9
Blacks, Asian Americans, Latinos, Native Americans, and Arab Americans. Treating race as
white vs. nonwhite is admittedly crude, but the small number of cases of people of color in this
study prevents a finer-grained analysis.
These tests are conducted to examine patterns among several cases of civic dialogue.
These data are not appropriate for generalizing about the outcomes of civic dialogue in general,
but are intended to probe for signs of attaining unity among participants in these rounds of
dialogue in Madison and Aurora. For that reason, I present standard errors, but do not focus on
statistical significance—the likelihood that relationships other than zero hold in the population of
dialogue as a whole.
Table 1 displays the results just among participants in Madison, and Table 2 displays the
results among Aurora participants. The results do not show the expected pattern, a positive
relationship between participating in a diverse group and feelings of closeness toward these
various groups. In both tables, we see that participating in diverse groups is generally associated
with feeling less close to these various social groups. Instead of positive coefficients on the
variable representing participation in a diverse group, the coefficients are almost always negative.
The exceptions are 1) among whites in Madison, a positive association between participating in a
diverse group and their feelings of closeness toward people of color
33
and Madisonians in
general; and 2) among Blacks in Madison, a positive relationship between participating in a
diverse group and closeness toward Blacks.
The control group enables a test of participation in general as well as participation in a
diverse group. To compare participants and waitlisters, I analyze respondents separately by round
to control for changes in the community information environment. I again use a weighted least
squares specification in which post-test scores are modeled as a function of pre-test scores, and
include controls for racial self-identification, and frequency of interracial talk when analyzing
closeness. The main effect for participation is a variable indicating whether the respondent was a
participant or not. The variable representing participation in a diverse group (which takes on a
value of 0 for waitlisters) represents the added effect of exposure to difference.
These models were estimated in two stages. Selection into the program off of the waitlist
was not random, but was determined partly on the basis of race (people of color were more likely
to be selected for participation in order to achieve diverse discussion groups), and availability.
Therefore, the variable representing participation was first modeled as a function of income,
parenthood of children under 19, race, efficacy, and political participation (community work and
writing letters to the editor). Results are displayed in Table 3.
Estimating these models shows that participating in these rounds in Madison was
generally associated with feeling closer to these groups (Tables 4 and 5). The one exception of
the pattern of a positive association between participation and closeness is the closeness of whites
toward Madisonians in general, in the Fall 2002 round. However, these models allow us to again
observe the effect of participating in diverse groups. Here we see that participating in diverse
groups at times weakens this effect for both whites and nonwhites. Whites in both rounds who
participated in diverse groups felt less close to Blacks as well as their own racial group.
Participating in diverse groups was associated with less closeness among nonwhites for all cases
in which there was ample valid data.
To briefly sum, we see mixed evidence at best that participating in these dialogues leads
to common ground in the form of a shared category. Participating in general is associated to
referring to the index of closeness toward racial groups other than whites, although the racial
categorizations used are identical.
33
When individual analyses are run among whites with respect to closeness to the groups comprising
“People of Color,” the signs on participating in a diverse group with respect to Asian Americans and Native
Americans are positive, while the signs with respect to the other groups (Blacks, Latinos and Arab
Americans) are negative. All are insignificant.
Convention