Know Thy Enemy: Education About Terrorism Improves Social Attitudes Toward Terrorists

Abstract

Hatred of terrorists is an obstacle to the implementation of effective counterterrorism policies—it invites indiscriminate retaliation, whereas many of the greatest successes in counterterrorism have come from understanding terrorists’ personal and political motivations. Drawing from psychological research, traditional prejudice reduction strategies are generally not well suited to the task of reducing hatred of terrorists. Instead, in two studies, we explore education’s potential ability to reduce extreme negative attitudes toward terrorists. Study 1 compared students in a college course on terrorism (treatment) with wait-listed students, measuring pro-social attitudes toward a hypothetical terrorist. Initially, all students reported extremely negative attitudes; however, at the end of the semester, treatment students’ attitudes were significantly improved. Study 2 replicated the effect within a sample of treatment and control classes drawn from universities across the United States. The present work was part of an ongoing research project, focusing on foreign policy and the perceived threat of terrorism; thus classes did not explicitly aim to reduce prejudice, making the effect of treatment somewhat surprising. One possibility is that learning about terrorists “crowds out” the initial pejorative associations—i.e. the label “terrorism” may ultimately call more information to mind, diluting its initial negative associative links. Alternatively, students may learn to challenge how the label “terrorist” is being applied. In either case, learning about terrorism can decrease the extreme negative reactions it evokes, which is desirable if one wishes to implement effective counterterrorism policies.

Full text

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
1
Running head: EDUCATION ABOUT TERRORISM CHANGES SOCIAL
ATTITUDES
Know thy enemy: Education about terrorism improves social attitudes toward terrorists
Jordan Theriault1, Peter Krause2, Liane Young1
Word Count: 7989
1Boston College!
Department of Psychology!
Chestnut Hill, MA, 02467
USA
2Boston College!
Department of Political Science!
Chestnut Hill, MA, 02467
Corresponding Author:!
Jordan Theriault!
140 Commonwealth Ave.!
Chestnut Hill, MA, 02467!
300 McGuinn Hall!
617-552-0240
Fax: 617-552-0523!
jordan.theriault@bc.edu!
© 2017, American Psychological Association.
This paper is not the copy of record and may not exactly replicate the final, authoritative
version of the article. Please do not copyor cite without authors
permission. The final article will be available, upon publication, via its
DOI: 10.1037/xge0000261

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
2
Abstract
Hatred of terrorists is an obstacle to the implementation of effective counterterrorism
policies—it invites indiscriminate retaliation, whereas many of the greatest successes in
counterterrorism have come from understanding terrorists’ personal and political
motivations. Drawing from psychological research, traditional prejudice reduction
strategies are generally not well suited to the task of reducing hatred of terrorists. Instead,
in two studies, we explore education’s potential ability to reduce extreme negative
attitudes toward terrorists. Study 1 compared students in a college course on terrorism
(treatment) with wait-listed students, measuring pro-social attitudes toward a hypothetical
terrorist. Initially, all students reported extremely negative attitudes; however, at the end
of the semester, treatment students’ attitudes were significantly improved. Study 2
replicated the effect within a sample of treatment and control classes drawn from
universities across the United States. The present work was part of an ongoing research
project, focusing on foreign policy and the perceived threat of terrorism; thus classes did
not explicitly aim to reduce prejudice, making the effect of treatment somewhat
surprising. One possibility is that learning about terrorists “crowds out” the initial
pejorative associations—i.e. the label “terrorism” may ultimately call more information
to mind, diluting its initial negative associative links. Alternatively, students may learn to
challenge how the label “terrorist” is being applied. In either case, learning about
terrorism can decrease the extreme negative reactions it evokes, which is desirable if one
wishes to implement effective counterterrorism policies.
Keywords: Prejudice, Education, Terrorism, Attitudes

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
3
Know thy enemy: Education about terrorism improves social attitudes toward terrorists
The most prominent and prolific terrorist groups today—from ISIS to Al-Qaeda
to Boko Haram—welcome our hatred as a key part of their strategy, inviting
indiscriminate retaliation that polarizes communities and drives up support for their
extreme ideologies and tactics (Kydd & Walter, 2006; Lake, 2002). States driven to
pursue these indiscriminate policies have either failed to eliminate terrorist attacks, or
have even increased them (Cronin, 2009). By contrast, most successful counterterrorism
and counterinsurgency campaigns have involved breakthroughs in understanding the
motivations, organization, and strategies of terrorist and insurgent groups. For instance,
in the late 1990s, India’s government and police were able to fragment and beat back
Islamist militants in Kashmir by recognizing cleavages in the insurgency and flipping
their former adversaries to fight on their side (Staniland, 2012). In 2006, the United
States located and killed the former leader of Al-Qaeda in Iraq (now ISIS)—Abu Musab
Al-Zarqawi—after American intelligence officers interviewed suspects and members of
the community to understand the motivations and social ties of the terrorist network
(Alexander & Bruning, 2011). The most dramatic example may be Northern Ireland,
where centuries of sectarian violence were effectively ended—not through the
extermination of all insurgents, but through the inclusion of Sinn Féin (the political wing
of the Irish Republican Army) in negotiations and the political process, culminating in the
Good Friday Agreement of 1998. To effectively combat terrorism, states must understand
their adversary as a rational actor who is sustained by recruits, funding, and sanctuary,
and who is motivated by political objectives; hatred of terrorists, in either policymakers
or the citizens that elect them, is an obstacle to this aim, and may lead to policies that are

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
4
the exact opposite of what effective counterterrorism strategy demands. Ironically then,
to combat terrorism, we must find ways to reduce prejudice against terrorists.
Techniques for Reducing Prejudice
Prejudice reduction is one of the most prolific areas of research in social
psychology; yet, applying it to terrorists—or, for that matter, many real world contexts
(Paluck, 2016; Paluck & Green, 2009)—is difficult for a number of reasons. “Implicit
attitudes” paradigms have revolutionized our understanding of prejudice, showing that
even those who reject explicit prejudice continue to show a measurable bias toward
outgroups (Greenwald, McGhee, & Schqartz, 1998; Lai et al., 2014). But this research
typically focuses on prejudice against marginalized groups (e.g., African Americans,
disabled people), where public displays of prejudice are already socially unacceptable.
Researchers have rarely turned their attention to the “fundamental challenge” of
“[discovering] ways of changing ‘hard-core’ [or extreme] prejudiced beliefs” (Monteith,
Zuwerink, & Devine, 1994)—such as explicit racist beliefs, or the extreme prejudice
instilled by the label “terrorist”. A similar problem is present in the “minimal-groups”
paradigm, where researchers instill, and then attempt to reduce, prejudice between
“teams” that were formed on the basis of irrelevant traits (and in actuality were randomly
assigned; Tajfel, 1970). This method provides tight experimental control, but the
prejudice being studied lacks a real world historical context; it occurs in the absence of
any competing or complicating influences. Furthermore, prejudice in the context of
minimal groups is typically defined as preference for one’s own team, rather than outright
intergroup hostility (Paluck & Green, 2009). Finally, classic approaches to prejudice
reduction suggest that prejudice can be reduced by facilitating contact with the outgroup

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
5
under ideal conditions (Allport, 1954; Pettigrew, & Tropp, 2006; Sherif, Harvey, White,
Hood, & Sherif, 1961). Unfortunately, for prejudice against terrorists, this approach is
practically and politically unfeasible—most citizens will never interact with a terrorist
personally; yet, their attitude toward them remains politically important.
Education, as a technique for prejudice reduction, has the potential to overcome
the limitations of the approaches above. It is uniquely positioned to reduce prejudice
where explicit antipathy is present, and where parties cannot be physically brought
together—this includes (but is not limited to) the case of prejudice against terrorists.
People can learn about, and change their attitudes toward, people that they may never
encounter. For instance, an education approach was implemented in the context of the
Israeli-Palestinian conflict: a class of Israeli students studied the history of conflict in
other countries (Lustig, 2002; Salomon, 2004). End of term essays, written by treatment
students, were more equitable to both sides of the Israeli-Palestinian conflict and more
likely to be written from the first-person perspective of Palestinians (although the
treatment had no effect on students’ explicit prejudice against Palestinians). This, and
other work (Gurin, Peng, Lopez, & Nagda, 1999; Schaller, Asp, Rosell, & Heim, 1996)
give some reason to believe that education-based prejudice reduction can be effective.
Another advantage of an education-based approach is that prejudice reduction is
both tested and implemented in the same context—the classroom. In such field-based
research, statistically significant effects are difficult to identify, but their ecological
validity can generally be trusted, as they must emerge from an environment full of noise
and competing influences. Many educational-based interventions are field studies, yet
few use well-controlled designs—including control groups, or ideally as-if

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
6
randomization—that allow for inferences about the causal effect of treatment. According
to a recent review, fewer than 12 of 207 quasi-experimental studies had designs that
licensed causal inferences (Paluck & Green, 2009; but see Broockman & Kalla, 2016).
Given the lack of field-based, experimental prejudice reduction research, conclusions
drawn from the present work may also have implications for prejudice reduction more
generally. Reducing antipathy toward terrorists (for the purposes of counterterrorism)
may be taken as a case study in reducing extreme and explicit antipathy, and it may be
that our findings can be applied to other cases (such as explicit racism, or sectarian
hatred).
Present Work
The present work was performed in the context of a larger ongoing research
project, exploring the impact of education on attitudes concerning terrorism and foreign
policy (thus, the majority of the survey was not focused on attitudes toward terrorists, and
classes generally focused on counterterrorism, as opposed to tolerance). In this context,
we had the opportunity to explore education’s potential role in reducing prejudice toward
terrorists. In Study 1, we performed an as-if randomized study, taking advantage of
randomized course registration times at our university. Study 2 replicated the effect in a
more representative sample, comparing treatment and control classes at 11 universities
across the United States.
Because our study was performed in the context of a larger ongoing project, the
courses had no explicit anti-prejudice aim. Main themes of the course used in Study 1
were: (a) the individual and group level causes and objectives of terrorism; (b) the
methods and mechanisms of terrorism; (c) discussion of recent and ongoing conflicts,

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
7
such as conflict with Al-Qaeda and the insurgencies in Iraq and Syria; and (d)
counterterrorism and counterinsurgency strategy (for a complete list of course readings in
Study 1, see Appendix A). Prior work has focused on teaching tolerance—e.g., teaching
white students about the positive role of intergroup conflict in democratic society, and
then tracking their attitudes toward students of color across their university tenure (Gurin
et al., 1999)—however, we had no intention of teaching students to tolerate terrorists.
Students were simply taught about terrorism, and completed surveys at the beginning and
end of the class, allowing us to track any changes in their attitude. It is possible then, that
students, or even professors, might show a confirmation bias (Haidt, 2001; Kuhn, 1991;
Kunda, 1990; Wason, 1960)—students might only learn, or professors might only teach,
information that is consistent with their initial view of terrorists (for an example in the
context of the Israeli-Palestinian conflict, see Gvirsman, Huesmann, Dubow, Landau,
Boxer, & Shikaki, 2016). For instance, political conservatives, who are generally more
threat-sensitive (for review see Jost & Amodio, 2012), may attend to the most threatening
information taught and resist any positive effect of treatment. Likewise, professors may
lead their students to adopt their personal viewpoint by consciously or unconsciously
presenting selective information about terrorists. In both studies we explore these
possibilities, examining biases based on political orientations, self-reported willingness to
learn, students’ initial attitudes, and even the views of the teaching professors.
Study 1
Methods
Participants. Fifty-eight students (MAge = 21.3, SDAge = 0.9, 34 female, 2
unspecified; Table 1) were given pre-class and post-class questionnaires at the beginning

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
8
and end of the semester (Qualtrics software). Thirty-five students completed coauthor
PK’s class “Terrorism, Insurgency, and Political Violence” at Boston College (Fall 2013,
n = 17; Spring 2015, n = 18); twenty-three students who were wait-listed for the same
class (Fall 2013, n = 8; Spring 2015, n = 15) formed an as-if randomized control group.
Wait-listed students had been randomized by the university to receive a later course
registration time and had emailed coauthor PK to enroll in the class after it had been
filled. The class was filled by 1:55pm (Fall 2013), and 9:32am (Spring 2015) on the first
of eight days of registration, making it unlikely that student interest drove their allotment
to the treatment or control group; put another way, it was reasonable to assume that wait-
listed students would be in a treatment class if they had not been randomly assigned a late
course registration time. The pre-class questionnaire was completed on the first day of
class, and the post-class questionnaire was completed three months later. Students were
included if they completed both the pre-class and post-class survey (response rate:
95.1%). In treatment classes, after both the pre-class and post-class survey, five
participating students were randomly awarded $10 Amazon.com gift cards. Students in
the control group who completed both surveys received $20 Amazon.com gift cards. The
Boston College Institutional Review Board approved the study, and informed consent
was obtained from all participants.

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
9
Table 1. Study 1 classes, response rates and demographics.
School
Professor
Class Name
Semester
Enrollment
Responses
Gender
Age
Political
Orientation
Boston
College
P. Krause
Terrorism,
Insurgency,
and Political
Violence
Fall, 2013
17
17
9 female
8 male
M = 21.1
SD = 0.8
M = 3.7
SD = 1.4
Spring, 2015
18
18
15 female
3 male
M = 21.6
SD = 0.5
M = 2.7
SD = 1.3
Wait-list
Fall, 2013
11
8
3 female
5 male
M = 20.6
SD = 1.3
M = 3.1
SD = 1.8
Spring, 2015
15
15
7 female
2 unspecified
6 male
M = 21.1
SD = 1.1
M = 2.5
SD = 1.2
Political Orientation was measured on a 7-point scale (1 - very liberal; 7 – very conservative). For control samples, enrollment is the number of students who
completed pre-class surveys, responses are the number of students who completed both pre- and post-class surveys.

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
10
Procedure and Measures. Pre-class and post-class surveys were identical. We
collected responses for dependent measures, covariates of interest, and other items that
were of interest for additional studies (see Supplemental Materials for a complete
description). Questions related to social affiliation made up a small percentage of the total
survey (one of six blocks, plus demographics), meaning that any attention drawn to
terrorists’ humanity was most likely diluted among questions about the threat, motives,
and effectiveness of terrorists, as well as the effectiveness of counterterrorism policies.
Relationships between social affiliation and the measures collected in the remaining
blocks were not examined, to avoid introducing unnecessary comparisons in our analysis.
Furthermore, while demand characteristics are always a concern, coauthor PK, who
taught the course, was not responsible for the inclusion of the social affiliation measures
and personally had no strong hypotheses about the direction of the effect (social
affiliation measures were proposed by coauthor LY). Despite this, we take a more direct
approach to combatting demand characteristics in Study 2, testing whether results depend
on the inclusion of data from PK’s classes.
Dependent measures. Questions related to social affiliation were asked on a
single page. Students read a brief introduction: “Suppose you met someone belonging to
a group that had carried out at least one terrorist attack,” and were then asked: “How
much would you like this person?” [“liking”]; “How similar would you be to this
person?” [“similarity”]; “How much would you get along with this person?” [“getting
along”]; and “How much would you like to interact with this person?” [“interaction”] (1
– “not at all”; 7 – “very much”). This set of four questions formed our measure of social
affiliation, provided that the questions were not differentially affected by treatment.

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
11
These questions showed good reliability (αPre = .78; αPost = .77); however, we
opted not to combine them into a scale in our analysis of treatment below. Our data were
hierarchical (e.g. multiple observations from each student; students were clustered within
classes), meaning that there was no simple way to model each observation as independent
from all others. Instead, we adopted a mixed effects approach, which allowed us to
respect this hierarchical design while also allowing that relationships among variables
may not be uniform across levels of the design. In particular, because students provided
four responses (at pre-treatment and at post-treatment), we could allow that pre-class
social affiliation may predict post-class social affiliation differently for each student (see
Statistical Methods and Random Effects structure below for more detail).
Covariates of Interest. Students were asked to rate: (a) their knowledge and (b)
interest regarding terrorism (“knowledge” and “interest”; 1 – “I have no knowledge
of/interest in the topic”; 7 – “I have a tremendous amount of knowledge about/interest in
the topic”); (c) the likelihood that they would change their opinions on terrorism
(“openness to change”; 1 – “very unlikely”; 7 – “very likely”); and (d) the confidence
they had in their opinions (“confidence”; 1 – “not confident at all”; 7 – “extremely
confident”). At the end of the survey students completed a brief demographics form.
Statistical methods. Not all samples collected were independent. Data were
collected across two semesters (Fall, 2013; Spring, 2015), meaning that groups of
students could be subject to cohort effects; likewise, each student provided multiple
measures of social affiliation. To address this, most analyses in this paper use linear
mixed effects analyses (Baayen, Davidson, & Bates, 2008; Bates, Kliegl, Vasishth, &
Baayen, 2015; Judd, Westfall, & Kenny, 2012), also commonly referred to as hierarchical

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
12
linear modeling. This technique allows us to model and test the significance of
dependencies within our sample—such as cohort effects, or the non-independence of
multiple data points from each student—and control for them when necessary. When
dependencies were a non-significant source of variance, they were removed from the
model to avoid overfitting, as per recent recommendations (Bates et al., 2015). We began
with a full factorial model of our random effects structure and winnowed it to a
parsimonious model using log-likelihood ratio tests, before testing for fixed effects of
interest. The parsimonious model is reported below, and necessary tests to derive it are
reported in supplemental materials (Table S1). We performed mixed effects analyses
using R (R Core Team, 2015) and the lme4 package (Bates, Maechler, Bolker, & Walker,
2015), and obtained p values for fixed effects using the Kenward-Roger approximation of
degrees of freedom, implemented in lmerTest (Kuznetsova, Brockhoff, & Christensen,
2015) and pbkrtest packages (Halekoh & Højsgaard, 2014). Following recent
recommendations (Cummings, 2014), for key results, we report bootstrapped 95%
confidence intervals (5000 resamples) in square brackets using the bias corrected and
accelerated method (BCa, Efron, 1987). We also use Welch’s unequal variance t-tests, in
lieu of traditional student’s t-tests, to avoid imposing the assumption that variance is
perfectly equal between groups (Moser & Stevens, 1992). Note that our results report
non-integer degrees of freedom; for mixed effects analyses this reflects corrections for
the non-independence of observations, and for Welch’s t-tests this reflects corrections for
unequal variance between groups.
Results

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
13
Pre-test scores. As-if randomization placed students into treatment and wait-list
(control) groups; however, it remained possible that the groups may differ on pre-class
measures. The groups did not differ on any attitudinal measures: “liking”, t(30.9) = 1.33,
p = .194; “similarity”, t(42.8) = 0.20, p = .843; “getting along”, t(39.4) = 0.60, p = 0.550;
“interaction”, t(41.3) = 0.46, p = -.646. Scores were generally low for all pre-class
attitudinal measures—MPre-liking = 1.79; MPre-similarity = 2.32; MPre-getting along = 1.93; MPre-
interaction = 2.79—and were all significantly below the scale mid-point: “liking”, t(56) =
15.93, p < .001; “similarity”, t(56) = 9.06, p < .001; “getting along”, t(56) = 14.21, p <
.001; “interaction”, t(56) = 4.42, p < .001. Thus, at the beginning of the semester,
attitudes were low, and equal between treatment and control groups. We also conducted
combined placebo tests, using the random effects structure described below for “Effect of
Treatment”. There was no interaction between Treatment and Question, F(3, 165.0) =
0.24, p = .868, so the parameter was removed from our model. In the resulting model,
Treatment and Control groups did not differ on the combined measure of pre-class social
affiliation, b = -0.23, t(54.9) = 0.73, p = .467.
We compared treatment and wait-list groups on a number of additional covariates:
“openness to change”, t(51.6) = 0.37, p = .711; “interest”, t(40.5) = 0.66, p = .514;
“confidence”, t(40.7) = 1.03, p = .307; and “knowledge”, t(49.0) = 0.78, p = .437. We
also compared treatment and wait-list groups on demographic measures: political
orientation (1–“very liberal”; 7–“very conservative”), t(41.9) = 1.10, p = .277; gender
(male = 0; female = 1), t(39.5) = 1.53, p = .135; and age, t(25.8) = 1.86, p = .075. The
marginal difference in age uncovered one potential limitation of our as-if randomization
procedure—although course registration times are randomized within each student year,

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
14
they are not randomized across them; college seniors are given priority above juniors,
sophomores and freshman in registration, meaning that our treatment group is biased to
contain more senior students (MAge:Treatment = 21.49; SDAge:Treatment = 0.66; MAge:Wait-list =
21.0; SDAge:Wait-list = 1.19). Given this, we report whether key results below are affected
by the inclusion of student year as a covariate.
Random effects structure. Before testing fixed effects (e.g. Treatment), we
created a random effects structure, also commonly called a hierarchical linear model
(Baayen et al., 2008). Each data point was nested within several levels—e.g., student,
semester—and by modeling each, when necessary, we could produce accurate estimates
of effects that also generalize to a sampled population (e.g. to a population of university
students). Effects may also vary across these levels; for instance, pre-class attitudes may
predict post-class attitudes better for some students more than for others. Working
backwards from a maximal model (Bates et al., 2015; Table S1), we arrived at the
following parsimonious model:
Attitude Post = 1 + (0 + Attitude Pre | Semester) + (1 + Attitude Pre | Student)
Within our sample, there was significant variability in: (a) the by-semester relationship
between pre-class and post-class social affiliation, (Attitude Pre | Semester),2(1) = 18.56,
p < .001; (b) the by-student relationship between pre-class and post-class social
affiliation, (Attitude Pre | Student), χ2(1) = 7.62, p = .006; and (c) by-student mean post-
class social affiliation, (1 | Student), χ2(1) = 27.57, p < .001. Thus, our model allows that
the relationship between pre-class and post-class social affiliation differs for each
semester and student, and that mean post-class social affiliation differs for each student.

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
15
Effect of treatment. We added fixed effects of interest to the random effects
structure described above. First, we examined whether treatment differentially affected
our four measures of post-class social affiliation (“liking”, “getting along”, “similarity”,
“interaction”); the interaction between treatment and question was non-significant, F(3,
154.0) = 1.37, p = .254, and so social affiliation was defined as the combination of the
four attitudinal measures. With the interaction term removed, there was a main effect of
question, F(3, 157.6) = 4.38, p = .005, where some questions received higher ratings than
others; however, critically, there was a main effect of Treatment, where treatment
students reported higher post-class social affiliation toward terrorists than wait-listed
students, F(1, 52.5) = 7.59, p = .008, b = .70, [0.21, 1.19] (Figure 1; Table S2). The main
effect of treatment remained significant after controlling for student year, F(1, 52.8) =
5.31, p = .025, b = .76, [0.13, 1.43]. Thus, treatment students, relative to wait-listed
students, reported having less extreme negative attitudes toward terrorists at the end of
the semester.
Note that this effect of treatment did not depend on the specification of our
random effects structure. In an ordinary least squares regression, predicting the average
of our four post-class social affiliation measures (α = .77), and including average pre-
class social affiliation (α = .78), and student class year (freshman / sophomore / junior /
senior / graduate) as covariates, the effect of treatment remained significant, t(49) = 2.83,
p = .007, b = .87, [0.25, 1.49].

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
16
Figure 1. Study 1 main effect of treatment. 95% confidence interval computed using BCa
method (Efron, 1987; 5000 resamples). The scale mid-point for post-class social
affiliation is marked with a dotted line. Percent increase represents the estimate for the
treatment group relative to the wait-list group. Error bars represent standard error of the
treatment coefficient.
Potential moderators. It was possible that treatment might affect some students
more strongly than others. As we had collected several measures of individual
differences, we explored the interaction between treatment and pre-class measures of (a),
knowledge, (b) interest, and (c) opinion confidence. In no case was the interaction with
treatment significant, ps > .350 (Table S3). Thus, there were no obvious individual
differences accounting for the effect of treatment—at the end of the semester, students
who completed a course on terrorism, compared to those who were wait-listed, reported
having less extreme negative attitudes toward terrorists.
Potential confirmation bias. Several covariates were of additional interest
because they may reflect confirmation bias on the part of students. Treatment may be less

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
17
effective for students who: (a) initially reported extreme hostility toward terrorists (i.e.
students with low pre-class social affiliation), (b) initially reported being unwilling to
change their minds about terrorists (i.e. low pre-class openness to change), or (c) were
more politically conservative. None of these covariates interacted with treatment, ps >
.280 (Table S4). Thus, in our sample, there was no evidence that treatment was affected
by confirmation biases.
Discussion
Study 1 provided causal evidence (through as-if randomization) that participation
in a course on terrorism improved students’ initial (strongly negative) attitudes toward
terrorists. This was surprising, as the course was not intended to teach students
tolerance—the survey itself was part of an ongoing project to study politically relevant
attitudes surrounding terrorism, and given this, coauthor PK had no explicit aim to reduce
prejudice (for an example of the course readings in Study 1, see Appendix A). To ensure
that the observed effects were not specific to PK’s class, Study 2 aimed to replicate the
effect within a larger sample, spanning professors, classes, and universities. Collecting a
larger and more diverse sample also provided an opportunity to revisit the potential
moderation of treatment by individual differences.
Study 2
Study 2 surveyed students at 31 classes, taught by 16 professors, at 11 universities
across the United States. As-if randomization was not possible in this case; instead,
treatment classes (classes teaching about terrorism; e.g. “Causes of Terrorism and
Political Violence”; “Chemical, Biological, Radiological, and Nuclear Threats to the
Homeland”) were compared with control classes (classes covering topics only indirectly

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
18
related to terrorism; e.g. “Causes of War”; “Theories of Peace and Conflict”). We were
interested in whether the main effect of treatment would replicate within this more
diverse sample.
Methods
Participants. Three hundred seventy-seven students (MAge = 22.0, SDAge = 4.9;
189 female, 12 unspecified, 176 male; for full crosstabs see Tables 2-3) completed a pre-
class and post-class survey, as described in Study 1. Students were recruited from classes
across the United States over a period of two years (NClasses = 31; NProfessors = 16;
NUniversities = 11; NSemesters = 4). We compared classes teaching about terrorism (treatment;
NStudents: Treatment = 249; NClasses: Treatment = 20; Table 2), with classes covering topics only
indirectly related to terrorism (control; NStudents: Control = 128; NClasses: Control = 11; Table 3).
To collect as many treatment and control classes as possible, we solicited
professors to participate and included all who responded, categorizing each as treatment
or control based on syllabus content. Coauthor PK and three other professors taught
classes at Boston College. Other classes were taught by professors who responded to a
request for participants, circulated through the START (Study of Terrorism and the
Prevention of Terrorism) professional listserver. Some of these professors were currently
teaching courses related to terrorism, and some were currently teaching other courses,
creating a natural control group of professors who were knowledgeable about terrorism
but not currently teaching it. Control classes at Boston College were selected to cover
material in related subfields that excluded terrorism (i.e. international relations, security).
Classes in which terrorism was studied for over three weeks were coded as treatment;
otherwise classes were coded as control.

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
19
All classes were taught within political science, history, and international studies
departments. No classes were taught within psychology departments or by psychology
professors, and only one class included psychological readings related to prejudice
reduction (“Psychology of Political Violence and Terrorism”; 7 students, comprising
1.9% of our total sample). Classes were generally small (28 of 30 classes had fewer than
36 students), and conducted as lecture-discussions. Treatment classes focused on topics
like the causes, strategies, and effects of terrorism, whereas control classes focused on
topics like the causes of war, crisis communication, and the politics of intelligence.
Thirteen students completed the survey in more than one class; entries beyond
their first were excluded, and if two entries occurred within the same semester then only
the treatment entry was retained. Within each class, after the pre-class and post-class
survey, five participating students were randomly awarded $10 Amazon.com gift cards—
except at Georgia Tech, where a state ban prohibits gambling via random incentives, so
all students received $20 gift cards for completing both surveys. Professors (with the
exception of coauthor PK) who completed the survey received a $50 Amazon.com gift
card. Institutional review board approval was obtained from each school, and informed
consent was obtained from all participants.
The average response rate in Study 2 (52.4%) was lower than that reported in
Study 1 (95.1%). This was not unexpected, as students would almost certainly be less
motivated to complete a survey for a professor they do not know personally. Detailed
comparisons between full and drop-off respondents were precluded, as informed consent
was collected with the post-class survey (to avoid alerting students to the purpose of the
survey). Critical for us however, there were no differences in response rate across classes

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
20
between the treatment and control group—Welch’s unequal variance t-test, t(25.8) =
0.26, p = .796. Likewise, there were no differences in class size between the treatment
and control group— Welch’s unequal variance t-test, t(28.8) = 0.40, p = .691. Thus,
although a higher response rate would be desirable, our treatment and control groups
were well matched.
Procedure and Measures. Pre-class and post-class surveys were identical to
those described in Study 1. Once again, measures of social affiliation (“liking”,
“similarity”, “getting along”, and “interaction”) showed good reliability (αPre = .78; αPost
= .79), and again, we opted to avoid combing them into a scale in most analyses, favoring
a mixed effects approach to estimate by-subject random slopes and intercepts (see
Statistical Methods and Random Effects structure below for more detail).
Statistical Analysis. As in Study 1, not all samples were independent: data could
potentially be clustered within students, professors, classes, universities, and semesters.
Linear mixed effects analyses allowed us to examine effects while controlling for this
variability when necessary (Baayen et al., 2008; Bates et al., 2015; Judd et al., 2012).
Once again, we began with a full factorial model of our random effects structure, and
winnowed it to a parsimonious model before testing fixed effects of interest (Table S5).
For key results, we also report bootstrapped 95% confidence intervals in square brackets
(5000 resamples; BCa, Efron, 1987). Non-integer degrees of freedom reflect corrections
for the non-independence of observations in mixed effects analyses, and for corrections
based on unequal variance between groups in Welch’s t-tests.

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
21
Table 2. Study 2 treatment classes, response rates and demographics.
School
Professor
Class Name
Semester
Enrollment
Responses
Gender
Age
Political
Orientation
American
University
Prof A
Causes of
Terrorism and
Political Violence
Fall, 2013
25
17
6 female
11 male
M = 20.3
SD = 0.9
M = 4.1
SD = 1.5
Prof B
Psychology of
Political Violence
and Terrorism
Fall, 2014
23
7
2 female
5 male
M = 23.3
SD = 1.8
M = 3.9
SD = 1.2
American
Military
University
Prof C
Chemical,
Biological,
Radiological, and
Nuclear Threats to
the Homeland
Fall, 2013
22
2
1 female
1 male
M = 27.5
SD = 0.7
M = 5
SD = 1.4
Senior Seminar in
Homeland Security
Fall, 2013
13
3
3 male
M = 38
SD = 11.3
M = 5.3
SD = 1.2
Introduction to
Homeland Security
and Defense
Fall, 2014
13
1
Omitted to protect anonymity.
Boston
College
Prof D
The History of
Terrorism
Fall, 2014
75
56
29 female
26 male
1
unspecified
M = 20.5
SD = 0.7
M = 3.3
SD = 1.3
Prof E
Terror and the
American Century
Spring,
2014
28
12
4 female
8 male
M = 20.8
SD = 0.8
M = 4.1
SD = 1.4
P. Krause
Terrorism,
Insurgency, and
Political Violence
Fall, 2013
17
17
9 female
8 male
M = 21.1
SD = 0.8
M = 3.7
SD = 1.4
Spring,
2014
13
11
6 female
5 male
M = 21.4
SD = 3.3
M = 2.8
SD = 1.3
Spring,
2015
18
18
(including
4 repeats)
11 female
3 male
M = 21.6
SD = 0.5
M = 2.9
SD = 1.4
Introduction to
International
Studies
Spring,
2014
32
15
5 female
10 male
M = 19.3
SD = 0.5
M = 3.5
SD = 1.7
International
Studies Senior
Seminar
Spring,
2014
12
12
(including
2 repeats)
6 female
4 male
M = 21.1
SD = 0.6
M = 2.9
SD = 1.4

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
22
Excelsior
College
Prof F
International
Terrorism
Fall, 2013
7
2
0 female
2 male
M = 42
SD = 2.8
M = 5.5
SD = 0.7
Georgia
Institute of
Technology
Prof G
The Challenges of
Terrorism
Fall, 2014
23
12
4 female
7 male
M = 21.7
SD = 2.3
M = 3.1
SD = 1.2
Northeastern
University
Prof H
Terrorism,
Violence, and
Politics
Spring,
2014
22
11
6 female
5 male
M = 29.1
SD = 12.6
M = 3.9
SD – 0.9
University of
Denver
Prof I
International
Terrorism
Spring,
2014
15
13
6 female
1
unspecified
6 male
M = 27.3
SD = 4.6
M = 4.2
SD = 1.5
University of
Maryland
Prof J
Asymmetric
Warfare
Fall, 2013
58
28
6 female
2
unspecified
20 male
M = 22.4
SD = 3.9
M = 3.7
SD = 1.6
Prof K
Motivations and
Intents of
Terrorists and
Terrorist Groups
Fall, 2014
10
6
1 female
5 male
M = 29.2
SD = 7.0
M = 2.5
SD = 1.0
Westwood
College
Prof L
Terrorism (Class 1)
Spring,
2014
23
5
0 female
4
unspecified
1 male
M = 22.8
SD = 2.8
M = 2.8
SD = 1.3
Terrorism (Class 2)
Spring,
2014
23
7
4 female
3 male
M = 23.6
SD = 5.4
M = 3.1
SD = 1.2
Political Orientation was measured on a 7-point scale (1 - very liberal; 7 – very conservative).

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
23
Table 3. Study 2 control classes, response rates and demographics.
School
Professor
Class Name
Semester
Enrollment
Responses
Gender
Age
Political
Orientation
American
University
Prof A
U.S. National
Security and Civil
Wars
Fall, 2014
13
6
2 female
4 male
M = 28
SD = 4.0
M = 2.5
SD = 1.0
Boston
College
Prof M
Causes of War
Fall, 2013
35
11
5 female
6 male
M = 20.6
SD = 0.8
M = 3.6
SD = 1.8
Fall, 2014
29
6
3 female
3 male
M = 20.8
SD = 1.0
M = 2.3
SD = 0.8
Intelligence and
International
Security
Fall, 2013
36
14
(including
2 repeats)
9 female
3 male
M = 21.2
SD = 0.4
M = 3.6
SD = 1.7
Modern Classics of
International
Relations
Fall, 2014
9
3
1 female
2 male
M = 20.3
SD = 0.6
M = 3.3
SD = 2.3
United Nations and
International
Security
Fall, 2014
36
14
(including
1 repeat)
6 female
1 unspecified
6 male
M = 21.1
SD = 0.9
M = 3.2
SD = 0.8
P. Krause
Research Methods
and National
Movements
Spring,
2014
16
14
(including
4 repeats)
6 female
4 male
M = 20.5
SD = 1.3
M = 3.4
SD = 1.6
University
at Albany,
SUNY
Prof N
Honors Course on
Political Violence
Fall, 2013
23
17
13 female
2 unspecified
2 male
M = 18.6
SD = 0.8
M = 3.2
SD = 1.4
Honors Course on
Political Violence
Fall, 2014
25
21
16 female
5 male
M = 20.9
SD = 7.0
M = 3.3
SD = 1.5
University
of
Maryland
Prof O
Crisis
Communication
Fall, 2013
33
13
11 female
2 male
M = 21.2
SD = 1.3
M = 3.2
SD = 1.7
Washington
College
Prof J
Theories of Peace
and Conflict
Fall, 2014
25
16
10 female
6 male
M = 21.1
SD = 3.6
M = 3.4
SD = 1.4
Political Orientation was measured on a 7-point scale (1 - very liberal; 7 – very conservative).

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
24
Results
Pre-test scores. While we aimed to collect a representative control group, it
remained possible that it might differ from treatment on pre-class measures. There were
no group differences for measures of social affiliation: “liking”, t(201.7) = 1.54, p = .125;
“similarity”, t(224.8) = 0.60, p = .547, “getting along”, t(227.7) = 0.95, p = .343,
“interaction”, t(244.8) = 0.63, p = .527. As in Study 1, pre-class attitudes were low for all
measures—MPre-liking = 1.71; MPre-similarity = 2.17; MPre-getting along = 1.90; MPre-interaction =
2.50—and were all significantly below the scale midpoint: “liking”, t(360) = 43.68, p <
.001; “similarity”, t(361) = 27.58, p < .001; “getting along”, t(358) = 35.41, p < .001;
“interaction”, t(361) = 15.6, p < .001. We also conducted combined placebo tests, using
the random effects structure described below for “Effect of Treatment”. There was no
interaction between Treatment and Question, F(3, 1070.8) = 0.15, p = .929, so the
parameter was removed from our model. In the resulting model, Treatment and Control
groups did not differ on the combined measure of pre-class social affiliation, b = -0.15,
t(326.0) = 1.21, p = .226. Thus, both treatment and control groups began the semester
with the same strong negative attitudes toward terrorists.
We compared treatment and control groups on the remaining pre-test covariates.
Groups did not differ in “openness to change”, t(236.8) = 1.00, p = .319, “knowledge”,
t(216.6) = 0.94, p = .348, or “confidence”, t(246.2) = 1.02, p = .308. Across groups, there
were significant differences in “interest”, t(266.1) = 4.86, p < .001, which was expected
given that treatment students chose to be in the course on terrorism. There were also
differences in gender, t(261.2) = 3.93, p < .001, and age, t(339.3) = 3.45, p < .001, and a
marginal difference in political orientation, t(248.2) = 1.78, p = .076, such that treatment

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
25
students were more likely to be younger, female, and (marginally more) liberal. Given
this, below we report final models that include age, gender, political orientation, and
“interest” as covariates, to ensure that key effects remain significant after controlling for
these pre-existing differences.
Random effects structure. As in Study 1, we created our random effects
structure by beginning with a maximal model and working backwards to remove non-
significant random-effects components (Bates et al., 2015; Table S5). We arrived at the
following model:
Attitude Post = 1 + (0 + Attitude Pre | Semester) + (1 + Attitude Pre | Student) + (1 | Professor)
Within our sample, there was significant variability in: (a) the by-semester relationship
between pre-class and post-class social affiliation (Attitude Pre | Semester), 2(1) = 133.7,
p < .001; (b) the by-student relationship between pre-class and post-class social
affiliation, (Attitude Pre | Student), |2(1) = 57.8, p < .001; (c) by-student mean post-class
social affiliation, 2(1) = 166.6, p < .001; and (d) by-professor mean post-class social
affiliation, (1 | Professor), 2(1) = 14.5, p < .001. Thus, our model allows that the
relationship between pre-class and post-class social affiliation differs for each semester
and student, and that mean post-class social affiliation differs for each student, and group
of students taught by a professor.
Sources of variability in this model were the same as in Study 1, with the addition
of the final term—(d) by-professor random-intercepts, which was not strictly necessary
for our purposes. The aim of Study 2 was to replicate Study 1 in a more diverse sample,
that is, it tested the claim that: treatment is generalizable beyond a sample of students,

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
26
within a novel sample of professors. Controlling for by-professor random-intercepts
actually tested an even stronger claim: that treatment is generalizable beyond a sample of
students and beyond a sample of professors. While the prospect of this outcome was
exciting, it was also unlikely that our sample of only 16 professors would allow for this
level of generalization. Thus, analyses of treatment including by-professor random
intercepts are reported in supplemental materials (Table S6), and the following random
effects structure was used in analyses below:
Attitude Post = 1 + (0 + Attitude Pre | Semester) + (1 + Attitude Pre | Student)
Effect of treatment. We added fixed effects of interest to the random effects
structure described above. Fixed effects included treatment, question (“liking”,
“similarity”, “getting along”, “interaction”), and pre-class social affiliation (to control for
any relationship not already captured by random effects). There was no interaction
between treatment and question, F(3, 972.3) = 1.75, p = .154. With the interaction term
removed, there was a main effect of question, F(3, 972.1) = 10.9, p < .001, and critically,
a main effect of treatment, F(1, 328.8) = 9.00, p = .003, b = 0.34, [0.12, 0.55] (Figure 2).
To ensure that these results did not depend on the inclusion of coauthor PK, we removed
his students from the sample (77 students, 20.4% of the total sample); the effect of
treatment remained significant, F(1, 251.11) = 4.89, p = .028, b = 0.27, [0.03, 0.50]. Still
excluding students taught by PK, treatment remained significant after controlling for age,
gender, political orientation, and pre-class interest, F(1, 239.7) = 3.92, p = 0.049, b =
0.26, [0.01, 0.52]. Thus, the effect of treatment observed in Study 1 was successfully
replicated in a novel sample of professors.
Once again, our effect of treatment did not depend on the specification of our
random effects structure. In an ordinary least squares regression, predicting the average

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
27
of our four post-class social affiliation measures (α = .79) and including treatment and
pre-class social affiliation (α = .78) as predictors, the effect of treatment was significant,
b = 0.32, t(334) = 2.93, p = .004. When school was added as a fixed effect (to control for
differences in the probability of assignment to treatment/control), treatment remained
significant, b = 0.29, t(324) = 2.11, p = .036. Thus, the effect of treatment was not
dependent on our specifying a random effects model.
Figure 2. Study 2 main effect of treatment. Treatment remained significant after
excluding students taught by coauthor PK (20.4 % of total sample), b = 0.27, [0.03, 0.50],
p = .028. The scale mid-point for post-class social affiliation is marked with a dotted line.
95% confidence interval computed using the BCa method (Efron, 1987; 5000 resamples).
Percent increase represents the treatment rating relative to the wait-list group. Error bars
represent standard error of the treatment coefficient.
Potential moderators. As in Study 1, we explored whether individual differences
moderated the effect of treatment, making it more or less effective. We explored

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
28
interactions between treatment and pre-class measures of a) knowledge and (b) opinion
confidence. Interactions were non-significant ps > ..71, (Table S7).
Potential confirmation bias. As in Study 1, we expected that some individual
differences might reduce our treatment’s effectiveness based on confirmation bias: (a)
students’ pre-class openness to change; (b) students’ and professors’ pre-class social
affiliation; and (c) students’ and professors’ political conservatism. None of these
covariates interacted with treatment, ps > .245 (Table S8). Thus, as in Study 1,
confirmation biases neither interfered with, nor accounted for, the effectiveness of
treatment.
General Discussion
Some of the most prominent terrorist groups today welcome hatred from opposing
states and citizens as a means of provoking indiscriminate retaliation against their own
communities (Kydd & Walter, 2006; Lake, 2002). This indiscriminate retaliation is at
best ineffective, and at worst counterproductive (Cronin, 2009); it runs counter to the
most effective counterterrorism policies, which stem from understanding terrorists as
rational agents, acting in pursuit of political goals. Hatred of terrorists, in either
policymakers or the citizens that elect them, is an obstacle for effective counterterrorism
strategies. Education, as a prejudice reduction technique is well suited to reducing this
hatred in this context. The present work tested whether students’ initial extreme negative
attitudes toward terrorists became less negative after they completed a college course on
the topic (treatment). Studies of education-based methods for prejudice reduction rarely
allow for causal inference (Paluck & Greene, 2009), making the use of as-if
randomization in Study 1 a particular strength of the present work. Study 1 demonstrated

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
29
that education about terrorists increased students’ social affiliation toward them: students
became more willing to say they would “like,” “get along with,” were “similar to,” and
would “interact with,” “someone belonging to a group that had carried out at least one
terrorist attack” (Figure 1). Study 2 replicated the effect within a sample of treatment and
control classes drawn from 31 classes, taught by 16 professors, at 11 United States
universities (Figure 2). Students’ attitudes did not become positive in either study (in
Figures 1 & 2, means and error bars are nowhere near the scale mid-point); we consider
this ideal—after treatment, students do not think positively of terrorists, but critically,
they no longer hate them as they once did.
People are known to have a confirmation bias; they selectively attend to and
remember information that reinforces their existing beliefs (Ghvirsman et al., 2016;
Haidt, 2001; Kuhn, 1991; Kunda, 1990; Wason, 1960). We initially hypothesized that
treatment would be influenced by the confirmation biases of either students (as measured
by their initial attitudes, their political orientations, or their self-reported willingness to
learn) or professors (as measured by professors’ political orientations or social affiliation
toward terrorists). However, we found no evidence that confirmation biases affected
treatment.
But, presumably, both students and professors do have confirmation biases—they
are a well-established effect in social psychology (Haidt, 2001; Kuhn, 1991; Kunda,
1990; Wason, 1960). It is reasonable to assume that if confirmation biases could have
exerted an influence then they would have, and their absence may assist speculation
about the psychological mechanisms responsible for reducing prejudice. One possibility
is that the effectiveness of treatment stems from general, rather than specific, knowledge;

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
30
that is, if there were some specific piece of knowledge, some silver bullet, that could have
changed a student’s mind about terrorists, then he or she could have chosen to ignore it,
or the professor could have neglected to teach it. By contrast, if the effectiveness of
treatment depends on general knowledge, then it should be more difficult for
confirmation biases to exert an effect—there is no specific piece of information that
students (or professors) could either ignore or latch on to. Consistent with this, all our
measures of social affiliation (e.g. “liking”) asked students about a generic terrorist, as
opposed to an individual from a particular group (e.g., ISIL, the IRA). If we had asked
about a particular group then treatment might depend on specific knowledge about that
group, such as the historical or social circumstances that motivated their attack.
But how exactly did education increase students’ social affiliation toward
terrorists? While prior work has reduced prejudice by providing positive examples of
stigmatized outgroups or of intergroup interactions (e.g. Gurin et al., 1999), given that
classes (particularly Study 1; see Appendix A) focused on counterterrorism and the
causes, objectives, and methods of terrorism, it is less likely that positive information
about terrorists was responsible for our effect. Professors taught their students about
terrorism—they were not explicitly interested in fostering students’ pro-social attitudes.
Given that most students did not receive positive information about terrorists, is it
possible that neutral information alone could dilute a strong initial prejudice?
Associative research provides a psychological framework that could account for
this effect (Greenwald, Banaji, Rudman, Farnham, Nosek, & Mellott, 2002). In this
framework, activating one concept calls associated concepts to mind, which are (to
varying extents) positively or negatively valenced. At the beginning of the semester,

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
31
students knew relatively little about terrorism; that is, the concept “terrorism” only
possessed a small set of associative links to (mostly negative) related concepts—e.g., Al
Qaeda, Osama bin Laden, ISIL, foreigners (Tuman, 2010). Thus, when “terrorism” was
called to mind, only these few negative associations came to mind with it. Treatment
classes, in our study, did not attempt to remove these initial negative associations, but
they may have flooded the concept “terrorism” with new associative links (e.g. IRA,
Weather Underground, specific political objectives of terrorist groups). Through learning
about terrorism, students may come to associate it with so much that its strong pejorative
connotations—the initial links—are diluted amongst the new associations they have
learned.
This mechanism, if confirmed in future work, would be promising for other anti-
prejudice interventions, particularly as an alternative to methods that focus exclusively on
positive counterexamples, where treatment can suffer from issues related to subtyping:
positive counterexamples are represented as distinct from the more general category, and
thus fail to reduce prejudice (Greenwald et al., 2002; Weber & Crocker, 1983).
Theoretically, our proposed mechanism—diluting pejorative links among new
associations—should be less likely to risk sub-typing, as the central concept is not
pressured by opposite positively and negatively valenced associations. Instead, the
intervention may avoid putting pressure on the concept “terrorism” to split, and it may do
this by using new associations that do not have a strong valence themselves (i.e. general
knowledge about terrorism). This finding is consistent with Salomon’s (2004)
interpretation of the intervention in Lustig (2002), where Israeli students studied external
conflicts, as opposed to their own. As in their intervention, the present work may allow

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
32
students to learn about the nature of terrorism in less immediate and emotionally charged
contexts. Consistent with this, Salomon notes that learning in this way could circumvent
defenses, such as entrenchment in existing beliefs—an outcome we also observed in the
present work.
Limitations and Extensions
While we favor this psychological explanation, we allow that other mechanisms
may also account for the effect of treatment. One possibility is that students learn to
challenge whether the label “terrorist” is properly applied, treating the term as no more or
less pejorative, but questioning whether its use is justified, or what its use actually tells
them about the targeted group. “Terrorist” is a nebulous term, and while its public usage
carries a clear negative connotation, its professional use is vigorously debated, to the
point that the formal definition varies even across government departments within the
United States (Hoffman, 2009). At the beginning of the semester, when students were
told that an individual’s group had committed a terrorist attack, they may have seen very
little ambiguity in the statement; at the end of the semester they might ask: what was the
attack (e.g., what did it target, what were the aims), and who declared it a terrorist attack?
In its public usage, to apply the label “terrorist” is to implicitly make a moral judgment
(Jenkins, 1985). In its professional usage, which students may have become familiar with,
it becomes valid to ask whether the label is being properly applied—does this group fit
the objective features that define terrorism? As students are exposed to a broader

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
33
academic understanding of terrorism, they may become less likely to blindly accept that
all usage of the term is appropriate*.
While the implications of this explanation may be more specific to terrorism,
rather than to anti-prejudice research more generally, its importance should not be
understated. Politicians have, at times, silenced meaningful debate by labeling their
opponents “terrorists,” and students may now see through this rhetorical strategy. Leaders
in Syria and Egypt today apply the label to much of their political opposition as a means
of justifying increased executive powers and repressive policies (Black, 2012; “Egypt’s
Muslim Brotherhood Declared ‘Terrorist Group’ ”, 2013); in the post-9/11 United States,
accused foreign terrorists can be held indefinitely without trial, (de Nies, 2011), while
extreme environmentalists who committed arson can be labeled terrorists and sent to
maximum security prisons for years (“Arsonists or Terrorists?”, 2011). Indeed, the label
of “terrorist” is one of the most powerful rhetorical tools in policy today, as invoking it
can shift the treatment of suspects and prisoners, the focus of the media, and government
funding and policies from a crime model to a war model (Miller & Gordon, 2014). Thus,
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
* Note that this explanation speaks to students’ knowledge of the term “terrorism”, and
their avoiding a blind acceptance of it as necessarily pejorative. Alternatively, the effect
could depend on which terrorist group students assumed to stand in for “a group that had
carried out at least one terrorist attack”, in our measures of social affiliation. For instance,
knowing few terrorist groups initially, students may think the question must refer to ISIS
or Al Qaeda, only to realize post-class that it could refer to many more groups. If this
were the case, then students who can name more terrorist groups (particularly Western
groups, such as environmental activists) should report more positive attitudes. Although
not analyzed above, students were asked to name up to 10 terrorist groups pre- and post-
class. Across our full Study 2 sample the post-class number of correct groups was
correlated with post-class social affiliation; however, the number of specifically western
groups was not. Furthermore, the number of correct groups did not eliminate the effect of
treatment when modeled as a covariate (see Supplemental Results). Thus, while this
explanation may describe a small component of treatment, it cannot completely account
for it.

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
34
shifting students’ understanding and interpretation of the label could have serious
political ramifications. Effectively, learning about terrorism may neuter it as a rhetorical
tool to inspire hatred.
We must also acknowledge that our discussion is framed in terms of “hatred”, yet
we did not explicitly measure hatred of terrorists. The questions that were asked were less
emotionally charged: specifically, whether students would “like”, “get along with”,
“interact with”, or were “similar to” someone who belongs to a terrorist group. Our
concern was that asking students about “hatred” might introduce demand characteristics
and prompt students to signal that they have the “correct” attitudes (which might be to
denigrate terrorists, or to renounce hate; in either case students would report extremes on
our scale and variance would be reduced). While our measures do not specifically ask
about “hatred”, they are reliable and collectively they can be used to measure
positive/negative attitudes. Even after treatment, students’ attitudes were significantly
below the scale mid-point, suggesting that they maintained their initial negative attitudes,
but that these attitudes were also less negative than they were before. Thus, students were
hardly ever willing to say that they liked terrorists, but they varied in the strength of their
objection to this prompt.
Finally, when we assert that the most effective counterterrorism strategies are
based around understanding the enemy, one might object: Couldn’t it be just as effective
to brutally repress a population, at least until it is incapable of engaging in terrorist
attacks? For instance, a major military effort defeated the Tamil Tigers in Sri Lanka;
likewise, Russia has faced several insurgencies and has successfully repressed the
majority of them, often through harsh measures such as mass deportation (Engelhardt,

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
35
1992). While authoritarian tactics can be effective in some cases, they also carry a
number of costs—beyond their moral repugnance—that make them less effective than
“democratic” methods (Byman, 2016). Furthermore, these authoritarian methods are
generally successful in counterinsurgency campaigns, where insurgents are organized,
have strong support from the local community, and are geographically confined. By
contrast, counterterrorism efforts must contend with loosely organized and
geographically dispersed groups—Paris cannot be bulldozed, or put under martial law,
until the threat of terrorism passes. The most proven counterterrorism methods rely on
human interaction and communication (Lyall & Wilson, 2009), and we believe that the
use of these methods will find more support when terrorists are less hated by the general
population.
Conclusion
The threat posed by terrorism is one to be taken seriously. However, the greatest
successes in counterterrorism have stemmed from an understanding of terrorists’ personal
and political motivations. Given this, hatred toward terrorists is an obstacle; it is actively
counterproductive and may even lead to policies that increase attacks (Cronin, 2009). We
found that learning about terrorism can decrease the extreme negative reactions it evokes.
This suggests that knowing our enemies is an effective step toward defeating them.

EDUCATION ABOUT TERRORISM CHANGES SOCIAL ATTITUDES!
36
Acknowledgements
The authors would like to thank Philip Collins, Katie Frake, Larisa Heiphetz, Joshua
Kertzer, Peiyan Liu, Damian Mencini, Kevin Miranda, Natalia Peña, Todd Rogers, James
Russell, Benjamin Seo, Joseph Young, Yael Ziera, and the National Consortium for the
Study of Terrorism and Responses to Terrorism (START) at the University of Maryland,
in addition to all participating students and professors. This work was funded by a
Research Across Departments and Schools (RADS) Grant from the Boston College
Provost’s Office.

37!
References(
Alexander, M, Bruning, J. R. (2011). How to Break a Terrorist: The U.S. Interrogators who used
Brains, not Brutality, to Take Down the Deadliest Man in Iraq. New York, NY: St. Martin’s.
Allport, G. (1954). The Nature of Prejudice. Reading, MA: Addison-Wesley
Arsonists or Terrorists? Film Details Rise, Fall of Earth Liberation Front Cell. (2011). Available at:
http://www.pbs.org/newshour/bb/environment-july-dec11-efp_09-01 [Accessed January 6,
2016].
Baayen, R. H., Davidson, D. J., & Bates, D. M. (2008). Mixed-effects modeling with crossed random
effects for subjects and items. Journal of Memory and Language, 59(4), 390–412.
http://dx.doi.org/10.1016/j.jml.2007.12.005
Bates, D., Kliegl, K., Vasishth, S., & Baayen, H. (2015). Parsimonious mixed models. ArXiv e-print;
submitted to Journal of Memory and Language, 2015. arXiv:1506.04967v1
Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed effects models using
lme4. Journal of Statistical Software, 67(1), 1–48. http://dx.doi.org/10.18637/jss.v067.i01
Black, I. (2012). Syrian President Assad Blames ‘Foreign Conspiracies’ for Crisis. Available at:
http://www.theguardian.com/world/2012/jan/10/syrian-president-assad-foreign-conspiracies
[Accessed January 6, 2016].
Broockman, D., & Kalla, J. (2016). Durably reducing transphobia: A field experiment on door-to-door
canvasing. Science, 353(6282), 220–224. http://dx.doi.org/10.1126/science.aad9713
Byman, D. (2016). ‘Death solves all problems’: The authoritarian model of counterinsurgency. Journal
of Strategic Studies, 39(1), 62–93. http://dx.doi.org/10.1080/01402390.2015.1068166
Cronin, A. K. (2009). How Terrorism Ends: Understanding the Decline and Demise of Terrorist
Campaigns. Princeton, NJ: Princeton Univ. Press.

38!
de Nies, Y. (2011). With Reservations, Obama Signs Act to Allow Detention of Citizens. Available at:
http://abcnews.go.com/blogs/politics/2011/12/with-reservations-obama-signs-act-to-allow-
detention-of-citizens [Accessed January 6, 2016].
Efron, B. (1987). Better bootstrap confidence intervals. Journal of the American Statistical Association,
82(307), 171–185. http://dx.doi.org/10.2307/2289144
Egypt’s Muslim Brotherhood Declared ‘Terrorist Group’. (2013). Available at:
http://www.bbc.com/news/world-middle-east-25515932 [Accessed January 6, 2016].
Engelhardt, M. J. (1992). Democracies, dictatorships and counterinsurgency: Does regime type really
matter? The Journal of Conflict Studies, 12(3), 52-63.
Greenwald, A. G., Banaji, M. R., Rudman, L. A., Farnham, S. D., Nosek, B. A., Mellott, D. S. (2002). A
unified theory of implicit attitudes, stereotypes, self-esteem, and self-concept. Psychological
Review, 109(1), 3–25. http://dx.doi.org/10.1037/0033-295x.109.1.3
Greenwald, A. G., McGhee, D. E., Schqartz, J. K. L. (1998). Measuring individual differences in
implicit cognition: the implicit association test. Journal of Personality and Social Psychology,
74(6), 1464–1480. http://dx.doi.org/10.1037/0022-3514.74.6.1464
Gurin, P., Peng, T., Lopez, G., Nagda, B. R. (1999). Context identity and intergroup relations. In D.
Prentice, & D. Miller (Eds.), Cultural Divides: The Social Psychology of Intergroup Contact (pp.
133–70), New York, NY: Sage.
Gvirsman, S. D., Huesmann, L. R., Dubow, E. F., Landau, S. F., Boxer, P., Shikaki, K. (2016). The
longitudinal effects of chronic mediated exposure to political violence on ideological beliefs
about political conflict among youths. Political Communication. 33(1), 98–117.
http://dx.doi.org/10.1080/10584609.2015.1010670
Haidt, J. (2001). The emotional dog and its rational tail: a social intuitionist approach to moral judgment.
Psychological Review, 108(4), 814–834. http://dx.doi.org/10.1037/0033-295x.108.4.814

39!
Halekoh, U., & Højsgaard, S. (2014). A Kenward-Roger approximation and parametric bootstrap
methods for tests in linear mixed Models — The R package pbkrtest. Journal of Statistical
Software, 59, 1–32. http://dx.doi.org/10.18637/jss.v059.i09
Hoffman, B. (2006). Inside Terrorism. New York, NY: Columbia Univ. Press.
Jenkins, B. (1985). The Future Course of International Terrorism. Santa Monica, CA: RAND.
Jost, J. T., Amodio, D. M. (2012). Political ideology as motivated social cognition: behavioral and
neuroscientific evidence. Motivation and Emotion, 36(1), 55–64.
http://dx.doi.org/10.1007/s11031-011-9260-7
Judd, C. M., Westfall, J., & Kenny, D. A. (2012). Treating stimuli as a random factor in social
psychology: A new and comprehensive solution to a pervasive but largely ignored problem.
Journal of Personality and Social Psychology, 103, 54–69. http://dx.doi.org/10.1037/a0028347
Kuhn, D. (1991). The Skills of Argument (Cambridge Univ. Press, Cambridge).
Kunda, Z. (1990). The case for motivated reasoning. Psychological Bulletin, 108(3), 480–498.
http://dx.doi.org/10.1037/0033-2909.108.3.480
Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B. (2015). lmerTest: Tests in linear mixed
effects models [Computer software manual]. http://CRAN.R-project.org/package=lmerTest. (R
Package version 2.0-25).
Kydd, A. H., & Walter, B. F. (2006). The strategies of terrorism. International Security, 31(1), 49–80.
http://dx.doi.org/10.1162/isec.2006.31.1.49
Lai, C. K., Marini, M., Lehr, S. A., Cerruti, C., Shin, J. E., Joy-Gaba, J. A. … Nosek, B. A. (2014).
Reducing implicit racial preferences: I. A comparative investigation of 17 interventions. Journal
of Experimental Psychology: General, 143(4), 1765–85. http://dx.doi.org/10.1037/a0036260
Lake, D. A. (2002). Rational extremism: Understanding terrorism in the twenty-first century. Dialogue
IO, 1(1), 15–29.

40!
Lustig, I. (2002). The effects of studying distal conflicts on the perception of a proximal one.
Unpublished master’s thesis, University of Haifa, Israel.
Lyall, J., & Wilson, I, (2009). Rage against the machines: Explaining outcomes in counterinsurgency
wars. International Organization, 63(1), 67–106. http://dx.doi.org/10.1017/s0020818309090031
Miller, S., Gordon, I. A. (2014). Investigative Ethics: Ethics for Police Detectives and Criminal
Investigators. Chicester, UK: Wiley-Blackwell.
Monteith, M. J., Zuwerink, J. R., Devine, P. G. (1994). Prejudice and prejudice reduction: Classic
challenges contemporary approaches. In P. G. Devine, & D. Hamilton (Eds.), Social Cognition:
Impact on Social Psychology (pp. 323–46), San Diego: Academic.
Moser, B. K., & Stevens, G. R. (1992). Homogeneity of variance in the two-sample means test. The
American Statistician, 56(1), 19–21. http://dx.doi.org/10.2307/2684403
Paluck, E. L. (2016). How to overcome prejudice: A brief conversation can have a lasting effect on
prejudice. Science, 353(6282), 147. http://dx.doi.org/10.1126/science.aaf5207
Paluck, E. L., Green, D. P. (2009). Prejudice reduction: what works? A review and assessment of
research and practice. Annual Review of Psychology, 60, 339–367.
http://dx.doi.org/10.1146/annurev.psych.60.110707.163607
Pettigrew, T. F., Tropp, L. R. (2006). A meta-analytic test of intergroup contact theory. Journal of
Personality and Social Psychology, 90(5), 751–83. http://dx.doi.org/10.1037/0022-
3514.90.5.751
R Core Team. (2015). R: A language and environment for statistical computing [Computer software
manual]. Vienna, Austria. Retrieved from http://www.R-project.org/
Schaller, M., Asp, C. H., Rosell, M. C., Heim, S. J. (1996). Training in statistical reasoning inhibits the
formation of erroneous group stereotypes. Personality and Social Psychology Bulletin, 22(8),
829–44. http://dx.doi.org/10.1177/0146167296228006

41!
Sherif, M., Harvey, O. J., White, B. J., Hood, W. R., Sherif, C. W. (1961). Intergroup Conflict and
Cooperation: The Robbers Cave Experiment. Norman, OK: Univ. Oklahoma Book Exchange.
Salomon, G. (2004). A narrative-based view of coexistence education. Journal of Social Issues, 60(2),
273–287. http://dx.doi.org/10.1111/j.0022-4537.2004.00118.x
Staniland, P. (2012). Between a rock and a hard place: Insurgent fratricide, ethnic defection, and the rise
of pro-state paramilitaries. Journal of Conflict Resolution, 56(1), 16–40.
http://dx.doi.org/10.1177/0022002711429681
Tajfel, H. (1970). Experiments in intergroup discrimination. Scientific American, 223(5), 96–102.
http://dx.doi.org/10.1038/scientificamerican1170-96
Tuman, J. S. (2010). Communicating Terror: The Rhetorical Dimensions of Terrorism. Thousand Oaks,
CA: SAGE.
Wason, P. C. (1960). On the failure to eliminate hypotheses in a conceptual task. Quarterly Journal of
Experimental Psychology, 12(3), 129–140. http://dx.doi.org/10.1080/17470216008416717
Weber, R. & Crocker, J. (1983). Cognitive processes in the revision of stereotypic beliefs. Journal of
Personality and Social Psychology, 45, 961–977. http://dx.doi.org/10.1037/0022-3514.45.5.961

42#
Supplemental Materials
Methods
Unabridged survey description, Studies 1 & 2. Pre-class and post-class surveys all began by
asking students to provide a 1-2 sentence definition of terrorism, to name up to ten terrorist groups that
had performed at least one attack, and to rate their knowledge and interest regarding terrorism
(“knowledge” and “interest”; 1 – “I have no knowledge of/interest in the topic”; 7 – “I have a
tremendous amount of knowledge about/interest in the topic”). Next, students completed blocks of
questions in a random order about: (1) common motives for individuals and groups to engage in
terrorism; (2) whether facts related to terrorists and terrorism are true or false; (3) whether terrorism is
effective and whether Al-Qaeda had succeeded; (4) the threat of terrorism to the United States and to the
student him or herself; (5) the student’s opinions about the efficacy of government counterterrorism
policies; and finally, (6) the student’s attitude toward terrorists (i.e. social affiliation). At the end of the
survey, students completed a brief demographic questionnaire and rated the likelihood that they would
change their opinions on terrorism (“openness to change”; 1 – “very unlikely”; 7 – “very likely”) and
the confidence they had in their opinions (“confidence”; 1 – “not confident at all”; 7 – “extremely
confident”).
Results
Effect of post-class number of terrorist groups named. As described above, in pre-class and
post-class surveys students were asked to name up to ten terrorist groups that had performed at least one
attack. These groups were coded by research assistants (unfamiliar with the study hypotheses) for
correctness, whether a group was Islamic or Middle Eastern, whether a group was Western, and whether
a group was based in the United States. In our dataset for Study 2, we tested the correlations between
these measures and average post-class social affiliation (α = .79). Post-class social affiliation was
correlated with the number of correct groups, r = .208, p < .001, and the number Middle Eastern groups,
r = .159, p = .004, but not with the number of Western or American groups (Table S9).

43#
We explored whether either correct groups or Middle Eastern groups interacted with treatment,
or eliminated the effect of treatment as a covariate when added to our Study 2 random effects model.
Attitude Post = 1 + (0 + Attitude Pre | Semester) + (1 + Attitude Pre | Student)
Neither interacted with treatment (correct groups*treatment, b = -0.03, t(337.9) = 0.74, p = .463; Middle
East groups*treatment, b = -0.03, t(352.2) = 0.60, p = .546), and treatment remained significant when
both were included as covariates, b = 0.289, t(341.3) = 2.31, p = .022.

44#
Table S1. Random effects structure in Study 1.
Model #
Random effects structure
Parameter tested
Log-likelihood ratio
1
(1 + Attitude Pre | Student) +
(1 + Treatment + Attitude Pre | Semester)
-
-
2
(1 + Attitude Pre | Student) +
(1 + Treatment + Attitude Pre || Semester)1
Parameter correlations within:
(1 + Treatment + Attitude Pre | Semester)
2(3) = 5.28, p = .152
3
(1 + Attitude Pre | Student) +
(0 + Treatment + Attitude Pre || Semester)
(1 | Semester)
2(1) = 0, p = 1
4
(1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
(Treatment | Semester)
2(1) = 2.27, p = .132
Final
Model
(1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
(1 | Student)
(Attitude Pre | Student)
(Attitude Pre | Semester)
2(1) = 27.57, p < .001
2(1) = 7.62, p = .006
2(1) = 18.56, p < .001
Following recent recommendations (Bates, Kliegl, Vasishth, & Baayen, 2015), we used log-likelihood ratio tests compare models with,
and without a given parameter; correlations within a level were removed to test parameters, but were returned to the final model.
1 The notation “||“ indicates that correlations between parameters are not to be calculated—e.g. (1 + Attitude Pre || School) indicates that
by-school random intercepts, and by-school Attitude Pre random slopes should be calculated, but not the correlation between them.

45#
Table S2. Effect of treatment in Study 1.
Variable of Interest
Model Specification / Fixed Effects
F test
b [95% CI]
Treatment
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment
Question
Treatment
F(3, 157.6) = 4.4, p = .005
F(1, 52.5) = 7.6, p = .008
Intercept:
2.27, [1.83, 2.71]
Interact: 0.30, [-0.06,0.67]
Like: -0.24, [-0.57, 0.10]
Similar: 0.32, [-0.02, 0.67]
0.70, [0.21, 1.19]
Treatment
Controlling for student
year, which biased
randomization into
treatment and control
groups.
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester) + (1 | Professor)
Fixed effects: Attitude Pre + Question +
Treatment + Student_Year
Question
Student_Year
Treatment
F(3, 157.6) = 4.7, p = .004
F(1, 58.7) = 0.71, p = .583
F(1, 52.8) = 5.3, p = .025
Intercept:
2.30, [1.86, 2.75]
Interact: 0.33, [-0.03, 0.70]
Like: -0.23, [-0.57, 0.10]
Similar: 0.34, [-0.01, 0.68]
Sophomore: -1.74, [-4.24, 0.76]
Junior: -1.46, [-3.95 , 1.05]
Senior: -1.72, [-4.15 , 0.76]
Graduate: -0.85 [-3.82 , 2.22]
0.76, [0.13, 1.43]
All degrees of freedom calculated using the Kenward-Roger approximation method using lmerTest (Kuznetsova, Brockhoff, &
Christensen, 2015) and pbkr packages (Halekoh & Højsgaard, 2014). Confidence intervals are calculated using the bias corrected
asymmetry method (bCA; Efron, 1987), with 5000 resamples. All continuous IVs are mean-centered; categorical Ivs are dummy coded.

46#
Table S3. Potential moderators in Study 1.
Variable of Interest
Model Specification / Fixed Effect
F test
b [95% CI]
Treatment *
Knowledge Pre
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment *
Knowledge Pre
Question
Treatment
Knowledge Pre
Treatment * Knowledge Pre
F(3, 157.4) = 4.4, p = .005
F(1, 51.2) = 6.9, p = .012
F(1, 50.3) = 0.48, p = .492
F(1, 49.5) = 0.88, p = .351
Intercept:
2.29, [1.85, 2.74]
Interact: 0.31, [-0.06, 0.67]
Like: -0.24, [-0.57, 0.10]
Similar: 0.33, [-0.02, 0.67]
0.68, [0.19, 1.17]
0.19, [-0.33, 0.71]
-0.30, [-0.90, 0.30]
Treatment *
Interest Pre
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment *
Interest Pre
Question
Treatment
Interest Pre
Treatment * Interest Pre
F(3, 157.4) = 4.47, p = .005
F(1, 49.2) = 7.9, p = .007
F(1, 48.0) = 0.60, p = .441
F(1, 49.3) = 0.29, p = .593
Intercept:
2.23, [1.78, 2.68]
Interact: 0.31, [-0.05, 0.68]
Like: -0.23, [-0.57, 0.10]
Similar: 0.33, [-0.02, 0.67]
0.73, [0.24, 1.23]
0.17, [-0.25, 0.61]
-0.16, [-0.74, 0.42]
Treatment *
Confidence Pre
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment *
Confidence Pre
Question
Treatment
Confidence Post
Treatment * Confidence Pre
F(3, 150.7) = 3.6, p = .015
F(1, 48.7) = 6.6, p = .013
F(1, 43.9) = 0.13, p = .716
F(1, 44.3) = 0.19, p = .663
Intercept:
2.29, [1.86, 2.73]
Interact: 0.23, [-0.13, 0.59]
Like: -0.25, [-0.58, 0.09]
Similar: 0.27, [-0.07, 0.60]
0.67, [0.18, 1.16]
-0.07, [-0.43, 0.29]
0.10, [-0.32, 0.51]
All degrees of freedom calculated using the Kenward-Roger approximation method using lmerTest (Kuznetsova et al., 2015) and pbkr
packages (Halekoh & Højsgaard, 2014). Confidence intervals are calculated using the bias corrected asymmetry method (bCA; Efron,
1987), with 5000 resamples. All continuous Ivs are mean-centered; categorical Ivs are dummy coded.

47#
Table S4. Potential confirmation bias in Study 1.
Variable of Interest
Model Specification / Fixed Effect
F test
b [95% CI]
Treatment *
Attitude Pre
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment *
Attitude Pre
Question
Treatment
Attitude Pre
Treatment * Attitude Pre
F(3, 157.1) = 4.3, p = .006
F(1, 52.2) = 8.9, p = .004
F(1, 4.4) = 3.8, p = .115
F(1, 35.0) = 1.2, p = .286
Intercept:
2.23, [1.78, 2.66]
Interact: 0.31, [-0.05, 0.68]
Like: -0.23, [-0.56, 0.11]
Similar: 0.33, [-0.02, 0.67]
0.76, [0.26, 1.26]
0.30, [0.02, 0.60]
0.19, [-0.14, 0.54]
Treatment *
Openness to Change
Pre
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment *
Openness to Change Pre
Question
Treatment
Openness to Change Pre
Treatment * Openness to Change Pre
F(3, 150.5) = 3.7, p = .014
F(1, 1, 49.0) = 7.3, p = .010
F(1, 46.4) = 0.34, p = .564
F(1, 47.2) = 0.16, p = .690
Intercept:
2.28, [1.85, 2.71]
Interact: 0.24, [-0.12, 0.59]
Like: -0.24, [-0.57, 0.09]
Similar: 0.28, [-0.06, 0.61]
0.69, [0.21, 1.18]
-0.10, [-0.45, 0.24]
-0.09, [-0.52, 0.34]
Treatment *
Political Orientation
1 –
“Very Liberal”
7 –
“Very Conservative”
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment *
Political Orientation
Question
Treatment
Political Orientation
Treatment * Political Orientation
F(3, 150.7) = 3.8, p = .012
F(1, 48.6) = 6.7, p = .013
F(1, 45.2) = 0.13, p = .719
F(1, 45.7) = 0.18, p = .674
Intercept:
2.29, [1.86, 2.73]
Interact: 0.22, [-0.14, 0.58]
Like: -0.24, [-0.58, 0.09]
Similar: 0.30, [-0.05, 0.64]
0.67, [0.18, 1.16]
0.05, [-0.20, 0.29]
-0.07, [-0.40, 0.25]
All degrees of freedom calculated using the Kenward-Roger approximation method using lmerTest (Kuznetsova et al., 2015) and pbkr
packages (Halekoh & Højsgaard, 2014). Confidence intervals are calculated using the bias corrected asymmetry method (bCA; Efron,
1987), with 5000 resamples. All continuous Ivs are mean-centered; categorical Ivs are dummy coded.

48#
Table S5. Random effects structure in Study 2.
Model #
Random effects structure
Parameter tested
Log-likelihood ratio
1
(1 + Attitude Pre | Student) +
(1 + Attitude Pre | Professor) +
(1 + Attitude Pre | Class: Professor)1 +
(1 + Treatment + Attitude Pre | Semester) +
(1 + Attitude Pre | School)
-
-
2
(1 + Attitude Pre | Student) +
(1 + Attitude Pre | Professor) +
(1 + Attitude Pre | Class: Professor) +
(1 + Treatment + Attitude Pre | Semester) +
(1 + Attitude Pre || School)2
Parameter correlations within:
(1 + Attitude Pre | School)
2(1) = 0, p = 1
3
(1 + Attitude Pre | Student) +
(1 + Attitude Pre | Professor) +
(1 + Attitude Pre | Class: Professor) +
(1 + Treatment + Attitude Pre | Semester) +
(1 | School)
(Attitude Pre | School)
2(1) = 0, p = 1
4
(1 + Attitude Pre | Student) +
(1 + Attitude Pre | Professor) +
(1 + Attitude Pre | Class: Professor) +
(1 + Treatment + Attitude Pre | Semester)
(1 | School)
2(1) = 0, p = 1
5
(1 + Attitude Pre | Student) +
(1 + Attitude Pre | Professor) +
(1 + Attitude Pre || Class: Professor) +
(1 + Treatment + Attitude Pre | Semester)
Parameter correlations within:
(1 + Attitude Pre || Class: Professor)
2(1) = 0.12, p = .724
6
(1 + Attitude Pre | Student) +
(1 + Attitude Pre | Professor) +
(1 | Class: Professor) +
(1 + Treatment + Attitude Pre | Semester)
(Attitude Pre | Class: Professor)
2(1) = 0, p = 1
7
(1 + Attitude Pre | Student) +
(1 + Attitude Pre | Professor) +
(1 + Treatment + Attitude Pre | Semester)
(1 | Class: Professor)
2(1) = 0.06, p = .803
8
(1 + Attitude Pre | Student) +
(1 + Attitude Pre | Professor) +
(1 + Treatment + Attitude Pre || Semester)
Parameter correlations within:
(1 + Treatment + Attitude Pre | Semester)
2(3) = 3.29, p = .349
9
(1 + Attitude Pre | Student) +
(1 + Attitude Pre | Professor) +
(0 + Treatment + Attitude Pre || Semester)
(1 | Semester)
2(1) = 0, p = 1
10
(1 + Attitude Pre | Student) +
(1 + Attitude Pre | Professor) +
(0 + Attitude Pre | Semester)
(Treatment | Semester)
2(1) = 2.52, p = .113
11
(1 + Attitude Pre | Student) +
(1 + Attitude Pre || Professor) +
(0 + Attitude Pre | Semester)
Parameter correlations within:
(1 + Attitude Pre | Professor)
2(1) = 1.92, p = .165
12
(1 + Attitude Pre | Student) +
(1 | Professor) +
(0 + Attitude Pre | Semester)
(Attitude Pre | Professor)
2(1) = 0.08, p = .772
Final
Model
(1 + Attitude Pre | Student) +
(1 | Professor) +
(0 + Attitude Pre | Semester)
(1 | Student)
(Attitude Pre | Student)
(1 | Professor)
(Attitude Pre | Semester)
2(1) = 116.6, p < .001
2(1) = 57.8, p < .001
2(1) = 14.5, p < .001
2(1) = 113.7, p < .001
Log-likelihood ratio tests compare models with, and without a parameter, to test whether it contributes a significant amount of variance
(Bates, et al., 2015).
1 The notation “:“ indicates that one level is embedded in another—e.g. (1 | Class: Professor) would calculate random intercepts for
classes within professors.
2 The notation “||“ indicates that correlations between parameters are not to be calculated—e.g. (1 + Attitude Pre || School) indicates that
by-school random intercepts, and by-school Attitude Pre random slopes should be calculated, but not the correlation between them.

49#
Table S6. Effect of treatment in Study 2.
Variable of Interest
Model Specification / Fixed Effects
F test
b [95% CI]
Treatment
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Attitude Pre + Question +
Treatment
Attitude Pre
Question
Treatment
F(1, 2.2) = 175.1, p = .004
F(3, 972.1) = 10.9, p < .001
F(1, 328.8) = 9.0, p = .003
Intercept:
2.14, [1.94, 2.33]
0.51, [0.44, 0.58]
Interact: 0.23, [0.09, 0.37]
Like: -0.10, [-0.23, 0.04]
Similar: 0.24, [0.10, 0.38]
0.34, [0.12, 0.55]
Treatment
Excluding students
taught by coauthor PK.
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Attitude Pre + Question +
Treatment
Attitude Pre
Question
Treatment
F(1, 1.4) = 99.5, p = .030
F(3, 755.4) = 8.5, p < .001
F(1, 251.1) = 4.9, p = .028
Intercept:
2.08, [1.87, 2.29]
0.49, [0.40, 0.58]
Interact: 0.23, [0.08, 0.39]
Like: -0.09, [-0.24, 0.06]
Similar: 0.22, [0.07, 0.37]
0.27, [0.03, 0.51]
Treatment
Excluding students
taught by coauthor PK.
Controlling for pre-
existing differences
between treatment and
wait-list groups:
Age
Gender (Female = 1)
Interest Pre
Political Orientation
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester) + (1 | Professor)
Fixed effects: Attitude Pre + Question +
Treatment + Age + Gender + Interest Pre +
Political Orientation
Attitude Pre
Question
Age
Gender
Interest Pre
Political Orientation
Treatment
F(1, 1.3) = 86.9, p = .035
F(3, 731.7) = 8.6, p < .001
F(1, 239.6) = 7.0, p = .009
F(1, 242.5) = 1.3, p = .258
F(1, 256.6) = 2.4, p = .121
F(1, 242.1) = 4.3, p = .040
F(1, 239.7) = 3.9, p = .049
Intercept:
2.15, [1.89, 2.42]
0.48, [0.39, 0.57]
Interact: 0.27, [0.11, 0.42]
Like: -0.07, [-0.22, 0.08]
Similar: 0.23, [0.08, 0.38]
-0.03, [-0.05, -0.01]
-0.14, [-0.38, 0.10]
0.09, [-0.02, 0.20]
-0.09, [-0.17, -0.01]
0.26, [0.01, 0.51]
Treatment
Including by-professor
random intercepts.
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester) + (1 | Professor)
Fixed effects: Attitude Pre + Question +
Treatment
Attitude Pre
Question
Treatment
F(1, 2.05) = 164.1, p = .005
F(3, 971.6) = 11.3, p < .001
F(1, 57.6) = 2.4, p = .124
Intercept:
2.12, [1.81, 2.42]
0.50, [0.09, 0.37]
Interact: 0.23, [-0.23, 0.04]
Like: -0.10, [-0.23, 0.04]
Similar: 0.24, [0.11, 0.38]
0.26, [-0.05, 0.57]
All degrees of freedom calculated using the Kenward-Roger approximation method using lmerTest (Kuznetsova et al., 2015) and pbkr
packages (Halekoh & Højsgaard, 2014). Confidence intervals are calculated using the bias corrected asymmetry method (bCA; Efron,
1987), with 5000 resamples. All continuous Ivs are mean-centered; categorical Ivs are dummy coded.

50#
Table S7. Potential moderators in Study 2.
Non-Significant Moderators
Variable of Interest
Model Specification / Fixed Effect
F test
b [95% CI]
Treatment *
Opinion Confidence Pre
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Attitude Pre + Question +
Treatment * Opinion Confidence Pre
Attitude Pre
Question
Treatment
Opinion Confidence Pre
Treatment * Opinion Confidence Pre
F(1, 2.2) = 170.9, p = .004
F(3, 965.7) = 11.0, p < .001
F(1, 324.8) = 8.6, p = .004
F(1, 332.7) = 0.05, p = .817
F(1, 329.9) = 0.01, p = .928
Intercept:
2.13, [1.93, 2.33]
0.51, [0.44, 0.58]
Interact: 0.23, [0.09, 0.38]
Like: -0.09, [-0.23, 0.04]
Similar: 0.24, [0.10, 0.38]
0.33, [0.11, 0.55]
-0.02, [-0.17, 0.14]
0.01, [-0.18, 0.19]
Treatment *
Knowledge Pre
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Attitude Pre + Question +
Treatment * Knowledge Pre
Attitude Pre
Question
Treatment
Knowledge Pre
Treatment * Knowledge Pre
F(1, 2.3) = 150.2, p = .004
F(3, 930.1) = 10.0, p < .001
F(1, 312.5) = 8.8, p = .003
F(1, 314.0) = 0.03, p = .856
F(1, 311.4) = 0.13, p = .714
Intercept:
2.13, [1.92, 2.34]
0.50, [0.42, 0.57]
Interact: 0.21, [0.07, 0.35]
Like: -0.10, [-0.24, 0.03]
Similar: 0.23, [0.09, 0.36]
0.35, [0.12, 0.58]
-0.02, [-0.22, 0.18]
0.05, [-0.19, 0.28]
All degrees of freedom calculated using the Kenward-Roger approximation method using lmerTest (Kuznetsova et al., 2015) and pbkr
packages (Halekoh & Højsgaard, 2014). Confidence intervals are calculated using the bias corrected asymmetry method (bCA; Efron,
1987), with 5000 resamples. All continuous Ivs are mean-centered; categorical Ivs are dummy coded.

51#
Table S8. Potential confirmation bias in Study 2.
Variable of Interest
Model Specification / Fixed Effect
F test
b [95% CI]
Treatment *
Attitude Pre
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment *
Attitude Pre
Question
Treatment
Attitude Pre
Treatment * Attitude Pre
F(3, 971.6) = 10.9, p < .001
F(1, 327.8) = 8.7, p = .003
F(1, 9.4) = 62.7, p < .001
F(1, 165.9) = 0.09, p = .767
Intercept:
2.14, [1.94, 2.34]
Interact: 0.23, [0.09, 0.37]
Like: -0.10, [-0.23, 0.04]
Similar: 0.24, [0.10, 0.38]
0.33, [0.11, 0.55]
0.53, [0.41, 0.65]
-0.02, [-0.17, 0.12]
Treatment *
Professor Attitude Pre
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment *
Professor Attitude Pre
Attitude Pre
Question
Treatment
Professor Attitude Pre
Treatment * Professor Attitude Pre
F(1, 2.2) = 167.7, p = .004
F(3, 950.8) = 10.8, p < .001
F(1, 321.9) = 7.5, p = .006
F(1, 332.8) = 0.63, p = .427
F(1, 342.0) = 0.42, p = .518
Intercept:
2.15, [1.95, 2.35]
0.51, [0.43, 0.58]
Interact: 0.22, [0.08, 0.36]
Like: -0.09, [-0.23, 0.04]
Similar: 0.25, [0.11, 0.39]
0.31, [0.09, 0.53]
0.08, [-0.11, 0.27]
-0.07, [-0.28, 0.14]
Treatment *
Openness to Change
Pre
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment *
Openness to Change Pre
Attitude Pre
Question
Treatment
Openness to Change Pre
Treatment * Openness to Change Pre
F(1, 2.2) = 163.5, p = .004
F(3, 966.0) = 11.1, p < .001
F(1, 322.8) = 8.6, p = .004
F(1, 326.8) = 0.39, p = .531
F(1, 324.1) = 1.4, p = .246
Intercept:
2.13, [1.94, 2.33]
0.50, [0.43, 0.57]
Interact: 0.24, [0.10, 0.38]
Like: -0.10, [-0.23, 0.04]
Similar: 0.24, [0.10, 0.38]
0.33, [0.11, 0.54]
0.04, [-0.09, 0.17]
0.10, [-0.06, 0.26]
Treatment *
Political Orientation
1 –
“Very Liberal”
7 –
“Very Conservative”
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment *
Political Orientation
Attitude Pre
Question
Treatment
Political Orientation
Treatment * Political Orientation
F(1, 2.2) = 170.1, p = .004
F(3, 971.8) = 11.1, p < .001
F(1, 324.1) = 10.3, p = .001
F(1, 322.5) = 1.4, p = .246
F(1, 323.1) = 0.01, p = .910
Intercept:
2.11, [1.92, 2.31]
0.50, [0.43, 0.57]
Interact: 0.23, [0.09, 0.37]
Like: -0.10, [-0.23, 0.04]
Similar: 0.24, [0.10, 0.38]
0.36, [0.14, 0.58]
-0.07, [-0.19, 0.05]
-0.01, [-0.15, 0.14]
Treatment *
Professor Political
Orientation
1 –
“Very Liberal”
7 –
“Very Conservative”
Random Effects: (1 + Attitude Pre | Student) +
(0 + Attitude Pre | Semester)
Fixed effects: Question + Treatment *
Professor Political Orientation
Attitude Pre
Question
Treatment
Professor Political Orientation
Treatment * Professor Political Orientation
F(1, 2.2) = 169.4, p = .004
F(3, 950.9) = 10.8, p < .001
F(1, 319.2) = 10.8, p = .001
F(1, 326.2) = 1.4, p = .238
F(1, 324.8) = 0.03, p = .858
Intercept:
2.10, [1.89, 2.30]
0.51, [0.44, 0.58]
Interact: 0.22, [0.08, 0.36]
Like: -0.09, [-0.23, 0.04]
Similar: 0.25, [0.11, 0.39]
0.38, [0.16, 0.60]
0.11, [-0.07, 0.29]
-0.02, [-0.22, 0.18]
All degrees of freedom calculated using the Kenward-Roger approximation method using lmerTest (Kuznetsova et al., 2015) and pbkr
packages (Halekoh & Højsgaard, 2014). Confidence intervals are calculated using the bias corrected asymmetry method (bCA; Efron,
1987), with 5000 resamples. All continuous Ivs are mean-centered; categorical Ivs are dummy coded.

52#
Table S9. Post-class Pearson’s correlations among # of reported terrorist groups and social affiliation.
Post-class
measure
Social
affiliation
Total
correct
groups
Middle
Eastern /
Islamic
groups
Western
groups
American
groups
Social affiliation
(α = .79)
Pearson's r
p-value
—
—
0.208
< .001
0.159
0.004
0.059
0.282
0.032
0.562
Total correct
groups
Pearson's r
p-value
—
—
0.715
< .001
0.385
< .001
0.141
0.009
Middle Eastern
/ Islamic groups
Pearson's r
p-value
—
—
0.013
0.803
-0.080
0.138
Western groups
Pearson's r
p-value
—
—
0.695
< .001
American
groups
Pearson's r
p-value
—
—

53#
Supplemental References
Bates, D., Kliegl, K., Vasishth, S., & Baayen, H. (2015). Parsimonious mixed models. ArXiv e-
print; submitted to Journal of Memory and Language, 2015. arXiv:1506.04967v1
Efron, B. (1987). Better bootstrap confidence intervals. Journal of the American Statistical
Association, 82(307), 171–185. http://dx.doi.org/10.2307/2289144
Halekoh, U., & Højsgaard, S. (2014). A Kenward-Roger approximation and parametric bootstrap
methods for tests in linear mixed Models — The R package pbkrtest. Journal of Statistical
Software, 59, 1–32. http://dx.doi.org/10.18637/jss.v059.i09#
Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B. (2015). lmerTest: Tests in linear mixed
effects models [Computer software manual]. http://CRAN.R-
project.org/package=lmerTest. (R Package version 2.0-25).

54#
Appendix A: Course Readings:
P. Krause; Terrorism, Insurgency, and Political Violence; Spring 2015
Class Plan
WEEK 1: What are Terrorism and Insurgency? Definitions and Cases Across History
September 1: Defining Terrorism and Insurgency: A New or Old Phenomenon?
September 3: Film “The Weather Underground” Part I
WEEK 2: Individual Level Causes and Objectives of Terrorism and Insurgency
September 8: Psychology, Economics, Education
September 10: Film “The Weather Underground” Part II
WEEK 3: Organizational, Strategic Level Causes and Objectives of Terrorism and
Insurgency
September 15: Religion, Gender, Ideology
September 17: Solidarity, Networks, and Numbers; Organizational Survival and Competition
WEEK 4: Methods and Mechanisms: Strategies of Terrorism and Insurgency
September 22: Political Grievances and Occupation; Failed States and State Sponsors
September 24: Strategies of Terrorism and Insurgency- Academics
WEEK 5: Methods and Mechanisms: Suicide Bombing and WMD
September 29: Strategies of Terrorism and Insurgency- Practitioners
October 1: Suicide Bombing and Weapons of Mass Destruction in Terrorism and Insurgency
WEEK 6: Morality and the Media
October 6: Morality, Emotions, and the Media in Terrorism and Insurgency
October 8: Exam #1
WEEK 7: The Impact and Effectiveness of Terrorism and Insurgency
October 13: Individual and Organizational Level Effects: Fear, Casualties, Support, Group Strength
October 15: Strategic Level Effects: Political Concessions, Military Withdrawals, New States
WEEK 8: Al-Qaeda
October 20: Al-Qaeda: The Past
October 22: Al-Qaeda: The Present and Future
WEEK 9: The Boundaries of Terrorism: Nonviolence and State Terror
October 27: Nonviolence and Non-Lethal Violence
October 29: States and Terrorism: Repression, Mass Violence, and Genocide
WEEK 10: The Insurgencies in Iraq and Syria
November 3: The Causes, Dynamics, and Effects of the Insurgencies
November 5: Foreign Fighters, ISIS, and Insurgent Rivalries
WEEK 11: Counterterrorism and Counterinsurgency I
November 10: Exam #2 and “If a Tree Falls”
November 12: How Terrorism and Insurgency End
WEEK 12: Counterterrorism and Counterinsurgency II
November 17: CT and COIN Debates: Hard & Soft Power, Democratization, Threat Inflation
November 19: The Freedom of Speech, Profiling and Airport Security, Torture
WEEK 13: Counterterrorism and Counterinsurgency III
November 24: Drones and Intelligence Agencies
November 26: *Happy Thanksgiving*
WEEK 14: The Boston Marathon Bombings

55#
December 1: Definitions, Causes, and the Media
December 3: Effects, Community Response, and the Dzokhar Tsarnaev Trial
WEEK 15: Terrorism, Insurgency, and Political Violence, Now and in the Future
December 7: Remaining Questions and Lessons Learned
Readings and Class Schedule
Before Classes Begin
By Monday, August 31 at noon (the day before the first class meeting), you must email Professor Krause your own 1-
2 sentence definition of “terrorism” without consulting any sources. Please send your definition to
peter.krause.2@bc.edu with the subject heading “PO352701 Terrorism Definition”. This assignment will be graded
for timely completion.
WEEK 1: What are Terrorism and Insurgency? Definitions and Cases Across History
Key Questions
How do scholars, governments, the media, and the public define terrorism and insurgency? Are
terrorism and insurgency distinct concepts? How are they similar and different?Is terrorism a new
or old phenomenon?What are some key cases of terrorism and insurgency?
Skills Introduced
Defining and comparing concepts
September 1: Defining Terrorism and Insurgency: A New or Old Phenomenon?
Required Readings
. Bruce Hoffman, Inside Terrorism, Ch. 1, pp. 1-41 
. John Gerring, “What Makes a Concept Good? A Critical Framework for Understanding Concept
Formation in the Social Sciences,” Polity, Vol. 31, No. 3 (1999), Table 1, pp. 367 
. David Rapoport, "Fear and Trembling: Terrorism in Three Religious Traditions," American Political
Science Review Vol. 78, No. 3 (1984) pp. 658-677 
. Alexander Spencer and Rohan Gunaratna, “Is the New Terrorism Really New?” in Debating
Terrorism and Counterterrorism, Ch. 1, pp. 1-34 
Recommended Readings 
 Alison M. Jaggar, “What Is Terrorism, Why Is It Wrong, and Could It Ever Be Morally
Permissible?” Journal of Social Philosophy, Vol. 36, No. 2 (2005) pp. 202-215 
 Colin Beck and Emily Miner, “Who Gets Designated a Terrorist and Why?” Social Forces Vol. 91,
No. 3, pp. 837-858 
 Leonard Weinberg, Ami Pedahzur, and Sivan Hirsch-Hoeffler, “The Challenges of
Conceptualizing Terrorism,” Terrorism and Political Violence, Vol. 16, No. 4 (2004) pp. 777-794
 Alex P. Schmid and Albert J. Jongman, eds., Political Terrorism: A New Guide to Actors, Authors,
Concepts, Data Bases, Theories, and Literature (New Brunswick, NJ: Transaction Publishers,
2005) 
 Martha Crenshaw, “Thoughts on Relating Terrorism to Historical Context,” in Martha Crenshaw,
ed., Terrorism in Context (University Park, PA: The Pennsylvania State University Press, 1995)

56#
 John Horgan and Michael Boyle, “The Case Against Critical Terrorism Studies,” Critical Studies on
Terrorism, Vol. 1, No. 1 (2008) pp. 51-64 
September 3: Film “The Weather Underground” Part I 
WEEK 2: Individual Level Causes and Objectives of Terrorism and Insurgency 
Key Questions 
Who are the key actors in terrorism and insurgency campaigns?
What are the levels of analysis for examining terrorism and insurgency?
Do mental illness, poverty, a lack of education, ideology, gender, or religion cause terrorism? 
How can an individual become radicalized? Is ‘radicalization’ necessary to commit terrorism?
Skills Introduced
How to read as a scholar and analyst: Identifying and critiquing argumentsWhat is political science?
Understanding variables, theories, predictions, tests, and evidence How to generate theories and
hypotheses
September 8: Psychology, Economics, Education
Required Readings
 “You Don’t Need a Weatherman to Know Which Way the Wind Blows,” “A Declaration of
War,” and “Headquarters” in Bernadine Dorhn, Bill Ayers, and Jeff Jones, eds., Sing a Battle
Song: The Revolutionary Poetry, Statements, and Communiqués of the Weather Underground 1970-1974
(New York: Seven Stories, 2006) 
 Ehud Sprinzak, “The Psychopolitical Formation of Extreme Left Terrorism in a Democracy: The
Case of the Weathermen,” in Walter Reich, ed., Origins of Terrorism (Washington, D.C.:
Woodrow Wilson Center Press, 1998) pp. 65-85 
 Diego Gambetta and Steffen Hertog, “Why Are There So Many Engineers Among Islamic
Radicals?” European Journal of Sociology, Vol. 50, No. 2 (2009) pp. 201-230 
 James A. Piazza and Karin von Hippel, “Does Poverty Serve as a Root Cause of Terrorism?” in
Gottlieb, Debating Terrorism and Counterterrorism, Ch. 2, pp. 35-68 
 “Global Terrorism Index 2014,” Institute for Economics and Peace, p. 59
Recommended Readings 
 John Horgan, “From Profiles and Pathways and Roots to Routes: Perspectives from Psychology
on Radicalization into Terrorism,” The Annals of the American Academy of Political and Social
Science, Vol. 618, No. 1 (2008) pp. 80-94 
 Jeff Victoroff, “The Mind of the Terrorist: A Review and Critique of Psychological Approaches,”
Journal of Conflict Resolution, Vol. 49, No. 1 (2005) pp. 3-42 
 Alan Krueger and Jitka Malečková, “Education, Poverty and Terrorism: Is There a Causal
Connection?” The Journal of Economic Perspectives, Vol. 17, No. 4 (2003) pp. 119-144 
September 10: Film “The Weather Underground” Part II 
WEEK 3: Organizational, Strategic Level Causes and Objectives of Terrorism and

57#
Insurgency

Key Questions

Are most terrorist attacks committed by unconnected individuals or organizations? 
What is collective action and when is it achieved?
When and why does organizations’ pursuit of strength and survival generate violence?
What political environments and government types make terrorism more likely?
Does military occupation cause terrorism? 
Skills Introduced 
Causal Inference: How do we know when X causes Y? 
September 15: Religion, Gender, Ideology 
Required Readings 
• Mark Juergensmeyer, “Soldiers for Christ,” in Terror in the Mind of God, 3rd Edition, (Berkeley:
University of California Press, 2003) pp. 19-43
 Mark Juergensmeyer, “Zion Betrayed,” in Terror in the Mind of God, 3rd Edition, (Berkeley:
University of California Press, 2003) pp. 45-60 
 Caron Gentry and Laura Sjoberg, “The Gendering of Women’s Terrorism,” in Women, Gender, and
Terrorism, Laura Sjoberg and Caron Gentry, eds. (Athens, University of Georgia Press, 2011),
pp. 59-70 
 Alexis Henshaw, “Taking Female Armed Rebels Seriously,” The Washington Post (April 11, 2015) 
 Dan Byman, “Five Myths About Violent Extremism,” The Washington Post (February 13, 2015)
Recommended Readings 
 Mia Bloom, Bombshell: Women and Terrorism (Philadelphia: University of Pennsylvania Press, 2011)

 Alessandro Orsini, Anatomy of the Red Brigades: The Religious Mindset of Modern Terrorists (Ithaca:
Cornell University Press, 2009) 
 Mattias Gardell, “Crusader Dreams: Oslo 22/7, Islamophobia, and the Quest for a Monocultural
Europe,” Terrorism and Political Violence, Vol. 26, No. 1 (2014) pp. 129-155 
 James A. Piazza, “Is Islamist Terrorism More Dangerous? An Empirical Study of Group
Ideology, Organization, and Goal Structure,” Terrorism and Political Violence, Vol. 21, No. 1
(2009) pp. 62-88 
 Richard Jackson, “Constructing Enemies: ‘Islamist Terrorism’ in Political and Academic
Discourse,” Government and Opposition, Vol. 42, No. 3 (2007) pp. 394-426 
 Cynthia K. Mahmood, Fighting for Faith and Nation: Dialogues with Sikh Militants (Philadelphia:
University of Pennsylvania Press, 1996) 
September 17: Solidarity, Networks, and Numbers; Organizational Survival and Competition
Required Readings 
 Mancur Olson, “Introduction” in The Logic of Collective Action: Public Goods and the Theory of Groups
(Cambridge: Harvard University Press, 1971) pp. 1-3 
 Marc Sageman, “Joining the Jihad” in Understanding Terror Networks (Philadelphia: University of

58#
Pennsylvania Press, 2004) pp. 99-135 
 Mia Bloom, “Outbidding, Market Share, and Palestinian Suicide Bombing,” Political Science
Quarterly, Vol. 119, No. 1 (2004) pp. 61-88 
 Peter Krause, “The Structure of Success: How the Internal Distribution of Power Drives Armed
Group Behavior and National Movement Effectiveness,” International Security, Vol. 38, No.
3, pp. 72-116
Recommended Readings 
 Paul Staniland, “States, Insurgents, and Wartime Political Orders,” Perspectives on Politics, Vol. 10,
No. 2 (2012) pp. 243-264 
 Shawn Flanigan, “Nonprofit Service Provision by Insurgent Organizations: The Cases of
Hizballah and the Tamil Tigers,” Studies in Conflict & Terrorism, Vol. 31, No. 6 (2008) pp.
499-517 
 Eitan Alimi, "Contextualizing Political Terrorism: A Collective Action Perspective for
Understanding the Tanzim," Studies in Conflict and Terrorism, Vol. 29, No. 3 (2006) pp. 263-
283 
 Wendy Pearlman, “Spoiling Inside and Out: Internal Political Contestation and the Collapse of
Intrastate Peace Accords,” International Security, Vol. 33, No. 3 (2008) pp. 79-109 
 Ghaith Abdul-Ahad, “How to Start a Battalion (In Five Easy Lessons),” London Review of Books,
Vol. 35, No. 4 (2013) pp. 13-14 
WEEK 4: Methods and Mechanisms: Strategies of Terrorism and Insurgency 
Key Questions 
What are the main strategies of terrorism and insurgency?
What is the causal logic of each strategy?
Under what conditions is each strategy most likely to succeed or fail?
Do observers’ assessments of terrorist strategies match with those of the perpetrators?
Skills Introduced
Identifying and explaining causal mechanisms
Linking theory and practice
Identifying gaps in scholarship
September 22: Political Grievances and Occupation; Failed States and State Sponsors
Required Readings
 Bruce Hoffman, Inside Terrorism, Ch. 2, pp. 43-62 
 Erica Chenoweth, “Terrorism and Democracy,” The Annual Review of Political Science, Vol. 16, pp.
355-375 
 Daniel Byman, Deadly Connections: States That Sponsor Terrorism (New York: Cambridge University
Press, 2005) pp. 10-15, 21-78 
Recommended Readings 
 Barry Posen, “The Security Dilemma and Ethnic Conflict,” Survival, Vol. 35, No. 1 (1993) pp. 27-
35 
 Paul Stern, “Why Do People Sacrifice for Their Nations?” Political Psychology, Vol. 16., No. 2 (1995)

59#
pp. 217-235 
 Mia Bloom, “Death Becomes Her: Women, Occupation, and Terrorist Mobilization,” PS: Political
Science and Politics, Vol. 43, No. 3 (2010), pp. 445-450 
 James A. Piazza, “Incubators of Terror: Do Failed and Failing States Promote Transnational
Terrorism?” International Studies Quarterly, Vol. 52 (2008) pp. 469-473, 481-485 
September 24: Strategies of Terrorism and Insurgency- Academics 
Required Readings 
 Andrew Kydd and Barbara Walter, "The Strategies of Terrorism," International Security Vol. 31, No.
1 (2006) pp. 56-80 
 David Lake, “Rational Extremism: Understanding Terrorism in the Twenty First Century,”
International Organization, Vol. 56, No. 1 (2002) pp. 15-29 
 Stathis Kalyvas, “Wanton and Senseless? The Logic of Massacres in Algeria,” Rationality and Society,
Vol. 11, No. 3 (1999) pp. 252-259 
Recommended Readings 
 Ian Lustick. "Terrorism in the Arab-Israeli Conflict: Targets and Audiences" in Martha Crenshaw,
ed., Terrorism in Context (University Park, PA: Pennsylvania State University Press, 1995) pp.
514-533 
 Martha Crenshaw, “The Logic of Terrorism: Terrorist Behavior as a Product of Choice,” Terrorism
and Counter Terrorism Vol. 2, No. 1 (1998) pp. 54-64 
 Ignacio Cuenca, “The Dynamics of Nationalist Terrorism: ETA and the IRA,” Terrorism and
Political Violence, Vol. 19, No. 3 (September 2007) pp. 289-206 
 Victor Asal and R. Karl Rethemeyer, “The Nature of the Beast: Organizational Structures and the
Lethality of Terrorist Attacks,” Journal of Politics, Vol. 70, No. 2 (2008) pp. 437-449 
WEEK 5: Methods and Mechanisms: Suicide Bombing and WMD 
Key Questions 
Are suicide bombing and WMD attacks major threats?
How can we assess intentions vs. capability?
Why do some groups choose to employ these methods and others do not? 
Skills Introduced 
Operationalizing variables and testing predictions 
September 29: Strategies of Terrorism and Insurgency- Practitioners
Required Readings
 Carlos Marighella, “Problem and Principles of Strategy,” and “Minimanual of the Urban
Guerrilla,” in James Kohl and John Litt, eds., Urban Guerrilla Warfare in Latin America
(Cambridge: MIT Press, 1974) pp. 81-86, 108-133 
 Menachem Begin, The Revolt (New York: Nash, 1977) pp. 47-58, 76-96 
 Mao Tse-Tung, Basic Tactics (New York: Praeger, 1967) pp. 51-68 
 Abu Bakr Naji, “The Management of Savagery: The Most Critical Stage Through Which the
Umma Will Pass,” trans. William McCants, pp. 18-20, 28-34 

60#
 Ayman Al-Zawahiri letter to Abu Musab Al-Zarqawi, July 9, 2005
Recommended Readings

• Che Guevara, On Guerrilla Warfare (NewYork:Praeger, 1961) 
October 1: Suicide Bombing and Weapons of Mass Destruction in Terrorism and Insurgency
Required Readings 
 Robert A. Pape, “The Strategic Logic of Suicide Terrorism,” American Political Science Review, Vol.
97, No. 3 (2003) pp. 343-361 
 Scott Atran, “The Moral Logic and Growth of Suicide Terrorism,” The Washington Quarterly Vol.
29, No. 2 (2006) pp. 127-147 
 Lindsey O’Rourke, “Behind the Woman Behind the Bomb,” The New York Times (August 2, 2008)
 Matthew Bunn and Susan Martin, “Is Nuclear Terrorism a Real Threat?” in Gottlieb, Debating
Terrorism and Counterterrorism, Ch. 6, pp. 172-199 
 Richard Danzig et al., “Aum Shinrikyo: Insights Into How Terrorists Develop Biological and
Chemical Weapons,” Center for a New American Security (2012) pp. 9-41 
Recommended Readings 
 Michael C. Horowitz, “Nonstate Actors and the Diffusion of Innovations: The Case of Suicide
Terrorism,” International Organization Vol. 64, No. 1 (2010) pp. 33-64 
 Greg Koblentz, “Pathogens as Weapons: The International Security Implications of Biological
Warfare,” International Security Vol. 28, No. 3 (Winter 2003/2004) pp. 84-122 
 Assaf Moghadam, “Motives for Martyrdom: Al-Qaida, Salafi Jihad, and the Spread of Suicide
Attacks,” International Security Vol. 33, No. 3 (Winter 2008/2009) pp. 46-78 
 Keir Lieber and Daryl Press, “Why States Won’t Give Nuclear Weapons to Terrorists,”
International Security Vol. 38, No. 1 (2013) pp. 80-104 
WEEK 6: Morality and the Media 
Key Questions 
How do feelings of rage, humiliation, fear, depression, revenge, and injustice impact terrorism?
How does the media impact the causes, mechanisms, and effects of terrorism?
How should the media balance profit, the public’s ‘need to know,’ and responsibility to society? 
October 6: Morality, Emotions, and the Media in Terrorism and Insurgency 
Required Readings 
 Eamon Collins, Killing Rage (London: Granta Books, 1997) pp. 1-29 
 Bruce Hoffman, “The Old Media, Terrorism, and Public Opinion,” and “The New Media,
Terrorism, and the Shaping of Global Opinion,” in Hoffman, Inside Terrorism, Ch. 6 and 7,
pp. 173-228 
 Gadi Wolfsfeld et al., “Covering Death in Conflicts: Coverage of the Second Intifada on Israeli
and Palestinian Television,” Journal of Peace Research, Vol. 45, No. 3 (2008) pp. 401-417 
• Agence France-Presse, “Paris Supermarket Hostages Sue Media Over Live Coverage” (April 3,
2015)
Recommended Readings
 Roger Petersen, Understanding Ethnic Violence: Fear, Hatred, and Resentment in Twentieth-Century Eastern

61#
Europe (Cambridge: Cambridge University Press, 2002) 
 Fredric Wehrey, “A Clash of Wills: Hizballah’s Psychological Campaigns against Israel in South
Lebanon,” Small Wars & Insurgencies Vol. 13, No. 3 (2002) pp. 53-74 
 Gabriel Weimann, “The Psychology of Mass-Mediated Terrorism,” American Behavioral Scientist,
Vol. 52, No. 1 (2008) pp. 69-86 
 James Sheehan et al., “Al-Shabaab’s Propaganda War and Alternative Media” (2012) pp. 29-39
October 8: *Exam #1*
WEEK 7: The Impact and Effectiveness of Terrorism and Insurgency

Key Questions

How many people are killed and wounded by terrorist and insurgent attacks?
Do terrorism and insurgency achieve the personal goals of the attackers?
How does the use of violence impact the strength and survival of organizations?
When and why does the public support terrorism and insurgency?
Do terrorism and insurgency generate political concessions? Do they win wars?
What is the economic and social impact of terrorism and insurgency? 
Skills Introduced 
Conceptualizing and measuring effects
Generating and analyzing competing arguments
Marshaling and analyzing relevant evidence 
October 13: Individual and Organizational Level Effects: Fear, Casualties, Support, Group Strength
Required Readings 
 Jennifer Lerner et al, “Effects of Fear and Anger on Perceived Risks of Terrorism: A National
Field Experiment,” Psychological Science Vol. 14 No. 2 (2003) pp. 144-150 
 Christophe Chowanietz, “Rallying Around the Flag or Railing Against the Government? Political
Parties’ Reactions to Terrorist Acts,” Party Politics Vol. 17, No. 5 (2011) pp. 673-698 
 Glenn Feldman, “Soft Opposition: Elite Acquiescence and Klan-Sponsored Terrorism in
Alabama, 1946- 1950,” The Historical Journal Vol. 40, No. 3 (1997) pp. 753-777 
 David Chalmers, Backfire: How the Ku Klux Klan Helped the Civil Rights Movement (Lanham: Rowman
and Littlefield, 2003) pp. 137-144 
Recommended Readings 
 “Global Terrorism Index 2014,” Institute for Economics and Peace 
 Jodi Vittori, "All Struggles Must End: The Longevity of Terrorist Groups," Contemporary Security
Policy Vol. 30, No. 3 (2009) pp. 444-466 
 Ethan Bueno De Mesquita and Eric Dickson, "The Propaganda of the Deed: Terrorism,
Counterterrorism, and Mobilization," American Journal of Political Science Vol. 51, No. 2
(2007) pp. 364-381 
October 15: Strategic Level Effects: Political Concessions, Military Withdrawals, New States
Required Readings 

62#
 Robert Pape, “Learning Terrorism Pays,” in Dying to Win (New York: Random House, 2006) p.
40, 61-76 
 Max Abrahms, "Why Terrorism Does Not Work," International Security Vol. 31, No. 2 (2006) pp.
42-52 
 Peter Krause, “The Political Effectiveness of Non-State Violence: A Two-Level Framework To
Transform a Deceptive Debate,” Security Studies Vol. 22, No. 2 (2013) pp. 259-294
 Timothy Wickham-Crowley, “A Qualitative Comparative Approach to Latin American
Revolutions,” International Journal of Comparative Sociology, Vol. 32, Nos.1 and 2 (January-April
1991), pp. 87-90, 99-105
Recommended Readings
 Andrew Kydd and Barbara Walter, “Sabotaging the Peace: The Politics of Extremist Violence,”
International Organization Vol. 56, No. 2 (2002) pp. 263-96 
 Kelly Greenhill and Solomon Major, "The Perils of Profiling: Civil War Spoilers and the Collapse
of Intrastate Peace Accords," International Security Vol. 31, no. 3 (Winter 2006/07) pp. 7-40 
 Kathleen Cunningham, “Divide and Conquer or Divide and Concede: How Do States Respond to
Internally Divided Separatists?” American Political Science Review Vol. 105, No. 2 (2011) pp.
275-97 
 Max Abrahms and Peter Krause Exchange on Krause’s Security Studies Article, H-Diplo (2013)
WEEK 8: Al-Qaeda 
Key Questions 
What are the origins of Al-Qaeda? What is its ideology and strategy?
What is transnational terrorism and what distinguishes it from other types?
Is Al-Qaeda a unique group, or do they share similarities with other organizations?
Is Al-Qaeda on the ropes, on the rise, or at an impasse? What is its future? 
Skills Introduced 
How to generate and frame general and specific research questions
What is a case? Case selection and research design 
October 20: Al-Qaeda: The Past 
Required Readings 
 Lawrence Wright, The Looming Tower, Ch. 5-7, 13-20 
 Osama Bin Laden, “Declaration of War Against the Americans Occupying the Land of the Two
Holy Places” (August 23, 1996) 
Recommended Readings 
 “Bin Laden’s Bookshelf” (Collection of Documents Captured During Raid on May 1, 2011) 
 Osama Bin Laden, “Letter to the American People” (November 24, 2002) 
 Steve Coll, Ghost Wars: The Secret History of the CIA, Afghanistan, and bin Laden, from the Soviet Invasion
to September 10, 2001 (New York: Penguin, 2005) 
 Thomas H. Kean, Lee Hamilton, et al., The 9/11 Commission Report: Final Report of the National
Commission on Terrorist Attacks upon the United States (Washington, DC: Government Printing
Office, 2005) 

63#
 Peter Bergen, Holy War, Inc.: Inside the Secret World of Osama bin Laden (New York: Simon & Shuster,
2001) 
 Fawaz Gerges, The Far Enemy: Why Jihad Went Global (Cambridge: Cambridge University Press,
2009) 
October 22: Al-Qaeda: The Present and Future 
Required Readings 
 Daniel Byman, “Buddies or Burdens? Understanding the Al Qaeda Relationship with Its Affiliate
Organizations,” Security Studies, Vol. 23, No. 3 (2014) pp. 431-470 
 Daveed Gartenstein-Ross, “Lone Wolf Islamic Terrorism: Abdulhakim Mujahid Muhammad
(Carlos Bledsoe) Case Study,” Terrorism and Political Violence, Vol. 26, No. 1 (2014) pp. 110-
128 
 “Al-Qaeda’s Use of Female Suicide Bombers in Iraq: A Case Study,” in Women, Gender, and
Terrorism, Laura Sjoberg and Caron Gentry, eds. (Athens, University of Georgia Press, 2011),
pp. 159-175 
• Rukmini Callimachi, “Paying Ransoms, Europe Bankrolls Qaeda Terror,” The New York Times
(July 29, 2014) 
Recommended Readings
 Anthony Lemieux et al, “Inspire Magazine: A Critical Analysis of its Significance and Potential
Impact Through the Lens of the Information, Motivation, and Behavioral Skills Model,”
Terrorism and Political Violence, Vol. 26, No. 2 (2012) pp. 354-371 
 Risa Brooks, "Muslim 'Homegrown' Terrorism in the United States: How Serious Is the Threat?"
International Security Vol. 36, No. 2 (2011) pp. 7-47 
 William McCants, “How Zawahiri Lost al Qaeda,” Foreign Affairs (November 19, 2013) 
 Bruce Hoffman, “American Jihad,” The National Interest No. 107 (May/June 2010) pp. 17-27 
 Thomas Hegghammer, “Should I Stay or Should I Go? Explaining Variation in Western Jihadists’
Choice between Domestic and Foreign Fighting,” American Political Science Review Vol. 107,
No. 1 (2013) pp. 1-15#
WEEK 9: The Boundaries of Terrorism: Nonviolence and State Terror 
Key Questions 
What is the same and different about the causes, mechanisms, and effects of insurgency and civil
war as compared to terrorism?
When and why is nonviolence more effective than terrorism and insurgency?
Is there ‘ecoterrorism’ and is it comparable to other forms of terrorism? 
Is there “state terrorism”? Should we adjust the common definition of terrorism to include it?
How many civilians do states and non-state actors kill? What are the causes of mass killing by states?
Skills Introduced 
Comparing cases 
October 27: Nonviolence and Non-Lethal Violence 
Required Readings 

64#
 Stefan Leader and Peter Probst, “The Earth Liberation Front and Environmental Terrorism,”
Terrorism and Political Violence Vol. 15, No. 4 (2003) pp. 37-58 
 Blythe Copeland, “5 Ways Sea Shepherd's Controversial Methods are Changing the World For
Whales,” Treehugger (February 23, 2011) 
 Sivan Hirsch-Hoefler and Cas Mudde, “Ecoterrorism: Threat Or Political Ploy?” The Washington
Post (December 19, 2014) 
 Gene Sharp, “The Intifadah and Nonviolent Struggle,” Journal of Palestine Studies Vol. 19, No. 1
(1989) pp. 3-13 
 Erica Chenoweth and Maria Stephan, "Why Civil Resistance Works: The Strategic Logic of
Nonviolent Political Conflict," International Security Vol. 33, No. 1 (2008) pp. 7-44
Recommended Readings 
 Fabio Rojas, “Social Movement Tactics, Organizational Change and the Spread of African-
American Studies,” Social Forces Vol. 84, No. 4 (June 2006) pp. 2139-2158 
 Andrew Mack, "Why Big Nations Lose Small Wars: The Politics of Asymmetric Conflict," World
Politics Vol. 27, No. 2 (1975) pp. 175-200 
 Ivan Arreguin-Toft, "How the Weak Win Wars: A Theory of Asymmetric Conflict," International
Security Vol. 26, No. 1 (2001) pp. 93-128 
 Victor Asal and R. Karl Rethemeyer, “Dilettantes, Ideologues, and the Weak: Terrorists Who
Don’t Kill,” Conflict Management and Peace Science, Vol. 25, No. 3 (2008) pp. 244-260 
 Leena Malkki, “Political Elements in Post-Columbine School Shootings in Europe and North
America,” Terrorism and Political Violence, Vol. 26, No. 1 (2014) pp. 185-210 
October 29: States and Rebel Governance: State Terrorism and Insurgents as State Builders
Required Readings
 Ruth Blakeley, “Bringing the State Back into Terrorism Studies,” European Political Science Vol. 6,
No. 3 (September 2007) pp. 228-236 
 Noam Chomsky, “The United States is a Leading Terrorist State,” Monthly Review, Vol. 53, No. 6
(November 2001) pp. 1-8 
 Paul Staniland, “States, Insurgents, and Wartime Political Orders,” Perspectives on Politics Vol. 10,
No. 2 (June 2012) pp. 243-264 
 Tim Arango, “ISIS Transforming Into Functioning State That Uses Terror as Tool,” The New
York Times (July 21, 2015) 
Recommended Readings 
 Martha Crenshaw, “The Effectiveness of Terrorism in the Algerian War,” in Terrorism in Context
(University Park: Pennsylvania State Press, 2007) pp. 473-513 
 Ben Valentino et al., “’Draining the Sea’: Mass Killing and Guerrilla Warfare,” International
Organization Vol. 58, No. 2 (2004) pp. 375-407 
 Alex Downes, “Draining the Sea by Filling the Graves: Investigating the Effectiveness of
Indiscriminate Violence as a Counterinsurgency Strategy,” Civil Wars Vol. 9, No. 4 (2007) pp.
420-444 
 Alexander George, ed., Western State Terrorism (Cambridge: Polity Press, 1991). 
 Paul Wilkinson, “Can A State Be ‘Terrorist’?” International Affairs Vol. 57, No. 3 (1981) pp. 467-
472 

65#
WEEK 10: The Insurgencies in Iraq and Syria 
Key Questions 
How and why did the insurgencies in Iraq and Syria begin?
How have the ruling regimes and foreign states responded?
How was ISIS created, and what explains its variation in strategy and effectiveness over time?
What explains the shifting alliances among insurgent groups across time and space? 
Skills Introduced 
Analyzing theories and cases: process-tracing and congruence testing 
November 3: The Causes, Dynamics, and Effects of the Insurgencies 
Required Readings 
 Wendy Pearlman, “Emotions and the Microfoundations of the Arab Uprisings,” Perspectives on
Politics Vol. 11, No. 2 (2013) pp. 387-409 
 Stathis Kalyvas, "The Paradox of Terrorism in Civil War," The Journal of Ethics Vol. 8, No. 1 (2004)
pp. 97-138 
 James Fearon and David Laitin, “Ethnicity, Insurgency, and Civil War,” The American Political
Science Review Vol. 97, No. 1 (February 2003) pp. 75-77 
 James Fearon, “Obstacles to Ending Syria’s Civil War,” Foreign Policy (December 10, 2013) 
November 5: Foreign Fighters, ISIS, and Insurgent Rivalries 
Required Readings 
 William McCants, “The Believer: How an Introvert with a Passion for Religion and Soccer
Became Abu Bakr al-Baghdadi, Leader of the Islamic State,” Brookings, September 1, 2015 
 Graeme Wood, “What ISIS Really Wants,” The Atlantic (March 2015) 
 Graeme Wood, “What ISIS Really Wants: The Response,” The Atlantic (February 24, 2015) 
 Christoph Reuter, “The Terror Strategist: Secret Files Reveal the Structure of Islamic State,”
Spiegel Online (April 18, 2015) 
 Thomas Hegghammer, “Syria’s Foreign Fighters,” Foreign Policy (December 9, 2013) 
 “Global Terrorism Index 2014,” Institute for Economics and Peace, pp. 50-52 
 Hasnain Kazim, “Interview with an Islamic State Recruiter: ‘Democracy Is For Infidels,’” Spiegel
Online (October 28, 2014) 
 Roula Khalaf and Sam Jones, “Selling Terror: How ISIS Details Its Brutality,” Financial Times
(June 17, 2014) 
 Ariel Ahram, “Sexual and Ethnic Violence and the Construction of the Islamic State,” Political
Violence @ a Glance (September 18, 2014) 
Recommended Readings 
 Baghdadi’s first speech after declaration of ‘caliphate’” (July 1, 2014)
https://www.youtube.com/watch?v=VOORW63ioY0 
 Charlie Winter, “The Virtual ‘Caliphate’: Understanding Islamic State’s Propaganda Strategy,”
Quilliam (2015) 

66#
WEEK 11: Counterterrorism and Counterinsurgency I 
Key Questions 
When, why, and how do terrorism and insurgency end?
What are the objectives and strategies of counterterrorism and counterinsurgency? 
November 10: *Exam #2 and “If a Tree Falls”* 
November 12: How Terrorism and Insurgency End (Proposal Due) 
Required Readings 
 Audrey Kurth Cronin, “How Al Qaida Ends: The Decline and Demise of Terrorist Groups,”
International Security Vol. 31, No. 1 (Summer 2006) pp. 7-48 
 John Horgan, Walking Away From Terrorism (London: Routledge, 2009) pp. 27-39, 50-62 
 Peter Neumann, “Negotiating With Terrorists,” Foreign Affairs, Vol. 86, No. 1 (2007) pp. 128-138

 Efraim Inbar and Eitan Shamir, “Mowing the Grass in Gaza,” BESA Center Paper No. 255, July
2014 
 Micah Zenko, “Terrorism is Booming Almost Everywhere But in the United States,” Foreign Policy
(June 19, 2015) 
Recommended Readings 
 Seth G. Jones and Martin C. Libicki, How Terrorist Groups End: Lessons for Countering al Qa'ida (Santa
Monica, Calif: RAND, 2008) 
 Ben Connable and Martin Libicki, "How Insurgencies End," (Santa Monica, CA: RAND, 2010) 
WEEK 12: Counterterrorism and Counterinsurgency II 
Key Questions 
What are the organizations involved in U.S. counterterrorism and counterinsurgency efforts?
Do counterterrorism and counterinsurgency work? Do they have unintended consequences? 
Skills Introduced 
Linking theory and evidence to policy, and vice versa 
November 17: CT and COIN Debates: Hard & Soft Power, Democratization, Threat Inflation
Required Readings
 John Mueller, “Six Rather Unusual Propositions about Terrorism,” Terrorism and Political Violence
Vol. 17, No. 4 (2005) pp. 487-505
 Richard Betts, Daniel Byman, and Martha Crenshaw, “Comments on John Mueller’s ‘Six Rather
Unusual Propositions about Terrorism’,” Terrorism and Political Violence, Vol. 17, No. 4 (2005),
pp. 507-521 
 Ayaan Hirsi Ali, “A Problem From Heaven,” Foreign Affairs (July/August 2015) 
 Williams McCants, “Islamic Scripture Is Not the Problem,” Foreign Affairs (July/August 2015) 
 Gregory Gause and Jennifer Windsor, “Can Spreading Democracy Help Defeat Terrorism?” in
Gottlieb, Debating Terrorism and Counterterrorism, Ch. 8, pp. 243-275

67#
Recommended Readings 
 Brigitte Nacos and Michael Rubin, “Counterterrorism Strategies: Do We Need Bombs Over
Bridges?” in Gottlieb, Debating Terrorism and Counterterrorism, Ch. 7, pp. 209-242 
 Williams McCants and Clint Watts, “U.S. Strategy for Countering Violent Extremism: An
Assessment,” Foreign Policy Research Institute (December 2012) 
 Robert Art and Louise Richardson, eds. Democracy and Counterterrorism: Lessons from the Past.
(Washington, D.C.: United States Institute of Peace Press, 2007) 
 Peter Krause and Stephen Van Evera, “Public Diplomacy: Ideas for the War of Ideas,” Middle
East Policy Vol. 16, No. 3 (Fall 2009) pp. 106-134 
 Cass Sunstein, "Terrorism and Probability Neglect," The Journal of Risk and Uncertainty Vol. 26, No.
2/3 (2003) pp. 121-136 
 Ami Pedhazur, “Struggling with the Challenge of Right-Wing Extremism in Democracies,” Studies
in Conflict and Terrorism, Vol. 24, No. 5 (September 2001) pp. 339-359 
 Ian Lustick, Trapped in the War on Terror (Philadelphia: University of Pennsylvania Press, 2006) 
November 19: The Freedom of Speech, Profiling and Airport Security, Torture
Required Readings 
 Conor Friedersdorf, “Why the Reaction Is Different When the Terrorist Is White,” The Atlantic
(Aug 8, 2012) 
 Abby Ohlheiser and Elahe Izadi, “Police: Austin Shooter was a ‘Homegrown American
Extremist’” The Washington Post (December 1, 2014) 
 Jonathan Turley, “The Biggest Threat To French Free Speech Isn’t Terrorism. It’s The
Government,” The Washington Post (January 8, 2015) 
 “Twitter suspends account of Hamas' military wing,” Haaretz (January 14, 2014) 
 Matthew Yglesias, “Two—But Only Two—Cheers for Blasphemy,” Vox (January 8, 2015) 
 Charles Kenny, “Airport Security is Killing Us,” Businessweek (November 18, 2012) 
 Justin Fishel et al, “Undercover DHS Tests Reveal Security Failures at US Airports,” ABC (June
1, 2015) 
 Michael Posner and Alan Dershowitz, “Is an Outright Ban the Best Way to Eliminate or
Constrain Torture?” in Gottlieb, Debating Terrorism and Counterterrorism, Ch.10, pp. 312-344

 John Yoo and David Cole, “ Counterterrorism and the Constitution: Does Providing Security
Require a Trade-off with Civil Liberties?” in Gottlieb, Debating Terrorism and Counterterrorism,
Ch. 11, pp. 345-379 
Recommended Readings 
 James McAllister, Jonathan Kirshner, Austin Long, Robert Pape, Joshua Rovner, “Forum on the
Senate Select Committee on Intelligence (SSCI) Report and the United States’ Post-9/11
Policy on Torture,” H-Diplo | ISSF Forum, No. 5 (2015) 
 Ron Hassner, “Fundamentalist Wrath,” Washington Post (January 12, 2015) 
 Darren W. Davis and Brian D. Silver, “Civil Liberties vs. Security: Public Opinion in the Context
of the Terrorist Attacks on America,” American Journal of Political Science, Vol. 48, No. 1
(January 2004) pp. 28-46 

68#
 Matthew Alexander, How to Break a Terrorist (New York: Free Press, 2008) 
 Gary Crowdus, “Terrorism and Torture in The Battle of Algiers: An Interview with Saadi Yacef,”
Cineaste Vol. 29, No. 3 (Summer 2004) pp. 30-37 
WEEK 13: Counterterrorism and Counterinsurgency III
Key Questions
Is terrorism a significant threat to you, your society, your country, and the world?
How does the U.S. justice system deal with terrorism and terrorist suspects and perpetrators?
What are the tactics of the Department of Homeland Security for counterterrorism?
Is the current U.S. approach to counterterrorism and counterinsurgency the right one?
How should the U.S. balance concerns of security and liberty in dealing with terrorism?
Skills Introduced
How to write a research paper
November 24: Drones and Intelligence Agencies
Required Readings
 Daniel Byman, “Why Drones Work” Foreign Affairs (July/August 2013) 
 Audrey Kurth Cronin, “Why Drones Fail” Foreign Affairs (July/August 2013) 
 Dylan Matthews, “Everything You Need to Know About the Drone Debate, in One FAQ,” The
Washington Post (March 8, 2013) 
 Jeremy Scahill and Ryan Devereaux, “Barack Obama’s Secret Terrorist-Tracking System, By the
Numbers,” The Intercept (August 5, 2014) 
 Barton Gellman and Laura Poitras, “U.S., British Intelligence Mining Data from Nine U.S.
Internet Companies in Broad Secret Program,” The Washington Post (June 6, 2013) 
 Daniel Solove, “Why Privacy Matters Even If You Have ‘Nothing to Hide,’” The Chronicle of Higher
Education (May 15, 2011) 
 Peter Bergen et al. “Do NSA's Bulk Surveillance Programs Stop Terrorism?” New America (2014)

 Michael Hirsh, “The Next Bin Laden,” National Journal (November 14, 2013)
Recommended Readings 
 Austin Long, “Whack-a-Mole or Coup de Grace? Institutionalization and Leadership Targeting in
Iraq and Afghanistan,” Security Studies Vol. 23, No. 3 (2014) pp. 471-512 
 Patrick Johnston, “Does Decapitation Work?: Assessing the Effectiveness of Leadership
Targeting in Counterinsurgency Campaigns,” International Security Vol. 36, No. 4 (2012) pp.
47-79 
 Jenna Jordan, “When Heads Roll: Assessing the Effectiveness of Leadership Decapitation,”
Security Studies, Vol. 18, No. 4 (December 2009) pp. 719-755 
 Ronen Bergman, “The Hezbollah Connection,” The New York Times (February 10, 2015) 
 M.S., “Why We Spy: The War on Terror is Obama’s Vietnam,” The Economist (June 10, 2013) 
November 26: Have a Happy Thanksgiving!
Assignment: Impress your family members with your newfound knowledge about terrorism and
political violence, then defeat them in political debates using said knowledge. 

69#
WEEK 14: The Boston Marathon Bombings 
Key Questions 
How do theory and history help us to explain what happened?
Were these attacks ‘terrorism’? What were the causes of the attacks? How did various types of media
cover them?
What aspects of this attack were common, anomalous?
Why did the government and the community react they way that they did? Should anything have
been done differently? What should be the outcome of Dzokhar Tsarnaev’s trial?
Skills Introduced
How to apply theory to current events
December 1: Definitions, Causes, and the Media#
Required Readings
 Jess Bidgood, “Link to Marathon Bombing Rattles City Known for Its Tolerance,” The New York
Times (June 4, 2013) 
 Jim Mackinnon, “Bill Ayers Defends Weather Underground Bombings,” Akron Beacon (May 4,
2013) 
 Ken Bensinger and Andrea Chang, “Boston Bombings: Social Media Spirals Out of Control,” The
Los Angeles Times (April 20, 2013) 
 “Rolling Stone Defends Cover Featuring Boston Marathon Bombing Suspect,” CBS News (July
17, 2013) 
December 3: Effects, Community Response, and the Dzokhar Tsarnaev Trial
Required Readings 
 David Montgomery et al, “Police, Citizens and Technology Factor into Boston Bombing Probe,”
The Washington Post (April 20, 2013) 
 Ayaan Hirsi Ali, “The Problem of Muslim Leadership,” The Wall Street Journal (May 27, 2013) 
 Wardah Khalid, “Day by Day: An American Muslim's Thoughts After the Boston Attack,” The
Huffington Post (April 23, 2013) 
 Conor Friedersdorf, “Falsely Accused in Boston: 3 Examples and What They Should Teach Us,”
The Atlantic (April 19, 2013) 
 Anonymous Security Professional, “Thoughts on Responding to the Boston Bombings” 
 Frank Bruni, “The Lesson of Boston,” The New York Times (April 27, 2013) 
 Peter Krause, “BC Should Respond to Attacks with Renewed Community Spirit,” The Heights
(April 15, 2013) 
 Austin Tedesco, “Students Organize ‘Last 5’ Walk, Vigil as Memorial Events,” The Heights (April
18, 2013) 
 Bill and Denise Richards, “To End The Anguish, Drop The Death Penalty,” The Boston Globe
Recommended Readings 
“Why Was Boston Strong? Lessons From the Boston Marathon Bombings” Harvard Kennedy
School (April 2014)

70#
WEEK 15: Terrorism, Insurgency, and Political Violence, Now and in the Future
Key Questions
What are the major lessons we learned in this course? What questions remain unanswered?
What is the future of terrorism, insurgency, and political violence?
Skills Introduced
How to read the news
How to generate policy implications
December 7: Remaining Questions and Lessons Learned
Required Readings
 Husna Haq, “Why #I’llRideWithYou Worked, and Other Muslims Hashtags Didn’t,” Christian
Science Monitor (December 15, 2014) 
 Arit John, “With the NAACP Bombing, the Media-Coverage Gap Went Viral,” Bloomberg (January
8, 2015)
 Linda Robinson, “The Future of Counterterrorism: Fewer Drones, More Partnerships,” The
Washington Post (October 28, 2013) 
 Antonia Blumberg, “Sikhs Mark Anniversary Of Temple Shooting With Community Service,”
Huffington Post (July 31, 2015) 
 One news article of your choice from the past week
!