Journalism Schools, relevance vs. cost
In the United States, J-Schools are too expensive when taking into account the economics of the profession. By and large, teaching media skills needs an urgent reinvention.
[This version has been updated wit some precisions in dates and numbers]
Should students pay $30,000 or even up to $100,000 to attend a journalism school in the United States? Is it realistic considering the starting salaries in the profession? Is teaching journalism schools actually needed?
An interesting discussion was prompted last week by Rafat Ali, about the cost of tuition in journalism schools in the United States. Rafat is a creator of Skift and a pioneer in digital journalism and media entrepreneurship. On February 21, he tweeted the following:
Anita Zielina, Rafat refers to, is director of innovation and leadership at the Craig Newmark Graduate School of Journalism at CUNY (and a former JSK Fellow at Stanford, class of 2012). The table is lifted from the tuition comparison calculator of the Craig Newmark school) which sums up the cost of some of the best J-schools in America:
UPDATE 03.08: CUNY has since corrected the table below that displayed a cost of $94,176 for Arizona State University. According to ASU, “The costs are for a three-semester program (Fall/Spring/Fall) taking at least 12 credit hours per semester. The tuition-only cost, as requested, is a total of $17,580 for Arizona residents and $47,100 for non-residents. These costs are based on academic year 2019–2020 rates.”
The table below reflects the correction:
On average, a student will pay $54,000 per year to attend one of these J-schools (living expenses not included). Compared to that, at $16,500 per year, the SciencesPo Journalism school where I teach in Paris looks like a bargain (students actually get generous aid: only 2% pay the full rate and 35% of the SciencesPo J-School pay zero tuition). For that price, you get the best journalism school in France, with some of the classes taught in English, and the curriculum also offers a dual degree with the Paris School of International Affairs — end of our commercial break.
During my years as a JSK Fellow, some of my classmates at the Stanford Graduate School of Business incurred $100k or $150,000 in student loans. When asked about it, most of them were not utterly concerned given the expected starting salary for the Stanford MBA: $200,000+ a year. For 2019, it amounted to $216,000, including a base salary of $152,000, a signing bonus of $28,000 and an expected performance bonus of $66,000.
That’s for the stellar word of MBAs.
Back to Earth with journalism salaries. According to Glassdoor, it looks like this:
Yep, an American journalist makes almost five times less than a junior Stanford MBA hitting the job market. Worth mentioning also is the rise of freelancing. Many Columbia graduates won’t even get a real contract for several years and will be either pure freelance or “permalance” where they work a regular gig but without the security or benefits.
It’s a double whammy for a young journalist attending Arizona State or Columbia: they will pay the equivalent of an MBA or a law school tuition for a salary that will make the burden of their student loan almost unbearable.
That’s my first point: top universities are making a huge margin on tuition and they should better align their fees with expected salaries of their students.
There is room for it. Look at this chart: in ten years, higher education costs grew at a rate of 184 percent vs cumulative inflation of 56 percent:
The longer the period, the wider the gap between inflation and tuition fees: between 1970 to 2018, the Consumer Price Index in the United States has been multiplied by 6.5x. During the same period, the average cost of higher education grew 19x for private institutions (to $30,282/year) and 22x for public education ( to $7054), according to the Bureau of Labor Statistics and the National Center for Education Statistics. Again, we are talking about averages here. No wonder why outstanding student loans have soared to $1.41 trillion in 2019, a 33% increase since 2014.
There are several consequences: first, journalism curriculums are less likely to attract people as they see their future financial condition — high cost of education met by low salary — degrading. Two, they are less likely to in journalism as the trade mingles with multiple commercial streams: production of branded content, all forms of “corporate journalism” and the various flavors of communication, all of them hungry for genuine editorial talent. As a senior business writer who crossed that Rubicon told me once: “Look, I’m 45, I have worked at Reuters and The Economist, do you know how much I need to live in Brooklyn with two kids? Now I’m doing fine, I have more time and more access for my reporting… But yeah, I can’t bash the client…”
Now let’s talk about how relevant it is today to join a journalism school. I still believe that it is necessary to give to those who will collect and shape information proper training and methods. I’m writing this from France where we are still squabbling with the Yellow Vest movement, which recently morphed into violent opposition to the pension reform. One of the key features of this populist unrest is the emergence of a new self-labeled “journalist” category, halfway between political activists and agitators, a banner in one hand a GoPro in the other. They present themselves as authentic reporters, bringing up an ethic that knowingly blends journalism and activism. In such a context, I think journalists taught with all the fundamentals of the trade are more necessary than ever.
Having said that, I think journalism teaching is in need of a profound change on two counts. One is bringing more expertise into the profession by making it attractive for people who have already accumulated a solid experience in a field that is particularly hard to cover (economics, science, law, medicine, you name it). Two, I don’t think organizing a curriculum around one or two years makes sense anymore. In the next decades, the media market will call for numerous qualifications that no one can even fathom today. New forms of teaching would, therefore, encompass flexible, lifelong-learning modules, certified “nano-degrees” corresponding to skills in demand at a given moment of a career, taught online or on short bursts of in-class courses, and fostering communities of learners.
My prediction is that in less than five years, the majority of journalism teaching will have made the switch to life-long, segmented learning. That future makes way more sense than asking someone to commit for a $60,000 or $100,000 tuition bill to be paid back with a $50,000 a year salary.
— frederic.filloux@mondaynote.com
Ask Deepnews (part 3): “How far can you go in terms of accuracy of your deep learning model?”
[Here, we reply to readers and users questions, a few weeks after the launch of Deepnews.ai’s professional newsletter system (the Distills).]
— The short answer is: we will reach an asymptote that will be below absolute accuracy. I can’t think of a Natural Language deep learning model that is 100 percent accurate.
The way we currently measure the accuracy of our main model (the Deepnews Scoring Model or DSM), is comparing the score of an article returned by the DSM to three evaluations made by humans. We did it for about 10,000 articles, a process that was important for the calibration of our model. But taking into account the law of diminishing returns, we would need to do many more comparisons to jump from, say, 86 percent to 90 percent accuracy.
Eventually, though we will. It would be great to have a batch of 100,000 or 300,000 articles tested in parallel by humans and the machine. Ideally in several languages… But it is a costly process (we need to pay qualified people to assess the stories, it can’t be done by Mechanical Turks), and we could easily consume half a million dollars in doing just that.
We also have so much on our plate right now, like testing new breeds of model, trying to understand what the neural networks “sees” in a story, transferring the DSM to other languages and the uncertainties that go along with that, adapting the Deepnews algorithm to specialized content, launching a new series of Distills in a few weeks, plus maybe doing some bespoke verticals for corporate customers as described last week.
2020 will be a busy year.
