August 09, 2007
Does Diversity Cause Us to Mistrust One Another?Issues
Via Ruchira Paul and 3QD, an article in the Boston Globe about the work of Robert Putnam, a Harvard University political scientist. The Globe summarizes the gist of the article as follows:
It has become increasingly popular to speak of racial and ethnic diversity as a civic strength. From multicultural festivals to pronouncements from political leaders, the message is the same: our differences make us stronger.
But a massive new study, based on detailed interviews of nearly 30,000 people across America, has concluded just the opposite. Harvard political scientist Robert Putnam — famous for “Bowling Alone,” his 2000 book on declining civic engagement — has found that the greater the diversity in a community, the fewer people vote and the less they volunteer, the less they give to charity and work on community projects. In the most diverse communities, neighbors trust one another about half as much as they do in the most homogenous settings. The study, the largest ever on civic engagement in America, found that virtually all measures of civic health are lower in more diverse settings.
“The extent of the effect is shocking,” says Scott Page, a University of Michigan political scientist.
The study comes at a time when the future of the American melting pot is the focus of intense political debate, from immigration to race-based admissions to schools, and it poses challenges to advocates on all sides of the issues. The study is already being cited by some conservatives as proof of the harm large-scale immigration causes to the nation’s social fabric. But with demographic trends already pushing the nation inexorably toward greater diversity, the real question may yet lie ahead: how to handle the unsettling social changes that Putnam’s research predicts. (link)
What makes this all more interesting is the fact that Robert Putnam is not himself a conservative, but a progressive-minded scholar who supports diversity. He didn’t expect these findings when he started this project, and has worked hard to make sure they are understood correctly — though anti-immigrant conservatives have definitely been eating this up.
I want to speculate a little on how South Asian immigrants might fit into the ‘diversity problem’ Putnam’s study raises, but before that it seems important to get into a little more detail about just what Putnam is saying. Please forgive the long quote:
The results of his new study come from a survey Putnam directed among residents in 41 US communities, including Boston. Residents were sorted into the four principal categories used by the US Census: black, white, Hispanic, and Asian. They were asked how much they trusted their neighbors and those of each racial category, and questioned about a long list of civic attitudes and practices, including their views on local government, their involvement in community projects, and their friendships. What emerged in more diverse communities was a bleak picture of civic desolation, affecting everything from political engagement to the state of social ties.
Putnam knew he had provocative findings on his hands. He worried about coming under some of the same liberal attacks that greeted Daniel Patrick Moynihan’s landmark 1965 report on the social costs associated with the breakdown of the black family. There is always the risk of being pilloried as the bearer of “an inconvenient truth,” says Putnam.
After releasing the initial results in 2001, Putnam says he spent time “kicking the tires really hard” to be sure the study had it right. Putnam realized, for instance, that more diverse communities tended to be larger, have greater income ranges, higher crime rates, and more mobility among their residents — all factors that could depress social capital independent of any impact ethnic diversity might have.
“People would say, ‘I bet you forgot about X,’” Putnam says of the string of suggestions from colleagues. “There were 20 or 30 X’s.”
But even after statistically taking them all into account, the connection remained strong: Higher diversity meant lower social capital. In his findings, Putnam writes that those in more diverse communities tend to “distrust their neighbors, regardless of the color of their skin, to withdraw even from close friends, to expect the worst from their community and its leaders, to volunteer less, give less to charity and work on community projects less often, to register to vote less, to agitate for social reform more but have less faith that they can actually make a difference, and to huddle unhappily in front of the television.” (link)
Wow — that’s a long list of problems associated with living in diverse communities! Personally, I’ve never felt the difference Putnam’s study finds, but for the most part I’ve mainly lived in relatively diverse places. I’ve lived in glum diverse places (Malden, MA; Bethlehem, PA) — where no one would give me the time of day or even stop and say ‘hi’ — and somewhat happier diverse places (Potomac, MD; Parsippany, NJ; New Haven, CT; Durham, NC; and my current town of Conshohocken, PA). Most places I’ve lived, though, I’ve felt that most people do “hunker down” and spend their evenings in front of the TV. I’ve never lived in the vibrant downtown of a big city (sigh), nor have I ever lived in a place that was really ethnically homogeneous — so perhaps I’ve only seen one side of this.
People interested in seeing more detail — and hearing it directly from Putnam, might want to check out the article in question here. For the most part it should be readable for non-academics (it helps if you know what he means by “social capital”), though Putnam does get into some statistical analysis that goes over my head.
The other big questions are 1) why could this be happening, 2) what can be done about it, and 3) is it a permanent problem, or merely a temporary phenomenon associated with recent immigration, which will dissipate over time?
One can easily speculate that the answer to (1) has to do with the natural mistrust produced when people have different ethnic and racial backgrounds, different cultural values, speak different languages, and so on. The answers to (2) and (3) are harder.
Again, thinking speculatively here, I’m not sure that anything can be actively done about (2), but I do feel quite confident on (3) that the mistrust and the lower “social capital” Putnam sees in more diverse communities is likely to dissipate over time — as immigrants acculturate and/or assimilate. Here, one’s experience as a second-gen desi comes into play. And the high levels of interracial dating and marrying out of one’s ethnic group seen among second and third generation Asian immigrants suggests that blending is already well under way. Another positive sign — which we dwell on a lot at Sepia Mutiny — is the growing involvement of immigrants and their children in civic and political life. Upper-middle class Desis aren’t just becoming doctors and engineers anymore; they’re becoming actors, and yes, even politicians.
Putnam himself agrees with that prognosis, and in his article, quotes Barack Obama to that effect. Obama has called for:
… an America where race is understood in the same way that the ethnic diversity of the white population is understood. People take pride in being Irish-American and Italian-American. They have a particular culture that infuses the (whole) culture and makes it richer and more interesting. But it’s not something that determines people’s life chances and there is no sense of superiority or inferiority… . [I]f we can expand that attitude to embrace African-Americans and Latino-Americans and Asian-Americans, then … all our kids can feel comfortable with the worlds they are coming out of, knowing they are part of something larger. (link)
Obama is in effect calling for “race” to start acting more like immigrant “ethnicity” — for it to be porous rather than a “hard” dividing line. Are South Asians a “race” or an “ethnicity”? Though I’m proud of my Indian heritage and proud of being both an Indian American and a practicing Sikh, I tend to agree with Obama on the value of thinking of oneself as part of “something larger,” and of not allowing one’s ethnic background to determine one’s “life chances.”
amardeep on August 9, 2007 10:03 AM in Issues · T·r·a·c·k·b·a·c·k address · Direct link · Email post
Ennis, One of those counties is Los Angels County, which has 7 million people from minority groups -- more than the entire populations of 38 states! (Washington Post, on the same census results).
Hm, perhaps this stuff should have been in the post itself...
I wonder whether the gaps will be closed somewhat in the next generation? I mean, I'm from a small white town, and civic engagement/social fabric is high just because everybody knows everybody else-- they all grew up together. (My grandparents' generation, in the same town, was much more divided by white ethnic identities-- the Poles, Italians etc. all had their own social clubs.) My husband grew up in Long Branch, NJ, and he says that while his parents stayed segregated into Italian, Jewish, Black and Puerto Rican communities, the kids all grew up playing together and now things are much less segregated. Maybe it's a matter of just knowing one another from an early age... maybe the diverse crop of kids now going to school together will think the current segregation is anachronistic and weird.
I hope so.
It is a mistake to assume that progressives would be more likely to embrace diversity versus conservatives. The two are equally likely. It is a characteristic of humans regardless of political ideology, race, or nationality. Take a look at the recently concluded YearlyKos convention in Chicago. All progressives yet the great majority of the attendees were white.
What a coincidence! I'm listening to On Point where Tom Ashbrook is talking to Robert Puttnam. Will write more later.
I'd like to think that the largeness of the community is as much of a factor as its diversity, if not the defining factor itself.
Amardeep, do you have any info about the civic engagement of smaller diverse communities vs. cities?
Here's another wrinkle. I temped for a handful of different companies when I lived in Mpls, and although the city itself had a variety of population groups, the employees with whom I worked were nearly exclusively white. Point being: if the city itself is "diverse" but a particular resident's life is structured in a way which does not reflect that diversity, how then does diversity become an issue in that particular resident's civic engagement?
That should be "Putnam."
It is a mistake to assume that progressives would be more likely to embrace diversity versus conservatives. The two are equally likely. It is a characteristic of humans regardless of political ideology, race, or nationality. Take a look at the recently concluded YearlyKos convention in Chicago. All progressives yet the great majority of the attendees were white.
Two thoughts -- first, with the reaction to Putnam I'm not talking about all conservatives but the anti-immigration conservatives, who explicitly want America to be more homogeneous.
Second, I see what you're saying about YearlyKoss -- I've been watching a lot of videos from the event at the Times website and places like TPM, and you're right: pretty much the only person of color around is Barack Obama. Some of that might be self-segregation, though: some of us here at SM (not all), for instance, *could* be Kossacks, but we've chosen to spend our blogging energy focusing on Desi issues on our own, smaller blog. Other progressive bloggers who are minorities also tend to focus on issues pertinent to their own respective groups. (One thinks of blogs like Angry Asian Man and RaceWire)
The challenge is to have a presence in both worlds. If I had time and energy, I would start a diary at Kos where I would cross-post posts like this one -- and see if the other Kossacks would invite me into the club.
More in the Guardian. The comments are worth reading.
There was also an interesting piece on Putnam's research in the NYT Magazine in June titled "The Way We Live Now: Idea Lab; Home Alone" by Erica Goode. (It is not available in it entirety online.)
1. Is there a "tipping point" at which diversity changes?
2. How much does within-area segregation factor into this?
3. Are there counties that are "exceptions" to this? If so, do they have different integration programs or services?
In development economics, many have argued that increased diversity (measured as ethno-linguistic diversity) leads to decreases in public expenditures. Ted Miguel did a study ("Tribe Or Nation?") a few years ago comparing and contrasting public goods spending and civic participation among two similar areas of Kenya and Tanzania. His results were really interesting, in my opinion -- the differences were statistically insignificant in Tanzania, whereas they were clear in Kenya. What was the difference? Both how "civic sense" or identity was built post-decolonization, and how ethnicity was leveraged (in good and bad ways) politically. what I took away from the study is that it isn't diversity per se that drives down participation, etc.; it's how this is then used to either exclude, include, or marginalize groups through policy, public goods disbursement, and political voice.
The link to Putnam's article is broken, or at least doesn't work for me. Says "cookie absent."
If anyone else has problems, this link will get you there, and if not that, this one.
Of course, now this means I actually have to read the entire article. ^__^
On the question of trust in societies with minorities, is anyone aware of any work done on how the fraction of a certain minority affects its perception? Is there some sort of critical mass that makes the group small enough to feel like a minority but big enough to get on the majority population's radar as a group identity? The reason I ask this is that extremely small groups (much less than 1%) (ethnic/racial/behavior-based) generally assimilate and sufficiently large groups (30% or more) don't feel any sort of group-based insecurity. Is there some sour spot in between, like in the 10% range, where group-based conflicts arise?
Also, w.r.t. the Guardian article and its comments, I do think that the issue of "uneven distribution of immigrants" is interesting, also. Some of the states where you see the highest number of immigrants (of all sorts) are also the states where you see some of the most racist and severe anti-immigrant sentiments. I wonder how much of this is about "diversity" and how much of it is really about xenophobia.
You also have to consider what role an individual has in affecting the level of diversity. In a workplace environment, hiring managers and other employees will have some say in who is hired. That feeling of control or participation in selected new members can be comforting. But is it really "diversity" if say, an American architectural firm hires top-notch talent from India, Spain, or Japan. After all, wouldn't their common profession weigh far more than their ethnicity?
OTOH, in your neighborhood, you have no say who your new neighbors will be. If you got particularly close to one set of neighbors, and then they move, to be replaced by someone you may not be as close to, you may resent it. This can be seen in both upper class neighborhoods such as Great Neck on Long Island, where an increasing number of Persian Jews has upset some of the older Jewish residents that are not Persian. Or in gentrifying neighborhoods, such as Chicago's Humboldt Park - where the Puerto Rican residents don't know how to feel about moneyed white yuppies moving into their neighborhoods and thus raising their rents.
Looking outside the U.S., if you look at countries where there is little or no ethnic diversity, such as Japan or the Scandanavian countries (recent Muslim immigrants notwithstanding), there is greater degree of trust among neighbors. By contrast, Indians often try to manage their diversity in their own way, such as setting up apartment buildings that are vegetarian only - or neighborhoods that are solely Marwari or Parsee.
As I've argued in the past, the phrase "Unity in Diversity" is about as meaningful as "Workers of the World, Unite." It's something that people say so as not to appear rude to others. Increasing ethnic diversity is neither a cost-free benefit nor a harbinger of cultural decay. A public policy response should aim to promote the benefits while discouraging the downsides.
Progressive is a loaded word and I think one that implies an orientation to the left of the Democrats. But in any case I will have to say Democrats & progressives are more likely than Republicans to agree publicly with the statement "Ethnic/racial diversity is a good thing". What people actually believe is a whole other matter, that a bunch of Ivy League educated candidates can consistently say stupid things about Indians implies some insularity. With the Kos convention or for example Burning Man, the overwhelming whiteness has more to do with self selection based on income, interests etc than any kind of anti-minority bad vibe
Looking outside the U.S., if you look at countries where there is little or no ethnic diversity, such as Japan or the Scandanavian countries (recent Muslim immigrants notwithstanding), there is greater degree of trust among neighbors. By contrast, Indians often try to manage their diversity in their own way, such as setting up apartment buildings that are vegetarian only - or neighborhoods that are solely Marwari or Parsee.
KXB, yes -- there is an interesting US/India comparative discussion of diversity at Ruchira Paul's blog; I thought this comment by Ruchira was particularly good.
At least to me, on the surface, this study points to what many economists have been implying all along that economic policy, which depends on sweeping social consensus (socialist economic policy) isn't effective for culturally diverse nations because it is very difficult to build such agreement for a top down, central system, that tries to serve this diverse population's needs on an individual level satisfactorily.
Seeing things for what they are, a culturally diverse nation has strengths just as one that is homogenous. Each needs a system that maximizes their abilities to prosper, while trying counteract the pitfalls. India is a classic example, with far more development and economic growth happening as socialist policy has slowly been discarded in favor of more open market competition (using the diverse talent to generate growth, rather than trying to solve a diverse set of challenges with 'one' method.) A crude analogy would be sports, where in American football you have several different positions with incredibly different levels of talent at specialized positions, while in soccer players can essentially play different positions (maybe not as effectively, but you can still play the spot). An offensive lineman is good at blocking, not throwing the ball 50 hards down the line with pin point accuracy, or running for a catch, or punting the ball, etc. I know, it's a crude analogy.?
comes from the refusal of many white progressive organizations/individuals to confront racism in their work
what would you want white progressives to do concretely? my own personal impression is that many white progressives engage in unctuous tokenism and symbolism on the theme of race. that being said, there's very little candor in talking about race in america today so most of these types just want to say the right things and move on with their lives.
(using the diverse talent to generate growth, rather than trying to solve a diverse set of challenges with 'one' method.) A crude analogy would be sports, where in American football you have several different positions with incredibly different levels of talent at specialized positions, while in soccer players can essentially play different positions (maybe not as effectively, but you can still play the spot).
can you be more concrete in how ethnic diversity plays into this? i.e., can you enumerate the strengths that various groups bring which makes us different from a nation like japan or china which are ethnically homogeneous?
The danger is that anti-immigrants will likely see a solution in reverting back to homogenous societies. In reality, we need to find out what can be fixed or worked on (institutional/structural racism, civic engagement) to increase social capital.
And Amen Camile@ 10: "what I took away from the study is that it isn't diversity per se that drives down participation, etc.; it's how this is then used to either exclude, include, or marginalize groups through policy, public goods disbursement, and political voice."
In reality, we need to find out what can be fixed or worked on (institutional/structural racism, civic engagement) to increase social capital.
diverse societies are so much work! i'm lazy, perhaps that's why i think we should reduce immigration levels quite a bit to allow for some homogenization.
At IALI's Democratic Dialogue this past Saturday, I was asked over and over again, why we weren't at YearlyKos. It reminded me of how we got shut out of CNN's election night watch party, which was held five minutes from my apt; "I guess you're not important enough, or you'd be here. We invited all the significant bloggers", their obnoxious intern said. I remember wanting to punch her out. I knew for a fact that our readership was higher than that of several of the bloggers "inside".
I went home and wrote this ridiculous, rage-fueled manifesto about it, right after walking away from her...and then my hard drive died. I mourned that post for two days. And I still want to punch that girl.
Of course, then there's Scott Page's argument about diversity:
Page shows how groups that display a range of perspectives outperform groups of like-minded experts.
this is certainly true, and anyone who has worked in a lab group understands this. in fact, there are arguments that one reason japanese under-perform in their creative science, their top-down lab culture doesn't leverage the full range of ideas and opinions (one must defer to seniors). there are japanese scientists who came to work in the USA specifically cuz they didn't think their ideas would get anywhere in their scientific culture.
p.s. the above of course doesn't necessarily imply diversity of race, ethnic group or religion. 80-90% of national academy of science members in the USA are atheists & agnostics for example. not very religious diverse.
Amardeep makes the point that the high level of intermarriage between Indian Americans and other ethnicities is one way that integration occurs and ethnic difference is reduced. Do you think that in a two or three generations there simply won't be any sizeable number of Indian Americans, because they will have become fully assimilated? As opposed to African and Hispanic Americans who are much larger in numbers and thus a more strong and pervasive culture that means even with intermarriage, a strong African American and Hispanic culture will exist that simply won't be there for fully assimilated Indians? In short, is the future of the South Asian American community simply being white?
Let me preface this by saying diverse talent brings strength. Diversity for the mere sake of it doesn't add anything to the mix. There has to be something of use to society that the newcomers bring to the local population that outweights the 'drag' that they put on it. I don't have any numbers or data off hand, so can't make a scientific argument beyond a hypothesis.
One industry off the top of my head stands out - music. Most of American music has its roots in African Amerian culture (Blues, Jazz, Rock n Roll, Rap, Hip Hop, Soul, Gospel) and it has flourished very well. However, diversity for the sake of being 'different' doesn't add squat in terms of value. You're adding more dissent to the mix for no other reason. But if you're adding distinct and diverse talent unavailable in the local population (the brain drain to the United States), it has benefits. The mass waves of immigration in the USA previously coincided with the availablilty and need to tap resources, which needed labor, that the local demographic could not supply and technology could not replace.
Maybe in today's global community, physical diversity is overrated since innovation and information can cross boundaries without the need of physical presence (in the global market), but years ago that was not the case. Japan has done well, but post WWII American influence in the form of a occupation definitely added some cultural change, even if the people aren't physically present anymore. Japan also has a shrinking pool of young folks. They've used technology as a crutch, but do you think they'll be able to maitain economic growth in the future the way they're going?
Pimping ain't easy, neither is being #1 by trillions. But I'm assuming there is a point of diminishing returns. As much as I like my free markets, you do need some social consensus in any society. Controlled and minor variation helps growth by adding distinct talent, variation out of control results in an unstable operation unless you hammer everyone into a homogenous mix (We are the Borg).
Razib: How do you reconcile your desire for homogenization of a population with an increase in socialist models?
Pali, speaking briefly, no. :) Please see
To be honest, it's not so much that I want tokenistic overtures, or even for them to "say the right thing." I think it's more about either being willing or unwilling to accept that there are often racial dynamics in "progressive" grassroots organizing. I'm thinking specifically in the Bay Area context, just for reference. Sorry if this isn't clear. It's easier to observe than it is to describe possible solutions, but on the other hand, I don't know why people of color are expected to be responsible for coming up with actions for white people to take to mitigate institutional racism. That said, I have plenty of white progressive friends who are awesome and great at incorporating anti-racism into their work and lives, and most of the time it's because they feel a direct connection or investment in issues of diversity. Sorry razib, I just spit out what I aws thinking, but I'm a little brain-dead today on backing it up =/It's a bit along the lines of "I'm liberal/open-minded/inclusive/progressive/whatever, so now all I have to do is sit back and wait for the diversity to show up."
Which it doesn't, because it's not that simple.
I'm in the middle of writing a post (for my own blog) about how this works in the theatre (a microcosm, of course, of the real world ^__^). So it's been on my mind lately.
Camille, I think your statement that "I think it's more about either being willing or unwilling to accept that there are often racial dynamics in "progressive" grassroots organizing"is right on, although I would add "willing or unwilling to notice."
As opposed to African and Hispanic Americans who are much larger in numbers and thus a more strong and pervasive culture that means even with intermarriage, a strong African American and Hispanic culture will exist that simply won't be there for fully assimilated Indians? In short, is the future of the South Asian American community simply being white?
i've argued this with manish & v-man before. the key is the immigrant stream which will replenish the community and the socioeconomic characteristics. i think if brown americans remain upper-middle-class then assimilation is more likely than if not. the chain migration of less intelligent siblings via family reunification is diminishing this profile, but not as much as some think.
Absolutely. I think it's direct experience that changes people... During the UPS strike in 1997 (about which my friend wrote an enlightening book) management tried to divide the workers by using racism to play them against one another. The workers realized what was going on and that they would have to come together if they were going to win-- and in the process a lot of white people who had held racist ideas had to let go of them.
The white progressives who don't confront race are the ones still laboring under the delusion that racism doesn't hurt them or affect their lives... not an easy stance to maintain if your friends, workplace and neighborhood are integrated.
My husband used to have a poster that quoted an Aboriginal activist who said 'If you have come to help me, you are wasting your time. If you have come because your liberation is bound up with mine, then let us work together.'
And with that I will stop making this discussion about white people :)
my own exp. as a center-right brown with my white liberal friends (ie., all of my friends aside from those i've met via SM) is two fold:
1) they think racism is operationally a conservative problem much of the time despite notional nods to structural differences of power
2) they have less problem being racially insensitive if they perceive you aren't on the same political page. i.e., you deserve it for being un-progressive.
2) they have less problem being racially insensitive if they perceive you aren't on the same political page. i.e., you deserve it for being un-progressive.razib, this is totally right on, and I think it demonstrates, also, that racism is kind of a "cross-political belief" issue. Whether or not it's given importance, I do meet way more people who will judge someone (in racist terms) for not being "down" yet is totally perpetuating racial stereotypes, hierarchies, etc., in their own daily actions. I think some folks feel a sense of "white guilt," and by projecting the blame outwards it helps them establish what they feel is their street cred. When I was a resident assistant in college this actually was a huge dividing point among the staff. About 30-50% of the staff were people of color, and we had a mandatory diversity training. Afterwards, many of our white colleagues expressed discomfort, unhappiness, didn't understand why race was being "rubbed in their faces," said they felt guilty, etc. It ended up being a "counsel the white folks for feeling unhappy" session, but the good thing that came out of it was that the POC folks on staff were pretty direct (and kind) about saying, "Don't feel guilty, do something about it, and if you need resources, we've got them!" It actually ended up, for most of us, to be a really great growing moment.
'Immigrant problem'?
It seems the problem is racism (towards immigrants, and others) and not immigrants.
This is like those who oppose lesbian and gay couples adopting children because the children would be stigmatized/teased. Well the problem is homophobia, not LGBT adoptions.
*haven't had a chance to read all the comments, so forgive me if this has already been said.
Self segregation is the essence of it. There is something inside of each of us that selects for certain things and eliminates others. Self-identification becomes stronger in diverse neighborhoods versus homogenous ones. We tend to want to hold onto those things that make us who we are. Imagine the reverse of this. If our ability to assimilate was strong, we would immediately abandon our culture and identity upon joining a new community. If this were the case, no culture would survive to be passed on to future generations.
Very interesting article.
Other than anecdotal evidence(and warm feelings on the part of the practitioners), was there any basis prior to this finding that diversity in itself led to greater civic sense ? Or is this the first study done on those lines ?
To clarify futher - ethnic diversity doesn't play into this directly, it just happens to be that along with immigrants of the particular era, came resources of a certain skill/talent level and they just happened to be [insert ethnic group here].
The 'tards who came up with study wouldn't know what a covariance matrix is if it hit 'em on the head. Just because cov(X,B) is positive is unsurprising if cov(A,X) and cov(A,B) are both heavily positive. X = diversity, B = stupid conclusion. A = something more highly correlated with the stupid conclusion and reasonably correlated with diversity (e.g. "Big City" or "Care-free"). Wake me up with when these "researchers" get whacked with a statistics cluestick, Singaporean-style.
Also, the definition of "diverse" is reverse-engineered from the conclusion. A french, a german, and an italian all living on a street are considered "the same" for the reason that they "get along" - introduce a gay person or a mexican and they're now "diverse" because there's somebody who (a) others don't trust, (b) is segregated or stratified apart, or (c) a combination of the previous. Rather than these ill-defined terms like "diversity", the report should show partial-derivatives with respect to well-defined dimensions like gender, sexual orientation, education, et cetera. I bet a "homogeneous" community of WASPs but stratified as union workers and yuppies would be just as dysfunctional as any other non-integrated community.
What bad joke is this setting up?
Seriously though, excellent points re statistics.
Ah, Simpson's paradox. Sst, Putnam's article is full of conceptual problems (the statistical problems are completely separate). In fact he could not answer questions about his original paper similar to the one's you are asking here at the annual meeting of the american political science association meet a couple of years ago. Also he takes "trust"(which is part of his definition of "social capital") as a static concept which creates as whole lot of problems. given all these problems (including the one you pointed out), if you put a normative spin on his logic, he seems to be saying that there is nothing "wrong" with certain behaviors that can--in some context--be called racist, because people are like that and it is fruitless to expect them to change.
Many typos. I meant "he could not answer questions similar to the one you are asking here about his original paper"
Cautiously, a few points (before I reread the paper 2 or 3 times; which is my usual habit before I attempt to critique a scholarly paper):
Most of the figures/data presented are ecological studies. In my field (health/medical/vet research), there is only so much grudging respect we give ecologic studies due to the problems with ecological fallacy.
'Table 3--Predicting trust in neighbors from individual and contextual variables' presents the results of a multivariate regression model. I have two concerns here. 1) They have gone on a fishing expedition predicting multiple dependent variables and then chosen one model to be presented. If you look at 20 things, the odds of at least one of the 20 offering you the answer you are looking for will be high. Moreover, the model that was chosen has a R-sq of only 0.26. That is, only 26% of the variability in trust in neighbors was explained by this model. My thought is that it is a weak model that tells you that there are several more explanatory variables that can and should be entered into the model.
I am struck by the first-person narrative style and the personal anecdotes (e.g., story about daughter/granddaughter) that abound in a scholarly paper. Is this common for political science and/or this journal, or is this because this appears to be a delivered lecture?
Mistake in terminology in my above post. It should be 'ecological studies' (synonym: aggregate studies; an accepted study design) not 'ecologic study' as I have written above. (i.e., not to be confused with studies in the field of ecology).
My mistake.
Where is the ecological fallacy in his work - I haven't read the paper, but from what I know of it, he has individual level data. The ecological fallacy comes from imputing individual level characteristics from aggregate behavior, that's not what he's doing here (if I understand it correctly).
Again, I haven't read this paper but I do know Bob in general. When he's good, he has checked most of the variables you have in mind and found similar results for models using all of them. I don't know how thorough he was this time, but his oral comments suggest that he may not have made that simple mistake:
Again, I'm speaking without examining the paper and the work underlying it, just having read this.
I have a hard copy of this on my desk and I read it once fully.
Look, I have a paper using aggregate level data myself. We all make do with the type of data available to us but we also appropriately caution that these are only exploratory analyses. I am sure your Bob has also done that since you say that he is a good scientist. A good scientist does not have the perfect data but she manages with the type of data she gets her hands on, and then brings up the limitations in methods and inferences.
I am glad I don't know him personally, because if I did I would not publicly critique his paper. Neither would I publicly critique any Harvard Prof if I was in his field, because I have absorbed appropriately that academia is a small world. For this same reason I have never publicly critiqued my colleagues' or my PhD advisor's weak papers. Not yet, at least. Maybe I never will except in the Introduction to a paper I would write myself, improving on the original study design.
Please read your Bob's paper and discuss with me if any of my points about the (methodological and quantitative) weaknesses in his paper are invalid in your opinion. Now that would be constructive criticism towards me and true respect towards Bob.
Why would you, a Political Science Prof, read only this and not even be aware of this?
And why are you jumping on me alone when there are at least two people before me who are also cautioning people to interpret the findings in this paper with some necessary nuance?
I have read the paper. I have some fundamental problems with studies of such kinds (so, to be fair to Putnam the following does not only apply to this paper), which mainly have to do with the (mis)use of statistical models. In many sciences the statistical models are derived from theory, and this is how it should be done. But in some social sciences (actually most of them), there is no underlying theory to speak of with clear quantitative specifications and functional form of variables. People don't seem to realize that statistical models make a series of assumptions about how the data is being generated and its functional form; such assumptions (such as the assumption about errors) are more or less metaphysical depending on the extent that they are actually tested. Based on such assumptions, models generate parameter estimates that predict a change in the dependent variable when independent variables are intervened on. All this is completely independent of the theoretical hypothesis being ostensibly tested. The causal theory might not warrant the statistical assumptions and the exact quantitative parameters estimated by a statistical model. So I think (I actually learned this from some good statisticians; I should say "they" think, and now that I understand them, I agree), regression is all right as long as it is used to summarize the data, i.e. used as descriptive statistics where the causal weight is carried by the verbal (or other) arguments, but it does not suffice as a causal argument when the interpretation of causality rests on inferential statistics and the statistical significance and size of coefficients (when the functional forms of variables are not specified by theory and the statistical properties of the error terms are not known).
ZING!
malathi..sigh!
is there any way to summarize your arguments for someone who had to take "introduction to social science statistics" over...and over...and over...and over...?
sigh!, I haven't read his full paper, but I do find myself wary of the design. I think it's generally difficult to apply a regression analysis to these kinds of things without setting very clear and reliable variables, etc., etc. I liked your point re: the descriptive, not authoritative, value of stats.
murali, if I may try to sum up the main critiques from sigh! and malathi, I'll give it a try (these kind of build on one another):
1. To effectively use statistics, you have to make a lot of assumptions about how you collected information, but also, about how people behave in the world.
2. That said, if you're going to try to see if something relates to something else, you have to look for all the factors that could also explain your question.
3. When you pick these factors, be careful that you didn't miss one, or that they're not interacting with one another.
4. If you collect everything, and your findings still don't explain most of what you "observed" (i.e. collected through interviews/surveys), then maybe your equation is not so good.
5. You cannot take individual observations and then try to "back them up" to explain the whole. To use a "logical" analogy, what's true of the whole is not always true of the parts, and what's true of the parts is not always true of the whole. For example, I could say my blood looks red. I have blood all over my insides. That does not mean that I am red.
I don't know if that helped at all :(
Camille, that helped me at least -- I too missed out on Stats 101, and have been regretting it ever since.
heaps actually. I find it easier to understand, however, the motivation to protect oneself in the world of academia, from the inevitable blowback that results from openly criticizing such a luminary as Putnam, or anyone with sufficient clout/leverage to derail your own career.
I think this is why i tucked my postcolonial tail between my legs and scurried over to the private sector, leaving my dreams of one-upping Ashish Nandy in the dust.
I'm glad I could help :) And just to clarify, if any of those 5 things is "missing" or "wrong," then basically your entire "model" (i.e. statistical equation/question) falls down. I think this is where sigh!'s comment about stats being a good description, but not necessarily good proof, comes in :)
Yeah, I don't think I could do academia, although sometimes watching professors duke it out through papers is like watching some kind of nerd-driven, crazy Academia DeathMatch! I'd rather be a spectator :)
Moreover, the model that was chosen has a R-sq of only 0.26. That is, only 26% of the variability in trust in neighbors was explained by this model. My thought is that it is a weak model that tells you that there are several more explanatory variables that can and should be entered into the model.
i have been told that r-sq of 0.26 isn't bad in a social science, where of course there are multiple variables.
D'oh. Discipline. Varies by discipline.
re: stats. they are a good guide and map which start you off on a deeper exploration of a topic. a bunch of regressions can be, i think, like scouts. they don't determine the outcome of the battle, but they allow you plan your moves. in any case, i do think that some of the skepticism for this result is a little much. plenty of people i've met through SM express the attitude that they are more comfortable around their own kind, people who could understand their values naturally because they were of the same ethnicity. e.g., the experience of being the "brown kid" in an all-white school. the basic human impulse to group and feel anomie seems obvious. how one organizes and perceives the world is of course relatively fluid. 100 years ago WASPs were terrified of irish catholics in the northeast, but today they are probably relatively comfortable with white ethnics (and even elected one to the presidency in 1960). so things change. in modern american society there is a promotion of diversity and its benefits, but a lot of the time it reminds me of theists who talk about how great god is. you know the answer before you start asking questions, so of course you will be skeptical or confirmatory in your biases depending on where you stand a priori. the same obviously goes for anti-immigration conservatives; they love putnam's conclusions, but they're less interested in comparing the scientific creativity of the united states and japan for obvious reasons.
D'oh. Discipline. Varies by discipline.
yes. i knew a physics student who took an ecological statistics class once. his first lab report was a war & peace sized analysis of error and variance in the data. didn't realize that that was just normal.
btw, i betcha diversity is a "robust" predictor.
Please forgive an outsider for listening in, but this is the most serious and thoughtful of the dozens of blog discussions of my paper that I've read so far. One or two comments from the author:
(1) It is entirely fair game to ask tough methodological questions of my paper (like anyone's paper)--individual-level vs. aggregate-level analysis, magnitude of Rsq, statistical controls, causal direction, endogeneity, and so forth. The last thing I would do (and if one or two of you know me, I think you'll confirm this) is "retaliate" against someone who found an error, because I really believe that science is a dialectical procress in which we catch and correct one another's errors. The only thing that I ask is that you actually read my paper carefully before offering methodological judgments of it. For example, someone asked if I cherry-picked the one significant result of out 20; if you read the paper, you'll see that I carefully report all the outcome variables--those that fit my summary conclusion and the exceptions. Please don't just guess what errors I might have made, but tell us ones that I actually did make.
(2) In the very interesting discussion about whether my findings are right or not, and if so, why, I slightly fear that the most important question got somewhat overlooked--how should those of us who see the net long-run benefits of diversity act (publicly and privately) so as to minimize the downside and maximize the upside, as someone above comments. The main point I wanted to make in my lecture was not about an "inevitable condition," but about a "potentially solvable problem." Comments on that issue will be gratefully accepted. Meanwhile, thanks for the thoughtful discussion.
Bob Putnam
Can't add anything intelligent about stats because I can't add- period. Camille is right on about non-profits in Cali. I deal with a number of them as an artist and curator- and some are very open to looking at race and the dialog about race that is implicit in Socal while some are more into humming happily together. I have this conversation with my students periodically and my 2 cents is basically this: I think the unease among various groups is temporary. What I tell my (ESL) students is that they basically see most of us as acting "American" in public, but in private, many of us are a mix of various things. The problem is what is "American" and I think that is shifting- the white mainstream is finding that it can no longer dictate or expect a specific, rather limited definition of, and that the flattening of identity that was common when I am younger is no longer being tolerated or even aspire to in larger, more diverse cities.
if it is ye, Mr. Putnam, please forgive the unfortunate insinuation I made--it was not really directed at you in particular, but rather at the less-than-considerate atmosphere at universities I attended, which tended to stifle the dialectical process you thankfully recognize as characterizing your discipline.
My glib response seems to be "make a diverse group of friends; develop genuine personal relationships with people of various backgrounds and include them in your social network."
The structural (e.g. "diversity training" in offices) will just be words and empty policy. IMHO it has to develop from genuine, personal connections.
Tricky to do, if we're being truly diverse. Because, as was mentioned somewhere earlier along the thread, even groups which include people of a variety of nationalities often finds itself segregated by economic strata or level of education.
First, thanks to Robert Putnam for responding/posting. About statistical issues, I might have a little more to add to what camille pointed out above later if I have time (briefly, there are two sets of assumptions that need to be justified, one has to do with the statistical model which should actually closely follow a theoretical model, which in the social sciences does not happen much. I might give you Gauss's original example tomorrow to show how it should ideally be done).
But more fundamentally (and conceptually), my bigger point is that if you asked a bunch of WASP New Yorkers similar questions in the 1920s, you would probably have come up with similar answers. Therefore the definitions of what constitutes "diversity" is subject to change (as some commentators including sst, have argued above). So I don't understand why we should be concerned about all this now (more than we already normally are).
But I'm very sleepy now...
Hmm. From your lips to Bob's ears.
Putnam's main new finding is that "diversity" doesn't just lead to inter-group conflict (that's old news) but also leads to intra-group mistrust (i.e. social anomie). Diversity leads to people hating everyone, not just the out-group.
Setting aside methodological issues, some thoughts:
1) I think some clarification of how "diversity" changes would help. Putnam's paper mentions how group definitions change over time (how the Irish became white, etc), but diversity can also change in an instant, once you add a new member to the nieghbourhood. For example, sometimes I find a (completely hypothetical) group made up of Punjabis, Gujuratis, and Tamils to be uncomfortably fractious and diverse. But throw a white guy in the mix and we're all on the same team. An all desi neighbourhood (from different parts of the various desh-es) is more diverse than a half-desi half-white neighbourhood. The statistical modell should (try) to account for that.
2) Putnam dismisses self-selection pretty quickly. He doesn't think that socially-isolated people would deliberately choose racially diverse neighbourhoods. But I pretty much hate community activities, and if given a choice between an all Desi neighbourhood, and a diverse one, I'd choose the latter so that i wouldn't have to meet all those freakin' aunties and uncles everyday. It's possible that misanthropist curmudgeons seek out diversity in order to avoid social contact. Putnam needs to do the statistical work before dismissing self-selection (this is also the Steven Derlauf critique published on isteve.com yesterday)
3) I'm also curious about what the "net long term benefits to diversity" are. I can see short term benefits (gains to a society from ethnic occupational specialization). But over the long term, if the American melting pot really works, diversity vanishes without continued immigration (excepting, always, African Americans). So if in the long-term we are all homogenous, what is the long term benefit of diversity?
4) How soon before this work is replicated in Canada, Britain, and Australia, all of whom have the same majority community but very different minority communities. Is ethnic diversity correlated with social anmie in those countries?
All that said, Putnam's essay, especially the last third, is a great discussion of the issues. I think everyone should "read the whole thing" before commenting. (yeah, fat chance!).
I read two really great articles yesterday, this one about Putnam's work being one of them. The other one was a book review of Gregory Clark's theories suggesting natural selection is part of what led to the Industrial Revolution. I know this might sound silly but if you read this article and then the one I just linked you get a glimpse of some grand unifying theory linking the two. Vinod has blogged about Clark before.
Professor Putnam,
Your five-year survey is a useful antidote to years and years of sentimental, happy-faced flapdoodle. But still, I take issue with the tone and knee-jerkishness of your essay. Some friendly words of advice:
(1) Don’t turn this issue into a mano-a-mano showdown between holier-than-thou liberals, on the one hand, and nativist yokels on the other. ‘Progressive’ is a name appropriated by those in power to describe themselves, and to portray dissenters as in some way unhinged or eccentric. It has no objective meaning. The current categories described as ‘liberals’, ‘progressives,’ ‘neoconservatives’ and ‘reformers’ all appear to me to be similar varieties of the same Fabian consensus.
(2) You’re a political scientist, not an economist. So stop rattling on about the supposed ‘benefits’ of mass, unfettered and especially illegal immigration. The consensus among the biggest names in the field – such as Thomas Sowell, Paul Krugman, Paul Samuelson, Gary Becker and George Borjas – reflected in the National Research Council’s ‘The New Americans’ back in 1997, is that low-to-semi-skilled migration does little for a $12 trillion economy. Please visit the Borjas blog and scan its archives, or alternatively, read Borjas’ academic papers and essays.
(3) And finally: Mexico is not going to ‘enrich’ us culturally. Their mores and manners, like mine, are too low-brow! As much as I enjoy the music of Jennifer Lopez and the tomfoolery of Bumble-bee Man (the Simpsons), it is not wise to manufacture more inequality within the Untied States for the sake of a few cheap thrills. As Charles Murray pointed out in his compendium extravaganza ‘Human Accomplishment,’ civilisational progress has depended on a few individuals of magnanimous ability. So a selective immigration policy is probably best.
Thank you for all the hard work.
MA Jones
Michael Jones,
I can only hope your last comment: "And finally: Mexico is not going to ‘enrich’ us culturally. Their mores and manners, like mine, are too low-brow!" was meant as a joke. If not, you may want to avoid places like California for the rest of your life because you might find a) the demographics, b) the "low brow" culture and c) the hybridity that's developing to be highly disturbing. Mexico has already enriched the US, if not you, in a number of ways from language to architecture to food to visual arts. Some of the most dynamic independent art is coming out of Tijuana (Nortec, Bulbo Collective). California architecture owes a great deal to Mexican tradition- without it - the arts and crafts style in Cali would have never come to its full expression. This country may indeed need some kind of overhaul of the immigration system but the type of selectivity you suggest, is a culture worthy of being allowed to immigrate is disgusting. Under that logic system, the Irish, for example, would have never been allowed in since the English considered them uncouth, uncivilized and speaking a bastard tongue.
Thanks for the link, Amardeep.
Having lived in very homogenous (as in white, in the US) and very mixed (both in the US and in New Delhi, India) neighborhoods, I still am not able to put my finger on what actually contributes to higher or lower levels of civic engagement.
There are probably a lot of things involved apart from changing demographics based on race alone. Not just in the US but also in India.
When I was young, most woman stayed home and sometimes they formed the bulwark of neighborhood watch groups and kept an eye on the upkeep of civic amenities. There was more social contact with one's neighbors. I have heard the same from my American friends who grew up in the 1950s and 60s about the state of affairs in the US. My own generation (and the younger ones) in India and the US is much more busy, with both spouses employed outside the home in many cases. I have noticed that more and more, socialization and activism both involve professional relations and concerns, moving away from neighborhood issues. Notably, this trend is more prevalent in Indian middle class neighborhoods (at least in Delhi, for which I can speak) than in the US. It is also noticeable, coincidentally or not, that the state of residential roads, drains and garbage pick-up in New Delhi, in the same neighborhood where I grew up, is a shambles. Any connection between socialization among neighbors and the civic amenities? I don't know. The ethnic mix of the area on the other hand, has not changed since my parents' days.
I will copy here something I wrote in the comments section of my own post to which Amardeep has linked, regarding my observations in the US:
You can slice this baloney any which way you like and make a compelling (if somewhat superficial) case for all scenarios. I live in a hugely diverse city. I also hear the words "third world culture" being thrown around a lot in regards to the way the city functions - often by those who themselves came from third world countries. The suspicion is that every population group is focused on its own narrow ethnic / cultural interests and there is no sense of ownership for the community at large. The suspicion is not entirely unfounded. But what is interesting is that the indifference to community duties or the lack of civic strength cannot be explained away entirely due to this "Tower of Babel" or the aggressive "salad bowl" culture of the city.
Many working class neighborhoods in Houston are ethnically segregated. We have distinctly African American, Vietnames and Latino neighborhoods, as also some low income white areas. But most affluent neighborhoods are very diverse. Guess where the civic amenities function well? It always comes down to how much is at stake for the citizens and what community they are invested in.
I think that the impetus to engage in civic activities is not rooted only in the expected outcome of our efforts. It is also connected to a "feel good" attitude - we want our friends to "appreciate" our commitment. It appears that there are several tangibles and intangibles as to why people will or will not do their civic duty like voting. The most important of course is how high the political stakes are. During acute financial, national or cultural upheavals voters turn out in large numbers despite severe obstacles of weather, mobility and time constraints. When not much is at stake on the political / national scene, the visible reason for voter lethargy is inconvenience - time away from work, driving, standing in line, bureaucratic calisthenics etc. But there are some other less obvious reasons why we participate in the political /civic process - we want to be SEEN by our neighbors and friends as conscientious citizens. A good deed done in anonymity is like a tree falling in the forest. So our tendency to withdraw when we do not know our neighbors is not just the suspicion of the "other" but also our own " lethargy and indifference" when we we consider ourselves invisible among strangers or away from the eyes of our friends.
An interesting example of this was when a Swiss canton (racially homogenous) tried to make it easier for its citizens and arranged for voting by phone from home. Percentage of votes cast actually went down! When voting from home via telephone took away the "look at me" incentive, many did not bother to vote.
OK, first of all, who gets to say what is low-brow? You? If you think that J-Lo (who is PUERTO RICAN) and Bumblebee Man (an American cartoon's joke about Mexican culture) are what Mexican culture has to offer, you have absolutely nothing to say.
Second, you're quoting the authority of Charles Murray, the renowned theorist of the racist right? His credibility disappeared entirely with The Bell Curve. Read Stephen Jay Gould's The Mismeasure of Man before you trust anything Murray has to say.
Camille, thanks for laying it out there for the stats-challenged among us.
And Bob Putnam, thanks for coming by to talk about this-- I agree with your emphasis on "minimizing the downside and maximizing the upside" and hope that you will be successful in counteracting the spin and oversimplifications that the right will no doubt impose on your results.
Zazou,
(1) ‘Mexico has already enriched the US, if not you, in a number of ways from language to architecture to food to visual arts.’
‘Enriched’ is a positive, feel-good adjective. But Mexico’s contribution to cultural life in the United States has either been neutral or negative. California’s main creative industries – Hollywood and Silicon Valley – employ few Latinos above the technician level. And although Mexicans are by far the biggest immigrant group, they don’t even rank among the top 20 immigrant groups in the U.S. in terms of patents awarded.
But even on the plus side, you get things muddled-up. Mexico’s main strength isn’t its visual arts or architecture. (Most architects reject populist Latin American styles for a Euro-minimalist look.) Rather, it’s literature – novelists such as Garcia Marquez and Vargas Llosa. But then again, La Raza's literary talent is unlikely to ‘enrich’ the US: 65% of Mexican-born newcomers don't finish high school.
(2) ‘California architecture owes a great deal to Mexican tradition- without it - the arts and crafts style in Cali would have never come to its full expression.’
That’s a lousy mea culpa for mass immigration. Mexico’s most valued contribution to California is Rey Mysterio – let’s face it! Asian-Americans have adapted well to middle-class American life because of their nerdy-knack for IT and business. Hispanics in America, in contrast, are 2.9 times more likely to end up in jail than white Americans and 13 times that of Asians. The only ‘enrichment’ coming out of la Raza is a snazzier selection of semi-automatic weaponry.
(4) ‘Under that logic system, the Irish, for example, would have never been allowed in since the English considered them uncouth, uncivilized and speaking a bastard tongue.’
So what are you suggesting? That we invite the laziest, most untrustworthy, least productive and crime-prone people on the planet to settle here just to prove how ‘enlightened’ we all are?
You would think, would you not, that the aim of letting in more people would be to benefit the already-existing population – not to make happy-clappy liberals feel better about themselves.
Troll alert!
Michael A. Jones, please take your racist commentary to another site. Your anti-Latino vitriol is not welcome here.
Dr. Putnam,
Thank you for patiently responding here. I feel uncomfortable about my semi-anonymity (as opposed to your open identity), but please be assured that had I been present at your lecture or talk I would have brought up exactly the same (above) questions/concerns during 'Q&A' and you would have been happy to answer me because you would understand that a few others may have had the same questions. You would have considered it an opportunity to put their minds to rest (or work further, as the case maybe).
I am a researcher myself, so I instinctively understand that no one knows the data as well as the person who has been looking at it for the past several months/years. In fact, no single person knows it as well as the grad student who lived/dreamed/bathed/breathed it 24/7.
Have you tested to see if the contribution of all ethnicity/homegenity variables is significant? In SAS you would do this by introducing a test command after the model statement. It would read 'test_ethnicity: test black, Latino, Asian, Index of Homogeneity, Spanish speaker;' A significant F-test would mean that the collective contribution of one or more of these variables is significant. Another way to think of this, is that there is a significant difference between a model with black, Latino, Asian variables as compared to a model without them, i.e., there is a signifcant difference between the 'full' model and the 'reduced' model.
If you have already done this and reported it in your paper, I apologize for missing it during my reading yesterday. The narrative structure/style of the paper is something I am not used to. But I respected you enough to try and read the paper once, putting aside my own work. Which is more than can be said for at least one of your own colleagues, apparently.
It may be a discipline-specific thing, but with a model that has a R-sq of 0.26 I (as an epidemiologist who also uses observational data from the real-world) would write a paper and get it published but would rarely be allowed (by my seniors) to let the results escape the halls of academia. Perhaps, that is my main, or only, concern. I respect the role even descriptive/exploratory analyses play in knowledge development/translation; I just worry about how they will be used prematuredly by those who don't recognize the (admittedly fluid) boundaries between descriptive and analytical approaches.
Thank you, once again, for your patience.
Malathi, I think you are better off treating the results as descriptive rather than inferential (so ignore significance etc.). The model is not really a controlled experiment where assumptions about errors can be made more reliably (and statistical models basically simulate controlled experiments); the statistical model has essentially been "bootstrapped" from the data with all the attendant BLUE assumptions (again I should mention that to be fair to Prof. Putnam, this is not only a criticism of his paper and applies to all such social science papers). But treating the data as descriptive would not necessarily detract from Putnam's verbal argument about causality (just that that statistics cannot carry the causal weight).