
A Change in Attitudes Toward Muslims? A Bayesian Investigation of Pre and Post 9/11 Public Opinion

 Unformatted Document Text:
Aleks, and Su 2008a). In addition, we assign noninformative priors Dirichlet(1) on
response categories κ’s so that each κ is augmented by an additional count. We pooled
the four datasets together and include year indicators of 2001, 2002, and 2005. This
way, we try to see if any of these years, compared to the base year of 2000, has an
eﬀect on the prediction of American’s attitudes toward Muslim.
We ﬁnd similar results in the analysis of the ethnocentric structure of Muslim
aﬀect. Under the Bayesian pooled ordered logistic model with year ﬁxed eﬀects,
Figure 2
1
tells us that the best predictor of attitudes toward Muslims is the favorability
of Jews. Individuals who like (dislike) Jews also like (dislike) Muslims when we control
the model for demographic characteristic
2
. The variables of our main interest are the
year eﬀects, or lack thereof. Neither years (2001, 2002, and 2005) have any eﬀect on
Muslim aﬀect. In other words, the ethnocentric structure of Muslim aﬀect persists
during both the pre and post 9/11 periods.
Next, we ﬁt separate ordered logistic regressions to the datasets of four diﬀerent
years (see Figure 3) to allow within variations. We want to see the structure of
cutpoints separately for each year. Overall, the estimates are similar to those of the
pooled model. To compare the estimates of four diﬀerent years, I set the predictors X
at the mean values. In other words, we are comparing nonwhite, and male Americans
in their midage who have mideducation, midincome, live in the North, and hold
neutral feeling toward Jews. Under this condition of Xβ = 0, we can compare four
sets of cutpoints across four diﬀerent year.
1
Figure 2 displays the estimates of the ordered logistic regression using the function bayespolrof R package arm (Gelman et al. 2008b)
2
The descriptive analysis of these demographic characteristics is in Figure A, Figure B, andFigure C. To facilitate interpretation and to ensure modeling stability, we rescale all the predictors, except of the binary ones, by subtracting the means and divided by 2 standard deviations(Gelman Forthcoming). Thus, zero becomes the mean values of the predictors that are notbinary.
12

 Authors: Kalkan, Kerem. and Su, YuSung. 




response categories κ’s so that each κ is augmented by an additional count. We pooled
the four datasets together and include year indicators of 2001, 2002, and 2005. This
way, we try to see if any of these years, compared to the base year of 2000, has an
eﬀect on the prediction of American’s attitudes toward Muslim.
We ﬁnd similar results in the analysis of the ethnocentric structure of Muslim
aﬀect. Under the Bayesian pooled ordered logistic model with year ﬁxed eﬀects,
Figure 2
1
tells us that the best predictor of attitudes toward Muslims is the favorability
of Jews. Individuals who like (dislike) Jews also like (dislike) Muslims when we control
the model for demographic characteristic
2
. The variables of our main interest are the
year eﬀects, or lack thereof. Neither years (2001, 2002, and 2005) have any eﬀect on
Muslim aﬀect. In other words, the ethnocentric structure of Muslim aﬀect persists
during both the pre and post 9/11 periods.
Next, we ﬁt separate ordered logistic regressions to the datasets of four diﬀerent
years (see Figure 3) to allow within variations. We want to see the structure of
cutpoints separately for each year. Overall, the estimates are similar to those of the
pooled model. To compare the estimates of four diﬀerent years, I set the predictors X
at the mean values. In other words, we are comparing nonwhite, and male Americans
in their midage who have mideducation, midincome, live in the North, and hold
neutral feeling toward Jews. Under this condition of Xβ = 0, we can compare four
sets of cutpoints across four diﬀerent year.
1
Figure 2 displays the estimates of the ordered logistic regression using the function bayespolr of R package arm (Gelman et al. 2008b)
2
The descriptive analysis of these demographic characteristics is in Figure A, Figure B, and Figure C. To facilitate interpretation and to ensure modeling stability, we rescale all the predic tors, except of the binary ones, by subtracting the means and divided by 2 standard deviations (Gelman Forthcoming). Thus, zero becomes the mean values of the predictors that are not binary.
12


Convention  All Academic Convention makes running your annual conference simple and cost effective. It is your online solution for abstract management, peer review, and scheduling for your annual meeting or convention.  Submission  Custom fields, multiple submission types, tracks, audio visual, multiple upload formats, automatic conversion to pdf.  Review  Peer Review, Bulk reviewer assignment, bulk emails, ranking, zscore statistics, and multiple worksheets!  Reports  Many standard and custom reports generated while you wait. Print programs with participant indexes, event grids, and more!  Scheduling  Flexible and convenient grid scheduling within rooms and buildings. Conflict checking and advanced filtering.  Communication  Bulk email tools to help your administrators send reminders and responses. Use form letters, a message center, and much more!  Management  Search tools, duplicate people management, editing tools, submission transfers, many tools to manage a variety of conference management headaches!  Click here for more information. 






