The great paywall tightening of 2017 continues. The New York Times said Friday that it will cut the number of free articles available to “most” non-subscribers each month from 10 to five, Bloomberg reported. The change is the most significant one the Times has made to its pay model since 2012, when it cut the number of monthly free articles from 20 to 10. (According to Bloomberg, “The Times may eventually offer a different number of free articles to non-subscribers based on how they arrive or their reading habits.”) Continue reading “The New York Times has halved its free monthly articles to 5, its most significant paywall change since 2012”
Hey students: Want to spend next summer working with Nieman Lab?
I’m very happy to say that we will again be one of the host organizations for the Google News Lab Fellowships. You can apply here, and the deadline is January 15. Here’s Google’s description: Continue reading “Students: Spend the summer working with Nieman Lab via the Google News Lab Fellowship”
Five years is a long time, especially in the media business. It was five years ago this week that Mark Thompson took on the top job at The New York Times Company. It was an enterprise still wobbling from the effects of the Great Recession, its new paywall only a year old. The Huffington Post was trumpeting that it had surpassed the Times in digital traffic — a recognition of Google’s market power and of Facebook’s emergence. The Times was a shrinking enterprise. It had shed revenues, profits, staff, and share price. It had also shed its previous CEO, Janet Robinson. Publisher Arthur Sulzberger’s pick of Thompson to replace her surprised many; despite having led the BBC’s ongoing transition to the increasingly digital world, Thompson had no publishing management experience. And he was a Brit, plucked out of London to head America’s flagship newspaper company.
It’s been five years since the daily Skimm email made itself part of the wakeup routine of — now — more than 6 million largely female readers. (Yes, that subscriber number has quadrupled since we wrote about the company a little over two years ago.) And it’s a year and a half since theSkimm app asked users to pay for convenience — not necessarily content. Now, theSkimm (henceforth: The Skimm)’s cofounders, Danielle Weisberg and Carly Zakin, have spread the company’s product reach even farther. Continue reading “With video and audio, The Skimm pushes further into the daily routines of its 6 million readers”
Democracy dies in dankness.
That’s not a typo in the Washington Post’s Reddit profile: The Washington Post account is an avid poster of some pretty good memes and gifs. It’s got jokes. It’s also a sharer of everything from polling stories to breaking national security stories to lifestyle columns to geeky features to fact-checks, and a facilitator of, and participant in, AMAs. The official publisher account has been live since April of this year, shortly after the platform began allowing public profiles, and appears to have broken through Reddit’s tough anti-brand, anti-paywall shell.
Over the summer, the Guardian Mobile Innovation Lab hinted at its next experiment: improving the experience of consuming news when offline. Now it’s revealed the trial product, a news app that incorporates location sharing, content and time customizations, and user data transparency — but is only available for the next few months.
The Lab introduced LabRdr — which can apparently be pronounced like “lab reader” or “Labrador” or another general squish of consonants — on Wednesday. Designed for the public transit commuter who may lose cell service on the subway, for example, and then be left with nothing to read, the app prepares a “package” of Guardian content based on the user’s previous reads in the app and the current stories of the morning or evening. It’s delivered twice a day via push alert, at the times the user has specified they’re commuting; each package contains an amount of content that the app determines will be readable within the duration of each user’s commute.
And yep, location factors in. “We’re experimenting with making offline news reading easier and more relevant, through automatic personalizations of your reading package based on signals like your interests or, possibly in the future, your location and what’s being read nearby,” Mobile Lab editor Sasha Koren noted in the Medium announcement:
LabRdr’s approach to offline news reading is experimental, and different from existing offline news apps in a few ways: Rather than give you all current stories on every topic, it delivers only a self-contained package of Guardian articles keyed to your interests, twice a day just in time for your commute, at times you can specify.
As you use it, it learns what you like to read and delivers you content keyed to your interests. (We’re setting aside important conversations about filter bubbles for now to learn something about personalization.) In addition, we show you how we use the data you share with us, in an effort to enhance trust through transparency…
What we’re looking to learn
What makes a good content recommendation system for news? A lot of the existing work about content recommendations are around e-commerce and we’re interested in what signals are particularly good for news organizations and news reading.
We’re also looking to gauge readers’ reactions to the utility of having a short package of news defined for them for a set period of time. Without the option to read a full spectrum of articles on many topics, will they feel better informed with those they do read, or have a sense of achievement at completing a few articles in a set?
As with all our experiments, we’ll report on what we learn in follow-up posts after the app has been running for a while and we’ve collected and analyzed data and reactions.
LabRdr isn’t the first attempt at improving offline news. Way back in 2012, reading apps News.me and Instapaper both endeavored to serve the offline reader and relied on location to do so, but News.me didn’t survive a Twitter API update. Other apps like Pocket or Evernote require the readers themselves to do the legwork of saving the content for later perusing, rather than having relevant material presented to them.
Another difference is that the Mobile Lab is making an effort to share the data it collects through LabRdr. In a section of the app called the Log, you can view the tracked reading and commute patterns. “The app is a really good first step for gathering information, using it in a respectful way, and seeing how people feel about that,” said Sarah Schmalbach, co-leader of the Mobile Lab with Koren and its senior product manager. She pointed out that readers might feel different about sharing personal information with a news organization than they do about sharing it with, say, Google Maps or Amazon.
“If we can deliver news in more contextually relevant moments, then [will] that content be more valuable to the user?” wondered Connor Jennings, the app’s developer, who came up with the idea during his own frustrating experience reading offline news during his commute.
The team hopes to share its findings about reader trust, habit formation, and more with news organizations; the Mobile Lab is funded by the Knight Foundation (disclosure: Nieman Lab also receives funding from Knight) to explore solutions for the mobile news experience. But its sample will likely be restricted to those who commute using public transit, rather than people who drive, bike, or walk to work.
“It’s pretty narrow. We’re not targeting people who don’t commute; we’re not targeting people who commute by car. There’s a whole range of people we’re not gearing this toward,” Koren acknowledged.
LabRdr provides a “targeted product until we get better and deeper insights,” Schmalbach said. “We’re confident that the audience is big enough to get a big read on this content.”
Like the Mobile Lab’s other experiments — such as real-time Guardian commentary on a U.S. presidential debate via push alert; live push notifications with the Wall Street Journal — LabRdr is a temporary project. It will be removed from the App Store (it’s iPhone-only) after a couple of months.
1/ Today we launch our most intricate experiment yet. LabRdr, an iOS app for offline reading is now in the App Store https://t.co/leQmyXqdoA
— Guardian Mobile Lab (@GdnMobileLab) October 25, 2017
2/ It’s designed with a mass transit commuter in mind, who has little to no cell or wifi and limited time to read.
— Guardian Mobile Lab (@GdnMobileLab) October 25, 2017
3/ Users get alerted to ‘packages’ of Guardian articles twice a day at times they determine. The packages geared to length of commute.
— Guardian Mobile Lab (@GdnMobileLab) October 25, 2017
4/ Packages will be customized over time, using topical interests for recommendations at first; possibly location & commute patterns later.
— Guardian Mobile Lab (@GdnMobileLab) October 25, 2017
5/ We’re also experimenting with alerts based on user’s location, and offering transparency into what data we’re collecting.
— Guardian Mobile Lab (@GdnMobileLab) October 25, 2017
6/ More detail, and background on what we’re hoping to learn and share –> https://t.co/KlNp7jC9QC
— Guardian Mobile Lab (@GdnMobileLab) October 25, 2017
7/ If you’re an iPhone user – and don’t commute behind a steering wheel – give it a try and let us know what you think! /end/
— Guardian Mobile Lab (@GdnMobileLab) October 25, 2017
Google, an advertising giant, has been making nice with news publishers by developing a series of tools they can use to more precisely attract and target paid subscribers. (It also ended the first-click-free policy this month, allowing subscription-based publishers to choose how many articles to show to readers for free without search-ranking consequence.)
Google’s nice comes at a small business price for any publishers who might want to use the planned subscription tools, but the details are still being ironed out with publishers.
“It will obviously come down to what we think that business relationship should be, but bottom line, I think [revenue sharing] will be exceedingly generous [to news publishers],” Google’s head of news Richard Gingras told the Financial Times on Sunday. “In our ad environment, the rev shares are 70 per cent-plus. The rev shares [for publishers] will be significantly more generous than that.” (Google’s AdSense offers around a 70-30 split for publishers who use it to place ads on their sites.)
Gingras made sure to distinguish Google’s tack from Facebook’s “walled garden” approach, telling the FT that “unlike other participants in the environment, we’re not trying to own the publisher. If there are cases where we do cause the subscription to happen, we don’t want to own the customer. None of this changes the marketplace economics, people will pay for what they value.”
That “other participant in the environment” on Friday formally announced its test of news subscriptions models within its Instant Articles format, through which it won’t take any cut of the revenue from subscription signups (the subscription transaction and payment processing will take place entirely on the publishers’ site). Facebook’s subscription tests are Android-only, as it’s been wrestling with Apple over the past few months over Apple’s default 30 percent cut of “in-app sales,” Recode reported.
— Greg Emerson (@emersongreg) October 20, 2017
— Tony Haile (@arctictony) October 22, 2017
— Benjamin Braun (@BJMbraun) October 23, 2017
A new era of personalized news products began, in earnest, as a reaction to horrific global news.
Today, a Google search for news runs through the same algorithmic filtration system as any other Google search: A person’s individual search history, geographic location, and other demographic information affects what Google shows you. Exactly how your search results differ from any other person’s is a mystery, however. Not even the computer scientists who developed the algorithm could precisely reverse engineer it, given the fact that the same result can be achieved through numerous paths, and that ranking factors — deciding which results show up first — are constantly changing, as are the algorithms themselves.
We now get our news in real time, on-demand, tailored to our interests, across multiple platforms, without knowing just how much is actually personalized. It was technology companies like Google and Facebook, not traditional newsrooms, that made it so. But news organizations are increasingly betting that offering personalized content can help them draw audiences to their sites—and keep them coming back.
Personalization extends beyond how and where news organizations meet their readers. Already, smartphone users can subscribe to push notifications for the specific coverage areas that interest them. On Facebook, users can decide — to some extent — which organizations’ stories they would like to appear in their news feeds. At the same time, devices and platforms that use machine-learning to get to know their users will increasingly play a role in shaping ultra-personalized news products. Meanwhile, voice-activated artificially intelligent devices, such as Google Home and Amazon Echo, are poised to redefine the relationship between news consumers and the news.
While news personalization can help people manage information overload by making individuals’ news diets unique, it also threatens to incite filter bubbles and, in turn, bias. This “creates a bit of an echo chamber,” says Judith Donath, author of The Social Machine: Designs for Living Online and a researcher affiliated with Harvard’s Berkman Klein Center for Internet & Society. “You get news that is designed to be palatable to you. It feeds into people’s appetite of expecting the news to be entertaining … [and] the desire to have news that’s reinforcing your beliefs, as opposed to teaching you about what’s happening in the world and helping you predict the future better.”
As data-tracking becomes more sophisticated, voice recognition software advances, and tech companies leverage personalization for profit, personalization will only become more acute. This is potentially alarming given the growth of websites — news-oriented and otherwise —inhabiting the political extremes, which on Facebook are easy to mistake for valid sources of news. When users can customize their news, and customize to these political and social extremes, civic discourse can suffer. “What’s important is how people use the news to have a discussion,” says Donath. “You may have friends or colleagues, and you read the same things in common. You may decide different things about it. Then you debate with those people. If you’re not even seeing the same news story, it leaves you with a much narrower set of people with whom you share that common ground. You’re losing the common ground of news.”
Quartz got some money from Knight last year to launch its own Bot Studio, building interactive tools/chat interfaces/general bot substrate for both itself and other newsrooms. (More here and here.) Today, Quartz announces the latest fruit of that effort — a Slack bot named Quackbot, built in collaboration with DocumentCloud:
Together we’re releasing Quackbot, which performs tasks useful to reporters, editors, and news producers right where so many of us work all day — inside Slack. In its first version, Quackbot can do a select few tricks that might prove handy in a modern newsroom…But we’re excited to collaborate with the rest of the journalism world to give Quackbot many more skills over time. Think of it as a fully hosted and friendly interface to open-source tools…
Journalist-programmers are an especially sharing lot. Sure, they’ll work night and day to scoop each other, but once the story’s published they’re happy to share how they did it — even sharing the tools they built. As a result, there are many dozens of useful tools available to programmers in newsrooms everywhere.
But there’s a catch: Not every newsroom has programmers. And even existing programmers might not have the time, skills, or resources to get a project’s code, put it on a server, and keep it working.
It’s in an early state, but a few of those launch features might still be useful to you:
1. It can take a screenshot of any webpage.
2. It will preserve any URL by telling the Internet Archive to save a copy of the page.
3. Given a topic, it can suggest some reliable sources of data.
4. If you provide Quackbot with a URL, it will identify any cringe-worthy clichés on that page.
Soon, Quackbot will also allow journalists to upload PDFs to DocumentCloud, extract text and charts from PDFs, monitor websites for changes, make quick charts, and more. We’re also inviting other journalists to bring their tools into Quackbot, making them readily available within Slack.
Once it achieves anatine maturity, “Quackbot will become a core feature of DocumentCloud, which will maintain the infrastructure and provide troubleshooting and support.”
The bot itself will be available for install this Thursday (such teases), and you can bug John Keefe about it at ONA.
“Six years is a long time in product,” said Martin Fallon, the Financial Times’ product manager for apps. Six years ago was the last time that the FT’s main app was in Apple’s App Store. In 2011, the company introduced a web app and, a couple months later, pulled its dedicated iOS apps because, reportedly, it did not want to give Apple a 30 percent cut of in-app subscription revenue and wanted more information about subscribers than Apple was willing to provide.
More than half a decade later, you can find the FT app in the App Store once again (alongside some of the company’s other apps, like FTChinese, that never left.) Apple will not, however, be getting a cut of subscription revenues; as The Wall Street Journal reported Monday:
The new iOS app will therefore only be accessible to existing FT subscribers. New readers won’t be able to purchase subscriptions from within the app itself, but must instead do so from the FT’s website before logging in.
This model means the FT can avoid giving Apple a cut of subscription revenue and will allow it to collect payment information and other valuable data directly from its subscribers. Spotify and other subscription-based services have taken a similar approach in recent years.
“We identified an iOS app as a key way to drive engagement,” Fallon told me. “We saw that readers who used our existing apps were much more engaged than those who didn’t. We also saw that we had a much higher app adoption rate on Android, where we had a native app, than on iOS.” (Engagement became a bigger part of the discussion in 2015, when the FT created an audience engagement team.) He also mentioned other benefits of an iOS app over a web app: Improved offline reading, persistent login, easier sharing, and — ultimately — push notifications and automatic update downloads, things that readers have been asking for but that weren’t possible with the web app.
“Another motivation is simple — customers asked us for it,” Fallon said.
Right now, the FT’s iOS app is aimed only at existing subscribers. When you open it, you’re prompted to sign in; if you don’t, you can see a homepage but can’t read a single story. The web app will continue to exist for now, but the FT intends to move the majority of its readers over to the iOS app. (As of Monday, I couldn’t find anything on the FT’s site to alert readers to the existence of the iOS app; it was still only promoting the web app.)
The Financial Times has nearly 870,000 total paying subscribers (up 9 percent from this time last year); of those, 666,000 are digital-only subscriptions (up 13 percent from this time last year). More than 50 percent of the FT’s digital subscribers already use its apps, and with the launch of the iOS app, that percentage is expected to increase.
Newsletter > Apple News > podcast > app: In terms of how digital-native news outlets get their information out, the newsletter wins. That’s according to a digital news fact sheet from Pew Research Center, released Monday. It looks at 36 news outlets that originated online and have at least 10 million unique visitors per month (list of outlets, from 247sports.com to Vox.com, here).
Sites do not appear to be increasingly building native apps: The percentage of top digital-native news sites with an app remained steady between 2016 and 2017, at 61 percent:
This does not appear to include responsive sites: “For mobile apps, researchers searched the Google Play and iOS App Store for each site,” according to the fact sheet’s methodology.
Other publishing methods are more popular:
Fully 97 percent of these outlets offer newsletters, and 92 percent have an official presence on Apple News. Three-quarters, meanwhile, release podcasts and 61 percent allow comments on their articles.
The full fact sheet is here. Pew also released two other new fact sheets — one on public broadcasting, one on Hispanic and African American news media — on Monday. (These fact sheets, with staggered releases, have taken the place of what was once Pew’s giant annual State of the News Media report.)
When the Bureau of Labor Statistics released its jobs report at the beginning of the month, news organizations unleashed their push notifications.
On Friday morning, the Wall Street Journal tested live mobile push alerts for their jobs coverage, working closely with the Guardian Mobile Innovation Lab, which has been for the past year tirelessly testing a range of ideas for distributing news that make the most of people’s phone-reading preferences.
Readers who arrived at the Journal’s mobile site or its Android or iOS apps were able to read its live coverage of the jobs numbers for July — but were also alerted with preview push notifications on updates as they read the existing analysis on the page (readers could dismiss and keep reading, or jump to the update from the push alert).
Journal developers built the infrastructure for the live notifications, and its markets team reported on the event and sent the pushes. The Mobile Innovation Lab provided guidance — based on learnings from its own past experiments and user testing — throughout the process, from evaluating design prototypes for the alerts to crafting an effective survey for users who encountered the Journal’s experiment.
The Journal has its own internal live coverage tool, built ahead of the Iowa Caucuses coverage in time for last year’s elections, but hadn’t dealt with live push notifications, according to Jennifer Hicks, deputy managing editor of digital at the Journal.
“We had a highlights feature where we could pin key posts, but we couldn’t notify readers within the live reading experience,” she said.
The Guardian’s Mobile Innovation Lab had been hosting some get-togethers and roundtables with various news organizations after the November 2016 election around news notifications, and the Journal expressed interest in trying out an experiment with the Lab. Work on this project started in June.
“There were lots of experiments the Guardian group was doing, so we talked about what we could bite off and pull off in a short amount of time,” Hicks said. “For us, it was also an opportunity to change our culture and talk directly to readers about testing a new feature.” (The Journal and Mobile Lab teams had a joint Slack channel going morning of the live notifications project for potential troubleshooting in implementation.)
The Journal plans to use the live notifications feature in future live coverage (with tweaks as necessary), according to Journal mobile editor Phil Izzo: “From jobs reports to the Olympics to terrorist attacks, we use live coverage a lot, and that’s one of the reasons we really wanted to build this out, since we knew there were so many use cases for it,” he said.
Both the Guardian and Journal teams emphasized the project’s experimental nature; it’s the first partnership of this kind for both organizations. The Lab is welcoming similar partnerships with other interested outlets.
“In the Lab we’re working for the industry and not just for ourselves — if we were to experiment in silence for two years and not share tips and tricks that we’ve experimented with, that wouldn’t be fulfilling the mission of the Lab,” Sarah Schmalbach, the Lab’s senior product manager, said. “We have been flexing our notifications muscle, then when we felt more confident in what we’d learned, we began to host events to ask other organizations what they were doing, where we’d then make a point to say, please come talk to us if there’s anything we can do to help, any data we can provide. Maybe we can launch something together.”
“We really relied on Sarah and [Mobile Innovation Lab editor] Sasha Koren to provide expertise in terms of, how do you talk to your audience directly, how do you conduct a real-time experiment, how do you offer a survey to audiences that gets you useful and actionable feedback,” Hicks said. “We had a lot of guidance on how to set up an experiment, which is not something we’ve done regularly at the Journal.”
Data points the Journal will evaluate for this jobs report experiment center around engagement, and include time spent on the live coverage, whether readers dismissed the notifications or clicked into the post, and bounce rate during the live event.
“Another thing we’re thinking about is, does this tell us anything about experimentation at the Journal?” Izzo said. “Did we make the job reports live blog better, because we put more attention to it, and should we push to do more things like this in the newsroom in general?”
A study comparing redesigns of homepages for news organizations in the United States and Canada found that contemporary designs with more images and less text resonate more with readers — but noted that experimenting with the redesign and doing before-and-after testing can reveal some helpful insights.
The Engaging News Project, an initiative of the University of Texas Austin, followed an unnamed major U.S. news organization and a similarly unnamed major Canadian news organization as they revamped their sites’ homepages. In each case, “two concurrent studies occurred. The first was an online survey-based experiment and the second was a live test conducted by the news organization.”
“Our results show that an online experiment can pick up on many of the same signals as a full launch of a site redesign,” Emily Van Duyn, a research associate for the Engaging News Project, said in a statement. “We believe that doing an online experiment could provide news organizations with a relatively inexpensive way to test out a redesign before a full launch.”
In the case of the Canadian redesign, both pageviews and the time spent by readers on the page were higher on the new site than on the old site. Readers were able to remember articles more effectively, potentially aided by a greater number of photos. The new site also replaced a grouping of 20 picture-less sections with nine sections labeled by topic and their own pictures.
The U.S. organization’s redesign did not result in higher visit times, but the bounce rate did rise. There was no major difference in recall rates.
The researchers considered potential reasons for improved article recall on the Canada site and plausible explanations for the U.S. site’s data.
(1) Pictures increase recall. Six articles were recalled more frequently on the new site than the old one. In 28 of 30 observations (6 articles across 5 different time periods), these articles were accompanied by a picture on the new site, but not on the old site.
(2) No differences in recall when articles equally prominent on the old and new sites. Three articles were recalled at a similar rate for both the old and new sites. In 12 out of 15 observations, the articles were equally prominent on both sites. The Trudeau article, for instance, appeared in the top third of the page, in the first column, and had a photo on both sites for all time periods analyzed.
(3) Column on the page affects recall. Three articles were recalled more frequently on the old site than the new one. It is more difficult to explain why these articles were recalled more frequently on the old site. The best explanation seems to be the column in which the story appeared. In nine of 15 instances, the old site had the article in the more prominent first column reading from left to right. In three instances, the story appeared in an equally prominent column. And in three instances, the pattern is the reverse, where the new site had the article in a more prominent column
(1) Pictures affect recall. The Putin article, better recalled on the new site, appeared with an image atop both the old and new sites. On the new site, however, there were no other images in the same row as the Putin story, while there was a competing image in the same row on the old site.
(2) The amount of scrolling matters. The Yellen article, better recalled on the new site, was featured at the top of the page on the new site, but was farther down the page on the old site. The rich banker, Indian overpass, and Syria stories required less scrolling on the old site compared to the new site. They also were better recalled on the old site.
(3) Column on the page affects recall. The Trump and China Xi stories, better recalled on the old site, appeared in a more left-hand column on the old site compared to the new.
(4) No differences in recall when articles equally prominent on the old and new sites. Just as we did on the Canadian site, we found on the U.S. site that the articles with no differences in recall (Trooper wounded, Megacopter) were similarly prioritized on the old and new sites.
Also, only 54 percent of participants in the Canada study knew what a hamburger menu was. (Hint: it’s the three horizontal lines that appear as a button to unleash the site’s list of section options. The question wasn’t asked on the U.S. side of the study.)
We have an opening for a staff writer here at Nieman Lab. If you’re interested, apply over here!
The job’s pretty easy to describe: You see all the stories on this website? The ones about journalism innovation — changes in how news gets reported, produced, distributed, discovered, consumed, and paid for? This job is about coming up with, reporting out, and writing those stories. There are some other duties, of course, like helping run our social media presence, but it’s a reporting job at its core. If you’ve ever thought I’d be good at writing Nieman Lab stories, I’d encourage you to apply.
This person will join our little five-person Harvard newsroom. She or he will also be joining the larger Nieman Foundation, which does a lot of exciting things for journalism and for journalists. (That’s our home, Walter Lippmann House, above; it’s nice.) For more details, see my writeup from a previous time we had an opening and, of course, the job listing.
One note about the position: To be considered for it, you’ll have to apply at the Harvard HR site linked above, where you should include a cover letter telling me why you think you’d be right for the job. (Don’t email me a resume directly; I’m not allowed to consider anyone who doesn’t go through the official HR process. But if you have specific questions about the job, feel free to drop me a line.)
I recently, shamefully fell for a photo plastered all over my timeline last week of Vladimir Putin sitting in a chair at G-20 as other world leaders, including Donald Trump, leaned in for what appeared to be an intense, whispered discussion. The photo was, as Gizmodo put it gently, totally fake.
— Matt Novak (@paleofuture) July 10, 2017
Fake headlines of the Pope-endorsing-Trump variety are just one part of the ecosystem of fakery online. There’s faked audio to worry about. Faked video. And of course, faked images.
It turns out people aren’t very good at identifying manipulated images, according to new research published Tuesday in the journal Cognitive Research by researchers Sophie J. Nightingale, Kimberley A. Wade, and Derrick G. Watson from the University of Warwick.
Participants were slightly better than random at picking out untouched versus manipulated photos, classifying 62 percent of the images in the study correctly. Participants also weren’t great at picking out where exactly a photo had been changed, even when they did accurately identify a photo as manipulated: they were able to identify an average of 45 percent of manipulations presented.
The study first tested participants on whether or not they could identify a manipulated image by showing them images of people in real-world scenes taken from a Google search, and manipulated versions of those images. In a second experiment, the researchers tested whether participants could pinpoint the region of the photo that had been changed.
People don’t necessarily appear to be better at pinpointing “implausible” manipulations (such as a shadow in the wrong place) than “plausible” ones (such as removal or addition of something into the photo), the researchers found:
Recall that we looked at two categories of manipulations — implausible and plausible — and we predicted that people would perform better on implausible manipulations because these scenes provide additional evidence that people can use to determine if a photo has been manipulated. Yet the story was not so simple. In Experiment 1, subjects correctly detected more of the implausible photo manipulations than the plausible photo manipulations, but in Experiment 2, the opposite was true. Further, even when subjects correctly identified the implausible photo manipulations, they did not necessarily go on to accurately locate the manipulation. It is clear that people find it difficult to detect and locate manipulations in real-world photos, regardless of whether those manipulations lead to physically plausible or implausible scenes.
Future research might also investigate potential ways to improve people’s ability to spot manipulated photos. However, our findings suggest that this is not going to be a straightforward task. We did not find any strong evidence to suggest there are individual factors that improve people’s ability to detect or locate manipulations. That said, our findings do highlight various possibilities that warrant further consideration, such as training people to make better use of the physical laws of the world, varying how long people have to judge the veracity of a photo, and encouraging a more careful and considered approach to detecting manipulations. What our findings have shown is that a more careful search of a scene, at the very least, may encourage people to be skeptical about the veracity of photos. Of course, increased skepticism is not perfect because it comes with an associated cost: a loss of faith in authentic photos. Yet, until we know more about how to improve people’s ability to distinguish between real and fake photos, a skeptical approach might be wise, especially in contexts such as law, scientific publication, and photojournalism where even a small manipulation can have ethically significant consequences.