Dudan sobre efectividad de medidas antiterroristas

THE EFFECTIVENESS OF
COUNTER-TERRORISM STRATEGIES
A Campbell Systematic Review
Cynthia Lum *
Leslie W. Kennedy †
Alison J. Sherley §
January, 2006
_________________________________
* George Mason University, Administration of Justice Program, 10900 University Blvd., MS 4F4, Manassas, VA
20110, clum@gmu.edu. Corresponding Author.
† Rutgers University, School of Criminal Justice and Center for the Study of Public Security, 123 Washington Street,
Room 568, Newark, NJ 07102, kennedy@andromeda.rutgers.edu.
§ Rutgers University, School of Criminal Justice and Center for the Study of Public Security, 123 Washington Street,
Room 568, Newark, NJ 07102, asherley@pegasus.rutgers.edu.
CAMPBELL SYSTEMATIC REVIEW SUMMARY
Title and Authors
The Effectiveness of Counter-Terrorism Strategies: A Campbell Systematic Review
Cynthia Lum (corresponding author), Leslie Kennedy and Alison Sherley
Objectives
• To determine the effectiveness of counter-terrorism strategies from the available social scientific research
literature using systematic review methods.
• To stimulate debate about the cost-effectiveness of expenditures on counter-terrorism measures.
Findings
• There is almost a complete absence of high quality scientific evaluation evidence on counter-terrorism strategies;
• What evidence there is does not indicate consistently positive results – some counter terrorism interventions
show no evidence of reducing terrorism and may even increase the likelihood of terrorism and terrorism-related
harm;
• The available evidence suggests:
The use of metal detectors in airports reduces hijackings; however, there may also be a substitution or
displacement effect of airport security on other types of terrorism (e.g. assassinations, bombings, hostage
taking, death and wounded events).
Fortifying embassies and efforts to protect diplomats do not appear effective in reducing terrorist attacks on
these targets.
Increasing the severity of punishment for hijackers does not appear to have a statistically discernible effect
on reducing skyjacking incidents although there is very limited research conducted in this area.
UN resolutions, without the implementation of metal detectors, have not been shown to reduce terrorism.
Retaliatory attacks (for example, the U.S. attack on Libya in 1986 or attacks by Israel on the PLO) have
significantly increased the number of terrorist attacks in the short run, particular against the United States,
the United Kingdom, and Israel.
The existence of intolerant political parties (to terrorism) and the end of the Cold War could increase
terrorism events although the findings in this review were uncertain.
In the U.S. alone the non-defence costs of homeland security have increased from $9 billion in 2000 to $32
billion in 2005. In light of the uncertain effectiveness of counter-terrorism measures the cost-effectiveness
of this expenditure is open to debate.
Caveats and Qualifications
• The available scientific evidence was drawn from only a handful of studies which use moderately rigorous
research designs; this limits the strength of evidence and the conclusions that can be drawn from it.
• It may be that further studies of a higher quality will provide stronger evidence of the effectiveness,
ineffectiveness, or harm of particular interventions
Implications for Decision Makers
• The evidence base for policy making, strategic thinking and planning against terrorism is very weak; there is an
urgent need to commission research and evaluation on counter-terrorism measures to determine whether these
strategies work.
• Scientists need to be included in counter-terrorism policy making, strategic thinking, planning and evaluation;
some of the secrecy surrounding counter-terrorism activities is unwarranted.
Implications for Research
• More of the research on terrorism and counter terrorism needs to be empirical and evaluative using scientific
principles and different types of methodology. Government agencies can assist in this work by providing
funding and access to conduct evaluations.
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
1
Contents
List of Tables 2
List of Figures 2
1 Review Summary 3
2 Introduction and Background for the Review 3
3 The General State of Terrorism Research 6
4 The Campbell Systematic Review of Counterterrorism Strategies 9
4.1 Criteria for considering studies for review 10
4.2 Search strategy 11
4.3 Selection process and criteria for final inclusion 13
4.4 Further exclusion of medical-related studies 16
4.5 Summaries of the final seven studies 18
4.6 Data extraction, methodology used, and special issues regarding
meta-analyses with time series data 20
5 Analysis and Findings 22
5.1 Overall effects of counterterrorism strategies 22
5.2 Effects of specific intervention categories 23
5.2.1 Metal detectors and security screening 24
5.2.2 Fortifying embassies and protecting diplomats 26
5.2.3 Increasing the severity of punishment 27
5.2.4 United Nations resolutions against terrorism 28
5.2.5 Military retaliation 29
5.2.6 Changes in political governance 30
6 Discussion 32
7 Recommendations to Decision Makers 33
7.1 To government agencies generating counter-terrorism policies 35
7.2 To government agencies funding terrorism research 35
7.3 To researchers studying terrorism 36
8 Plans for Updating the Review 36
9 Acknowledgements 37
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
2
10 References 37
11 Appendices 43
A. Organizations conducting terrorism research 43
B. Data sources available 47
List of Tables
Table 1. Distribution of Subject Matter in Terrorism Research 8
Table 2. Studies Included in the C2 Systematic Review (Last Column) and
Final Exclusion Steps 17
List of Figures
Figure 1. Yearly Distribution of Terrorism Publications as a Percentage
of Total Publications 7
Figure 2. Effect Sizes and 95% Confidence Intervals (if available) for
All Findings 23
Figure 3. Increased Detection: Metal Detectors and Security Screening 25
Figure 4. Increased Protection: Fortifying Embassies and Protecting Diplomats 26
Figure 5. Increased Length of Incarceration and/or Severity of Punishment
for Skyjackings 27
Figure 6. Resolution Interventions: United Nations Resolutions Against Terrorism 28
Figure 7. Military Retaliation by States: Israeli Military-led Retaliation Attacks
on the PLO and Lebanon and United States Retaliatory Raids on Libya 29
Figure 8. Political Governance: Socialist Party in Power or Post Cold War 30
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
3
1 Review Summary
Since September 11th, there have been massive increases in personal, commercial, and
governmental expenditures on anti-terrorism strategies, as well as a proliferation of programs
designed to fight terrorism. These increases in spending and program development have focused
attention on the most significant and central policy question related to these interventions: Do
these programs work? To explore research evidence regarding this question, we conducted a
Campbell systematic review on counter-terrorism strategies to determine the scope and strength
of evaluation research in this area.
In the course of our review, we discovered that there is an almost complete absence of
evaluation research on counter-terrorism strategies. From over 20,000 studies we located on
terrorism, we found only seven which contained moderately rigorous evaluations of counterterrorism
programs. We conclude that there is little scientific knowledge about the effectiveness
of most counter-terrorism interventions. Further, from the evidence we were able to locate, it
appears that some evaluated interventions either didn’t work or sometimes increased the
likelihood of terrorism and terrorism-related harm.
The findings of this review dramatically emphasize the need for government leaders,
policy makers, researchers, and funding agencies to include and insist on evaluations of the
effectiveness of these programs in their agendas. These agendas would include identifying ways
to overcome methodological and data challenges often associated with terrorism research,
increasing funding to evaluate existing programs through methodologically rigorous evaluation
designs, and paying attention to existing evaluations of programs when implementing them.
Further, programs should be assessed to establish if they cause more harm than good or if they
create unanticipated consequences.
2 Introduction and Background for
the Review
Although actual figures of federal, state, municipal, international, and private
expenditures are difficult to calculate, there has clearly been a massive increase in government
and private spending on anti-terrorism strategies since September 11th (Congressional Budget
Office, 2002; 2005; Guinnessy and Dawson, 2002; Issues in Science and Technology, 2002;
Macilwain, 2002; Silke, 2004, generally). In 2002, the Congressional Budget Office estimated
that since 1998, there would be almost a doubling of “Appropriations for Combating Terrorism
and Protecting Critical Infrastructure”, given President Bush’s 2002 request, from US$7.2 to
$13.6 billion dollars (Congressional Budget Office, 2002). Their most recent report states that
these appropriations have further increased to $88.1 billion in 2004 (Congressional Budget
Office, 2005).
More remarkably, these estimates represent only a subset of United States defense
spending and do not include the billions of dollars expended since September 11th on counterterrorism
measures in other U.S. sectors and around the world. Spending has taken a wide array
of forms, including the creation of entire government bureaus (such as the Department of
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
4
Homeland Security), improving security at airports and borders, increasing research related to
responding to biological or chemical agents, creating and enforcing new laws, building or
operationalizing prisons to hold newly defined terrorists, improving medical emergency
responses, providing aid to foreign countries and developing municipal disaster plans. The House
Budget Committee of the U.S. Congress has recently estimated non-Department of Defense
funding for homeland security to have risen from $9 billion in 2000 to $32 billion in 2005.1
This massive increase in personal, commercial, and governmental expenditures on
anti-terrorism strategies, as well as the proliferation of programs designed to fight
terrorism, raise arguably the most significant policy question related to these counterterrorism
interventions: Do these programs work? More specifically, do they reduce the
likelihood of, or damage from, terrorism events or discourage individuals from acquiring
motivations to carry out political violence?
This Campbell systematic review has discovered that the answer to this question is,
disturbingly, “we don’t know”. Of the thousands of studies we located that focused on terrorism,
we found only seven which contained evaluations of anti-terrorism strategies (often examining
the same strategies) using at least moderately strong evaluation designs to ensure some internal
validity and believability in the findings.
Nor did we find any evidence that the minuscule proportion of government counterterrorism
spending related to research has been directed towards evaluating these programs.
Only recently, in February of 2004, the Committee on Governmental Reform of the United
States House of Representatives probed the issue in a hearing on “Effective Strategies Against
Terrorism” before the Subcommittee on National Security, Emerging Threats and International
Relations (U.S. House of Representatives, 2004).2 During those hearings, Representative
Christopher Shays stated:
Scientists remind us the plural of “anecdote” is not “data.” In the realm of
national security, a similar axiom would hold the proliferation of
counterterrorism strategies does not necessarily mean we are any safer. Only if
those strategies guide us inexorably and measurably toward clearly articulated
goals will they secure our liberty and property against the threat of a new and
dangerous era. (p. 3)
Yet, despite this promising start, the hearing diverted away from scientific evaluation of
anti-terrorism strategies. Ultimately, like many government assessments, the Government
Accountability Office (which was asked to conduct the “evaluation” of counter-terrorism
programs) was charged to find out whether national strategies had “fundamental characteristics
of a coherent strategic framework; one that clearly states a purpose, assesses risk, sets goals,
defines needed resources, assigns responsibilities and integrates implementation” (p. 4). Such an
assessment lacked the most fundamental requirement of evaluation – connecting programs with
measurable outcomes to determine the success of strategies (e.g., reduction in terrorist acts, fear
of terrorism, terrorist recruitment, etc). Nor did it appear that the assessment conducted by the
1 See https://www.house.gov/budget/ .
2 See https://www.mipt.org/pdf/Effective-Strategies-Against-Terrorism.pdf .
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
5
Government Accountability Office was informed by existing evaluation research on antiterrorism
programs.
At least the House Subcommittee had the right idea. With the decisions to commit funds
to anti-terrorism strategies comes the responsibility to know whether these programs are
effective. Perhaps even more importantly, as Andrew Silke (2005) has recently pointed out, we
not only need to determine if measures are ineffective, but also if policies produce harmful
effects (e.g., increasing terrorism). As examples in social policy have shown, programs are often
not effective and can sometimes increase the problem or cause further social negatives (see
McCord, 2003). Further, evaluations of both effectiveness and harm are necessary for policy
that involves phenomena like terrorism that can generate irrational thinking and hasty responses.
Responsible public policy for these situations requires a more rational approach, especially since
the use of terrorism is clearly not a passing fad and continues to be a major challenge. In
particular, the choice of programs on which to spend money should depend on whether we have
some evidence that they can be successful, or, at the very least, we should seek to evaluate these
programs for evidence of effectiveness.
In the scientific community, this type of research has become associated with the
evidence-based policy movement. Evidence-based policy suggests that choices to implement
intervention programs, like those which attempt to counter terrorism, are based on what we know
is effective or that policies should be subjected to evaluations of effectiveness (Cullen and
Gendreau, 2000; Davies et al., 2000; MacKenzie, 2000; Nutley and Davies, 1999; Sherman,
1998; Sherman et al., 2002; Weisburd et al., 2003). Evidence-based policy can provide a
moderating effect on knee-jerk responses to crises that have become more influenced by fear and
moral panics (see Cohen, 1972) than by reason or facts. In both the medical and social sciences,
this movement has led not only to a call for increased scientific evaluations on the effects of
intervention or treatment programs (and funding for these evaluations), but also for metaanalyses
and systematic reviews which serve to make generalizations from multiple studies of
similar programs.
To assess the evidence on the effectiveness of counter-terrorism measures, the Campbell
Collaboration3 (see Boruch et al., 2000; Farrington and Petrosino, 2001), through its Crime and
Justice Coordinating Group,4 approved our protocol to conduct a systematic review of research
related to anti-terrorism strategies and to comment on the state of counter-terrorism evaluation
research. Specifically, this study systematically reviews all terrorism research and literature,
searching for evaluations of anti-terrorism policies using the Campbell review framework5 and
meta-analytic techniques. What follows is our analysis and findings. In particular, we highlight
not only the research that we did find, but also what we didn’t find, and make policy suggestions
to government agencies and scientists in terms of what is needed.
We anticipated a number of problems and controversies in conducting such a review.
First, we expected that despite the proliferation of counter-terrorism and terrorism prevention
programs since September 11th, there would be little evaluation research on these policies,
making it difficult to locate quality studies to include in the review. Additionally, terrorism as a
phenomenon has not been easily or clearly defined (Crenshaw, 1992; Merari, 1991; Wilkinson,
3 See www.campbellcollaboration.org .
4 See https://www.aic.gov.au/campbellcj/ .
5 See https://www.campbellcollaboration.org/Fraguidelines.html .
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
6
1986) and therefore, types of strategies could cover a wide range of phenomena and seek a
variety of outcomes. Further, there may be strategies that work but might be legally
questionable, violate international law, challenge democratic society, or displace problems
elsewhere. Other strategies may seem effective but may be in gross violation of human rights
norms, such as the inhumane treatment of prisoners or other corrective policies that may not be
acceptable in many societies.
Additionally, terrorism and political violence may involve multiple points at which
interventions may be effective and where outcomes might be measured. For example, antiterrorism
strategies may include prevention and alleviation of early risk factors, situational
prevention of actual events, or post-event responses. Furthermore, because of the rare nature of
terrorism events, it also may be difficult to determine whether these strategies actually worked.
Thus, researchers may be interested in detecting secondary effects of prevention tactics, such as
the reduction of the fear of terrorism or the adequacy and efficiency of the response after an
event occurs. Other strategies may focus on detecting potential terrorism events or other high
risk situations.
Despite these controversies and challenges, we sought the most comprehensive review
possible not only to learn about what we know “works” in terms of reducing terrorism, but also
to learn about the state of terrorism research more generally to guide future evaluation research
agendas. To do this, we begin by briefly reporting our findings of a comprehensive, more
general review of terrorism research. We then report, in detail, the Campbell Systematic Review
of evaluation research on counter-terrorism strategies and describe the methodology used to find
and analyze evaluations related to anti-terrorism measures.
3 The General State of Terrorism
Research
A general review of terrorism research is a helpful beginning for a more systematic
review for a few reasons. Because of the anticipated difficulty in locating evaluations of
counter-terrorism strategies, an understanding of the distribution of terrorism research can help
to focus on these evaluations and target certain areas of research. Secondly, we expected no one
academic discipline to have a monopoly on terrorism research and therefore we needed to
familiarize ourselves with a variety of perspectives. Similarly, we anticipated that different types
of research methodologies might be used to understand subjects related to terrorism which also
necessitated a more general survey to understand the depth of the research.
From time to time, a few attempts at reviews of terrorism research have been undertaken
(see, e.g. Halkides, 1995; Hoffman, 1992; Miller, 1988; Romano, 1984; Schmid and Jongman,
1988), although September 11th has certainly necessitated the need for an updated review. To
conduct this general overview and gain a sense of terrorism research for our systematic review,
we collected any article in published, unpublished, peer-reviewed, non peer-reviewed, academic
and non-academic sources which mentioned terms related to terrorism and political violence.
We conducted this search across seventeen separate literary databases,6 many of which extend
6 The databases used were Academic Search Premier, ArticleFirst (OCLC), Contemporary Women’s Issues,
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
7
back to research conducted since the early 1960s. Although books, government and technical
reports, online documents, and web information would be included in our systematic review to
be discussed shortly, we limited this initial search to articles in order to gain an overall sense of
the literature. We found that there has been much written on terrorism and we located 14,006
articles published between 1971 and 2003 from these databases after excluding duplicate
records.
The first, most unique finding compared to past literature reviews was that the events of
September 11th have had an enormous effect on terrorism research. As Figure 1 indicates,
among the entire 14,006 works located, approximately 54% were published in 2001 and 2002.7
When only examining articles from peer-reviewed sources (many of the articles were opinioneditorials,
news reports, advertisements or general bulletins), this same proportion was found.
Neither the Oklahoma City bombing in 1995 nor any other significant terrorist event, including
the World Trade Center bombing in 1993, has been followed by this much research interest on
terrorism.
Figure 1. Yearly Distribution of Terrorism Publications as a Percentage of Total Publications
5
10
15
20
25
30
35
40
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
Year
Percentage
All Articles (N=14,006)
Peer-Reviewed Sources Only (N=6,041)
Our interest for the Campbell review would ultimately be empirical evaluations of
counter-terrorism programs. Thus, to gain a sense of the extent of empirical work on terrorism,
Criminal Justice Abstracts, EbscoHost, EconLit, Educational Abstracts, Electronic Collections Online,
ERIC(OCLC), GEOBASE, Humanities Abstracts, Ingenta, ISI Web of Science, MEDLINE, National Criminal
Justice Reference Service, PAIS International Articles Only, PUBMEDLINE, Social Science Abstracts, Sociological
Abstracts. The time periods covered by each of these databases can be obtained at
https://www.lib.neu.edu/gateway/databasestrifold.pdf .
7 This general review was initiated by the authors in 2003; hence, only literature up to the end of 2002 is
represented.
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
8
we read each abstract of the peer-reviewed articles to see to what extent empirical analysis had
been conducted using a very liberal definition of the term “empirical”. Studies deemed empirical
had abstracts which indicated that any analysis (either quantitative or qualitative) had been
conducted on terrorism data. Thought pieces, on the other hand, were articles where authors
discussed an issue theoretically or offered an opinion, while case studies (as deemed by the
author) examined a particular situation from a (usually) historical approach.8
From this categorization, we found that only 3% of the articles from peer-reviewed
sources appeared to be based on some form of empirical analysis. Approximately 1% were
categorized as case studies and the rest (96%) were thought pieces. The scarcity of any
empirical analysis (whether evaluative or not) on terrorism-related research supported our initial
hypothesis that we would find only a small amount of evaluation research on anti-terrorism
strategies.
To examine the distribution of specific subject matters studied in terrorism, we grouped
each of the articles whose abstract we read into common subject areas. Thirty-five general
groups emerged which we collapsed into the seventeen categories shown in Table 1. Table 1
also reports the distribution of these categorizations for studies conducting some form of
empirical analysis.
Table 1. Distribution of Subject Matter in Terrorism Research
Peer- reviewed
sources (N=4,458a)
Empirical only
(N=156 a)
Weapons of mass destruction (biological, chemical, nuclear) 18.1% 10.3%
Article on a specific issue such as the IRA, Al Qaeda or incident b 12.2% 5.1%
Political responses to terrorism (war, politics, international relations) 9.5% 1.9%
Causes, motivations, psychology, trends of terrorism 8.7% 18.1%
Impacts of terrorism (political, social, economic) 7.7% 5.2%
Non-political responses to terrorism(medical, social, economic) 5.5% 3.9%
Victimology, coping mechanisms, psychological effects of terrorism 5.4% 25.8%
Other (nationalism, intelligence issues, democracy and vulnerability) 5.4% 3.9%
Legal issues surrounding terrorism 5.2% 0.6%
The media and public attitudes towards terrorism 4.6% 18.7%
How to define terrorism 4.2% 1.3%
Non-conventional, cyber and narco-terrorism 3.0% 0.6%
Religion and terrorism 2.6% 1.3%
State-Sponsored terrorism 2.6% 1.3%
Law enforcement responses to terrorism (airports, police) 2.5% 0.6%
Research/science of studying terrorism 2.1% 0.6%
Domestic terrorism 0.6% 0.6%
a Excluding book reviews and articles where not enough information was given to be categorized.
b If could not be placed into any other category.
8 It should be noted that some case studies were based on empirical data, but we categorized them as case studies if
specifically indicated by the author.
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
9
Generally, issues related to weapons of mass destruction represented the largest
proportion of articles (18.9%) followed by articles which focused on a specific issue, such as the
Israel-Palestinian conflict, the problems in Northern Ireland, Al Qaeda, or September 11th (if
they could not be categorized elsewhere). Additionally, political responses to terrorism and the
sociology of terrorism (causes, motivations, explanations, definitions) clearly dominate the
research agenda of these studies. These emphases most likely represent the strong political
science influence in the study of terrorism.
Surprisingly absent from the literature were articles regarding subjects that are more
likely to affect individuals, their fear, the ramifications of anti-terrorism efforts, or the evaluation
of policy effectiveness. Generally, there is little written about the effectiveness of law
enforcement and other non-political responses to terrorist events. This is also true for issues
related to coping with terrorism events, legal concerns, and the general victimology of terrorism.
Additionally, when thinking about future threats, the relationships between terrorism and
religion, socio-economic factors, and political responses have yet to be made.
When examining those articles preliminarily deemed to be based on the analysis of
empirical information, the findings are both encouraging and discouraging. A quarter of the
empirical work appeared to be conducted on victimology, a subject seemingly relevant to our
search for evaluation studies. While political response literature makes up 9.5% of the literature,
only 1.9% of the empirical literature seems to empirically analyze those responses. There
appears to be an overrepresentation of empirical literature on the cause and sociology of
terrorism, rather than on programs designed to combat it. The vast majority of empirically-based
studies did not seem to evaluate the effects of anti-terrorism strategies.
These preliminary findings regarding terrorism research support a number of
justifications for a Campbell systematic review. Certainly, as Figure 1 indicates, the study of
terrorism is not simply a passing fad of little interest to scientists and evaluation researchers.
Given the proliferation of counter-terrorism strategies after September 11th, there is even more
reason to increase evaluation research on these programs. Yet, there is a dearth of empirical
research on counter-terrorism interventions. This is not to say that the current literature is not
useful; the more we know and understand about terrorism and those who wield it, the better.
However, much of this literature does not address the effectiveness of counter-terrorism
strategies. Nor do we have a grasp on whether measures might be harmful. To explore these
concerns, we now turn to the findings from the Campbell review of evaluation research related to
anti-terrorism strategies.
4 The Campbell Systematic Review of
Counter-Terrorism Strategies
While the general review of terrorism research described above helps to determine the
scope and nature of terrorism research, our task was to locate and more specifically examine
research evaluating the effectiveness of anti-terrorism strategies. To do this, we narrowed our
focus by establishing criteria for the consideration of studies for the review. We then engaged in
a systematic search strategy in choosing evaluations deemed at least moderately
methodologically rigorous. In the final phase, we extracted information from each study and
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
10
used meta-analytic techniques to examine research findings. Each step is now described.
4.1 CRITERIA FOR CONSIDERING STUDIES FOR REVIEW
As the preliminary review of terrorism literature indicated, the objects of study, the
research methods used, and perspectives related to terrorism are wide-ranging. The definition of
terrorism, and therefore the interventions and measurable outcomes of interventions related to
this definition, can be subjective and value-driven. Because of this, a systematic search for
evaluations of strategies to counter terrorism can provide a number of challenges not present in
other reviews.
Therefore, to provide structure for this review, we centered our search strategy on three
organizing constructs. First, we considered what might be deemed terrorism for the purposes of
finding evaluations of programs designed to reduce it. This proved challenging because of
continuing disagreement over what constitutes terrorism and how it is defined (Crenshaw, 1992;
Merari, 1991; Wilkinson, 1986). In terms of definitions, we learned from our general review of
terrorism research that what researchers considered as terrorism could range from political
violence to domestic violence, rape, or child abuse. As our goal was to initially be as inclusive
as possible, we not only considered any study whose concern was politically motivated violence
as normally or officially defined,9 but also any study in which the author(s) referred to the
program or outcome as terrorism or terrorism-related, even if not normally or traditionally
considered as terrorism or even if their perception of terrorism contradicted official definitions.
The second organizing construct centered on what types of interventions would be
considered as counter-terrorism measures. Like interventions intended to reduce criminality or
crime, anti-terrorism programs may not directly address the reduction of terrorism events, but
rather focus on the reduction of related risk factors. We learned from our preliminary review of
the general terrorism literature that what might be considered as an anti-terrorism strategy can
include a wide variety of subjects, including political, social, legal, law enforcement, economic,
preventative, reactive, or after-care responses. Limiting our search to law enforcement
responses, for example, may exclude those responses that focus on psychological coping
interventions for victims or situational crime prevention efforts. Additionally, lessons learned
from other crime-related systematic reviews (e.g., Sherman et al., 1997; Sherman et al., 2002),
show that social intervention programs (e.g., crime prevention programs) are often not carried
out by one type of social institution (e.g., the police) but can be carried out by many different
institutions (e.g., schools, churches, community groups, etc.). Furthermore, like other types of
crime, there may be different time periods prior to an actual terrorist event where interventions
might be applied (see Schmid, 1983; Walter, 1969).
Because of the breadth of possible interventions, we initially took the most
comprehensive approach possible and considered for our review any study evaluating the effects
of programs generally designed to prevent, detect, manage, or respond to terrorism events and
related incidents. Prevention strategies can include a wide range of programs designed to deter
9 For example, the United States Department of State, which has described terrorism since 1983 (Title 22 of the
United States Code, Section 2656f[d]) as “premeditated, politically motivated violence perpetrated against
noncombatant targets by sub-national groups or clandestine agents, usually intended to influence an audience.”
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
11
future events, such as improving embassy security, placing metal detectors in airports, or
increasing penalties for crimes deemed to be terrorism. Prevention strategies might also include
early risk prevention, which could include diplomatic efforts, foreign aid, programs to reduce
political or religious fundamentalism, or employment opportunities. Research on detection
strategies (which can also be preventative) might include evaluations of measures which improve
the scanning of shipping containers, border-related strategies, immigration policy or other tactics
used to detect people, places, or situations involved in terror-related activity. Examples of
evaluations of management strategies might include examining the government’s threat level
warning system, new hospital procedures designed to address issues of health security, or
programs put in place to help people who have been impacted by a terrorist event,
psychologically or otherwise. Evaluations studying response strategies might focus on law
enforcement responses to suicide bombing or the effects of economic sanctions on terroristharboring
states. While many of these categories overlap, we felt that prevention, detection,
management, and response strategies provided us with the most inclusive search construct.
The third construct used in considering studies eligible for this review involved deciding
what might be viewed as a measurable outcome of anti-terrorism strategies. Because we
intended our intervention range to be as comprehensive as possible (prevention, detection,
management and response), we broadened the types of acceptable outcomes we would consider
from evaluations. The most straightforward outcome measure would be how the intervention
affected (hopefully lowered) events related to terrorism. These “events” could be incidents of
terrorism, the number of groups who engaged in the use of terrorism, or the frequency of other
activities related to terrorist groups. However, other measurable outcomes might include the
reduction of the public’s fear of terrorism or their feelings of safety from it, the increased ability
to respond to certain events, the reduction in the level of risk of an event occurring, or the ability
to detect mechanisms of terrorism (for example, the ability to detect anthrax spores in mail).
Outcomes measured might include general manifestations (i.e., “terrorism”) as well as specific
groupings (i.e. “skyjackings”, “hostage-taking”, “casualty” or “non-casualty”).
Outcomes that were too far removed from measurable outcomes with regard to terrorism
were not included in the final review. For example, we found one study which used stock
market performance as a measure of the effectiveness of assassinations as a counter-terrorism
strategy (Zussman and Zussman, 2005a; 2005b). We believe that the use of stock market
performance is too indirect a measure of the effectiveness of targeted assassinations of political
leaders in reducing terrorism. Further, the study did not offer evidence to support a more direct
connection and therefore was not included in the final analysis.
4.2 SEARCH STRATEGY
Using these three general constructs we pursued a systematic and comprehensive search
strategy to locate evaluation research. Again, we conducted the most inclusive search possible,
given that our preliminary review of terrorism literature suggested a number of challenges in
locating research that evaluated anti-terrorism programs. Since we had already read the abstracts
for thousands of articles from peer-reviewed sources from our more general review, we
examined these studies as a first step in locating evaluations. At this point, we included all
studies that made any reference to the evaluation of a program, no matter the methodological
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
12
quality of that evaluation. We then went back to the entire database of 14,006 articles we
collected and ran multiple keyword searches on their abstracts and titles for words related to
evaluations. Any study that included the following key words was extracted:
assess
assessment
evaluate
evaluation
effect/effectiveness
empirical
intervention
policy
program
works
Because the initial collection of 14,006 terrorism studies only included studies up through
January 2003 and did not include research from other mediums (books, government and
technical reports, online documents, websites, unpublished material), we re-ran our initial search
through December 2004 and also extended our search across multiple mediums, including books,
government and technical reports, online documents, websites, and unpublished material. Again,
the databases that were searched were:10
10 The time periods covered by each of these databases can be obtained at
https://www.lib.neu.edu/gateway/databasestrifold.pdf.
Academic Search
ArticleFirst (OCLC)
Contemporary Women’s Issues
Criminal Justice Abstracts
EbscoHost
EconLit
Educational Abstracts
Electronic Collections Online
ERIC(OCLC)
GEOBASE
Premier Humanities Abstracts
Ingenta
ISI Web of Science
MEDLINE
National Criminal Justice Reference Service
PAIS (Public/government documents)
PUBMEDLINE
Social Science Abstracts
Sociological Abstracts
And, the keywords (and derivations of these words) we used for our initial search included:
anti-Semitic
black supremacist
bio-terrorism
bombing
emergency response
ethnic violence
hijacking
homeland security
militia group
national security
political crime
political violence
riots
skyjacking
suicide bombing
terrorism (all derivatives of the word)
weapons of mass destruction
white supremacist
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
13
Further, we included combinations of these words with our evaluation keywords listed above to
locate evaluation research related to terrorism. We also used labels of significant terrorism
events (for example, “September 11th”).
Additionally, we searched the internet, as well as terrorism-related organizations and data
sources of terrorism research, for studies that might be reported online or unpublished. The
complete list of websites and available databases we researched are listed in Appendices A and
B. Additionally, when promising articles were located, we searched within articles for possible
references to other evaluations. Using this search strategy, we initially located 290 articles,
reports, internet publications, and other published and unpublished materials that hinted that
some evaluation had been done or was of interest to the study author(s).
We also extended our key word search to multiple book databases using ENDNOTES
connection files.11 We initially discovered 6,415 books that used our terrorism related keywords
in their titles or abstracts and after running a key word search on evaluation terms, identified 64
books that might have evaluations within them or refer to evaluations. Combined with our 290
other documents, a total of 354 studies were located that satisfied our initial search for
empirically-based evaluation studies. Thus, of the over 20,000 reports regarding terrorism
that we located, only about 1.5% of this massive literature even remotely discussed the idea
that an evaluation had been conducted of counter-terrorism strategies.
There were a number of limitations in this search approach. First, was the authors’
dependence on English-language reports. We recognize this is a constraint that cannot be easily
overcome in systematic reviews. We did not have contacts within every country to conduct
subsequent reviews in multiple languages. Secondly, we did not have access to government
reports or data labeled as secret and therefore may have missed classified evaluations (if any) in
our search.
4.3 SELECTION PROCESS AND CRITERIA FOR FINAL INCLUSION
The next step was to narrow these 354 works to only those which evaluated anti-terrorism
strategies using scientific methods. To do this, we more closely examined the abstracts, titles
and notes of each of the 290 non-book studies to see if an evaluation had been conducted,
discussed or even proposed or mentioned. In cases where we felt the abstract provided
insufficient information, we examined the actual article. For the 64 books, we expected the
abstracts to most likely provide insufficient information to help us determine if an evaluation had
been conducted or discussed. Thus, we chose to examine each of the 64 books individually at
this stage. As with general literature related to intervention programs, it is often difficult to tell
from reading the abstracts whether an evaluation had actually been conducted, whether the
article was proposing that an evaluation should be conducted, or whether the article simply
discussed the process of an intervention (without an evaluation). Thus, using the most liberal
approach, we kept any article or book at this stage that even hinted to either conducting or
discussing an evaluation of a counter-terrorism strategy.
11 Using Endnotes version 6, we employed all university library connection files within that version. While this list
is too numerous to display here, it can be found at www.endnote.com.
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
14
Of the 64 books initially located through our keyword search, 38 books were physically
found when searching four separate university libraries and their affiliates.12 Of these 38 books,
only one book indicated that some evaluation of an anti-terrorism program had been undertaken
(Wilkinson, 1977). Of the 290 non-book studies initially chosen for examination, 94 abstracts
appeared to suggest that the study was at least remotely connected to an evaluation designed to
prevent, detect, manage, or respond to terrorism. We were able to physically locate 79 of the 94
non-book studies (84%), again by searching multiple libraries, databases, and the internet, asking
for assistance from colleagues and graduate students at different universities, and contacting
authors themselves for articles that could not be located. We now had 80 studies that seemed,
upon closer examination, to involve some minimal form of evaluation.
The next stage of our selection process involved making a methodological judgment
about these 80 studies to determine if they satisfied the minimum requirements to be considered
an evaluation. To do this, we chose a two-step process. First, we conducted an initial reading of
the full text of each of the eighty studies and chose only those studies which appeared to at least
attempt to connect an outcome or effect with a program through a minimally rigorous scientific
test. For example, studies which used simple correlation statistics were retained. While this step
was a conservative approach, we felt that the nature of terrorism studies (for example, that events
are rare or the difficulty in collecting information on terrorism) warranted careful consideration
of possible studies to include.
We identified 21 from these 80 studies that satisfied this minimum criterion. Reasons for
excluding the other 59 studies were that the study:
– was not an evaluation of a program or policy designed to detect, prevent, respond to,
or manage terrorism;
– described the process of a program, but did not evaluate it;
– made claims that a program was effective without any empirical test of that claim;
– was a news article reporting on individuals claiming effectiveness;
– was a review of non-evaluation literature;
– advocated that an evaluation should be done but did not conduct one;
– surveyed individuals about their feelings as to a program’s effectiveness;
– surveyed individuals about how prepared they felt for another attack;
– determined the effects of terrorism, rather than the effectiveness of a program;13
– examined attrition for participation in a program, but not the effects of the program;
– made suggestions for treatment of terrorism related injuries; and/or
– described criteria as to what effective policy should look like, but did not evaluate
anything.
The second step of selecting the final studies for analysis was to only include studies in
which authors used more rigorous methodological designs. Not all evaluations are created equal;
12 Graduate assistants helped search for books at Harvard University, Northeastern University, Rutgers University
and the University of Maryland at College Park.
13 One anonymous reviewer of this review questioned why we did not consider studies examining the effectiveness
of terrorism itself. We believed that this literature was generally inappropriate for this review unless the study
authors framed their interest in the effectiveness of terrorism in terms of evaluating how terrorism might thwart a
counter-terrorism strategy.
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
15
they vary in both internal and external validity which lends to the believability of their findings
(Cook and Campbell, 1979; Farrington, 2003; Shadish et al., 2002). The minimal requirements
of the first stage only indicated that a weak test was conducted connecting the program with the
desired outcome. For example, a program designed to treat post-traumatic stress disorder
(PTSD) given to victims of terrorist acts may determine that the program worked because scores
for indications of PTSD reduced over time after an intervention. However, this finding could be
entirely spurious; those individuals who were not given the program may have improved
similarly in the same amount of time.
On the other hand, only including randomized controlled experiments in our review
would be impractical. Not only would we have nothing to report, but the use of experimentation
on rare event outcomes may be too strict a criteria. Thus, as we initially outlined in our protocol,
we were guided by what Sherman et al. (1997) describe as a moderate or mid-level scientifically
rigorous design. In Sherman et al., they devised a five point “Scientific Methods Scale” (SMS)
to score the methodological quality of evaluation research.14 For them, the highest quality
evaluation method, the randomized controlled experiment, was ranked a “5” while simple
correlational studies were scored as “1” or “2”. We recognize there have been critiques of this
scale, and we used it only as a general guide to assist with our selection process.
We used the middle SMS score of “3” as a general guide to exclude those studies which
were not at least moderately rigorous, but warn the reader that moderately rigorous designs are
less believable than more rigorous ones. Sherman et al. describe an SMS = 3 as “a comparison
between two or more units of analysis, one with and one without the program”. Throughout
Sherman et al. (1997), there are also elaborations on the meaning of a “3” by different authors,
suggesting that a methodological score of “3” can also point to studies with multiple units who
receive or do not receive the program and attempt to control for other factors.
Time series methods were also coded a score of “3” in Sherman et al.15 However, as
Shadish et al. (2002) point out, time series can be a powerful research design. They equate
quality time series designs with quasi-experimental studies. In this study, the vast majority of
findings were derived by time series methods and they are viewed as moderate to stronger
research designs.
Using this guidance, we decided that our final inclusion of studies from the 21 minimally
acceptable studies included those that at least:
– evaluated two or more units of analysis, comparing some with and without the
counter-terrorism intervention;
– made some attempt to provide for controls within a statistical analysis;
– conducted an interrupted time series or intervention analysis to indicate some
temporal ordering of effects.
After re-reading each of the articles, we determined that 10 studies in Table 2 satisfied our
criteria of having used a moderately rigorous methodology in determining the effects of a
program on a desired outcome. In Table 2 below, the first column, entitled “Moderately rigorous
14 The Scientific Methods Scale can be found in Chapter 2 of the Report and can be found at
https://www.ncjrs.org/works/chapter2.htm.
15 This is mentioned in Sherman et al. (1997) on page 2-19, note 7.
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
16
method used (Scored at least a “3” on the SMS or equivalent)” denotes these 10 studies.
4.4 FURTHER EXCLUSION OF MEDICAL-RELATED STUDIES
The 10 studies we located that were at least of moderately strong methodological design
included three studies from the medical field – Halloran et al. (2002), Lallement et al. (1998),
and Quinn et al. (2002). Halloran et al. (2002) created and tested a simulation of the release of
smallpox and the effectiveness of targeted and mass vaccinations. They found that smallpox
epidemics could be more effectively prevented and controlled with well-timed mass vaccination
(prior to smallpox introduction or immediately thereafter) in comparison to targeted vaccination
if there is no preexisting immunity. Targeted vaccination, which prevented more cases per dose,
was the preferred method if vaccines were limited. Lallement et al. (1998) analyzed the effects
of gacyclidine (GK-11) as a therapy for nerve agent poisoning using monkeys in a laboratory
setting. They found positive effects of the therapy on reducing the chances of sickness and death
associated with nerve agent poisoning. Finally, Quinn et al. (2002) analyzed the effects of an
ELISA (enzyme-linked immunosorbent assay) for immunoglobulin G antibodies to anthrax
infections. They found that the ELISA was sensitive in diagnosing the presence of anthrax
toxins.
Each of these studies used moderate to strong evaluative methods in testing for (or
simulating) the effectiveness of these interventions. However, we decided to exclude these three
studies for two reasons. First, while our search revealed only three medically-oriented
evaluations related to terrorism, we knew that this was not representative of the medical field’s
research on nerve agents, smallpox or anthrax (or other issues related to terrorism), but rather
just what appeared in our search specific to terrorism. For example, a Cochrane review related to
Anthrax (Jefferson et al., 1998) already exists. Findings related to these treatments could not be
summed up by the three studies we found that happen to focus more specifically on their use in
terrorism. As an example, we came across one article during our search – Gillespie et al. (2002)
– that didn’t make our final cut of having a moderately rigorous design but is a good example,
nonetheless. This article discussed the effectiveness of cognitive behavioral therapy on posttraumatic
stress disorder (PTSD). Had this article been methodologically rigorous enough to be
included, its inclusion in our study would occur only because it directed its study towards victims
of terrorism. However, there has been a large literature on the treatment for post-traumatic stress
disorder (Bisson and Andrew, 2005; Rose et al., 2002; Stein et al., 2000), and to include one
article without including others simply because one referenced terrorism victims would be unfair
to the literature.
Secondly, while we generally understood the methodological strategies used by these
researchers, the context by which the findings were discussed were much removed from the
expertise of the authors. As Lipsey and Wilson (2001) have pointed out, research synthesis often
demands a wide range of expertise across fields, and in this particular case, terrorism can extend
out towards many different areas of expertise. To interpret the meaning of outcomes measured
may require more medical expertise than the authors have. After excluding these three studies,
we were left with seven articles in our final selection. Again, we refer to Table 2, which in the
last column indicates the final seven studies that were included.
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
17
Table 2. Studies Included in the C2 Systematic Review (Last Column) and Final Exclusion
Steps
Study listed if it was at least some form of a weak evaluation (SMS=1 or equivalent)
Moderately rigorous method used
(Scored at least a “3” on the SMS or
equivalent)
Moderately rigorous and not a
medical evaluation
Barros, C. (2003). An Intervention Analysis of Terrorism: The Spanish Eta Case. Defence and
Peace Economics, 14(6): 401-412. X X
Bozzette, S. A., Boer, R., Bhatnagar, V., Brower, J. L., Keeler, E. B., Morton, S. C. et al. (2003).
A Model for Smallpox-Vaccination Policy. New England Journal of Medicine, 348(5): 416-425.
Brophy-Baermann, B., and Conybeare, J. A.C. (1994). Retaliating Against Terrorism: Rational
Expectations and the Optimality of Rules Versus Discretion. American Journal of Political
Science, 38(1) (Feb): 196-210.
X X
Cauley, J. and Im, E. (1988). Intervention Policy Analysis of Skyjackings and Other Terrorist
Incidents. The American Economic Review, 78(2):27-31. X X
Chauncey, R. (1975). Deterrence: Certainty, Deterrence, and Skyjacking. Criminology,
12(4):447- 473.
Enders, W. and Sandler, T. (1993). The Effectiveness of Antiterrorism Policies: A Vector-
Autoregression-Intervention Analysis. The American Political Science Review, 87(4): 829-844 X X
Enders, W., Sandler, T., and Cauley, J. (1990). UN Conventions, Terrorism, and Retaliation in
the Fight Against Terrorism: An Econometric Evaluation. Terrorism and Political Violence,
2(1):83.
X X
Enders, W. and Sandler, T. (2000). Is Transnational Terrorism Becoming More Threatening?
Journal of Conflict Resolution, 44: 307-332. X X
Gillespie, K., Duffy, M., Hackmann, A. and Clark, D. (2002). Community Based Cognitive
Therapy in the Treatment of Posttraumatic Stress Disorder Following the Omagh Bomb
Behaviour Research and Therapy, 40: 345-357.
Halloran, M. E., Longini Jr., I. M., Nizham, A. and Yang Y. (2002). Containing Bioterrorist
Smallpox. Science, 298: 1428-1432. X
Johnston, R. G., Garcia, A. R.E., and Pacheco, A. (2002). The Efficacy of Tamper Indicating
Devices. Journal of Homeland Security, April. Los Alamos National Laboratory, Vulnerability
Assessment Team. Available online at: https://www.homelandsecurity.org
Lallement, G., Clarencon, D., Masqueliez, C. et al. (1998). Nerve Agent Poisoning in
Primates: Antilethal, Anti-epileptic and Neuroprotective Effects of GK-11. Archives in
Toxicology, 72: 84-93.
X
Landes, W.M. (1978). An Economic Study of U.S. Aircraft Hijackings, 1961-1976. Journal of
Law and Economics, 21:1-31. X X
LeVine, V. T. and Salert, B. A.. (1996). Does a Coercive Official Response Deter Terrorism?
The Case of the PLO. Terrorism and Political Violence, 8(1): 22-49.
Martz, H. and Johnson, M. (1987). Risk Analysis of Terrorist Attack. Risk Analysis, 7(1): 35-
47.
Prunckun, H. and Mohr, P. (1997). Military Deterrence of International Terrorism: An
Evaluation of Operation El Dorado Canyon. Studies in Conflict and Terrorism, 20:267-280.
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
18
Quinn, C., Semenova, V., Elie, C., et al. (2002). Specific, Sensitive, and Quantitiative Enzyme-
Linked Immunosorbent Assay for Human Immunoglobulin G Antibodies to Anthrax Toxin
Protective Antigen. Emerging Infectious Diseases, 8(10):1103-1110.
X
Smith, B. L., Damphousse, K. R., Jackson, F. and Sellers, A. (2002). The Prosecution and
Punishment of International Terrorists in Federal Courts: 1980-1998. Criminology and Public
Policy, 1(3): 311-338.
Smith, B. L. and Orvis, G. P. (1993). America’s Response to Terrorism: An Empirical Analysis
of Federal Intervention Strategies During the 1980’s. Justice Quarterly, 10(4): 661-681.
Wilkinson, P. (1977). Terrorism and the liberal state. New York: John Wiley and Sons.
Zussman, A. and Zussman, N. (2005). Assassinations: Evaluating the effectiveness of a
counterrorism policy using stock market data. Cornell University, Department of Economics.
4.5 SUMMARIES OF THE FINAL SEVEN STUDIES
Below are brief summaries for each of the seven reports that satisfied our criteria. As
will be detailed, within each of these reports were multiple findings on multiple interventions.
However, we summarized the studies here by author(s) as an overview. Additionally, with the
exception of Landes (1978) and Barros (2003), the studies primarily use some version of
Mickolus’s chronologies of terrorism (see Mickolus, 1980; 1982; Mickolus et al., 1989;
Mickolus et al., 1989; 1993), which were later used to create the ITERATE databases.
LANDES (1978): Landes’ study of United States Federal Aviation Administration data on
skyjackings is the earliest evaluation study of counter-terrorism strategies that we could locate
and is often referred to in subsequent evaluations. Using analytic techniques common to the
field of economics, Landes examined the effects of changes in laws and security measures
through the increased probability of apprehension, incarceration, longer sentences, and being
killed (by authorities) on the quarterly rate of domestic hijackings and the number of days and
flights between successive hijackings for the period between 1961 and 1976. Landes also
controlled for other variables, including the number of air flight operations per quarter, civilian
unemployment rates, population rates, and personal consumption. He used a variety of ordinary
least squares regression techniques to ascertain relationships between these variables. We chose
to use Landes’ Table 5 (p. 17) as opposed to his Table 3 as it appeared the most complete model
in terms of its inclusion of foreign hijackings as a control variable.
CAULEY AND IM (1988): Cauley and Im offered an interrupted time series analysis (also
known in economics as intervention analysis) of incidents occurring between 1968 and 1979. In
their study, they examine the effectiveness of increased airport security screening measures that
occurred in 1973, increased security at embassies and other diplomatic missions in 1976, and the
U.N. convention on preventing crimes against diplomatic personnel enacted in 1977. Cauley and
Im not only analyzed multiple interventions, but also the effects of these interventions on
different outcomes, including skyjackings and non-skyjacking incidents, such as hostage taking,
barricades, and attacks on diplomats.
ENDERS, SANDLER AND CAULEY (1990): This evaluation of multiple interventions and
The Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
19
outcomes used an interrupted time series approach for events between 1968 and 1988. Enders et
al. examined the effects of metal detectors in airports in 1973, the UN Convention on the
Prevention and Punishment of Crimes Against Internationally Protected Persons Including
Diplomatic Agents in 1977, United Nations resolutions against General Assembly and Security
Council hostage taking (1985), United Nations resolutions against aerial hijacking (1969-1970),
as well as the United States retaliatory raid on Libya in 1986.
ENDERS AND SANDLER (1993): Enders and Sandler continued to contribute to the evaluation
literature with the use of intervention analysis (and improving upon specific techniques within
that analysis) in this 1993 evaluation of the substitution effects that policies may have between
different types of terrorism. Like Enders et al. (1990), they examined the effects of metal
detectors and resolutions during the period of 1968 – 1988, but also examined security
fortification measures taken on U.S. embassies. However, this study differs from their 1990
paper in that they are analyzing interactions and substitution/displacement effects of different
interventions across different types of terrorism. Of the many models that Enders and Sandler
put forth, we chose to use Model 2 (p. 839) for our review because it had a richer set of variables
than Model 1 and did not just include the United States (as did Model 3, p. 841).
BROPHY-BAERMANN AND CONYBEARE (1994): Brophy-Baermann and Conybeare also
employ an interrupted time series/intervention analysis approach to determine the effectiveness
of six Israeli military-led retaliation attacks on reducing terrorism from the PLO and Lebanon.
These retaliations began in September of 1972 in response to the killings of Israeli athletes at the
Munich Olympic Games, of which five more retaliations followed through 1988.
ENDERS AND SANDLER (2000): In 2000, Enders and Sandler expanded their analysis to
terrorism events between 1970 and 1996. They study the effectiveness of metal detectors,
embassy fortification, and the Libyan raid, as well as the reduction in totalitarian governments
that occurred after the end of the Cold War. Additionally, rather than study the effects of these
interventions on different types of terrorism (or their substitutions) as they did before, they used
outcomes which measure the type of person-based destruction. These outcomes included death
events, wounded, and non-casualties per quarter. It should also be noted that we considered
including a related article in our review (Enders and Sandler, 2002). However, the 2002 article
explores alternative time series technical approaches but does not present complete results. The
2002 article is more of an exploration into alternative methods, rather than an evaluation of antiterrorism
interventions.
BARROS (2003): Barros deviates from the use of ITERATE datasets and employs intervention
analysis specifically on Spanish terrorist organization, ETA, terrorism using information
collected by Abadie and Gardeazabal (2001). The data include assassinations and kidnappings
conducted by ETA between 1968 and 2000. He studied the effects of different political
ideologies in power, police and military expenditures as well as increases in foreign investment
on incidents of kidnappings and assassinations conducted by the Spanish ETA. Barros suggested
that the most parsimonious model used a vector autoregression (VAR) time series framework,
and thus, we report findings from his Table 6 (p. 410) only. This model only tested for the
effects of having a socialist party in power, which he suggests represents “hard line politicalThe
Effectiveness of Counter-Terrorism Strategies
___________________________________________________________________________
20
oriented policy” (p. 412), as the i

Deja una respuesta

Tu dirección de correo electrónico no será publicada. Los campos obligatorios están marcados con *