BY KEVIN MOE
When it Comes to Desires, People Find Media Use Hard to Resist
At a research meeting in the Netherlands, Marketing Professor Kathleen Vohs was talking with a colleague, Will Hofmann of the University of Chicago Booth School of Business, about what people feel and desire in their everyday lives. As they both knew of no research on the topic, they decided they needed to start studying it.
Their research, published as “What People Desire, Feel Conflicted About, and Try to Resist in Everyday Life” (Psychological Science, 2012), included data from 205 adults who provided 7,827 reports of their desires over the course of a week.
Vohs says she did not expect how much people walk around in a state of desire. “Fifty percent of the time is spent wanting something! That itself was fascinating,” she says. Vohs was also surprised, even though she studies self-control, at how effective self-control is. “If people have a desire and try to resist giving in, most of the time they are hugely successful in restraining themselves,” she says. “There was a four-fold drop in the likelihood that a person would engage in a behavior if he or she had resisted doing it. So the power of self-control was quite a surprise and a welcome one.”
The findings show that desires for sleep and sex were experienced the most intensively, while desires for tobacco and alcohol, although considered addictive, were found to have the lowest average strength. When it came to what desires people try to manage, they found—surprisingly—that desires for work and media use, including social media, were the most difficult to control.
Forty-two percent of these desires were enacted by the study’s participants even when they had attempted to resist. Vohs speculates that media desires, such as social networking, checking emails, and surfing the web, might be hard to resist in light of the constant availability, huge appeal, and apparent low costs of these activities.
Media desires were most often felt in the evenings, showing that people are more vulnerable to succumbing to (often unrelated) impulses that arise later in the day to the extent they restrained themselves earlier from enacting their desires. This also indicates that the aftereffects of using self-control accumulate over longer time spans than previously thought.
The results of this study, including the findings of what people desire the most, how often, and when, and how they try to resist, provide much food for thought for companies, most of which are in the business of stimulating desires. “Gyms, financial planning services, and all manner of self-improvement firms will find our data interesting,” Vohs says.
How Does Diversity Impact Online Collaboration?
As Information and Decision Sciences Assistant Professsor Yuqing Ren was reading The Wisdom of Crowds by James Surowiecki, she was fascinated by its argument that aggregating opinions of a good number of ordinary people can lead to better and more accurate decisions than those made by a small number of elite experts.
“A key condition to harvest the wisdom of crowds though, is diversity. That is, you need to solicit inputs from people with different backgrounds and perspectives,” she says. “Pitting that against decades of research on diversity in offline groups, I saw a paradox in the role that diversity plays in online groups.”
On one hand, diversity provides a rich pool of information that can be integrated to generate better work outcomes. On the other hand, it can trigger social categorization and cause conflict when members use individual differences to form subgroups and favor in-group members. “I wanted to find out exactly what effects diversity has on performance of online volunteer groups such as Wikipedia projects,” she says. “The original goal was to generate insights that can be used to create and manage online volunteer groups with the ideal level of diversity.”
“The Impact and Evolution of Group Diversity in Online Collaboration” (Management Science, under review) by Ren and Jilin Chen and John Riedl of the University’s Department of Computer Science and Engineering takes an in-depth look at this matter. Ren and her colleagues examined two types of diversity over time—tenure diversity (newcomer versus old-timer) and interest diversity (members with interest in different domains such as art, science, or religion) in 648 WikiProjects. Findings show that high-performing Wikipedia projects have a moderate level of tenure diversity but high interest diversity.
“The most surprising finding was the evolution of diversity over time,” Ren says. “Different from work groups in organizations, there is no or little managerial oversight that is at play in creating or staffing Wikipedia projects.” Ren found that members voluntarily join and exit the project as they see fit or find opportunities to contribute. Over time, the level of both types of diversity converged to a level that is optimal for project success, meaning moderate levels of tenure diversity and high interest diversity.
A key takeaway from this research is how it shows the power of self-organizing. Today’s organizations are entrenched to have managers oversee and control lots of things, including staffing and managing project work. Yet, many teams fail miserably or under-perform because of lack of expertise, low member participation, or malfunctioning team dynamics. Often, issues ensue shortly after a team is created and there is no low-cost mechanisms for members to move in and out of the team—and they get stuck with the team and people whom they don’t know how to work with or don’t like, Ren says.
“The Wikipedia project model shows an alternative way—a more powerful way—in which project work can be organized in organizations,” she says. “When individuals have the autonomy to initiate, join, and leave project teams, the teams go through what is like a natural selection process under which new people join and those who don’t fit leave. As a result, the team evolves toward an optimal level of composition.
“We have seen signs of such autonomy being offered to employees in some companies such as Google’s 20 percent rule and 3M’s 15 percent rule, which allows employees to use part of their time to work on passion projects that are not necessarily related to their job descriptions. But we have yet to see more drastic changes in how work is organized in corporations and there is much organizations can learn from online production efforts like Wikipedia.”
Trust Experiments in Online Social Networks
Recent advances in social computing have resulted in a billion people globally connected to each other via the major online social networks (Facebook, Twitter, and other niche networks). These online social platforms offer researchers access to their data and processes through what are called APIs (application programming interfaces). This access allows researchers to build applications that can be used for experimentation in a real-world setting that have very strong external validity in their inferences. This paradigm, being pioneered at the Carlson School, is indicative that a golden era of research at the intersection of social sciences, information systems, and network computing is at hand.
“Trust, Reciprocity, and the Strength of Social Ties: Economic Experiments on an Online Social Network” (working paper, 2012) by Information and Decision Sciences Professors Ravi Bapna and Alok Gupta, along with colleagues Sarah Rice of the University of Connecticut and Arun Sundararajan of New York University, is the first proof of concept of this new research approach.
“We took a famous economic experiment that was originally designed to be played in the lab with student subjects to measure trust and reciprocity, and created a Facebook app that not only replicated the original anonymous game, but also extended to a non-anonymous version,” Bapna says. The research question was to link trust between pairs of individuals, some of whom were friends on Facebook, to traditional measures of social ties such as the number of common friends between them, and to newer “revealed preference” measures that could be derived from their online activities, such as how many times they posted on their partner’s wall or how many times they were tagged on a photograph together.
“We believe this is important because trust is the fundamental construct that underlies all models of social influence and contagion, and if we can, in real-time, dynamically measure trust between pairs of individuals, then it has huge implications for a variety of new business models that are emerging on platforms such as Facebook,” Bapna says. “For example, AirBnB is a Facebook application allowing peer-to-peer room rentals in various cities around the world, but would you want somebody you do not trust in your bedroom?”
A key finding from the research is that each wall post made on a friend’s wall resulted in a 21 percent increase in trust. Also, Bapna and Gupta found that each photo two friends jointly appeared in, which shows social affinity and physical world ties, is associated with a 5.1 percent increase in trust. Interestingly, a correlation between “number of common friends” and trust is not significant. “In online social networks, there are a lot of spurious friendships, so the common friends measure as a predictor of trust is less reliable than the revealed preference style measures such as wall post interactions and being tagged together in a photograph,” Bapna says. Gupta says one interpretation of the results is that people with smaller friend circles online may have larger influence over their friends than people with a large number of friends.
“Whether these findings can be used to manipulate behaviors is an open question. We plan to do experiments that can potentially establish casual relationships between actions that an application designer or social media site can take to generate desired reactions from the users,” Gupta says.
Can Online Information Undermine Brand Performance?
Advertising can be a large part of businesses’ budgets as they are trying to establish familiarity and trust with their customers. However, consumers can get information elsewhere, such as from their friends and family, and nowadays, on the internet. It is possible that information gleaned from these sources can inform their opinions a lot more strongly than advertising ever could. In “Does Information Undermine Brand? Information Intermediary Use and Preference for Branded Web Retailers” (The Journal of Industrial Economics, December 2006), Professor Joel Waldfogel, along with his former student Lu Chen, took a look at this possibility.
“I started this project relatively early in the rollout of internet commerce,” Waldfogel says. “At the time, people were talking about ‘frictionless’ commerce, the idea that low entry barriers would make all of retailing perfectly competitive. This was a frightening vision for retailers, since it meant that margins would shrink and nobody would make any money.”
Then a funny thing happened. A few retailers, such as Amazon, came to dominate internet retailing and it looked like traditional ideas like brand were as important in the frictionless online world as they had been in the real, offline world. “This struck me as a potentially premature conclusion,” Waldfogel says. “The right question is not whether brand still matters—although that is surely an interesting result. The right question is whether the kind of information available online can cause brand to matter less.”
Waldfogel and Chen looked at a 13-month panel dataset on shopping at branded and unbranded retailers and the use of information intermediaries by more than 30,000 online households. The specific question that was addressed is if consumers patronize branded online retailers less after taking up the use of information intermediaries such as Bizrate.com. Sites such as these include price comparisons and survey-based information about whether a retail site is reliable.
The results show that consumers’ information use weakens the pull of brand. Most individuals using price comparison sites substantially increased their level of shopping at unbranded retail sites as they shopped less at branded sites; in the case of this study, Amazon.
“This study relates to other work I’ve done on the evolution of industrial concentration,” Waldfogel says. “The internet and related technologies have substantially reduced the cost of entering many businesses, including retail, media, and others. Econ 101 would predict an explosion of entrants, all of whom would operate at small scale. Yet, in retail and in media, that’s not quite what we see.” As Amazon is a dominant force in retailing, a small number of media firms, such as CNN and The New York Times, are dominant in their areas. “These facts raise the important question of what factors determine whether markets will be dominated by a few big players versus having lots of small players, none of which are making much money,” he says.
An important lesson for businesses from this study is that some of the initial fear of the internet was overblown. “Online and off, customers are reluctant to deal with complete strangers. Hence, traditional ways of winning customers’ trust remain important,” Waldfogel says. “While one of the most successful players—Amazon—is a pure internet entity, it now has something in common with competitors who began offline. That is, it has a reputation.” The vast majority of the big players in online retailing are companies with pre-existing offline presence. So traditional retailers like Target and Best Buy have something they can leverage online.
“As is usually the case, the glass is both half empty and half full: The internet is both a threat and an opportunity for retailers,” Waldfogel says.