The Twitter world versus the world of academic research

A recent conversation about Twitter and my own acknowledgement of my lack of Twitter participation pushed me to think about the differences between Twitter conversations and academic research. These rough thoughts may be obvious to many but I think they are helpful to enumerate as we think about good information and data.

apple applications apps cell phone

Photo by Tracy Le Blanc on Pexels.com

1. The speed of Twitter, and social media, is quite fast. An opinion expressed or a “story” (link/video/article/”text”) can generate a lot of feedback but the process usually happens over just a few days. In contrast, academic studies routinely take years as researchers develop good questions, collect and analysis data/evidence/”texts”, and respond to hypotheses, theories, and existing conversations while also thinking about the implications of their findings and then go through the publication process. Academic work can sometimes go quicker but that length also theoretically allows time for more reflection.

2. Posts on Twitter are limited to a certain number of characters through tweet threads or good conversation following a post can effectively convey a longer argument or set of information. Academic studies also have space restrictions – while there are indeed examples of very lengthy books or articles, journals tend to have proscribed word count or page limits depending on their audience and the format of papers – but there is more space to make and develop an argument.

3. Twitter offers more immediate feedback, possibly much more, compared to academic works. When students ask me how many people read academic studies, it is hard to know: we have citation counts (which suggest at least those citing the work read it or are familiar with it) and journal websites now often offer the ability to see how many times an article has been viewed. But, how to count students who read pieces for class or projects, researchers who access material through databases and repositories, and other means of accessing academic work? However, I would assume the viral posts of Twitter gain more readers in a shorter amount of time than almost all academic works.

4. Those with Twitter accounts can post or access tweets. Those who publish in the academic world are a small subset of the population generally with advanced degrees and specialized knowledge. Yet, the percent of the American population who engage regularly on Twitter is low.

5. Tweets are validated or not by likes, retweets, and comments made by other users. Academics have more formal processes to vet work including peer review and presentations at conferences, lectures, and colloquia plus responses from audience members. A published academic work likely has had multiple eyes on it; tweets do not require this.

On the whole, academic research involves a longer, more formal process to get to publication and information available to other academics and the public. In contrast, Twitter offers speed, quick feedback, and is easier for both readers and posters to access. Thus, when the two worlds collide – academic posting about research or Twitter users engaging with academic work – it can seem as if two worlds with different processes and rules are trying to engage. The overlap can go successfully but it does not always; the advantages each system has do not necessarily complement what the other side does well. And the two systems may influence each other: the world of Twitter may prompt academics to speed up research and/or communicate their work differently while academics participating in Twitter may engage in deeper and longer conversations since much study involves nuance and complexity.

I do not spend much time on Twitter. It can be used effectively to quickly gather or share information. And if you follow engaging Twitter users in a particular subject area or field, there is much to be learned. I am grateful there are academics who can effectively use Twitter to engage audiences regarding their research and knowledge. But, the speed of the conversation can gloss over the depth of the issues at the heart of conversations or leave little room for the important context and background knowledge of phenomena.

(An aside: attempts to find a middle ground between such universes are worth thinking about. TED Talks seem to offer some compromises: an expert on the subject gets roughly 10-20 minutes to share out of their vast expertise. The videos are easy to follow and digest and they tend to come from people with advanced experience or education. The visual format has some appeal as opposed to text-based communication on Twitter and in academic writing. Podcasts could offer some similar benefits: there is more space for the storyteller to share but the audio cannot go on too long.)

Ascertaining the popularity of the tiny house movement via Twitter

HomeAdvisor looked at tweets about tiny houses and examined geographic patterns:

Top 10 States for Tiny Living

The best states for #tinyliving living – https://www.homeadvisor.com/r/off-the-grid-capitals/

The methodology:

To create these visualizations, we collected data by “scraping” it. Scraping is a technique that gathers large amounts of data from websites. In this case, we wrote a custom script in Python to get the data for each hashtag. The script collected information including the number of likes, number of comments, location, etc. for posts with each of the three lifestyle hashtags. The python script also collects data that human users can‘t see, like specific location information about where the post was published from.

We didn’t include posts without location information. We also didn’t include posts outside of the United States. We then standardized the city and state data. Then, we grouped the posts by city and by state, tallying the number of posts for each hashtag. This gave us our top locations.

And more details on the state and city level data:

TinyLivingHomeAdvisorJun20

On one hand, this is interesting data. California, in particular, stands out though this may not be that surprising given its size, good weather, and high housing prices. The rest of the top ten seem to match similar characteristics including scenic areas and good weather (New York and possibly Colorado winters excluded). The city level data compared to the state numbers provides some insights – major cities can account for a large percentage of tweets for a whole state –  but there are not many cases in any particular city.

On the other hand, it is hard to know what exactly this Twitter data means. There are multiple issues: how many Americans are on Twitter or are active on Twitter and does this overlap with those who like and have tiny houses? Some of the tweets about tiny houses did not have location data – is the data missing at random or does it intersect with the patterns above? Does the #tinyliving hashtag capture the tiny house movement or a part of it?

Because of these issues, I still do not have a better idea of whether the tiny house movement is sizable or not. Having some denominator would help; of the California tweets, how does this compare to the number of single-family homes or apartments in the state? Portland, Oregon leads the way with 695 cases but over 650,000 people live in the city. How do these tweet numbers compare to people tweeting about HGTV shows or single-family homes?

There is a lot that can be done here and making use of data publicly available on websites and social media is smart. Figuring out which questions can be asked and answered with such data and then collecting good data is a challenging and possibly rewarding task.

The difficulties for public institutions and spaces after COVID-19

Reopening and repopulating public spaces during and after COVID-19 might provide difficult:

Yet can you reopen a society — particularly a republic built on openness and public interaction — without its physical institutions at full capacity, without public spaces available for congregation?…

Something else unites these places. In each, the woman on the next bench, the man ahead in the checkout line, the family down the pew are suddenly potential vectors — or potential victims. So we’re assessing the public realm in the way we assess a salad bar when we walk into a restaurant…

“Democracy depends to a surprising extent on the availability of physical, public space, even in our allegedly digital world,” John R. Parkinson writes in “Democracy and Public Space: The Physical Sites of Democratic Performance.”

“How do you define the ‘public realm’ when an enormous percentage of the American public spends the majority of its day in its pajamas?” Stilgoe says.

This piece raises great questions for a COVID-19 world. The emphasis on how architecture and design shapes public behavior as well as how others in those spaces can be trusted or not is right on. At the same time, there are several elements I would add to this analysis:

1. The definitions of “institution” and “spaces” are pretty broad. Some of the listed locations, like shopping malls, colleges, and grocery stores, are not public spaces. They are owned by private groups that can and do dictate how the space can be used. Some of the other locations, like parks and squares, are public spaces. Government buildings are generally more open to all. Americans privilege private space even though we need some of the private spaces – grocery stores, workplaces – to survive. But, the same rules or expectations do not apply in each of these spaces. We saw this in the Occupy Wall Street protests where gatherings in what looked like public spaces could be ended when they spaces were actually owned by private groups or the government pushed people out. We actually do not have that many public spaces where people regularly gather; many of our “public spaces” are actually privately owned and this matters. The private public spaces require both private groups and the public to cooperate – and they may not always do so.

2. Even before COVID-19, it is not clear that many Americans value public spaces or use them regularly. As noted in #1, Americans like their private spaces. Homes may be less attractive when you are trapped in them but we have a society where success is owning your own suburban single-family home. Add to this declining trust in numerous institutions and it may be hard to make the case that we should put more resources and effort into creating and maintaining public spaces.

3. More broadly, many would argue a thriving society and democracy depends on regular interaction between people. And face-to-face interaction provides benefits that online communication does not regarding communicating clearly and building relationships. Yet, again, this has been on a decline for a while now. Twitter is not a good approximation of public conversation nor a good medium (at least as currently constructed or experienced) for public conversation. Telecommuting may provide efficiencies and allow people more private lives but something will be lost. See my earlier thoughts on sociologist Eric Klinenberg’s Palaces for the People where he takes up these issues (Part One, Part Two, Part Three, Part Four).

Reminder: “Twitter Is Not America”

A summary of recent data from Pew provides the reminder that Twitter hardly represents the United States as a whole:

In the United States, Twitter users are statistically younger, wealthier, and more politically liberal than the general population. They are also substantially better educated, according to Pew: 42 percent of sampled users had a college degree, versus 31 percent for U.S. adults broadly. Forty-one percent reported an income of more than $75,000, too, another large difference from the country as a whole. They were far more likely (60 percent) to be Democrats or lean Democratic than to be Republicans or lean Republican (35 percent)…

First, Pew split up the Twitter users it surveyed into two groups: the top 10 percent most active users and the bottom 90 percent. Among that less-active group, the median user had tweeted twice total and had 19 followers. Most had never tweeted about politics, not even about Twitter CEO Jack Dorsey’s meeting with Donald Trump.

Then there were the top 10 percent most active users. This group was remarkably different; its members tweeted a median of 138 times a month, and 81 percent used Twitter more than once a day. These Twitter power users were much more likely to be women: 65 percent versus 48 percent for the less-active group. They were also more likely to tweet about politics, though there were not huge attitudinal differences between heavy and light users.

In fancier social science terms, this suggests what happens on Twitter is not generalizable to the rest of Americans. It may not reflect what people are actually talking about or debating. It may not reflect the full spectrum of possible opinions or represent those opinions in the proportions they are generally held throughout the entire country. This does not mean that is no value in examining what happens on Twitter, but the findings are limited more to the population that uses it.

In contrast, the larger proportion of Americans who are on Facebook might appear to suggest that Facebook is more representative of the American population. But, another issue might arise, one that could dog social media platforms for years to come: how much content and interaction is driven by power users versus the percent of users who have relatively dormant accounts. I assume leaders of platforms would prefer more users become power users but this may not happen. What happens to any social media platform that has strong bifurcations between power users and less active users? Is this sustainable? Facebook has a goal to connect more people but this is unlikely to happen with such disparities in use.

This is why discussing or confirming trends seen on social media platforms might require more evidence from other sources or longer periods of time to verify. Even what might appear as widespread trends in social media could be limited to certain portions of the population. We may know more about smaller patterns in society that were once harder to see but putting together the big picture may be trickier.

 

Living the suburban teenage life through local Twitter

In The Levittowners, sociologist Herbert Gans said the suburban community was “endsville” for teenagers. But, suburban teenagers today can take to Twitter:

A decade ago, if you were a bored teen looking to post about suburban life, relationship problems, Starbucks, or fleeting thoughts like, “The holidays are approaching and being single sucksssss lol,” you might turn to Facebook. But of course, today’s teens don’t use Facebook. Instead, they take their most #relatable thoughts to Twitter, often racking up hundreds of thousands of retweets and faves in the process.

Twitter is full of tribes: gay Twitter, stan Twitter, politics Twitter, media Twitter, weird Twitter. The mostly white, well-adjusted suburban teens who share stale platitudes of the kind that some internet users might call “basic” are part of a tribe known as local Twitter.

Though most users do mainly follow people from their hometowns, local Twitter has more to do with what you tweet than where you live. The typical local Twitter user is a teen who is “in their own bubble of simple life pleasures and desires,” doesn’t live their entire life online, “and uses Twitter to connect to their real-life friends like they used to do on Facebook,” explains Raeequaza, a 22-year-old in New York…

Local Twitter teens are townie-like in the sense that their world mostly revolves around life in their hometown, though most will probably grow up and eventually leave for college. Some older local Twitter users might actually be townies, but the majority of local Twitter—particularly the part that has the power to make local tweets go viral—is made up of teens.

The ongoing plight of American teenagers in suburbia continues: they feel cut off from the exciting outside world, the suburbs are dull and do not feature spaces for teenagers, the suburbs represent conformity and middle-of-the-road values, and daily life revolves around school and family. Compared to Gans’ time where being able to drive represented freedom for teenagers, now teenagers can escape the suburbs (or live the ironic suburban life) through Internet and social media connections that theoretically can connect them to any person or place in the world.

Two additional quick thoughts:

  1. How many of these local Twitter users will end up living in suburbs as adults?
  2. Are the local Twitter users more perceptive about their local surroundings or are they just willing to tweet about their observations?

Can we expect an authenticity backlash after report of fake online followers?

“The Follower Factory” in the New York Times details how many public figures and social media users purchase followers.

The Times reviewed business and court records showing that Devumi has more than 200,000 customers, including reality television stars, professional athletes, comedians, TED speakers, pastors and models. In most cases, the records show, they purchased their own followers. In others, their employees, agents, public relations companies, family members or friends did the buying. For just pennies each — sometimes even less — Devumi offers Twitter followers, views on YouTube, plays on SoundCloud, the music-hosting site, and endorsements on LinkedIn, the professional-networking site.

The actor John Leguizamo has Devumi followers. So do Michael Dell, the computer billionaire, and Ray Lewis, the football commentator and former Ravens linebacker. Kathy Ireland, the onetime swimsuit model who today presides over a half-billion-dollar licensing empire, has hundreds of thousands of fake Devumi followers, as does Akbar Gbajabiamila, the host of the show “American Ninja Warrior.” Even a Twitter board member, Martha Lane Fox, has some.

At a time when Facebook, Twitter and Google are grappling with an epidemic of political manipulation and fake news, Devumi’s fake followers also serve as phantom foot soldiers in political battles online. Devumi’s customers include both avid supporters and fervent critics of President Trump, and both liberal cable pundits and a reporter at the alt-right bastion Breitbart. Randy Bryce, an ironworker seeking to unseat Representative Paul Ryan of Wisconsin, purchased Devumi followers in 2015, when he was a blogger and labor activist. Louise Linton, the wife of the Treasury secretary, Steven Mnuchin, bought followers when she was trying to gain traction as an actress.

Devumi’s products serve politicians and governments overseas, too. An editor at China’s state-run news agency, Xinhua, paid Devumi for hundreds of thousands of followers and retweets on Twitter, which the country’s government has banned but sees as a forum for issuing propaganda abroad. An adviser to Ecuador’s president, Lenín Moreno, bought tens of thousands of followers and retweets for Mr. Moreno’s campaign accounts during last year’s elections.

The incentives to do this are high: not only can these purchased followers act on the behalf of the purchaser, media accounts regularly highlight the number of friends or followers a user has. These counts are one of the important social markers of status online. If you are not actively trying to boost these counts by multiple means, you are falling behind.

If so many public figures then have purchased followers, then will we see an authenticity backlash? Imagine a scenario where Twitter or LinkedIn offers a special badge that all of your friends and followers are authentic people. Or, public profiles will include an estimate of how many followers are actual users. Then, it is not only about how many followers you have but rather how many are “real” people. The irony may be that even if you have “real” followers, the sort of interactions you have with them in the online realm can be quite different than offline interactions.

Would the public care to have such metrics? News of paid followers has been available for years. (For example, see earlier posts here and here.) Would they act differently toward certain users or profiles if they knew where they came from? In a world full of paid or compensated online reviews, fake followers, and who knows what else (targeted Facebook ads? Google search results just for you?), perhaps we are already past the point of no return.

Creating a quirky and warm Twitter personality for your on-the-market housing unit

Here may be a new housing trend: personifying your for-sale home in a Twitter account.

Bob the House — a three-bedroom, one-and-a-half-bath ranch in the Chicago suburb of Mount Prospect — has been tweeting about his journey on the market since October. You’ll find Bob to be a rather inspirational house, tweeting messages of positivity and hope on a regular basis, along with fanfare for the Chicago Cubs and humorous updates about his search for the right family. (“Six showings today! You like me, you really like me!”)

Here’s the open secret: It’s not really Bob who’s doing the tweeting. It’s his “handler,” Rich Burghgraef, an account executive with sales consulting firm Randolph Sterling, Inc. Burghgraef writes in a blog post that he created Bob, whose name comes from the street the house lives on, Robert Drive, as a way to get away from typical advertising tactics. It seems Bob was a hit, with showings of the home going from one or two a weekend to six and eight after the Twitter account debuted. It may have even been responsible (or at least contributory) to the ultimate happy ending, as Bob tweeted on March 1: “My new family moved in on Friday. Thank you all for taking an interest.”…

Burghgraef says that he used Bob to make a personal connection with buyers, not just to throw marketing messages of “buy me!” at them. Bob would tweet about the school that taught him to tweet (so there’s a good school in the neighborhood!); his friend, the stop sign (safety first!); and his stepson, the swing set (don’t you see your family here?). The home’s Twitter account gave buyers a new way to “fall in love with him even before stepping in for a showing,” Burghgraef says.

A clever way to use social media. The several accounts I read of this phenomena did not provide much evidence regarding the effectiveness of this tactic. Of course, social media attention is one of the currencies of today’s social realm so why not leverage it to help sell your home?We can’t be too far away from someone automating this process so that every new housing unit on the market could take advantage of Internet available information about the neighborhood and surrounding area to develop a winning personality.

I think Burghgraef is right in suggesting that this could be particularly effective with real estate since there is a high level of emotional investment. While we could imagine all sorts of consumer goods having their own online personalities, not all of those goods might have the same emotional connections to their owners.

Mining Twitter for ratings of mass transit and what the agencies can do in response

A new study examined Twitter comments about mass transit in the United States and Canada and came up with a ranking of those invoking the most positive and negative sentiments:

The results of her study, published this month in the Journal of the American Planning Association, ranked 10 of the largest public transit agencies in the US and Canada by how well regarded they are on Twitter. Based on Schweitzer’s “mean sentiment score” and more than 60,000 tweets collected between 2010 and 2014, Twitter was nicest to Vancouver’s Translink, which was followed by Portland, Oregon’s TriMet and Toronto’s TTC. The harshest tweets concerned systems in Chicago, Philadelphia, Boston, and New York. For comparison, Schweitzer calculated scores for public figures (the sentiment score ranged from William Shatner to Osama Bin Laden), airlines, police departments, and welfare programs (the full chart is at the bottom of this post).

Schweitzer used text mining to pick out positive and negative words from the tweets (and manually added terms including brokedown, wtf, scam, epicfail, pervy, and unsuck). Machine learning helped spot things like parody accounts and unusually frequent tweeters. Schweitzer and her graduate students also analyzed some 5,000 tweets by hand, to ensure they lined up with the computer system’s interpretations. Reasons for complaint included delays, facilities, staff conduct, public mismanagement, and the class, race, and gender of other riders.

Here’s the funny thing: The transit system’s scores don’t line up with service quality (judged by on-time performance). But the unsurprising fact that public griping doesn’t necessarily match reality doesn’t make the data useless. Because Schweitzer did find one factor that predicts “mean sentiment”—the way the transit agencies themselves behave on Twitter…

So what’s the takeaway? If you’re looking for a low investment way to improve your public image on Twitter, use Twitter as a tool for conversation, not one-way communication. It may seem that someone complaining to 18 followers that their train is late doesn’t matter, but Schweitzer makes the point that social media does influence broader public perceptions.

Engaging in public relations on social media is not new. However, the idea that government agencies or infrastructure organizations need to may be more recent. On one hand, Americans expect government to be responsive. On the other hand, mass transit is one of those areas that seems monolithic: leaders in those organizations are not elected and infrastructure faces its own kind of difficulties (aging, weather issues, particular funding sources, a sort of permanence that is difficult to change quickly). But, at least the disgruntled might feel heard if there is social media interaction even if their complaints are not fixed immediately.

Possible next steps: would major mass transit groups make policy decisions based on Twitter? Remember, a small percentage of Americans use Twitter regularly but those users can be pretty vocal and/or well positioned in society.

US unemployment figures distorted by lack of response, repeated takings of the survey

Two new studies suggest unemployment figures are pushed downward by the data collection process:

The first report, published by the National Bureau of Economic Research, found that the unemployment number released by the government suffers from a problem faced by other pollsters: Lack of response. This problem dates back to a 1994 redesign of the survey when it went from paper-based to computer-based, although neither the researchers nor anyone else has been able to offer a reason for why the redesign has affected the numbers.

What the researchers found was that, for whatever reason, unemployed workers, who are surveyed multiple times are most likely to respond to the survey when they are first given it and ignore the survey later on.

The report notes, “It is possible that unemployed respondents who have already been interviewed are more likely to change their responses to the labor force question, for example, if they want to minimize the length of the interview (now that they know the interview questions) or because they don’t want to admit that they are still unemployed.”

This ends up inaccurately weighting the later responses and skewing the unemployment rate downward. It also seems to have increased the number of people who once would have been designated as officially unemployed but today are labeled as out of the labor force, which means they are neither working nor looking for work.

And the second study suggests some of this data could be collected via Twitter by looking for key phrases.

This generally highlights the issue of survey fatigue where respondents are less likely to respond and completely fill out a survey. This hampers important data collection efforts across a wide range of fields. Given the enormity of the unemployment figures for American politics and economic life, this is a data problem worth solving.

A side thought: instead of searching Twitter for key words, why not deliver survey instruments like this through Twitter or smartphones? The surveys would have to be relatively short but they could have the advantage of seeming less time-consuming and could get better data.

When anti-government forces can control the public narrative about drone strikes in Yemen

While social media was praised in helping the Arab Spring movement, the new availability of Twitter in Yemen has changed who gets to control the public narrative about violence:

The result: AQAP and the Yemeni public have left the government far behind in an information war made possible by the spread of the Internet in the Arab world’s poorest nation. Authorities can no longer shape the narrative of counterinsurgency, particularly when it comes to controversial drone strikes…But the number of Internet users in the country increased nearly tenfold between 2010 and 2012, according to government figures, although even with that rapid expansion, less than a quarter of Yemenis have regular internet access.

Most drone strikes, which are believed to be US operations, target the most impoverished and isolated parts of Yemen where AQAP operates. The region’s remoteness plays into the group’s hands; it also makes it easy for the government to suppress any negative information, including civilian casualties from drone strikes and other aerial attacks.

But now Yemenis can easily, quickly share on-the-ground information. Last December, an airstrike targeted a wedding convoy, killing roughly a dozen civilians. The government initially identified the casualties as militants, but locals soon began posting photos of the dead on Facebook and tweeting the names of victims, directly challenging the government’s obfuscation.

Sounds like quite a change in a short amount of time. The availability of the Internet and social media threaten all sorts of traditional institutions that have relied on controlling information. All of the sudden, alternative viewpoints are available and regular citizens can pick and choose which to follow, believe, and propagate.

What does this do for American foreign policy? We generally disapprove of regimes that crack down on Internet availability (think China) but this is usually because we want to get our messages through. What happens when the same technologies are used to counter American narratives?