Essay: Human Intelligence in Federal Terrorism Cases

Abstract

The September 11th attacks changed everything. Domestic policy alterations were among the most significant changes after 9/11. Specifically, the 2002 Ashcroft Attorney General Guidelines and the 2008 Mukasey Attorney General Guidelines were at the forefront of post 9/11 policy changes. These Attorney General Guidelines changed the way the Federal Bureau of Investigation (FBI) investigated terrorism in the United States in addition to making counterterrorism the FBI’s top priority. One of the ways the FBI prevents terrorism is through the use of undercover agents and confidential informants. How the 2002 and 2008 Attorney General Guidelines influenced the FBI’s use of human intelligence in terrorism cases remained to be seen. This study reviewed the issue of policy changes post 9/11 and their influence on the FBI’s use of undercover operatives in terrorism cases.

HUMAN INTELLIGENCE IN FEDERAL TERRORISM CASES
I. Introduction
The terrorist attacks on September 11, 2001, killed nearly 3,000 people. During the aftermath of those attacks, the media and the public looked for someone to blame, and among those who received blame, the intelligence community faced heavy criticism. In the decade that ensued, policymakers amended guidelines upon which the intelligence community operated. The changes in guidelines were intended to help members of the intelligence community prevent terrorist attacks, such as those that occurred on September 11, from happening again.
The intelligence community in the United States encompasses sixteen different agencies and offices, each handled by the Office of the Director of National Intelligence (Kaplan, 2012). Within these sixteen agencies are six program managers who are responsible for gathering and examining various types of intelligence information (Kaplan, 2012). Amid the managerial agencies is the Federal Bureau of Investigation (FBI). As a member of the intelligence community, the FBI is responsible for warning policymakers of potential threats to the nation’s security (Schalch, 2003). Moreover, the FBI is the lead authority in investigating terrorism in the United States. In order to warn policymakers of potential threats, the FBI obtains and examines intelligence information gathered via human intelligence, also known as confidential informants and undercover agents.
The procedures governing the FBI’s domestic investigative operations are known as the Attorney General Guidelines. Since the attacks on September 11, 2001, there have been two changes in these guidelines: the 2002 John Ashcroft Attorney General Guidelines, and the 2008 Michael Mukasey Attorney General Guidelines.
The Ashcroft Guidelines expanded the scope of investigative procedures. After implementation of the 2002 guidelines the FBI began operating as more of a proactive, rather than reactive agency. The 2008 Mukasey Guidelines bolstered the FBI’s investigative abilities by imposing fewer restrictions (Shields, 2012).
Changes to the guidelines over the past decade not only increased the FBI’s power, but they also changed its ability to use confidential informants and undercover agents. Few empirical studies have analyzed the FBI’s use of confidential informants and undercover agents in domestic terrorism incidents. Of those that have, there have been several key findings. First, it is known that just over 40% of domestic terrorism cases use human intelligence (Greenberg, 2011). Second, there was nearly a 40% drop in the use of informants and over a 25% drop in the use of undercover agents in the immediate years following 9/11 (Shields, 2012). Finally, confidential informants and undercover agents represent some of the most successful tactics in detecting terrorist plots and in identifying and apprehending perpetrators in completed terrorist attacks (Hewitt, 2014). Thus, limited research on human intelligence in terrorism cases has shown the significance of undercover operatives.
To better understand the use of human intelligence in terrorism a theoretical approach from structural contextual theory will be used. Structural contextual theory is the idea that components of the justice system normally work autonomously of one another and that the political environment of the time may change normal criminal justice operations to come together to target a specific type of crime and criminal for prosecution (Smith & Damphousse, 1998). In the past, structural contextual theory has been used to help explain responses to terrorism in terms of sentencing outcomes. Smith and Damphousse (1998) hypothesized that if terrorism caught the public’s attention the justice system would work together rather than independently of one another. Collaboration throughout the justice system would then limit discretion among prosecuting and sentencing terrorists. Thus, the more seriously terrorism is viewed by the public, the less difference there should be in sentencing outcomes in terrorism cases. Smith and Damphousse (1998) found nearly four times more explained variance in sentence outcomes for terrorists than nonterrorists, supporting structural contextual theory. Using Structural Contextual Theory, this paper will examine the use of confidential informants and undercover agents post-9/11 to address two research questions. The first research question will examine how the Ashcroft and Mukasey Guidelines changed the use of confidential informants and undercover agents used in terrorism cases. The second research question will examine what impact, if any, the changes in the use of undercover agents and confidential informants have had on case processing and outcomes in terrorism cases.
As mentioned, there is a shortcoming in empirical research on the use of confidential informants and undercover agents in domestic terrorism cases. Prior literature is limited to the year 2004 concerning research on undercover agents and limited to the year 2011 concerning research on confidential informants (Shields, 2012; Greenburg, 2011). Additionally, the most recent study (Hewitt, 2014) involving human intelligence did not examine the nature of terrorism case outcomes in cases with informants and undercover agents. My contributions build off Shields (2012) as well as Greenberg (2011) and Hewitt (2014) by extending the time frame of analyzing human intelligence to 2014 and by examining how the use of confidential informants and undercover agents varies across category of terrorism, intended target type and conviction rate in cases with human intelligence vs. cases without human intelligence among other factors.
Examining confidential informants and undercover agents through this study will add to existing literature while also being socially relevant. After 9/11 the FBI was criticized for not connecting the dots concerning information they possessed that potentially could have prevented the attacks from that day. One criticism was that the FBI was too reactive and suffered from a structural weakness (Zegart, 2007). FBI agents were trained to be reactive, case driven, and conviction oriented- to investigate past crimes and not to prevent future ones (Zegart, 2007). In response to this criticism, Attorney General John Ashcroft changed the FBI from a reactive agency to a proactive agency with his 2002 Attorney General Guidelines by bolstering investigative procedures. While Ashcroft changed the structure of the FBI into a proactive agency, recent criticism has suggested that the FBI is being too proactive, especially through sting operations using confidential informants and undercover agents. One example comes from the 2014 HBO documentary The Newburgh Sting, which depicts a 2009 terrorism plot where four men were allegedly coaxed by an FBI informant to attack U.S. military planes in New York. Findings from the current study may be relevant for homeland security policy, as socially relevant human rights issues have arisen from cases such as The Newburgh Sting.
This study is set up as follows: First, I provide a historical account of policy changes and federal terrorism investigations. Second, I review extant literature on confidential informants and undercover agents. Third, I lay out research questions followed by a methods section containing hypotheses. Fifth, I present analytic results and finally end with a discussion.
II. Policy Change and Terrorism Investigations
Background
Terrorism occurs in practically every country throughout the world’s history, and its causes have been extremely varied over time and space (Mahan & Griset, 2013). In the history of American terrorism, the Oklahoma City bombing and the attacks of 9/11 are the two events upon which all other events are compared (Damphousse & Shields, 2007). The Oklahoma City bombing awakened America to the threat of domestic terrorism while 9/11 demonstrated the threat of international terrorism. The culmination of these attacks brought vast changes to anti-terrorism policies and the agencies that operate under them.
The FBI was most influenced by these changes, given its primary responsibility to investigate domestic terrorism. While most government and intelligence agencies define terrorism differently, the FBI defines terrorism as ‘the unlawful use of force or violence against persons or property to intimidate or coerce a government, the civilian population, or any segment thereof in furtherance of political or social objectives’ (FBI, 2014). The procedures to determine who the FBI investigates, and under what circumstances, are in many ways, as important as the FBI’s definition of terrorism (Smith, 1994). The procedures that determine the FBI’s investigations are administered under the ‘Attorney General’s Guidelines for Domestic FBI Operations’ (Mukasey, 2008).
Attorney General Edward Levi created the first Attorney General Guidelines in 1976. Levi developed the guidelines as a response to the free speech and privacy violations uncovered during the Nixon Administration, with the intent to prevent future abuses and protect American citizens while allowing the FBI to conduct intelligence operations (Jones, 2009). While the Levi guidelines were created with restrictions on the FBI’s investigative authority, events over time would cause the revision of previous guidelines with fewer constraints.
One of those events was the robbery of an armored truck by the May 19th Communist Organization that left a security guard and two police officers dead. Later that same year, the United Freedom Front killed a New Jersey state trooper during a routine traffic stop (Smith, 1994). In response to these attacks and others, Attorney General William Smith issued a new set of guidelines in 1983, giving FBI field offices greater flexibility in the investigative process (Smith, 1994). It wasn’t until after 9/11; however, that any major changes in the Attorney General Guidelines were implemented. The 2002 Ashcroft Guidelines heralded new revisions including allowances for longer investigation periods, centralizing fieldwork analysis at FBI headquarters, and eliminating the need to gain permission from FBI headquarters to open investigations (Shields, 2012).
While the Ashcroft guidelines were being formed, Congress passed the USA PATRIOT Act in late 2001. The PATRIOT Act strengthened the FBI’s investigative authority in terms of surveillance and wiretapping. Additionally, the PATRIOT Act loosened standards for obtaining warrants and issuing subpoenas in investigative matters (Jones, 2009). After creation of the PATRIOT Act there would be no significant policy changes until the 2008 Mukasey Guidelines were issued which further enhanced the FBI’s power to investigate cases. In addition to outlining the purposes and procedures for three types of investigations to be discussed below, the Mukasey Guidelines permitted these investigations on either the violation of federal statutes or ‘threats to national security’ (Jones, 2009).
While the Ashcroft Guidelines were the first significant changes to the Attorney General Guidelines since 1983, they are especially significant to this study because of changes to the Confidential Informant Guidelines (CI Guidelines, 2002) and Undercover Operations Guidelines (UCO Guidelines, 2002). Both of these guidelines provide detailed rules regarding the use of confidential informants and undercover agents. A 2005 OIG special report reviewing the Confidential Informant Guidelines noted that FBI personnel ranging from new agents to the Director stated that the paperwork associated with opening and operating informants was excessive, burdensome, and time-consuming. The report also noted that some FBI agents had become reluctant to open informants because of these and other administrative and operational burdens (OIG, 2005).
III. Policy Changes
The events of 9/11 significantly influenced counterterrorism policy. The 9/11 Commission scrutinized the FBI for being rooted in law enforcement and not equipped to accomplish intelligence assignments (Shields et al, 2009). Changes in policy after these attacks substantially altered the way the federal government investigated and prosecuted those individuals suspected to be involved in terrorism (Shields et al, 2009). Prior to the specific 2002 policy changes regarding confidential informants and undercover agents and the general changes to the FBI’s investigation of terrorism, domestic antiterrorism policy focused on infiltrating and beheading terrorist organizations (Shields et al, 2009). After the 2002 policy changes, the FBI’s focus as mandated by Attorney General Ashcroft was to intervene early and investigate aggressively. Ashcroft stated that, ‘our philosophy today is not to wait and sift through the rubble following a terrorist attack. Rather, the FBI must intervene early and investigate aggressively where information exists suggesting the possibility of terrorism, so as to prevent acts of terrorism’ (Ashcroft, 2002). The 2002 Ashcroft Attorney General Guidelines, the USA Patriot Act and the 2008 Mukasey Attorney General Guidelines altered the FBI’s approach to combating terrorism. The most dramatic of all the post 9/11 policy changes, though, were the Ashcroft Guidelines. These guidelines swiftly and singlehandedly transitioned the FBI from being a proactive agency to a reactive agency.
Ashcroft Guidelines
On May 30, 2002, Attorney General John Ashcroft issued the first Attorney General’s Guidelines on General Crimes, Racketeering Enterprise and Terrorism Enterprise Investigations after 9/11. At the time these were issued, Ashcroft and FBI Director Robert Mueller stated that the revised guidelines were necessary to abolish departmental barriers limiting field agents and their superiors to handle terrorist threats, while simultaneously directing day-to-day activities of federal law enforcement departments within legal and constitutional boundaries (OIG, 2005). The 2002 guidelines authorized several new tools for the FBI. The first was the authority to start specific types of investigations with fewer evidentiary thresholds and without approval from FBI Headquarters (OIG, 2005). Previously, the Smith guidelines required FBI Headquarters to approve all terrorism investigations before initiation, and the director was to supervise the investigation in 180-day intervals (Shields et al, 2009). The new Ashcroft Guidelines provided FBI field offices with the authority to commence investigations unilaterally, and allowed those investigations to proceed up to a year before reporting them to FBI Headquarters. Second, the Ashcroft Guidelines extended authority to the FBI to use undercover techniques in criminal intelligence investigations (racketeering enterprise and terrorism enterprise investigations) that were previously only allowed in general crimes investigations (OIG, 2005). The guidelines stated that, ‘In obtaining the foregoing information, any lawful investigative technique may be used’ (Ashcroft, 2002, at 17). A third authorization involved public places and events. The guidelines maintained that, ‘for the purpose of detecting or preventing terrorist activities, the FBI is authorized to visit any place and attend any event that is open to the public, on the same terms and conditions as members of the public generally’ (Ashcroft, 2002, at 22). Lastly, the guidelines allowed the FBI to conduct online searches, and access online forums and sites, just as the public may for purposes of preventing or detecting terrorism or other criminal activities (Ashcroft, 2002). In a speech addressing the revisions to the guidelines, Ashcroft summed up that a key objective for the FBI was to prevent terrorism by intervening early and investigating aggressively (2002).
Confidential Informant Guidelines
The Confidential Informant Guidelines define confidential informant as, ‘any individual who provides useful and credible information to a Department of Justice Law Enforcement Agency (JLEA) regarding felonious criminal activities, and from whom the JLEA expects or intends to obtain useful and credible information regarding such activities in the future’ (CI Guidelines, 2002: 2). The 2002 Ashcroft guidelines contained three minor revisions to the Confidential Informant Guidelines. The first change involved the verbatim reading of instructions to informants. Under previous guidelines, agents working with confidential informants were required to read, verbatim, specific instructions concerning the boundaries set on the CIs’ activities (OIG, 2005). The 2002 revision to this rule removed the verbatim reading requirement, stating that, ‘at least one agent of the JLEA, along with one additional agent or other law enforcement official present as a witness, shall review with the CI written instructions’ (CI Guidelines, 2002). FBI Director Robert Mueller deemed this change necessary because, ‘the verbatim instructions, written in often intimidating legalese, were proving to have a chilling effect, causing confidential informants to leave the program’ (Oversight Hearing, 2003).
The second alteration to the CI Guidelines permitted agents to adapt the instructions ‘ including instructions that safeguard the confidentiality of the informant’s identity ‘ to the informant’s distinct situation (OIG, 2005). The final modification to the informant guidelines involved promising immunity from prosecution. Prior guidelines required agents handling CIs to instruct them that investigative agencies could not promise immunity from prosecution (OIG, 2005). The 2002 Guidelines adamantly clarified agents’ roles stating that, ‘whether or not this instruction is given to a CI, the JLEA does not have any authority to make any promise or commitment that would prevent the government from prosecuting an individual’and a JLEA agent must avoid giving any person the erroneous impression that he or she has such authority’ (CI Guidelines, 2002: 12).
Undercover Operations Guidelines
The Undercover Operations Guidelines defines an undercover employee as, ‘any employee of the FBI, or employee of a Federal, or local law enforcement agency working under the direction and control of the FBI in a particular investigation, whose relationship with the FBI is concealed from third parties in the course of an investigative operation by the maintenance of a cover or alias identity’ (UCO Guidelines, 2002: 1). The 2002 Ashcroft Guidelines also provided four major counterterrorism-related revisions to the Undercover Operations Guidelines. The first change simply placed an emphasis on terrorism prevention as a legitimate goal of undercover operations (OIG, 2005). The 2002 guidelines stated that, ‘The use of undercover techniques’is essential to the detection, prevention, and prosecution’ of terrorism’ (UCO Guidelines, 2002: 1).
The second amendment to the guidelines made explicit and emphasized the FBI’s authority to use undercover techniques towards its criminal intelligence investigations goals. The guidelines stated that, ‘In criminal intelligence investigations ‘ i.e., racketeering enterprise investigations and terrorism enterprise investigations ‘ these methods may be used to further the investigative objective’ of detection, prevention, and prosecution of the criminal activities of the enterprise’ (UCO Guidelines, 2002: 2).
The third alteration to the UCO Guidelines gave Special Agents in Charge (SAC) more authority. The guidelines stated that, ‘the SAC may approve an undercover operation when’the initiation, extension, or renewal of an operation is necessary to avoid the loss of a significant investigative opportunity’ (UCO Guidelines, 2002: 15). The fourth and final revision to the UCO Guidelines added one additional factor that SAC’s must take into account before approving an operation. The additional factor was, ‘[t]he risk of invasion of privacy or interference with privileged or confidential relationships and any potential constitutional concerns or other legal concerns’ (UCO Guidelines, 2002: 3). In sum, the 2002 revisions of the UCO Guidelines overall gave the FBI more authority in undercover operations.
USA PATRIOT Act
Like the Attorney General Guidelines, the passing of the USA PATRIOT Act by Congress in 2001 significantly enhanced law enforcement’s investigative abilities. Among new powers the PATRIOT Act gave to the FBI was the capability to obtain a warrant and conduct investigations and surveillance without first notifying the individual, in addition to delaying notification given proper conditions (Jones, 2009). As stated in Section 213 of the Act, providing immediate notification could be delayed if the court found reasonable cause that execution of the warrant may have an adverse result (USA PATRIOT Act, 2001). Furthermore, the PATRIOT Act allowed the government to issue a subpoena and acquire information from targets without conferring with a court of law (Jones, 2009). Significantly, the PATRIOT Act stated that, ‘ ‘foreign intelligence information’ means information that relates to the national defense or the security of the United States’ (USA PATRIOT Act, 2001). These semantics relaxed boundaries of the procedure and scope of intelligence operations (Jones, 2009). This permitted the FBI to administer FISA in any way it believes applies to the context of national security (Jones, 2009). Other noteworthy changes resulting from the PATRIOT Act included enhanced surveillance procedures resulting in new wiretapping and pen register and trap and trace means. This provided the government more authority to monitor telephone numbers received and called from specific telephone lines.
Mukasey Guidelines
On December 1, 2008, Attorney General Michael Mukasey further modified the Attorney General Guidelines for terrorism investigations by consolidating FBI investigative guidelines under a single rubric. The issuance of the 2008 guidelines represented the pinnacle of the evolution of the FBI towards an intelligence agency rather than strictly being a law enforcement organization. Specific to terrorism, these guidelines primarily altered the investigation of cases. Three types of investigations are outlined in the 2008 Guidelines- Assessments, Predicated Investigations, and Enterprise Investigations.
Assessment investigations are to be used for the purpose of detecting, obtaining information about, or preventing or protecting against threats to national security or federal crimes (Mukasey, 2008). More specifically, they are to be used to detect and interrupt criminal activities at their early stages (Mukasey, 2008). Regarding authorization, assessments do not require any specific factual predication, only an ‘authorized purpose’ and do not warrant approval by supervisors (Shields et al, 2009).
Predicated Investigations are more restricted than assessments in that they require approval from a Special Agent in Charge or by an FBI Headquarters official (Mukasey, 2008). Additionally, predicated investigations require predication ‘ ‘allegations, reports, facts or circumstances indicative of possible criminal or national security threatening activity’ (Mukasey, 2008, at 18). Predicated investigations are partitioned between preliminary investigations and full investigations. Preliminary investigations can be commenced on any information or allegation indicative of possible national security or criminal threatening activity whereas full investigations require a more factual predication (Mukasey, 2008). Conversely, full investigations have no time limit, whereas preliminary investigations terminate after six months unless a SAC approves a six-month extension or FBI Headquarters approves an extension greater than one year (Mukasey, 2008).
The final type of investigation approved in the 2008 Guidelines is the Enterprise Investigation. Enterprise investigations are a form of full investigations, which generally examine the structure, scope, and nature of specific groups and organizations (Mukasey, 2008). While enterprise investigations require a factual predication, they are distinct in their focus on organizations and groups that may be involved in racketeering activity, terrorism, or other threats to national security (Mukasey, 2008).
IV. Literature Review
Research containing empirical terrorism analyses until recently has been limited, especially concerning human intelligence. Nonetheless, a couple of studies have examined confidential informants and undercover agents. Shields (2012) and Greenberg (2011) looked at the prevalence of informants and agents while Hewitt (2012) studied factors involved in preventing terrorism attacks and apprehending terrorist perpetrators. Confidential informants and undercover agents are additionally brought up in literature concerning successful terrorist attacks and thwarted terrorism plots (Strom, 2010; Dahl, 2011).
While empirical studies on the use of undercover agents and confidential informants within the context of terrorism investigations are scarce, Shields’ (2012) study of terrorism investigations and case outcomes found that the use of confidential informants dropped significantly after 9/11. In particular, he found that 58% of court cases used confidential informants before 9/11, but only 20% of cases did so after 9/11 (2012). In addition, the average number of confidential informants used per case decreased from 4 informants pre 9/11 to 1.2 informants post 9/11. The average level of assistance provided by confidential informants also decreased pre and post 9/11 from 2.71 to 2.00 (Shields, 2012) .
Interestingly, only one out of 25 terrorism investigations examined by Shields (2012) relied on an undercover agent in the three years following 9/11. Shields did note, however, that in cases filed prior to 9/11 there was an average of 2.67 undercover agents used in cases where at least one undercover agent was used (2012). Shields noted that these findings could significantly change when the remaining post 9/11 cases were coded and analyzed, but that the findings do suggest that there was a significant shift in the pursuit of suspected terrorists by the government after 9/11 (2012).
The proportion of cases with confidential informants pre-9/11 was 69 out of 120 event-linked cases. Post 9/11 there were only 5 out of 25 event-linked cases that used a confidential informant. Pre 9/11, 36 out of 118 event-linked cases included an undercover agent. For post 9/11 cases, Shields found only one event-linked case with an undercover agent out of a total of 25 cases.
A 2011 study by Greenberg provides another look at the use of confidential informants in terrorism cases. Using all federal court cases that the Department of Justice labels as terror-related that were inspired by jihadist ideas, this study examined the use of informants from 2001-2009. Greenberg found that since 9/11, 41% of terrorism cases have involved confidential informants (2011). Similar to Shields findings, Greenberg found that from 2002 to 2003, the same time that Attorney General Ashcroft’s guidelines took effect, that there was a consistent decrease in the use of informants (2011). From 2003 to 2007, Greenberg found the proportion of terrorism cases involving an informant increased from 10% to 70%. That number decreased to just fewer than 30% from 2007-2008. The year following implementation of the Mukasey Guidelines (2008) the percentage increased back up to 70%. Finally, the 2009-2011 timeframe reveals that the percentage of cases involving informants went back to the 2001 level of 10%.
More recently, Hewitt (2014) examined law-enforcement activities in 20 terrorism cases since 1968 and 38 cases of terrorism prevention since 9/11. The twenty completed acts of terrorism all resulted in at least one death. Hewitt used a seven-fold classification system of police actions to analyze the most successful tactics in identifying and apprehending perpetrators in these twenty cases (2014). These factors included: crime scene, witness, routine policing, informers, surveillance, tips from the public, and rewards. Of the twenty completed terrorism cases examined by Hewitt (2014), ten were perpetrated by organized groups and ten cases were perpetrated by unaffiliated individuals, whether lone wolves or members of autonomous cells. Informants accounted for the second highest total of significant factors in identifying and apprehending terrorists in organized groups and the fourth highest total in identifying and apprehending terrorism by unaffiliated individuals (Hewitt, 2014). In total, informers had the second highest percentage (45%) of significant factors in identifying and apprehending terrorists for both organized and unaffiliated cases.
For the 38 terrorism plots thwarted after 9/11, Hewitt used a six-fold classification system of factors involved in the detection of plots (2014). These factors included: routine policing, rewards, tips from the public, informants, surveillance, and undercover agents. The top two factors involved in detecting plots were undercover agents and informants, respectively. Together these two factors equaled 81% of the factors involved in detecting plots (Hewitt, 2014). Overall, Hewitt (2014) found that organized terrorist groups were most vulnerable to informers and surveillance and the most successful terrorist preventions involved undercover agents and informants.
As previous research shows, knowledge about the use of undercover agent’s and confidential informants is scarce. While several studies have analyzed the use of confidential informant’s, the level of assistance provided by them is unknown after 2004. Additionally, the impact of the Mukasey Guidelines on the use of confidential informants is unknown. Regarding undercover agent’s, there is a gap in knowledge in their overall use post 2004.
Success vs. Prevention
Since 9/11, the intelligence community has strived to refine ways to expose and prevent domestic terrorist schemes before they occur (Strom, 2010). The aforementioned Attorney General Guidelines and USA PATRIOT Act were measures taken by the government to assist the intelligence community in the prevention of terrorism after the 9/11 attacks. Since the goal of using undercover agents and confidential informants is to ultimately prevent terrorism incidents from occurring, it is important to examine completed and prevented terrorism cases.
Strom’s 2010 study examined 86 prevented and completed terrorist plots against the U.S. from 1999 to 2009 to determine the activities and the kind of information that either led to or could have led to their discovery. Results from this study found that over 80% of thwarted terrorist schemes were uncovered by law enforcement or general public observations. Nearly one in five schemes were prevented ‘accidentally’ through investigations of outwardly unassociated crimes, underscoring the need to understand when regular crimes might be associated with terrorism (Strom, 2010). Approximately 40% of the foiled plots were the result of tips from confidential informants and the public. Overall, Strom found that investigating leads, along with information sharing among agencies, led to the prevention of the majority of terrorist schemes in his study. All 86 cases contained executed or planned acts of violence in which the defendants intended to cause casualties or catastrophic damage to critical infrastructure (Strom, 2010).
In 2011, Dahl examined 176 terrorist plots from 1987-2010 against American targets that were prevented or otherwise failed. The most significant finding from this study was that human intelligence collected from informants and tips received from members of the public were the most successful counterterrorism tools for breaking up domestic plots (Dahl, 2011). Of the 89 domestic cases that were prevented, 66 involved either undercover agents or informants. Dahl found that most plots were foiled because officials had precise, tactical-level intelligence, often from human sources, on what the plotters were up to (2011).
As the aforementioned literature indicates, empirical studies regarding undercover agents and confidential informants are scarce. What is known from available literature about these forms of human intelligence is that policy changes resulting from the events of September 11 may have significantly altered the use of undercover agents and confidential informants. Shields’ early investigations into the use of undercover agents and confidential informants in the years following 9/11 show a decline in the use of human intelligence. Greenberg’s examination of informant’s post 9/11 revealed similar findings to Shields’, highlighting a decrease in use of informants right after September 11th. Additionally, Greenberg found an increase in informants towards the end of the decade after 9/11. The usefulness of human intelligence was shown in Strom and Dahl’s studies. Both Strom (2010) and Dahl (2011) found that a large percentage of prevented terrorism incidents resulted from cases involving either an undercover agent or confidential informant.
Since empirical research on human intelligence is scarce, there are several shortcomings of prior literature on this topic. First, previous research is limited to descriptive research up to 2004 on undercover agents (Shields, 2012) and to 2011 on confidential informants (Greenberg, 2011). Additionally, while Hewitt (2014) examined human intelligence in preventing terrorism plots, there is a dearth in analyses on the nature of terrorism case outcomes in cases involving human intelligence.
My contributions to this literature on undercover agents and confidential informants as well as thwarted terrorism plots will come from expanding the time frame to analyze cases involving human intelligence to 2014, a full decade beyond Shields’ work. Additionally, I will examine the category of terrorism, the intended target type, the convictions rates, length of prison sentences, and amount of weapons, drugs, and ammunition seized in cases involving undercover agents and/or confidential informants.
Theoretical Framework
Paul Wilkinson once quipped that, ‘Fighting terrorism is like being a goalkeeper. You can make a hundred brilliant saves but the only shot that people remember is the one that gets past you.’ The American intelligence community has prevented hundreds of terrorism plots, however, the attack on September 11th remains the infamous event that failed to be thwarted. After 9/11, the FBI was criticized for having deficiencies and other weaknesses that prevented it from seamlessly sharing intelligence across organizational units (Zegart, 2007). While the application of social theory plays an integral role in advancing our understanding of the way the world works, terrorism scholars have thus far focused mostly on issues of defining terrorism (Boyns & Ballard, 2004). Nonetheless, a few studies have applied criminological theory to causes of terrorism (Clare & Newman, 2007; 2009), and the criminal justice response to terrorism (Smith & Damphousse, 1998; Shields, 2012). Two related theories that will be used to explain possible changes in the intelligence community and government responses to 9/11 are John Hagan’s structural-contextual theory and Samuel Walker’s net-widening theory.
Structural Contextual Theory
A key premise of structural-contextual theory is that the criminal justice system is comprised of components that operate autonomously of one another (Hagan, 1989), or as a ‘loosely coupled’ system. Under normal circumstances components of the justice system compete for resources and pursue different goals. However, Hagan suggests that when political power is directed towards particular types of crime, the justice system tightens through proactive techniques (Hagan, 1989). The justice system may also direct political power towards specific types of crime in what Smith & Damphousse refer to as a ‘proactive political environment (Smith & Damphousse, 1998: 71). A proactive political environment is described as, ‘contexts where the surrounding political environment has mandated departures from normal criminal justice operations’ (Hagan, 1989: 130). Thus, the criminal justice system moves from a reactive to a proactive system ‘targeting the prosecution of a particular form of crime and criminal’ (Hagan, 1989: 130). Proactive political environments ultimately lead to the process of net-widening. Both the proactive political environment and net-widening occur after critical and high-profile events such as the Oklahoma City Bombing or the September 11th attacks. Samuel Walker has furthered a net-widening perspective to explain the implementation of ‘get tough’ laws and other responses to crime. The general idea behind these laws is that getting tougher on criminals will increase public safety by reducing crime victimization (Walker, 1998). Walker has argued, however, that get tough responses to crime also ‘widen the net’ of those who might fall under the authority of the criminal justice system. In other words, get tough policies and practices target lower level offenders who would normally not be arrested and adjudicated under usual circumstances (Walker, 1998).
Organizational Resistance to Change
The events of 9/11 brought numerous policy changes to the American intelligence community. Among these changes were the Ashcroft Guidelines, USA Patriot Act and Mukasey Guidelines. Most notably were the 2002 Attorney General Ashcroft Guidelines that focused on changes to aid the FBI in its then new goal to prevent terrorism. Overall, the Ashcroft Guidelines expanded the FBI’s authority to investigate terrorism, as Attorney General Ashcroft believed the goals of the FBI were the driving force behind any new powers given to the agency (Ashcroft, 2002). The dramatic, overnight shift of the FBI from a reactive to a proactive agency was the highlight of the Ashcroft Guidelines.
Despite these significant policy changes, there may be reasons to expect that the FBI would revert back to proven methods it’s used before the implementation of the Ashcroft Guidelines. This process can be explained by what Lipsky (1980) refers to as Street-level bureaucracy, which recognizes that street level actors, such as police, are active participants in policymaking and make choices based on experience and limited resources that may run counter to stated policy objectives. Likewise, Maynard-Moody (2003) found that street level bureaucrats routinely engage in activities that ‘rub against’ policies and rules because they believe them to be ineffective.
The Department of Homeland Security has spent millions of dollars towards developing intelligence-led policing agencies (Jackson & Brown, 2007). Intelligence-led policing is an information-organizing process that allows law enforcement agencies to better understand their crime issues and account for available resources that will aid the decision making process as to what tactic or strategy will prevent crime, i.e. terrorism (Ratcliffe & Guidetti, 2008). Nonetheless, twenty-first-century technology and advanced analysis programs cannot substitute the significance of human relationships (Taylor & Russell, 2012). Information collected by law enforcement through informants, undercover operatives and contacts with general citizens are significant components in aiding law enforcement agencies in fighting terrorism (Taylor & Russell, 2012).
If Lipsky and Maynard-Moody are correct, then we might expect to see SAC’s and FBI field offices faithfully adhere to the policy changes implemented by the Ashcroft and Mukasey Guidelines while the justice system is in a state of heightened coupling, but then revert back to tried and proven methods when the justice system begins to return to normal. In other words, if the AG Guidelines initially force the FBI to engage in more proactive policing techniques at first, sidelining the use of human intelligence, street-level bureaucracy within the FBI will eventually revert back to developing human intelligence leads.
Applying Theory to Terrorism
Regarding terrorism, Smith and Damphousse (1988: 73) suggest, ‘[w’hen a criminal act is officially designated by the polity as an act of terrorism, that designation sets in motion proactive law enforcement and prosecutorial techniques.’ After a major event like 9/11, there is increased scrutiny from the media and public policy officials on members of the criminal justice system (Damphousse & Shields, 2007). Structural contextual theory would suggest that the events of 9/11 brought the criminal justice system from a group of loosely coupled components to a tightened group of components. These components may have previously worked independently of one another, but after 9/11, pressure from the war on terror would force agencies to work together towards the goal of terrorism prevention.
With regard to terrorism policy after 9/11, Attorney General Ashcroft changed the way the FBI handled terrorism investigations. Prior to 9/11, the FBI used confidential informants and undercover agents to penetrate and take down terrorist groups from the inside. After 9/11, Ashcroft redirected domestic antiterrorism policy to intercept and disrupt terrorist organizations before their members could launch attacks (Shields et al, 2009). This resulted in the casting of a wider net by law enforcement to arrest terrorists on less serious crimes such as documentation and financial fraud, which before 9/11 was never a priority. The results of this proactive environment and net-widening were demonstrated with the swiftly implemented policy changes to the Attorney General Guidelines in 2002. Nonetheless, street-level bureaucrats typically participate in activities that ‘rub against’ rules and policies because they believe them to be unsuccessful (Maynard-Moody, 2003). Thus, I expect that FBI agents would abide by the Ashcroft Guidelines when the justice system is tightly coupled, but then default back to successful tactics and strategies of preventing terrorism through the use of human intelligence when the justice system returns to a more loosely coupled state. Since the Mukasey Guidelines did not change the FBI’s investigative mandate and only bolstered investigative powers, it may be expected that the FBI’s tactics of using human intelligence to prevent terrorism will increase after the 2008 guidelines implementation. Beginning in 2004 with the Intelligence Reform and Terrorism Prevention Act of 2004, which called for increased emphasis on human intelligence gathering, and continuing with the FBI’s push to expand the use of confidential informants in 2007 (FBI, 2007), I expect the proportion of confidential informants to increase in the middle of the decade, despite diminished impact of tightened coupling as fear decreases, and despite the adoption of the Mukasey guidelines, which bolstered the FBI’s investigative power, but did not change its nature.
V. Research Questions
Research Question 1
The way the FBI conducted terrorism investigations was dramatically altered due to post 9/11 policy changes. The fear from 9/11 created a tightly coupled justice system and forced law enforcement personnel to engage in net-widening. Additionally, changes where made regarding the use of one of the FBI’s most successful counterterrorism tools, the use of human intelligence. This purpose of this study is to examine whether policy changes post 9/11 influenced the FBI’s use of human intelligence and if so, what changes has that had on terrorism cases.
1. How did the Ashcroft and Mukasey Attorney General Guidelines change the use of undercover agents and confidential informants?
The Ashcroft Guidelines changed the FBI from a reactive agency to a proactive agency. Previously, the FBI’s strategy was to penetrate and take down the leaders of terrorist groups, whereas the 2002 Guidelines shifted the focus to targeting subordinate group members. The FBI accomplished this new goal by arresting and prosecuting terrorists earlier and more often to prevent attacks from being planned or attempted (Shields, 2012). In doing so, ostensibly, the FBI would have less time to infiltrate groups and gain human intelligence. By the time the 2008 Mukasey Guidelines came into effect, fear from 9/11 that caused the early intervention and arrest mandates should have subsided, and as noted above, new policy initiatives had been put in place focusing more attention on human intelligence, returning infiltration strategies to pre-9/11 levels. To test whether this is the case using structural contextual theory, I developed the following hypotheses:
H1= The proportion of cases by year that used undercover agents and/or confidential informants will be lower in the Ashcroft Guidelines era (2002-2007) than in the pre-Ashcroft Guidelines era (1980-2001) or Mukasey era (2008-present).
H2= The average number of undercover agents and/or confidential informants used per case will be lower in the Ashcroft Guidelines era (2002-2007) than in the pre-Ashcroft Guidelines era (1980-2002) or Mukasey era (2008-present).
If tightened coupling is not a permanent effect, as suggested by Damphousse and Shields (2007), then Lipsky might be correct. Again, Lipsky maintains that street-level bureaucrats change back to tried and proven methods that may run contrary to policy changes. If that is the case, then we would expect the proportion of cases involving confidential informants and undercover agents to increase in the latter half of the Ashcroft guidelines (2006 and 2007). To examine the effect of street-level bureaucracy, I will test the following hypothesis:
H3= The proportion of cases using undercover agents and/or confidential informants will increase in the latter half of Ashcroft Guidelines era (2006-2007).
Given the fear that gripped America following the attacks on 9/11, which would, in theory, create a proactive political environment, and given the policy shift towards proactive law enforcement, we would expect the government to have less time to develop human intelligence in the post-Ashcroft Guideline era. That should negatively impact the proportion of cases using confidential informants and undercover agents overall. However, as the attacks were perpetrated by Islamic extremists, we might expect the government to devote scarce resources to this particular threat compared to right-wing, environmental, and leftist groups. Therefore, the proportion of Islamic extremist cases using confidential informants and undercover agents should be higher in the post-Ashcroft era than before 9/11 or after implementation of the Mukasey guidelines, we would expect the government to foster more relationships in Islamic Extremist cases than non-Islamic Extremist cases. Therefore, I will test the following hypothesis:
H4= The proportion cases that make use of confidential informants and/or undercover agents will be higher among Islamic Extremist cases than among non-Islamic Extremist cases during the Ashcroft Guidelines era (2002-2007) compared to prior and subsequent eras.
Prior to 9/11, Al-Qaeda and associated movements typically attacked American embassies, warships, and military bases (Bergen et al, 2011). After 9/11 these targets increased security measures, thus forcing Al-Qaeda to attack ‘soft’ economic and business targets that are easier to hit and ever-present (Bergen et al, 2011). By the time the 2008 Mukasey Guidelines came into effect, fear for the security of government and military targets should have subsided whereas fear for the safety of ‘soft’ targets should have increased. Based on this information, I developed the following hypothesis:
H5= The proportion of cases involving undercover agents and/or confidential informants with planned attacks targeting government and military facilities will be lower in the Mukasey era (2008-present) compared with the Ashcroft Guidelines era (2002-2007) and the pre-Ashcroft Guidelines era (1980-2001), while planned attacks against economic and business targets will be higher in the Mukasey era (2008-present) compared with the Ashcroft Guidelines era (2002-2007) and the pre-Ashcroft Guidelines era (1980-2001).
The Ashcroft Guidelines caused the FBI to aggressively arrest and prosecute terrorists at a quicker rate than had ever been done. This essentially gave the FBI less time to use undercover agents and/or confidential informants to infiltrate groups. This also likely allowed for undercover agents and/or confidential informants to have less time to collect evidence against terrorists due to quick prosecution demands. By the time the 2008 Mukasey Guidelines came into effect, the impact of tightened coupling might have subsided and street-level bureaucracy may have altered policy, giving undercover agents and confidential informants more time to collect evidence. To test this reasoning, I created the following hypothesis:
H6= The average level of assistance provided by undercover agents and/or confidential informants will be lower in the Ashcroft Guidelines era (2002-2007) than in the pre-Ashcroft Guidelines era (1980-2001) or Mukasey era (2008-present).
The Ashcroft Guidelines mandated that the FBI intervene early and prosecute aggressively. This mandate gave the FBI less time to infiltrate terrorist groups with human intelligence. Moreover, in cases where the FBI had time to infiltrate a group, they likely had less time to collect evidence and set up stings in order to maximize prosecution chances of a conviction with numerous charges from defendants being caught with drugs or weapons. By the time the 2008 Mukasey Guidelines came into effect, the impact of tightened coupling might have subsided and street-level bureaucracy may have altered policy. With loosened time restrictions, undercover agents and confidential informants should have been able to collect evidence and set up sting operations that maximized the charges prosecutors can use against defendants who might be in possession of illegal drugs and weapons at the time of arrest. Based on this information, I developed the following hypothesis:
H7= The average number of weapons, drugs, and ammunition seized in cases that used undercover agents and/or confidential informants will be lower in the Ashcroft Guidelines era (2002-2007) than in the pre-Ashcroft Guidelines era (1980-2001) or Mukasey era (2008-present).
Research Question 2
Assuming the FBI and the criminal justice system followed Attorney General John Ashcroft’s guidelines to intervene early and prosecute aggressively, cases that used undercover agents and confidential informants may have been handled differently than cases not involving human intelligence. To analyze this issue, I created the following research question.
2. What impact, if any, have the changes in use of undercover agents and confidential informants had on case processing and outcomes?
The FBI’s change from a proactive to a reactive agency affected both the goals and the abilities of the FBI. Most significant was the early intervention strategy to prevent terrorism. To accomplish this, law enforcement agents had to arrest terrorists quicker and prosecutors had to convict terrorists sooner than before the Ashcroft Guidelines were put in place. Quicker arrest rates are likely to result in less serious crimes prosecuted as criminals will be caught in the early stages of terroristic activities. The prosecution of less serious crimes should be reflected through lower count severities. With less serious charges to prove a defendant was guilty beyond a reasonable doubt, prosecutors should be able to more easily convict defendants of less serious crimes. To examine this, I created the following hypothesis:
H8= Cases with undercover agents and/or confidential informants will result in higher conviction rates than cases without undercover agents and/or confidential informants and conviction rates will be higher in the Ashcroft Guidelines era (2002-2007) than in the pre-Ashcroft Guidelines era (1980-2001) or Mukasey era (2008-present).
The Ashcroft Guidelines called for the early arrests and prosecutions of terrorists. To make this happen, the FBI had to arrest terrorists on less serious crimes such as documentation and financial fraud. The prosecution of less serious crimes such as these should result in less serious sentences. Given the expectation of shorter prison sentences in cases post-Ashcroft Guidelines, however, the use of undercover agents and confidential informants typically involves more serious situations where the agent or informer is able to gather evidence against the defendant(s). Therefore, it is likely that cases involving undercover agents and confidential informants will feature a greater amount of evidence and will result in longer prison sentences than cases without human intelligence. Based on this reasoning, I developed the following hypothesis:
H9= Cases with undercover agents and/or confidential informants will result in longer prison sentences than cases without undercover agents and/or confidential informants and cases during the Ashcroft Guidelines era (2002-2007) will have shorter prison sentences than cases in the pre-Ashcroft Guidelines era (1980-2001) or Mukasey era (2008-present).
The ultimate purpose of undercover agents and confidential informants is to aid in the prevention of terrorism, whether that is through gathering information or collecting evidence. Since agents and informants put their lives at risk with these duties, and numerous resources including time and money are invested towards their efforts, the government wouldn’t use human intelligence if there were not a high chance of being successful. Moreover, undercover agents and confidential informants are embedded in terrorist groups where they gain knowledge of future plots, just as any other member would. This knowledge should allow the infiltrator to alert authorities before any successful plots are attempted. To analyze this reasoning, the following hypothesis was created:
H10= Cases with undercover agents and/or confidential informants will result in a higher percentage of prevented incidents than cases without undercover agents and/or confidential informants.
VI. Methods
Data
This study uses data from the American Terrorism Study (ATS), housed in the Fulbright College at the University of Arkansas (Smith & Damphousse, 2000; Smith, 2001). The ATS was created to collect information specifically on American terrorism in order to form a database that could be empirically tested to enrich criminologists’ and policy makers’ understanding of terrorism. The Federal Bureau of Investigation’s Terrorist Research and Analytical Center provided the name of persons indicted under the FBI’s counterterrorism program dating back to 1980. Using these, and subsequent lists of cases, the ATS is comprised of federal court cases upon which at least one person was investigated under a terrorism investigation by the Federal Bureau of Investigation and later indicted in federal court. Thus, the primary source of the American Terrorism Study’s data is federal court documents. Numerous ATS government funded grant projects have examined a variety of terrorism issues including geospatial and pre-incident indicators of terrorist activities, prosecutorial and defense strategies in terrorism trials, and geographic concentrations of violent extremism and terrorism in the United States. Data from the ATS are housed in an Oracle relational database. While the ATS currently contains 443 court cases, this study focuses on the 156 cases that have been infiltrated by undercover agents and/or confidential informants.
Sorting Variables
To conduct the majority of my analyses, I created two sorting variables from which I separated the data into different temporal samples. The first sorting variable I called Time Period. Time Period is a categorical variable that separates all ATS cases into three groups. The first group is composed of cases indicted before the implementation of the Ashcroft Attorney General Guidelines in 2002. The second group contains cases that resulted in indictment between 2002 and 2007, the time period in which the Ashcroft Guidelines were in effect. The third sample includes cases resulting in indictment between 2008 and 2014, when the Mukasey Guidelines superseded the Ashcroft guidelines. The second sorting variable I created is called Ashcroft Time Period. Ashcroft Time Period measures the proportion of cases using confidential informants and undercover agents throughout the Ashcroft era based on two samples. The first sample represents the earlier Ashcroft Guidelines era (2002-2005) and the second sample represents the later Ashcroft Guidelines era (2006-2007).
Undercover Operative Variables
As a central focus of this study is on undercover agents and confidential informants, I have several variables related to undercover operatives. First, to measure the year a case involving undercover agents and confidential informants occurred, I used the variable Infiltrated. Infiltrated is a dichotomous variable that measures whether cases used an undercover agent and/or a confidential informant. I coded cases that used an undercover agent and/or a confidential informant as 1, while all other cases were coded as 0. Second, I measured undercover agents and confidential informants using the variables Undercover Agent and Informant, respectively. Undercover Agent is a dichotomous variable that measures whether a government agent infiltrated the group. Informant is a dichotomous variable that measures whether the government had a confidential informant who was a member of the group or closely associated with the group. I coded cases involving an undercover agent as 1, while all other cases were coded as 0. I coded cases in which a confidential informant was used as 1, and all other cases were coded as 0. Third, I used the variables Informant Number and Undercover Number to measure how many confidential informants and undercover agents were involved in each case, respectively. Informant Number is a ratio level variable and measures how many confidential informants the government used. Undercover Number is also a ratio level variable and measures how many undercover agents the government used. I analyzed only cases in which at least one confidential informant and/or undercover agent, and I coded both variables 1-10. Fourth, in order to measure the average level of assistance provided by undercover agents and confidential informants, I used the variables Undercover Assistance and Informant Assistance. Undercover Assistance measures the level of assistance provided by a government agent. It is an ordinal level variable (provided some information, but no recording or testimony (1), provided recorded conversations (2), provided sworn testimony (3), provided both sworn testimony and recorded conversations (4)). Informant Assistance is also an ordinal level variable and is coded the same way as Undercover Assistance.
Case Processing and Outcome Variables
I have three variables (Convicted, Prison Sentence, Prevented Incident) that measure how court cases were processed and what their outcomes were. To measure conviction rates I recoded the ATS variable case result, which assesses the overall case result, into a dichotomous variable that I called Convicted. In this new variable, cases were coded as 1 where a conviction of any type occurred and all other instances were coded as 0. To measure prison sentences I created the variable Prison Sentence. Prison Sentence is a ratio level variable that measures the sentence length in months of prison that a defendant receives. This variable was recoded to give the death penalty and life sentences numerical values. I ran a frequency distribution to determine the longest sentence in months of defendants in my sample, which was 2880 months. I then added 1 month to that sentence for the death penalty value and life sentence value to give each of those the values of 2881 and 2882, respectively. In order to assess the prevention of incidents I used the variable Prevented Incident. Prevented Incident is a dichotomous variable that measures whether an attack was prevented or not as result of human intervention. I coded cases that contained no prevented incidents as 0 and cases that contained a prevented incident as 1.
Other Variables
In relation to my outcome variable Convicted, I have two control variables that were used. These variables are Count Severity and Prosecution Strategy. Count Severity and Prosecution Strategy were used to determine the conviction rate in infiltrated cases across the three Attorney General Guidelines eras. Count Severity is an interval level variable that measures the severity of every count a defendant is charged with. It is coded as 1-29 on a scale that increases in severity. Prosecution Strategy is an ordinal level variable that measures the prosecution strategy used against a defendant. There are three prosecution strategy categories: conventional criminality, political innuendo and explicit politicality, coded as 1, 2 and 3 respectively. Conventional Criminality involves cases where the defendant is not linked to a terrorist group or a terrorist act and is charged with conventional criminal charges (Shields, 2012). Political Innuendo involves cases where the defendant is linked to a terrorist group or terrorist act and is charged with conventional crimes. The last prosecution strategy, explicit politicality- involves cases where the defendants motive for committing a crime is questioned and where the defendant is outright linked to a terrorist group (Shields, 2012).
I also created a variable called Category. This variable measures the category of terrorism and is a nominal level variable, (Environmental (1), Far-left (2), Far-right (3), Islamic Extremist (4)).
Finally, I created the variable Intended Target. Intended Target is the primary target that a group or individual intended to attack, however, the target was not hit. Intended Target is a nominal level variable, and was recoded into a categorical variable (financial (1), government (2), military (3), business (4), private property (5), transportation (6), and other (7)).
Analyses
For my first hypothesis I used a cross-tabulation test to determine the proportion of cases by year that used undercover agents and/or confidential informants across the three temporal samples. I used a one-way ANOVA with the categorization variable Time Period to determine the average number of confidential informants and undercover agents that were used across the three temporal segments for my second hypothesis. For my third hypothesis I used a cross-tabulation test to find the proportion of cases by year that used undercover operatives in the early half of the Ashcroft Guideline era (2002-2005) and the latter half of the Ashcroft Guideline era (2006-2007). On my fourth hypothesis I used a cross-tabulation test to measure the proportion of cases that used undercover agents and confidential informants across different categories of terrorism. My fifth hypothesis was analyzed using a cross-tabulation test to find the proportion of different intended targets across the three time periods in cases infiltrated by undercover operatives. I analyzed my sixth hypothesis using a one-way ANOVA test to determine the level of assistance provided by undercover agents and confidential informants across the three time periods. I was not able to examine my seventh hypothesis measuring the amount of firearms, ammunition and drugs seized in infiltrated cases due to an absence of cases (n=3). I analyzed my eighth hypothesis using three different analyses. First, I ran an independent samples t-test to determine the conviction rate in infiltrated cases vs. non-infiltrated cases. Second, I ran a one-way ANOVA test to determine the conviction rates in infiltrated cases across the three time periods. Third, I ran a logistic regression test to predict the conviction rate in infiltrated cases controlling for count severity and prosecution strategy. For my ninth hypothesis, I ran two statistical tests. First, I ran an independent samples t-test to find the average prison sentence length in months in infiltrated cases vs. non-infiltrated cases. Second, I ran a one-way ANOVA test to find the average prison sentence length in months in infiltrated cases across the three time periods. For my final hypothesis I ran a cross-tabulation to measure the proportion of cases with prevented incidents in infiltrated cases vs. non-infiltrated cases, and limited the sample to only those cases with an identified target. I chose to limit the sample because many of the cases in the post-9/11 era were not linked to a planned act of terrorism (immigration fraud and financing schemes), and their inclusion would skew the results.
VII. Results
Analysis and Findings
The findings for this study are presented below and organized by the two research questions and the ten corresponding hypotheses. Table 1 represents descriptive statistics for the study. Frequency distributions from Table 1 show that the American Terrorism Study database contains 443 cases. I divided those cases into eras corresponding to the different Attorney General Guidelines. This resulted in a sample of 397 cases, as it was not known if an undercover operative was used in 46 cases of the original sample. The Pre-Ashcroft Era contained 150 cases, the Ashcroft Era contained 172 cases, and the Mukasey Era contained 75 cases. Furthermore, in the sample of 397 Attorney General Guidelines Era cases, 156 cases (39.3%) contained a confidential informant, an undercover agent or both. The Pre-Ashcroft era comprised the majority of cases infiltrated by operatives (65%) followed by the Ashcroft (24%) and Mukasey (23%) eras, respectively.
Table 1: Descriptive Statistics
Variables N (Percent)
American Terrorism Study Cases 443
Variables (Continued)
Attorney General Guidelines Eras

Pre-Ashcroft
Ashcroft
Mukasey N (Percent)

397

150 (38%)
172 (43%)
75 (19%)

Cases Infiltrated By Operatives

Pre-Ashcroft
Ashcroft
Mukasey 156

97 (65%)
41 (24%)
17 (23%)

Cases with Informants

Pre-Ashcroft
Ashcroft
Mukasey
127

79 (62%)
36 (28%)
12 (9%)

Cases with Undercover Agents

Pre-Ashcroft
Ashcroft
Mukasey 54

38 (70%)
6 (11.%)
10 (18%)
Ashcroft Era Divided

2002-2005
2006-2007 172

144 (84%)
28 (16%)

Category of Terrorism in Infiltrated Cases

Environmental
Far-Left
Far-Right
Islamic
88

29 (33%)
1 (1%)
33 (38%)
25 (28%)
Variables (Continued)
Intended Targets in Infiltrated Cases

Government
Military
Business
Private Property
Transportation
Other
N (Percent)
96

42 (44%)
5 (5%)
16 (17%)
4 (4%)
9 (9%)
20 (21%)
Conviction Rate: Infiltrated vs. Non-Infiltrated Cases

Non-Infiltrated Cases
Infiltrated Cases 840 Indictees

349 (87%)
491 (76%)
Prison Sentence in Months: Infiltrated vs. Non-Infiltrated Cases

Non-Infiltrated Cases
Infiltrated Cases 364

222 (73.71 months)
142 (336.11 months)

Research Question 1
How did the Ashcroft & Mukasey Attorney General Guidelines change the use of undercover agents and confidential informants?
Hypothesis 1
Hypothesis one was created to test the use of undercover operatives throughout different time periods. Specifically, I hypothesized that the proportion of cases that used undercover agents and/or confidential informants would be lower in the Ashcroft Guidelines era (2002-2007) than in the pre-Ashcroft Guidelines era (1980-2001) or the Mukasey era (2008-2014). Findings from testing this hypothesis are presented in Table 2, and partially support my hypothesis. Results indicate that the proportion of cases with undercover operatives was significantly lower in the Ashcroft era than the pre-Ashcroft era, however, the proportion of cases with undercover operatives in the Ashcroft era was nearly identical to the proportion of cases with undercover operatives in the Mukasey era. The pre-Ashcroft era had 64.7% of cases containing operatives that dramatically decreased to 23.7% of cases in the Ashcroft era (p <.05) and then remained stable at 22.7% of cases in the Mukasey era (p <.05). Table 2: Crosstabulation of AGG Eras in Infiltrated Cases Era No Operative Operative Used % Of Cases with Operative Pre-Ashcroft 53 97 64.7 % Ashcroft 131 41 23.8 % Mukasey 58 17 22.7 % Chi-Square = 66.54 Df = 2 Sig. = .000 Hypothesis 2 Table 3 represents ANOVA results comparing the average number of confidential informants and undercover agents used in cases throughout the pre-Ashcroft, Ashcroft and Mukasey eras. I hypothesized that the average number of undercover operatives would be the lowest in the Ashcroft Guidelines era. Counter to my hypothesis, the average number of informants was the highest during the Ashcroft era. The average number of undercover agents during the Ashcroft era was in-between the other two eras. The findings did not support my hypothesis, but the findings were statistically significant. Table 3: ANOVA of Avg # of Informants and Undercover Agents Per Case Era N Mean Sig. Df Informant Pre-Ashcroft 79 2.02 .000 2 Ashcroft 36 4.05 Mukasey 12 1.08 Undercover Pre-Ashcroft 38 1.94 .010 2 Ashcroft 6 1.16 Mukasey 10 1.00 Hypothesis 3 In hypothesis three I examined the Ashcroft era in depth by dividing it into two time periods, 2002-2005 and 2006-2007. I hypothesized that the latter half of the Ashcroft Guidelines era would feature a greater proportion of cases that used undercover operatives. Results supported the hypothesis and were statistically significant. Using cross-tabulations, I found that the proportion of cases using operatives dramatically increased from the 2002-2005 time frame at 16% to 64.3% during the 2006-2007 time frame. Table 4: Crosstabulation of Ashcroft Era Divided in Infiltrated Cases Era No Operative Operative Used % Of Cases with Operative 2002-2005 121 23 16.0 % 2006-2007 10 18 64.3 % Chi-Square = 30.319 Df = 1 Sig. = .000 Hypothesis 4 Table 5 represents crosstabulation results examining different categories of terrorism with the level of infiltration experienced across different categories. I hypothesized that Islamic Extremist cases would have the highest proportion of cases involving undercover operatives compared to cases in other categories of terrorism. The findings did not support hypothesis four. Statistically, Far-Left terrorism contained the highest proportion of cases that used undercover operatives, at 100%, however, there was only one Far-Left terrorism case in the sample. Islamic Extremist cases contained a nearly identical proportion of cases using undercover operatives as Far-Right cases at 24.0% and 24.2%, respectively. Notably, 62.1% of Environmental cases made use of undercover operatives. Table 5: Crosstabulation of Category of Terrorism in Infiltrated Cases Category No Operative Operative Used % Of Cases with Operative Environmental 11 18 62.1 % Far-Left 0 1 100 % Far-Right 25 8 24.2 % Islamic Extremist 19 6 24.0 % Chi-Square = 13.554 Df = 3 Sig. = .004 Hypothesis 5 In hypothesis five I examined the intended target type in infiltrated cases across the three Attorney General Guidelines Eras. I hypothesized that the proportion of cases involving undercover operatives (both confidential informants and undercover agents) with the planned attacks targeting government and military facilities would be the lowest during the Mukasey era, while planned attacks against economic and business targets would be the highest during the Mukasey era. The results did not support the hypothesis. The Mukasey era did have the lowest proportion of infiltrated cases with the intended target being government or military facilities as well as the highest proportion of cases with business targets; however, the findings were not statistically significant at the .05 level. Table 6: Crosstabulation of Intended Targets in Infiltrated Cases Era Government Military Business Pre-Ashcroft 46.8 % 5.2 % 15.6 % Ashcroft 33.3 % 8.3 % 16.7 % Mukasey 28.6 % 0 % 28.6 % Chi-Square = 9.069 Df = 10 Sig. = .526 Hypothesis 6 Table 7 represents ANOVA results comparing the average level of assistance provided by undercover agents across the three Attorney General Guidelines Eras. In addition, I performed a second ANOVA to measure the level of assistance provided by confidential informants. I hypothesized that the average level of assistance provided by either undercover agents or confidential informants would be the lowest during the Ashcroft Guidelines. The findings partially supported the hypothesis. I found that informants provided the highest level of assistance during the Ashcroft era and that this result was non-significant. In support of the hypothesis, I found that undercover agents had the lowest level of assistance during the Ashcroft era and that this was statistically significant. Table 7: ANOVA of Informant and Undercover Assistance Era N Mean Sig. Df Informant Pre-Ashcroft 74 2.39 .520 2 Ashcroft 36 2.55 Mukasey 9 2.11 Undercover Pre-Ashcroft 38 3.52 .012 2 Ashcroft 5 2.40 Mukasey 7 2.57 Hypothesis 7 In hypothesis seven I planned to examine the average number of weapons, drugs and ammunition seized in cases that used undercover operatives and compare it to cases that did not contain undercover operatives in which these same items were seized. However, there were not enough cases (N = 3) to run statistical tests on these variables. Research Question 2 What impact, if any, have the changes in use of undercover agents and confidential informants had on case processing and outcomes? Hypothesis 8 I examined conviction rates using three analytical analyses. I hypothesized that cases with undercover operatives would result in higher conviction rates than cases without undercover operatives and that conviction rates among the three Attorney General Guidelines eras would be highest in the Ashcroft era. I first used an independent samples t-test to find the conviction rate percentage in both types of cases (See Table 8a). The t-test results did not support the hypothesis and showed that the conviction rate in cases with undercover operatives was 75%, whereas the conviction rate in cases without undercover operatives was 86% (p < .05). Next, I ran a one-way ANOVA test to compare the conviction rate in infiltrated cases across the three Attorney General Guidelines eras (See Table 8b). The ANOVA results, however, were not statistically significant between Attorney General Guidelines eras (p >.05). I then ran logistic regression controlling for count severity and prosecution strategy to measure the significance of undercover operatives on conviction rates (See Table 8c). The Hosmer/Lemeshow Chi-Square indicated that there were no fitness problems, but the percentage of explained variance did not increase beyond the initial model. Count severity was not significant when controlling for prosecution strategy and infiltration, and infiltration was not significant when controlling for count severity and prosecution strategy, though it was very close (.059) with a negative impact on the likelihood of conviction. The only significant variable was prosecution strategy; consistent with prior research, the more politicized the case became the greater the log odds that there would not be a conviction. The Nagelkerke value shows that only 6.5% of the variation in conviction outcomes in my model can be explained by count severity, prosecution strategy and infiltration.
Table 8a: Independent Samples T-Test of Conviction Rates in Infiltrated Cases
N Mean Std. Deviation
No Operative 349 .86 .338
Operative Used 491 .75 .430

t = 4.073 Df = 838 Sig. = .000

Table 8b: ANOVA of Conviction Rates in Infiltrated Cases
Era N Mean Sig. Df
Pre-Ashcroft 168 .78 .070 2
Ashcroft 19 .63
Mukasey 10 1.00

Table 8c: Logistic Regression of Conviction Rates in Infiltrated Cases
B Df Sig. Exp(B)
Count Severity -.004 1 .620 .996
Pros Strategy 1 2 .001
Pros Strategy 2 -.681 1 .018 .506
Pros Strategy 3 -1.009 1 .000 .364
Infiltrate -.406 1 .059 .666
Cox & Snell R2 = .041
Nagelkerke = .065
Hosmer/Lemeshow Chi-Square = 9.656, Df = 8, Sig. = .290

Hypothesis 9
Table 9a represents independent samples t-test results of the average prison sentence length in months in cases with undercover operatives compared to cases without undercover operatives. I hypothesized that cases with undercover operatives would result in longer prison sentences than cases without undercover operatives (See Table 9a). The results support the hypothesis and are statistically significant, but I report these findings with a caveat. In order to properly measure sentence lengths, prior research indicates that I should control for overall case severity and prior criminal history. Ideally, I would have done so, but very few cases in the post-9/11 are coded for these variables. I found that cases with undercover operatives have an average prison sentence of 336.11 months whereas cases that do not use undercover operatives have an average prison sentence of 73.71 months. Additionally, I ran a one-way ANOVA test to compare the prison sentence lengths in infiltrated cases across the three Attorney General Guidelines eras (See Table 9b). I hypothesized that cases during the Ashcroft era would have the shortest prison sentences among the three eras. The ANOVA results were statistically significant between Attorney General Guidelines eras with defendants in the Ashcroft era having the lowest average prison sentence length in months with an average of 72.52 months (p <.05). Table 9a: Independent Samples T-Test of Prison Sentences in Infiltrated Cases N Mean Cases No Operative 302 73.71 Cases With Operative 433 336.11 t = -3.844 Df = 733 Sig. = .000 Table 9b: ANOVA of Prison Sentences in Infiltrated Cases Era N Mean Sig. Df Pre-Ashcroft 464 265.01 .000 2 Ashcroft 265 72.52 Mukasey 90 115.71 Hypothesis 10 In hypothesis ten I examined prevented incidents in both cases with and without undercover operatives. I hypothesized that cases with undercover operatives would result in a higher percentage of prevented incidents than cases without undercover operatives. The findings support the hypothesis and were statistically significant (See Table 10). I found that 23% of cases without undercover operatives contained a prevented incident whereas 27.6% of cases with undercover operatives contained a prevented incident. Table 10: Crosstabulation of Prevented Incidents and Infiltrated Cases No Operative Operative Used Prevented Incidents 56 43 % of Cases Prevented 23 % 27.6 % Chi-Square = 44.626 Df = 2 Sig. = .000 VIII. Discussion The events of 9/11 were the impetus needed to change America's counterterrorism efforts. As a response to that infamous day, some of the major changes the U.S. government made were policy related, most significantly the Ashcroft Attorney General Guidelines. The FBI's investigative scope was expanded through the Ashcroft Guidelines, ultimately changing the mission of the FBI from a reactive to a proactive agency and making counterterrorism its top priority. The purpose of this study was to explore the relationship of policy changes on the use of undercover operatives. Specifically, this study sought to determine whether the Attorney General Guidelines that were implemented after 9/11 changed the use of undercover agents and confidential informants in federal terrorism cases. Additionally, this study had the goal of examining what impact, if any, the changes in use of undercover operatives had on case processing and outcomes. I created two research questions with ten hypotheses stemming from prior empirical research on undercover operatives in terrorism. The results of my analyses were mixed in providing support for my hypotheses. Research Question 1 My first research question was designed to explore the influence of the 2002 and 2008 Attorney General Guidelines in terrorism cases that used undercover agents and confidential informants. I hypothesized that the Ashcroft Guidelines era would have the lowest proportion of cases that used undercover operatives compared to the pre-Ashcroft Guidelines era and the Mukasey era. I found a statistically significant decrease in the proportion of cases that used undercover operatives from the pre-Ashcroft era to the Ashcroft era (64.7% to 23.8%). However, the Ashcroft era contained a slightly higher proportion of cases that used undercover operatives than the Mukasey era (23.8% to 22.7%). These findings suggest that the FBI's counterterrorism goals via the Ashcroft Guidelines to arrest and prosecute terrorists earlier and more often were successful. These results do not suggest, however, that by the time the Mukasey Guidelines came into effect, fear, driven by the 9/11 attacks that caused the early intervention and arrest mandates, had subsided returning infiltration strategies and levels to what they were pre 9/11. While the justice system may still be in a state of tightened coupling causing a low proportion of Mukasey era cases to contain undercover operatives, the findings may be reflective of an absence of cases that await coding in the American Terrorism Study database. I ran a frequency distribution of the number of infiltrated cases by year. There were 9 infiltrated cases in 2010 and 6 infiltrated cases in 2011. These high numbers of infiltrated cases suggest that the Mukasey era may actually contain a higher proportion of cases that used undercover operatives than the Ashcroft era as there are numerous cases from the last five years still being coded and entered into the ATS database. Regarding the average number of undercover agents and confidential informants used per case, I hypothesized that the Ashcroft era would have the lowest average number of undercover operatives used among the three eras. The results indicated cases in the Aschroft era had neither the lowest average number of informants nor agents. In fact, Aschroft era cases had the highest number of confidential informants as compared to the Mukasey era. After closer examination of the cases during this time period, I found 14 related cases where informants were used in 2006, and each of those cases featured eight confidential informants'the same informants. These 14 cases are collectively known as the 'ELF Family,' and involved numerous defendants in multiple cases, some of whom turned states evidence and testified against their fellow Earth Liberation Front (ELF) members to mitigate their sentences. Thus, the ELF Family case study likely skews the number of informants per case, which makes the analysis of undercover agents a more accurate depiction of the average number of operatives used during the guidelines. Nonetheless, the Ashcroft era contained an average of 1.16 agents whereas the Mukasey era contained an average of 1.00 agents indicating a minimal change between time periods, at least among those cases currently coded. While the Ashcroft era did not contain the lowest average number of undercover agents, it was significantly less than the average number of agents used in the Pre-Ashcroft era (1.16 to 1.94 agents). This decrease from nearly 2 agents to 1 agent per case is likely the result of early arrest and prosecution strategies brought on by the justice system being tightly coupled. Turning to the category of terrorism, I hypothesized that Islamic Extremist cases would have the highest proportion of cases with undercover operatives compared to other categories of terrorism during the Ashcroft era. Not counting the Far-Left typology, as there was only one valid case, Environmental cases had the highest proportion, at 62%, whereas Far-Right and Islamic Extremist cases were proportionate to one another at 24.2% and 24.0%, respectively. It could be argued that the ELF Family cases skew results for this hypothesis as well. However, even by excluding the 14 ELF Family cases, the proportion of Environmental cases that used undercover operatives is still the highest at 26.6%. A possible explanation for why Environmental terrorism had the highest proportion of cases involving undercover operatives may be a reflection of the organizational structure and tactics of the Earth Liberation Front. ELF operates under a leaderless resistance model and the Ashcroft Guidelines changed the FBI's tactics from targeting group leaders prior to 9/11 to targeting subordinate group members after 9/11. With little to no established groups and/or group members to maintain loyalty to, it is likely easier for defendant's to turn states evidence on any other environmental terrorists they may know to receive a better sentence in court. Likewise, it is likely also easier for the FBI to establish informants to single individuals to gain their trust rather than attempting to gain the trust of an entire terrorist group. Thus, the FBI's post 9/11 counterterrorism strategy directly aligned with how environmental terrorists operated. Structural contextual theory would suggest that the FBI would have established undercover operatives in a greater proportion of Islamic extremist cases since Islamic extremists conducted the 9/11 attacks. However, the Ashcroft Guidelines demand for early intervention in terrorism cases aligned with the tactics of environmental terrorists more so than Islamic extremists, represented by the higher proportion of environmental terrorism cases. Concerning the intended target type in infiltrated cases, I hypothesized that the Mukasey era would have the lowest proportion of infiltrated cases that featured government and military facility targets, and that the Mukasey era cases would have the highest proportion of business targets. While there were no statistically significant differences between the three eras (sig. = .526) regarding actual targets, the Mukasey era did have the lowest proportion of infiltrated cases where the target in the case was a government or military facility. The Mukasey era cases also contained the highest proportion of cases where the target in the case was a business target. These findings, while not significant, coincide with prior literature on terrorist group tactics post 9/11. As Bergen et al (2011) found, terrorists have shifted from attacking hard targets such as American embassies, warships, and military bases, which are better defended after 9/11, to attacking soft targets such as economic and business organizations. While structural contextual theory does not explain terrorist behaviors, it may suggest why terrorist tactics have changed after 9/11. The fear from 9/11, with the Pentagon being attacked and potentially the nation's capitol being targeted, left the justice system believing that it was only a matter of time before the next government or military facility was attacked. Thus, when the justice system was tightly coupled, it worked together to increase security on both government and military facilities while security measures for economic and business organizations remain relatively unchanged. When I examined the average level of assistance provided by confidential informants and undercover agents, I hypothesized that the Ashcroft era would have the lowest level of assistance for both types of undercover operatives. Regarding informants, there was no statistical difference between eras. Undercover agents, however, did provide the lowest average level of assistance during the Ashcroft era, supporting my hypothesis, and providing evidence of structural contextual theory in effect after 9/11. This finding represents another example of how fear from the 9/11 attacks caused the criminal justice system to become tightly coupled. Undercover agents provided less assistance during the Ashcroft era representing a change in FBI investigations, which stemmed from the Ashcroft Guidelines demanding arrests earlier and more often, thus giving undercover agents less time to collect evidence. Research Question 2 My second research question was created to measure what impact, if any, the changes in the use of undercover operatives had on case processing and outcomes. Looking at conviction rates, I hypothesized that cases with undercover operatives would have higher conviction rates than cases without operatives, and that the Ashcroft Guidelines era would have the highest rate of convictions among the three eras. None of my findings supported this hypothesis. Conviction rates were 11% higher in cases that did not use undercover operatives. Furthermore, results were insignificant (p >.05) concerning the conviction rates between Attorney General Guidelines eras. Additionally, when I examined conviction rates controlling for count severity and prosecution strategy I found that undercover operatives were not a significant factor. As previously stated, these findings are consistent with previous research in that the prosecution strategy is the most significant factor in determining conviction rates. Logically, as Shields (2012) suggests, if prosecutors have more evidence, in this case, via confidential informants and undercover agents, they may be more likely to pursue more politicized charges. This would explain the negative relationship I found between level of assistance provided and conviction rates. While the ‘early and often’ arrest and prosecution strategies were brought on by the justice system being in a state of tightened coupling from fear from 9/11, it appears that quicker and more frequent arrests and prosecutions did not significantly impact conviction rates, at least not in my sample of infiltrated cases. Thus, while the justice system being in a state of tightened coupling did not influence conviction rates in infiltrated cases, as prior research suggests, tightened coupling was in effect as the overall amount of plea rates and conviction rates increased after 9/11.
Turning to prison sentence lengths, I hypothesized that cases with undercover operatives would have longer prison sentences than cases without operatives and that among the three Attorney General Guidelines eras the Ashcroft era would have the shortest prison sentences. As mentioned in my findings section, I would have preferred to run a multivariate analysis controlling for count severity and prior criminal history, but the cases are not coded for those two variables after 9/11, due in large part to changes in the way courts prepare and report judgments. With this limitation in mind, the first part of my hypothesis was supported. Cases with undercover operatives had a statistically significant greater prison sentence lengths than cases without operatives (336 months compared to 73 months, p <.05). This finding supports the notion that undercover operations involve more serious situations in which the undercover agent or informant is able to gather a greater amount of evidence to use against the defendant, ultimately resulting in a longer prison sentence. While this finding was significant, it does come with the caveat of needing to be tested with multivariate analysis. The ATS database contains a large proportion of cases that do not involve intended or actual targets. Instead, these cases involve individuals who have been prosecuted for what are referred to as 'paper crimes', crimes involving immigration violations and financial fraud. These paper crime cases contain less severe charges that result in shorter prison sentences. Cases such as these in the database show the necessity to code cases for overall case severity to run further statistical analyses. My last finding regarding prison sentences showed that the difference in prison sentence lengths between eras was significant with the Ashcroft Era containing the shortest average prison sentence length in months. Thus, it appears that structural contextual theory and the net-widening theory are helpful in predicting prison sentence lengths. Of course while my models derived from these theories are statistically significant, the theories themselves may indirectly influence statistical findings through the type of cases that are brought through the justice system. If the criminal justice system is in a state of tightened coupling with a wider net casted on crimes being investigated, then cases with undercover operatives will only continue to have greater prison sentence lengths than cases without operatives. My last hypothesis analyzed case outcomes in terms of whether an incident was prevented in that case. Specifically, I hypothesized that cases with undercover operatives would result in a higher percentage of prevented incidents than cases without undercover operatives. The results were statistically significant and supported my hypothesis. There was a greater percentage of prevented incidents in cases with undercover operatives (27.6%) than there were in cases without operatives (23%). As presented, this finding shows that the time and resources put into undercover operations are successful in the FBI's highest priority of countering terrorism. However, this may not be the entire story. When coding prevented incidents, I used a conservative measure that only included cases that had no actual targets. The sample of prevented incidents was comprised of cases that only had intended targets that were never physically attacked, as apposed to cases that contained both intended targets and actual targets. By including cases with both intended and actual targets (mixed cases), the difference may be greater than my results indicate because mixed cases mask the true number of prevented attacks. Nonetheless, cases with undercover operatives result in a higher percentage of prevented incidents than cases without operatives. This finding, along with my previous findings that prison sentences in cases with operatives are longer than prison sentences in cases without operatives, suggests that undercover operations are not only more successful in preventing terrorism incidents, but they are more successful in putting defendants behind bars for a longer period of time. IX. Conclusion & Future Research The purpose of this study was to add to the limited body of empirical research that has been conducted on undercover operatives in federal terrorism cases. This study contributed to prior literature by focusing on the impact of policy changes on the use of undercover operatives. Specifically, the 2002 Ashcroft and 2008 Mukasey Attorney General Guidelines served as a defining line between when undercover operatives were used and how policy changes influenced undercover operations throughout different temporal periods. My first research question asked, 'How did the Ashcroft and Mukasey Guidelines change the use of undercover agents and confidential informants'? I found, as expected, that the proportion of terrorism cases that used undercover operatives had significantly dropped (over 40%) since September 11th, 2001. Although the proportion of cases using undercover operatives did not increase in the Mukasey era as expected, this finding may change once more cases are coded and entered into the American Terrorism Study database. A frequency distribution of infiltrated cases by year showed that the last two years to have infiltrated cases in the database were 2010 and 2011 and that these years had a higher than average number of infiltrated cases. I also found that the average number of undercover operatives used and their average level of assistance have decreased since 9/11. These findings further highlight the impact of the Attorney General Guidelines changes to terrorism investigations, suggesting that after 9/11 time was of the essence. The last notable point from research question one involves the category of terrorism cases that were infiltrated. Unexpectedly, environmental terrorism cases contained the highest proportion of cases with undercover operatives. Given the fear in America after 9/11, I expected that because Islamic Extremists conducted the 9/11 attacks that the proportion of Islamic Extremist cases that were infiltrated would be higher than cases in other categories of terrorism after 9/11. Logically, however, my findings make sense in relation to the changes brought by the Ashcroft Guidelines. The Ashcroft Guidelines changed the FBI's counterterrorism strategy from targeting group leaders prior to 9/11 to targeting low-level group members after 9/11. Since environmental terrorists usually operate under a leaderless resistance, they were all susceptible to being targeted by the FBI. My second research question asked, 'What impact, if any, have the changes in use of undercover agents and confidential informants had on case processing and outcomes'? I found that undercover operatives do not increase the odds of the case resulting in conviction and that consistent with previous research, prosecution strategy is still the most significant factor in predicting convictions. I did, however, find significant results when examining prison sentence lengths. While I was not able to run the most appropriate statistical test for this model because of an under-coded variable, I found that cases that were infiltrated by undercover operatives had a significantly higher sentence length than cases that did not use operatives. This finding is relative to my last result of analyzing prevented incidents. I found that cases with undercover operatives result in a higher percentage of prevented incidents than cases without operatives. These last two findings suggest that undercover operations warrant the resources put into them, if not justifying more resources into these types of FBI investigations. Future research on undercover agents and confidential informants in federal terrorism cases has numerous possible avenues to explore. Expanding this study's findings, the addition of more cases into the ATS database would allow for more advanced statistical analyses into examining prison sentence lengths among infiltrated and non-infiltrated cases. Additionally, adding cases to the ATS database would allow greater insight into the Mukasey era and its influence on undercover operatives as the most recent year with infiltrated cases in the database is 2011. Another avenue for future research would be to examine the length of investigations in cases with undercover operatives versus cases without undercover operatives. By analyzing how long the government takes from the first day of the investigation to the day of arrest, the FBI may be able to better allocate time and resources in a more appropriate manner by balancing budget and safety issues. A final possibility for future research would be to examine undercover operatives in federal terrorism cases through a different theoretical approach than John Hagan's Structural Contextual Theory or Samuel Walker's Net-Widening. An example would be to use Ronald Clarke's Situational Crime Prevention theory or Rachel Boba's EVIL DONE approach to terrorism. These approaches would do away with policy influences and instead focus more on the intended and actual targets of terrorism cases and how, if at all, the targets influence the way the FBI investigates cases and uses undercover operatives. X. References: Attorney General's Guidelines for Domestic Security Investigations (Smith Guidelines): Hearings Before the Subcommittee on Security and Terrorism of the Senate Committee on the Judiciary, 98th Cong. 1 (1983) Boyns, D., & Ballard, J. (2004). Developing A Sociological Theory For The Empirical Understanding Of Terrorism. The American Sociologist, 5-25. Clarke, R. V., & Newman, G. R. (2007). Situational Crime Prevention and the Control of Terrorism. NATO SECURITY THROUGH SCIENCE SERIES E HUMAN AND SOCIETAL DYNAMICS, 21, 285. Clarke, R. V., & Newman, G. R. (2009). 'Reducing the Opportunities for Terrorism; Applying the principles of situational crime prevention.' Pp. 86-105 in Terrorism and Torture: An Interdisciplinary Perspective. Edited by Werner G. K. Stritzke. Cambridge University Press. Dahl, E. (2011). The Plots That Failed: Intelligence Lessons Learned From Unsuccessful Terrorist Attacks Against The United States. Studies in Conflict & Terrorism, 621-648. Damphousse, K., & Shields, C. (2007). "The Morning After, Assessing the Effect of Major Terrorism Events of Prosecution Strategies and Outcomes." Journal of Contemporary Criminal Justice 23(2): 154-174. Department of Justice. (2002). Remarks of Attorney General John Ashcroft. Retrieved from the United States Department of Justice website: http://www.justice.gov/archive/ag/speeches/2002/53002agpreparedremarks.htm

Federal Bureau of Investigation. (2007). FY 2008 Authorization and Budget Request to Congress. Retrieved from https://www.documentcloud.org/documents/238034-33-fbi-se-2.html

Greenberg, K., & Quatrone, S. (2010). Terrorism Trial Report Card: September 11, 2001-September 11, 2011. The Center on Law and Security, New York University School of Law

Hagan, J. (1989). ‘Why is there so little criminal justice theory? Neglected macro-and micro-level links between organization and power’. Journal of Research in Crime and Delinquency 26: 116-135

Hewitt, C. (2014). Law Enforcement Tactics and Their Effectiveness in Dealing With American Terrorism: Organizations, Autonomous Cells, and Lone Wolves. Terrorism and Political Violence, 58-68.

http://www.fbi.gov/albuquerque/about-us/what-we-investigate (2010, November 5). Retrieved September 23, 2014.

Jackson, A. L., & Brown, M. (2007). Ensuring efficiency, interagency cooperation, and protection of civil liberties: Shifting from a traditional model of policing to an intelligence’led policing (ILP) paradigm. Criminal Justice Studies,20(2), 111-129.

Jones, A. (2009). 2008 FBI Guidelines: Contradiction of Original Purpose, The.BU Pub. Int. LJ, 19, 137.

Jones-Brown, D., & Shane, J. M. (2011). An Exploratory Study of the Use of Confidential Informants in New Jersey

Kaplan, E. (2012). ‘OR Forum ‘ Intelligence Operations Research: The 2010 Philip McCord Morse Lecture’. Operations Research. 60(6): 1297-1309.

Lipsky, M. (1980). Street-level bureaucracy: Dilemmas of the individual in public services. New York: Russell Sage Foundation.

Mahan, S., Mahan, S. S. G., & Griset, P. P. L. (2012). Terrorism in perspective. Sage publications.

Moody, S., & Musheno, M. (2003). Cops, teachers, counselors: Stories from the front lines of public service. Ann Arbor: University of Michigan Press.

Oversight Hearing on Counterterrorism: Hearing Before the Senate Committee on the Judiciary, 107th Cong. 187 (2003) (Responses by Robert S. Mueller, III, Director, Federal Bureau of Investigation, to Committee’s written questions).

Office of the Inspector General. (2005). A Review of the FBI’s Handing of Intelligence Information Related to the September 11 Attacks. Retrieved from United States Department of Justice website: http://www.justice.gov/oig/special/s0606/final.pdf

Ratcliffe, J. H., & Guidetti, R. (2008). State police investigative structure and the adoption of intelligence-led policing. Policing: An International Journal of Police Strategies & Management, 31(1), 109-128.

Schalch, M. E. (2003). Intelligence Reform: The” Phoenix” of 9/11?. Army War Coll Carlisle Barracks Pa.

Shields, C., Damphousse, K., & Smith, B. (2009). ‘How 9/11 Changed the Prosecution of Terrorism.’ Pp. 125-144 in The Impact of 9/11 and the New Legal Landscape: The Day that Changed Everything?, edited by M. J. Morgan. New York, NY: Palgrave MacMillan.

Shields, C. (2012). American terrorism trials prosecutorial and defense strategies. El Paso: LFB Scholarly Pub. LLC.

Smith, B. (1994). Terrorism in America: Pipe Bombs and Pipe Dreams. Albany, NY: State University of New York Press.
Smith, B., & Damphousse, K. (1998). Terrorism, Politics, And Punishment: A Test Of Structural-Contextual Theory And The “liberation Hypothesis”*. Criminology, 67-92.

Strom, K. (2010). Building on Clues: Examining Successes and Failures in Detecting U.S. Terrorist Plots, 1999-2009. North Carolina: Institute for Homeland Security Solutions.

Taylor, R. W., & Russell, A. L. (2012). The failure of police ‘fusion’ centers and the concept of a national intelligence sharing plan. Police Practice and Research, 13(2), 184-200.

The Attorney General’s guidelines on general crimes, racketeering enterprise and terrorism enterprise investigations. (2002). Washington, D.C.: Office of Legal Policy.

The Attorney General’s guidelines regarding the use of confidential informants. (2002). Washington, D.C.: U.S. Dept. of Justice.

The Attorney General’s guidelines on Federal Bureau of Investigation undercover operations. (2002). Washington, D.C.: Office of Legal Policy.

The Attorney General’s guidelines for domestic FBI operations. (2008). Washington, D.C.: Dept. of Justice.

Uniting and Strengthening America by Providing Appropriate Tools Required to Intercept and Obstruct Terrorism (USA PATRIOT Act) Act of 2001, Pub. L. No. 107 ’56. 115 Stat. 272 (2001).

Walker, S. (1998). Sense and nonsense about crime and drugs: A policy guide (4th ed.). Belmont, CA: West/Wadsworth Pub.