Your current web browser is outdated. For best viewing experience, please consider upgrading to the latest version.


Send a question or comment using the form below. This message may be routed through support staff.

Email Article

Main Error Mesage Here
More detailed message would go here to provide context for the user and how to proceed
Main Error Mesage Here
More detailed message would go here to provide context for the user and how to proceed

Manhattan Institute

Close Nav
Share this commentary on Close

‘The Human Network’ Review: Like Goes With Like


‘The Human Network’ Review: Like Goes With Like

The Wall Street Journal April 3, 2019
OtherCulture & Society

A Stanford economist examines the ways in which one’s social position can determine power, beliefs and behaviors.

Ever have the sneaking suspicion that your friends are more popular than you? Turns out it’s probably true—and not just because you may be insufferable at cocktail parties. Why it’s true is fairly complicated, but in his book “The Human Network” Stanford economist Matthew O. Jackson entertainingly analyzes this and other mysteries. Drawing on the academic discipline known as network theory, Mr. Jackson aims to introduce and popularize a powerful way of understanding some of modern society’s central challenges.

Start with the “friendship paradox”: Imagine drawing a network of everyone from your high school—a circle representing each person and a line between circles representing a friendship. First, assign each circle a “popularity score” equal to the number of other circles connected to it (i.e., that person’s number of friends). Then, for each circle, calculate the average of the popularity scores of the circles connected to it; now each circle has a “popularity score” and a “popularity-of-friends score.”

Here’s the paradox: A typical person will have a “popularity-of-friends” score higher than his or her own “popularity score.” Most people’s friends really are more popular than they are. Precisely because the most popular people have more friends, they show up on the most lists of others’ friends. Thus insecure high-school students comparing their popularity to those in their social circle are not looking at a random sample, but one that overweights the most popular kids in school.

This effect of networks overrepresenting the already-popular helps explain a lot of adolescent behavior. More socially active teenagers tend toward more extreme behaviors, leaving everyone else with a mistaken impression about those behaviors’ frequency. College students overestimate how much a typical student drinks because the average level of drinking in an individual’s own network likely exceeds the actual average.

“The Human Network” presents an intriguing model for fads, in which cool kids set trends not because everyone aspires to be like them, but simply because everyone aspires to conform. Starting with a hypothetical network of 12 people, in which the most popular four come to school wearing solids one day while the other eight wear plaids, Mr. Jackson explains how all 12 will have converted to solids by day five. Individuals who observe their own few friends will conclude solids are more popular, change to conform, and the effect will cascade. All this happens even though, in the original state, plaid was actually more popular.

The early sections of Mr. Jackson’s book are entertaining, but in many cases the primary application will be to improve the reader’s cocktail-party repartee. The fun and games end with the introduction of “homophily,” which the author defines as “the general tendency of people to interact with others who are similar to themselves.” Importantly, he emphasizes, the phenomenon is common to almost all societies and “occurs along many dimensions including gender, ethnicity, religion, age, profession, [and] education level.” Racism and sexism are unnecessary to explain even highly segregated networks—in Africa’s Great Rift Valley, nomadic hunter-gatherers exhibit homophily on dimensions that include height, weight and strength.

Homophily’s precise effects depend upon the nature of an interaction. Where a single contact is sufficient for transmission—in the case of a pandemic, say—homophily matters little. With even very few connections between groups, everyone still ends up infected. But, homophily “has a much more pronounced impact on more nuanced things like beliefs” and behaviors, Mr. Jackson writes, where influence occurs across many contacts with many connections.

Take the decision some make to drop out of the labor force. Mr. Jackson’s model here begins with a network of two homophilic groups with few connections between them, but a rule that if most of a person’s friends drop out, he will too. Just a couple drop-outs in one group, we can observe, might trigger a cascading behavior through that group—but never transmit to the other group. You can prevent the cascade from starting, though, by adding just a few extra connections between the groups.

This way of describing behavior has the potential to reshuffle the political orthodoxy. It shows how inequality really might matter, if rising inequality creates rising homophily, leaving socioeconomic groups less connected to one another, and causing harm to those left behind. At the same time, it implies that the so-called “culture of poverty” is real and that the redistribution of resources offers no substitute for changing norms and behaviors. None of these insights, unfortunately, prompts easy policy solutions, and Mr. Jackson, clearly outside of his comfort zone, presents a muddled discussion about what we might do. As a framework for describing problems and evaluating solutions, however, his application of network theory is important.

Mastering the many concepts Mr. Jackson introduces requires careful attention. But “The Human Network” rewards the effort, especially as the author turns to what he calls “supported” relationships. People’s behavior, we learn, is guided not only by whom they connect to and how the people they connect to behave, but also by relationships that don’t include them—that is, relationships between their connections. Supported relationships, in which the two people share other friends in common, exhibit higher levels of trust and tend to be much stronger. “How one person treats another involves looking beyond the relationship to the network in which they are embedded,” Mr. Jackson observes.

Yet as technology enables us to choose our relationships from a much larger pool and homophily encourages us to search far and wide for those most like ourselves, we are eroding the local clustering that is crucial to community health and equality of opportunity. “Do we form the ‘right’ networks—ones that are best for our community or society?” asks Mr. Jackson. The book he has written suggests not.

This piece originally appeared at The Wall Street Journal


Oren Cass is a senior fellow at the Manhattan Institute and author of the new book “The Once and Future Worker.” Follow him on Twitter here.

Photo by iStock