- Berghs Betraktelser

I senaste Political Studies finns en upplyftande tillitsstudie:

When Does Diversity Erode Trust? Neighborhood Diversity, Interpersonal Trust and the Mediating Effect of Social Interactions av Stolle, Soroka och Johnston.

Ett robust resultat i dylika studier är att etnisk mångfald brukar samvariera med lägre tillit. Så även i denna studie. Men - och författarna lyckas inte helt dölja sin förtjusning över resultatet - interaktionstermen mellan att ha grannar från etniska minoriteter och att prata med sina etno-grannar är signifikant positiv:

Respondents who have diverse neighbors and talk to them on a regular basis are significantly more trusting than those who have diverse neighbors and do not talk to them. (p. 68)

Författarna menar till och med att studien talar emot att det skulle röra sig om omvänd kausalitet, dvs att tillitsfulla råkar vara de som pratar med sina etno-grannar. De är väl inte helt övertygande, men de ägnar åtminstone ett stycke åt frågan. (Nix, ingen fancy IV-skattning för att slå spiken i kausalitetskistan, det är ju Political Studies vi pratar om :-)

En annan kul sak med studien: De har INTE anävänt den vanliga tillitsfrågan om huruvida man kan lita på folk i allmänhet, de har frågat folk hur de bedömer sannolikheten att en tappad plånbok kommer tillrätta.

Abstract:
This article contributes to the debate aboutthe effects of ethnic diversity on social cohesion, particularlygeneralized trust. The analysis relies on data from both the"Citizenship, Involvement, Democracy" (CID) survey in the US and the"Equality, Security and Community Survey" (ESCS) in Canada. Ouranalysis, one of the first controlled cross-national comparisons ofsmall-unit contextual variation, confirms recent findings on thenegative effect of neighborhood diversity on white majorities acrossthe two countries. Our most important finding, however, is that noteveryone is equally sensitive to context. Individuals who regularlytalk with their neighbors are less influenced by the racial and ethniccharacter of their surroundings than people who lack such socialinteraction. This finding challenges claims about the negative effectsof diversity on trust - at least, it suggests that the negative effectsso prevalent in existing research can be mediated by social ties.

Update on 25 feb 2008, kl 11:10 by

bergh

Så här diskuteras kausaliteten i uppsatsen:

An obvious alternative interpretation of this finding involves reverse causation.
How do we know it is the talking that matters here? The reverse could also be
true – trusting respondents are perhaps the ones who are more talkative with their
neighbors, whereas distrusting respondents remain quiet.

As it happens, the
correlation between the two variables talking and trusting is only 0.04, and it is
not significant in the ethnic-majority sample. In other words, trusters are not
necessarily talkers. Perhaps distrusters are less talkative in adiverse context than
trusters? Although it is generally true that there is a little less talking in the diverse
neighborhood context, the propensity to talk does not vary by levels of trust.That
is, trusters and distrusters alike talk least when the racial or ethnic background of
neighbors is mostly or entirely different from their own.

Finally, in racially homogeneous contexts there is essentially nodifference in trust levels between talkers and non-talkers. Accordingly, we have a strong primafacie case for the causal mechanisms we describe.

(p.70-71)