess what we are doing.

I’ll share more specific information about our project as it develops, but for now, I want to share some specific questions that I’m working through about integrating computational thinking and integrate it into journalism classes.

What’s the right learning environment to support computational thinking in journalism?  One of the posts that I wrote for a 2010 series about my own early exposure to skills that are currently classed as computational thinking began with this prologue:

“There are, at least, two approaches to education: the mimetic approach and the mathetic approach. The mimetic approach emphasizes memorization and drill exercises and is most efficient in inculcating facts and developing basic skills [Gar89, p. 6]. The mathetic approach stresses learning by doing and self exploration; it encourages independent and creative thinking [Pap80, p. 120]. In the mimetic framework, creativity comes after the mastery of basic skills. On the other hand, proponents of the mathetic school believe that self discovery is the best, if not the only, way to learn…”

Educational Outlook,”

Sugih Jamin, Associate Professor, EECS, University of Michigan

Whether taught in a classroom or newsroom, journalism education tends to be mimetic, while approaches to engaging novices in computing tend to be mathetic. We introduce students to specific routines and rigors of reporting, emphasizing adherence to rules of attribution, AP style, divisions of genre and structure (hard news, features, inverted pyramids, nut grafs, and so on.)  We stress the importance of getting the story right the first time, and then admit that there will likely be corrections and emendations as a breaking news story develops. We do these things for good reason: flubbing the fundamentals can not only get a reporter fired, it can lead to lawsuits, or in extreme cases, endanger innocent lives and reputations. Consequently, journalism students and professionals learn to think of every thing they do in highly instrumental terms, especially when it comes to learning what they need to know to ensure that they will get or keep a job.

By contrast, programming environments for novices such as Scratch or Alice are very successful at making introductory programming concepts more accessible. However, their strategy for engaging learners emphasizes play in ways that can be off-putting to journalism students who feel a need to quickly learn how to assemble a professional product. In the past, I’ve used Scratch in two ways – as a first step in learning Flash (something I’ve abandoned since Adobe made Mindy McAdams’ Flash Journalism text obsolete, and experts such as Mark  Luckie began pooh-poohing it as an important skill for journalists.) I’ve had some success teaching Scratch in game design courses, and I may think about using Alice for this purpose in the future, since its most recent iteration is specifically designed to give students a leg up Java, and that can be useful to aspiring app developers.

Do we need a journalism-specific programming environment to engage novice journalism students?

There are other, more mimetic, web-based learning environments for learning to code, such as Udacity.com’s CS `101 course, which focuses on Python and teaches students how to build a web scraper. There is an appeal to that approach because it has students build something that has obvious practical use in journalism. However, that course is arguably vulnerable to the criticism made by Bret Victor of platforms such as Khan Academy and CodeAcademy – that is, that they emphasize rote skills, while programming is “a way of thinking.”

Might it make sense to create a hybrid learning environment that combines the low barriers to entry of Scratch or Alice, with the goal orientation of something like Udacity? Will we begin to succeed at teaching programming as a way of thinking if we can more closely articulate between these learning environments and our broader journalism education curricula? (Here I am speaking of curricula not only for the classroom, but also for professional training.) Will novice programmer journalists be more motivated to learn in an environment where they can see direct connections between what their growing computing knowledge, the specific journalism artifacts they are learning to create, and the marketable skills they are developing? If so, what is the best way to create these linkages?

Is learning scripting really a gateway to computational thinking? The notion that journalism students should learn to “code” has gained increasing acceptance, but what that means and how one learns to do it are not universally understood. For several years, I’ve taken a position similar to the one that Miranda Mulligan took in a September 5, 2012 essay for NiemanLab:

I am not arguing that every single writer/editor/publisher who learns some programming should end up becoming a software engineer or a refined web designer. The end goal here is not programming fluency. However, there’s a lot of value in understanding how browsers read and render our stories. Reporting and writing a story, writing some code (HTML, CSS, Javascript), and programming complex applications and services are all collections of skills. A fundamental knowledge of code allows for:

  • More significant conversations about digital presentation, ultimately leading to better, more meaningful, online storytelling. Understanding your medium makes you better at your craft.
  • Deeper thought and understanding of data. Learning more about what goes into writing and programming software teaches you to think in terms of abstractions, functions, parameters, components, frameworks, object classes, templates, and more.

What Mulligan is referring to here as code (html, css, javascript – or more likely, jquery) is not programming, but web scripting, and as Mindy McAdams noted earlier, doesn’t get students digging into data. Having taught html and css for several years in our Writing for Interactive Multimedia class, my TCNJ colleagues and I can attest to all of the challenges that McAdams cites.

But there may be an additional unexamined assumption here, that learning scripting leads to the kind of computational fluency that, as Mulligan puts it, “teaches you to think in terms of abstractions, functions, parameters, components…”  I would submit that we need data to support this hypothesis. I certainly agree with her intuitively, but we need to know. These are some of the things we hope to learn in our research project, but there is lots of good work to be done to understand what, if any correlations exist between learning to script and learning to think computationally about the creation of journalism artifacts.

What do we know about the success of CAR courses that teach Excel,  SPSS, Access and SQL? The one place in the journalism curriculum that has come closest to teaching something like computational thinking has been in Computer Assisted Reporting classes (which these days, of course, is arguably a redundant term.)  A syllabus repository for some of these courses is here. We’ve had a required CAR course at TCNJ for 10 years. Many of these classes required that students minimally learn to use Microsoft Excel and Access (something I required when I taught it in the early 2000s). Some also incorporated SPSS and SQL. I don’t know of anyone who has studied these courses to assess the degree to which they affect students’ computing efficacy, programming skill, or acquisition of computational thinking concepts such as abstraction, decomposition, data structures, etc.

We could also use some research on the viability of such classes as points of articulation with emerging computational journalism curricula in computer science. One hopeful example is the work done by my TCNJ colleagues Donna Shaw and Emilie Lounsberry on the development of a database manager, GUMSHOE, that tracked the  disposition of gun-related arrests through the Philadelphia courts, ultimately contributing to an award winning story package on endemic problems in the Philadelphia court system.

These are just some of the questions that I think could lead to fruitful education research. I have others, such as questions about the possible role of stereotype threat on the achievement gap issues that Michelle Johnson cited, and whether learning science might help us better illuminate the real gaps in understanding and engagement that have many of us classroom teachers worried. As I’ve learned from talking to learning scientist  Deborah Tatar, making assumptions about why whole groups of people aren’t grasping particular concepts is often a big mistake.

Much, much more to be learned. I’m hoping that what has been, until now, an understandably ad hoc and organic effort develops into an area of systematic study.