Ethnic Diversity and Social Cohesion

Reposted from “Alternative Hypothesis”

Ethnic Diversity and Social Cohesion

April 15, 2016 Francis Uncategorized 1


Summary: Research shows that people like those who are similar to them. Moreover, the more heritable a trait is the more important similarity in it is to friends and spouses. Thus, people enjoy the company of those who are genetically similar to them. This may explain why people to marry, befriend, and live near, members of their own race for more than randomness would predict. Moreover, a rich literature of correlational, longitudinal, and experimental, data shows that people get along better with members of their own race and that ethnic diversity damages social cohesion. 

People like People Who Are Similar To Them

If you take a class in social psychology, one of the first things you will learn is that people like people who are similar to themselves. Similar people get along better, make better friends, and better marriage partners. This viewpoint is supported by two kinds of evidence: correlational and expeirmental.

The correlational evidence consists of the repeated finding that friends and spouses are more similar, in just about every way imaginable, than average (Bryne, 1971Myers, 2008 pg 399). Moreover, the more similar spouses are the more satisfied they tend to be with their marriage and the less likely they are to get a divorce (Luo and Klohen, 2005Bryne, 1971Caspi and Herberner, 1990).

The experimental evidence is pretty straightforward: researchers measure various traits of individuals before they meet and then see whether pre meeting similarity predict who ends up liking who. The finding, repeated over dozens of studies, is that the more similar people are the more they end up liking each-other (Bryne, 1971NewComb, 1961Lee, 1996Myers, 2008 pg 400).

These findings contradict the idea that “opposites attract”, and that makes a lot of intuitive sense. If someone doesn’t like to do the same activities as you, doesn’t have the same sense of humor as you, etc., why wouldn’t you be less likely to like them?

People Like People Who Are Genetically Similar To Them

Perhaps less obvious than the fact that similarity breeds attraction is the role that genetics plays in this. Specifically, both friends and marriage partners are more genetically similar than average (Christakis and Fowler, 2013; Domingue et al, 2014; Rushton 1989). Now, you might think that this is to be expected, since such friends and spouses are more phenotypically similar than average, and all phenotypes are partly heritable. But there is more to it that that: the more heritable a trait is, the more similar friends and spouses tend to be in it (Rushton 1989). Moreover, similarity in highly heritable traits is a better predictor of marital success than similarity in less heritable traits is (Russel and Wells 1991). Further still, if you ask someone to imagine a person who is similar to them in various ways, the more heritable the traits involved in the similarities are the more that person will predict that they will like the person they are imagining (Abraham 1993).

People of the Same Race are More Similar Than Average

So, people not only get along people with people who are similar to them, they especially get along better with people who are genetically similar to them. Now, let’s relate this to race: people of the same race are, on average, more genetically similar than people of different races are (Witherspoon et al, 2007). In fact, if you measure a large enough portion of the genome, the probability that a same race pair of people will be more genetically similar than a different race pair of people nears 100%. As is covered elsewhere on this site, the races also differ in terms of demographic variables, psychological variables, psychological variables, and political views. Given this, on average, members of the same race can be expected to be more similar than members of different races on just about any measure.

Given these two facts, we might expect that people would tend to marry and befriend members of their own race and we’d be right: most marriages and friendships are between members of the same race. For instance, Ingraham (2014) reported on a study done by the Public Religion Institute which found the following:

Similarly, according to a report by Pew:

Among adults who are white with no other race in their background, fully 81% say that all or most of their close friends are white. Among single-race blacks, 70% say that all or most of their close friends are black. And among single-race Asians, 54% say all or most of their close friends are Asian.

The same report, which was based on census data, found that about 88% of new marriages in America are to members of the same race.

People’s racial preferences don’t stop at these close relationships either. Studies also show significant racial segregation in terms of the neighborhoods people live in and the churches they attend (Pew, 2014Lipka, 2014). Moreover, preferences for members of one’s own race begin to appear in infacnt (Kelly et al., 2005Vogel et al., 2012).

Ethnic Diversity Kills Social Cohesion 

So, people tend to get along with people who are a similar to them. Like co-ethnics. People typically befriend and marry co-ethnics, live near co-ethnics, and go to church with co-ethnics. And this preference for coethnics begins to form in the first few months of life. In other words, left to their own devises people primarily associate with members of their own race. Given this, it shouldn’t be surprising that a wealth of social science evidence shows that ethnic diversity hurts the degree to which people get along.

The most well known study in this literature is Putnam (2007). Putnam, a Harvard political scientist, analyzed a set of over 40 regions across the united states and looked at how ethnic diversity related to various measures of “social capital”. He built a regression model that held regional differences in age, poverty, crime, and other variables, constant and found that the more ethnically diverse a region was the lower its level of social capital was. In particular, Putnam found that the more diverse a community was the less trust residents reported having in their neighbors, the less people trusted local government and media, the less people voted, the less people gave to charity, the less people worked on community projects, and the fewer friends people had. Perhaps most strikingly, people in diverse communities were less happy overall and less satisfied with their lives.

Laurence and Bentley (2015)  replicated Putnam’s findings in England and, utilizing data on an 18 year period, found that the more diverse a neighborhood became the less people reported liking their neighborhood and engaging with their neighbors.

Similar, Lancee and Drunkers (2008) looked at social capital and ethnic diversity within the Netherlands. They utilized data from a danish survey which asked questions about how much people trusted their neighbors. After controlling for differences in gender, educational attainment, marital status, and income, the study found that the more ethnically diverse a person’s postal code was the less trust they had in their neighbors.

To be fair, not all studies turn out this way. Meer and Tolsma (2014) meta-analyzed 90 previous papers and found that 50% of the studies carried out in the U.S. replicated Putnam’s findings while 20% failed to do so. The remaining 30% were classified as having offered “mixed evidence”. By contrast, only 26% of studies in Europe replicated Putnam’s findings while 15% failed to do so and the remaining 59% of studies offered mixed evidence.

These replication rates aren’t great, but they are also easy to explain. First, many of these studies measured ethnic diversity in a way that would call a half White half Black neighborhood and a half British American half German American neighborhood equally diverse. More specifically, ethnic diversity was often measured as the probability that two randomly selected two people from the same region would be of different ethnics groups were the list of possible ethnic groups included not 3-6 “races” but, rather, 100+ ethnic groups.  Thus, many of these studies looked at ethnic, rather than racial, diversity. This is problematic because ethnic groups are far more genetically (and phenotypically) similar than races are. The second problem with this meta-analysis is that many of the studies referenced controlled for the mechanisms by which ethnic diversity might damage social cohesion. For instance, many studies controlled for income inequality and crime. The results of such an analysis will be misleading because ethnic diversity might cause a decrease in social cohesion by causing an increase in crime and income inequality. Thus, many of these studies controlled for the effects of diversity and then concluded that diversity had not effect.

The worst of these are studies that actually controlled for racial diversity. For instance, Aizlewood and Pendakur (2005) and Andersen and Millligan (2011) both found that the proportion of a neighborhood which is non white negatively correlated with its level of social cohesion but that ethnic diversity had no, or a positive, impact on social cohesion, after controlling for the effects of variation in the size of the population that isn’t white!  Several of these studies were counted as “failed replications” by Meet and Tolsma.

Given the degree to which the deck was stacked against successful replications , the fact that less than 25% of studies clearly failed to replicate Putnam’s finding should be seen as more impressive than it might at first seem.

The negative relationship between ethnic diversity and social cohesion has also been demonstrated in various “micro contexts”. For instance, Martinez and Doughtery (2013) looked at 75,000 people across nearly 400 church congregations and found that being a member of the dominant racial group of the congregation predicted a greater sense of belonging in the church, having more friends in the church, and participating in more church activities. Similarly, Montoya and Briggs (2013) found that people reported having a better experience in firms when the customer service representative they dealt with was a member of their own race. Finally, Dinesen and Sonderskov (2015) found that ethnic diversity within an 8o meter radius of a person’s home  related to less social trust.

Along with the correlational and longitudinal data already presented, the negative relationship between pro social behavior and ethnic diversity has been demonstrated experimentally. For instance, Glaeser et al (2000) had participants play an economic “game” in which one person sent another person a sum of money of their choosing. That money was then doubled money they were sent and the receiver had a chance to send some money back to the person who gave them the initial sum of money. This is a very basic measure of altruism, fairness, and trustworthiness. The researchers found that the receivers sent back far less money when they were paired with someone of another race. In fact, over 90% of the cases in which no money was sent back took place with racial diverse pairs of people.

Studies on the benefits of diversity in work environments also show us that diversity makes people trust people less. This was the finding of Levine and Stark (2015) who argued that reason that people in diverse groups came up with better solutions to hypothetical problems in there studies was that they were less likely to trust/like one another and, so, less susceptible to group think and conformity. While this may be seen as a benefit to problem solving, it is also obviously socially detrimental.


In conclusion, the totality of evidence clearly shows that ethnic diversity damages how well people get along, how much they trust one another, and how satisfied they are with their lives. These findings should be kept in mind when thinking about any form of racial diversity, regardless of the races involved.


This site uses Akismet to reduce spam. Learn how your comment data is processed.