Academic spin

Earlier in the week, The Australian published a story about Harvard academic Robert Putnam‘s research into ethnic diversity and trust. It reported that:

His extensive research found that the more diverse a community, the less likely were its inhabitants to trust anyone, from their next-door neighbour to their local government. People were even more wary of members of their own ethnic groups, as well as people from different backgrounds.

Now this in itself is hardly suprising. It is intuitively plausible, since the less you know or can infer about someone, and the less you are able to deliver social sanctions through social networks, the less rational it is to trust them. Andrew Leigh (who has worked with Putnam in the past) has already written a good study of it, reporting some international empirical work and adding Australian evidence. This story should just have been telling us that we were about to get some interesting extra detail. But instead it suggests that Putnam himself should be treated with some intellectual distrust.

The original Financial Times report said:

Professor Putnam told the Financial Times he had delayed publishing his research until he could develop proposals to compensate for the negative effects of diversity, saying it

32 thoughts on “Academic spin

  1. I agree with you that it doesn’t look good. Also that Andrew Leigh is a fine scholar (and nice guy too). While I’m always happy to think ill of lefties, we should really wait for the book to come out before judging Putnam. What he has said to the FT might not look good, but there is nothing wrong with ‘sitting’ on results and pondering them – or wanting to flesh out the details and implications of research findings. (Indeed, many University codes of research practice now require acdemics to consider the downstream consequences of their research – I’m not condoning this sort of self-censorship, but employees have a duty to their employers wishes).


  2. Andrew, Andrew Leigh admits he organised “closed-door” sessions in Australia to discuss Putnam’s findings. That is the same type of arrogant behaviour you commented on in the case of Putnam.

    Contrary to your view, it raises questions as to what else the publicly funded Leigh considers should only be discussed behind closed doors.


  3. It would be interesting to know the full reasons why he didn’t publish the work are rather than just those in the FT article (they might be quite reasonable as SD notes), and whether the previous problems Lawrence Summers (now “resigned”) caused had anything to do with it. It would be sad if even a university like Harvard is turning into Stalinland for free speech.


  4. well I’d like to know more about the study first because I’m not even convinced the results are sound (maybe it’s just the libertarian ideologue in me). Are they controlled for socioeconomic status, for instance? Campbelltown may well be an area of high mistrust because
    1) poor refugees making their start can’t afford to live anywhere else
    2) so they move into low socioeconomic status areas where there are already enough tensions because of economic uncertainty.

    On the other hand go to Chatswood where there are lots of rich Hong Kong migrants – is anyone there really more mistrustful of each other than anywhere else simply because you hear the occasional Cantonese conversation?


  5. Like Jason, I have some difficulty believing these results. Sounds very similar to the Amy Chua story ( I once did a back of the envelope regression that showed ethno-linguistic fractionalisation was not important once you controlled for good legal environment and good government and the like. When I get some time, I want to try expand that result and see if its robust etc.


  6. Sinclair/Conrad – I appreciate that there are many reasons for not releasing academic results (though mainly that they are not convincing or not interesting), and I should acknowledge that Putnam is now planning to release them. But he did put political reasons high on his list for not doing so in the past, and it would not have been difficult to construct a list of positives from ethnic diversity or to argue that ‘trust’ is not always an indicator that we need to worry about.

    Winifred – Andrew had, I think, by that time already released a working paper on this issue. He was simply meeting Putnam’s condition for speaking. There is nothing unusual or wrong about researchers being told about other people’s unpublished work on a not for citation until published basis.


  7. Andrew, usually that’s because the results are not finalised, and it concerns individual researchers.

    The issue here was that the secret meeting was a meeting of people, not a single researcher, and that the (publicly funded) researchers decided the public couldn’t be trusted with the results.

    There are serious questions raised by Leigh’s behaviour.


  8. btw the tenour of Putnam’s work has always seemed to me more nostalgist conservative than progressive anyway. Like the way his work on trust probably doesn’t recognise the importance of online networks and who cares if people substitute individual leisure for going to the bowling alley. I have the same suspicions about this latest stuff. to repeat – is the correlation between lack of trust and ethnic diversity simply a reflection of the fact that the first port of call for the types of immigrants the US gets, namely unskilled and semi-skilled Hispanics, is likely to be in the poorer neighbourhoods? what happens if you leave out the Hispanics and look at Silicon Valley which presumably has a high concentration of Chinese and Indian IT workers for instance?


  9. Jason – Since all we have is a few lines from a newspaper, we have no idea what Putnam’s research on this really says. There is a well-known relationship between low SES and low trust, but it would be truly surprising if there was no relationship between other forms of diversity and lower trust. There is analysis that does the fancy regressions to try to control fr other factors. It’s not just about migration, though migration gets extra attention because it is relatively amenable to political control.


  10. Winifred – Putnam is not funded by the Australian public, and indeed as he works for the world’s richest private university he may not be publicly funded at all. Funding is a red herring.

    People who attend seminars generally have no right to disclose other people’s research. I would never do so without their permission. The only issue is in covering up your own, politically inconvenient, research. Even here, other commenters have raised reasonable objections to my stance on Putnam. The case against Andrew L. is non-existent.


  11. It still sounds to me that the correlation insofar the sample is derived from US conditions is going to be piffle. Another reason – blacks.

    There is a whole history behind the black-white relations starting with slavery down to Jim Crow down to demoralisation by welfare which is going to colour (no pun intended) the feelings of trust between these two groups – and this is going to be picked up in his analysis.


  12. Most studies of trust try to control for the length of time people have been in proximity. That is, in rapid turn over areas, trust is low because nobody stays very long and the whole ambience is transient. If these areas also have high numbers of new immigrants (which would figure) the maths to pull these out is complex. It is easy to deduce that it is the cultural distance between groups that prodcues didtance. However social distance cannot act as a proxy for trust – and the trust issue is further confused by the bonding/bridging capital model that Putnam uses. Overall I find Putnam less than helpful in complex multi-factoral environments, where the issue of apprehension (or fear of difference and worry about hostile rejection) overlays many other potential relationships of sharing and inter-dependence.


  13. Andrew, thanks for the generous words. However, you won’t be surprised to learn that I think you’re being a tad harsh on Putnam. There are probably only a dozen academics in the world who could credibly say that they planned to come up with some good policy proposals. Bob is one of them.

    Jason, my Australian paper controls for neighbourhood income. Income is positively correlated with trust, but even controlling for it, diversity matters. As to Putnam’s work, he’s making more subtle points than you credit him with. Bowling Alone contains a chapter on the Internet and social capital. As to diversity, he’s not saying not that blacks distrust whites in diverse communities, but that in the presence of diversity, blacks distrust other blacks (and whites other whites). That’s new, I think.

    Winifred, there was nothing secret about the social capital symposium. In fact, I even posted the invitation on my blog. Putnam’s discussion was not for quotation (the Chatham House rule). That’s very standard in academia. Indeed, I’ve probably presented half a dozen papers this year at university seminars where I’ve asked the audience not to quote the findings until the work is final.


  14. I teach my students that they are responsible for the research they do and its consequences. This is particularly important in social research where nuances of questions, issues of timing and sampling can mean quite different responses. Where results can feed into societies that have been subjected to lots of dog whistle anxiety stimuli about crime and immifgration, it is hard to be sure that you are measuring distrust of diversity or broader distrust fiuelled by anxieties about inequalities, crime and corruption that flow into generalised distrust. As others have commented, USA areas of diversity are often transient and low income, so the measures need to be calibrated againts these factors.

    I suspect also the lack of bridging social capital in such diverse areas may also be fired by too much neo-liberal emphasis on markets that place risk on individuals, encourages self interest and undermines social cohesion. Generosity and respect for difference may be casualties of too feral levels of individual freedom.


  15. In terms of immigration policy. There are costs and benefits from diversity. Migrants who despise our democratic institutions, who are religious bigots, who treat half the human race as subhuman, who see violence as a reasonable way of resolving political differences and who are more than likely to live off the social security system (and crime) do not promote the diversity we seek. You don’t have to give too hoots about Eva Cox’s concerns to hold these views.

    Betts is mainly right. Australian society is not a social experiment conducted by elites. Current Australian citizens have the right to choose who they wish to have live in this country. Exercising this right is not xenophobia or racism.

    Putnam’s worries are appalling but mirrored in most Australian immigration debates. Many of the multicultural lobby regard Australian society as an empty slate that any form of riff-raff diversity will necessarily enrich. It isn’t true – we have a pretty good society with decent values, low levels of political violence and substantial cohesion. We have the obligation to be choosey and to refuse entry to those we don’t like and who we think don’t want to fit in.


  16. I’m struggling to see what the problem is here. Is it that political correctness about cross cultural sensitivities might be preventing the release of some report that confirms that poor people don’t like other poor people? Or is it a desperate search for some confirmation that racial segregation is good and should be returned to? Is the right so desperate to clutch at whatever straws are available to scientifically justify bigotry?

    People of all backgrounds will continue to want to come here because it’s a stable, affluent society, not necessarily because they want to embrace our values.

    I’ll tell you one thing though from personal experience: their *kids* want to embrace the values as long as you don’t keep pushing these people into racially segregated ghettoes, and the kids have a massive influence on the parents no matter what culture they come from. If somebody is brave/stupid enough to get on some leaky boat and make landfall, when they are escaping cultural, political or religious tyrranny, we should welcome them as heroes. If they are making an attempt to escape poverty, we need to open our houses to them.

    We don’t need cultural, political and religious tyranny here, although the conservative forces in this country are attempting to scare it into life. Society didn’t collapse when the first pointless pentecostal church opened in Australia (a seriously dark day), and it won’t collapse when they raze the house next door to yours to make a mosque.

    HC, if you want to start being choosey, I choose you to leave imy egalitarian, big-hearted Australia if you don’t like it.


  17. David, really. My basic contention is that good policy relies on careful analysis of the best information we can find, whether that conforms to our hopes – or our personal experience – or not. Weber’s ethic of responsibility and all that.


  18. What’s with ‘eva cox’ anyway?

    “Generosity and respect for difference may be casualties of too feral levels of individual freedom.”

    Which freedoms, I wonder, should we be considering cutting back on? Freedom of speech? How would restricting free speech increase generosity or respect for difference?


  19. The puzzling thing about Putnam’s statement to the Financial Times is that he was TALKING about this research at least as early as 2001. When I saw the Australian’s article I googled to see what I could find, and one of the items I came up with was an article in the Ottawa Citizen from 6 December 2001 which said he was about to present the findings at a conference the following day, and which quoted from an interview he gave. I would have thought that if he really believed it was irresponsible to publish without developing counteracting policy proposals, it was even more irresponsible to publicise them through the mediation of journalists.


  20. excessive freedom to pusue self interest, promote unfounded fears, compete unfairly and impose too many individual risks on others undermine trust.


  21. I don’t trust academics that think they know better advocating restrictions on freedom too.

    So be specific, Eva.


  22. Speaking as a journalist… if I were an academic, my first worry in publishing potentially controversial research is that I would be misrepresented by a media eager to concoct scandal and controversy… and that any labels thus placed on me during the ensuing three-day flurry of heat and light would last a lot longer than those three days!

    So I can see why Putnam might want to ensure that he has a balanced view to present before going public with it.


  23. Hans – Though as Ron Brunton points out, Putnam has now done the opposite – released enough for us to think that there is a problem, but not enough information to know whether it is a small or large problem.


  24. Andrew, perhaps. However, I’d still suggest that if I had research findings of this kind and I presented them on their without any proposed policy remedies, people (journalists) would assume that my proposed solution was to simply stop migration — that it was a straight critique of immigration and multiculturalism, rather than something more nuanced. Confronted with that probable outcome when contacted by a journalist, I’d try to come up with something on the hop (which is possibly what he did in saying we should “construct a new us”). Unsatisfactory.

    Also, we are assuming that he went to the media with his research. That’s quite likely; but it’s also possible that he’s been quietly discussing it in academic circles, as he develops the ideas and their consequences (and is there anything wrong with that? I would want to workshop my ideas before making them broadly public), and a journalist got wind of it and knew he/she could beat a story out of it. And Putnam made the decision, for whatever reason (naivete, politeness, opportunism), to talk to the journalist.


  25. Putnam has now published his findings for the first time (my take here).

    On the charge of “intellectual concealment and spinning”, I’m curious to know whether you’d now acquit him, or convict him of a second offence?


  26. Andrew – I have printed out the article (having seen the link on your blog) but reading it will have to wait until tonight. A verdict will follow….


  27. Andrew – It’s a good piece of analysis, though only the finding that trust of one’s own group is negatively associated with diversity is surprising. However when he told the FT that:

    “he had delayed publishing his research until he could develop proposals to compensate for the negative effects of diversity, saying it would have been irresponsible to publish without that”

    he need not have bothered with delay, because his ideas in this regard are just a few dot points with none of the evidence he supplies for the initial analysis.


  28. Your lack of surprise may reflect your priors. I would guess that if we polled the US and Australian social science organisations, a majority would take the view that ethnic diversity breeds tolerance and trust.

    And isn’t it the very nature of policy proposals that they’re not grounded in evidence, since it’s harder to evaluate things we haven’t yet done than things we’ve already done?


  29. “I would guess that if we polled the US and Australian social science organisations, a majority would take the view that ethnic diversity breeds tolerance and trust.”

    Andrew – If so, it says more about the prevailing group ideology than it does about their judgment as social scientists. Its hard to see the mechanism by which ethnic diversity would increase trust, since trust is presumably based on expectations that people will behave in various ways not inconsisent with one’s own interests, which requires knowledge of group norms and behaviour patterns (less likely to be the case with new migrant groups) and preferably enforcement mechanisms (more likely to develop in longer-standing communities where social sanctions can be applied).

    Ethnic diversity probably does often increase tolerance, but in the sense that it creates greater need for it since people are confronted with views and behaviour of which they may not approve. But it does not necessarily create ‘comfort with diversity’ – the evidence on this is very mixed, as Putnam’s article says.

    That’s the problem with his policy proposals – they assume that a little more mixing will make a big positive difference, despite the mixed evidence on this.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s