Friday, May 1, 2015

Addressing the Challenges and Limitations of Utilizing Data to Measure Serial Homicide

The following was presented to the Homicide Research Working Group on June 11, 2015:

Enzo Yaksic a and Michael Aamodt b

a Serial Homicide Expertise and Information Sharing Collaborative. Boston, MA

b Department of Psychology, Radford University. Radford, VA

According to Beasley (2004), the study of serial murder should be objective, standardized and include as many subjects as possible. Although researchers have been collecting serial homicide data for decades, it was done disparately. Hinch and Hepburn (1998) categorize past attempts as “dealing with narrowly defined acts and the most sensational cases.” Lack of reliable data has contributed to the slowing of research on serial crime (Petee and Jarvis, 2000) and is identified as a key obstacle for the future (Dowden, 2005). Leyton (1996) and Skrapec (2001) have advised scholars to abandon egos, academic ‘turf wars’, and self-serving agendas in search of a greater understanding of the phenomenon. In an update on the ‘Evolution of Serial Murder as a Social Phenomenon in American Society’, Hickey (2014) acutely summarizes the plight endured by those delving into this work, acknowledging that “being a researcher in this area has required getting past misinformation and sensationalism, finding others who share similar research interests, and dredging up reliable data.”

Overcoming these barriers is an ongoing endeavor, undertaken by the efforts of those involved in the Serial Homicide Expertise and Information Sharing Collaborative (SHEISC), a network of one-hundred academic and law enforcement professionals with the goal of coming to a greater understanding of the serial murderer (Boyne, 2014). In 2015, after a combined total of thirty-seven years amassing data on serial homicide offenders, Michael Aamodt and Enzo Yaksic began recruiting scientists to analyze the output in unique ways. Relationships were built with Clare Allely, author of Neurodevelopmental and Psychosocial Risk Factors in Serial Killers and Mass Murderers (2014), Ronald Hinch, author of Researching Serial Murder: Methodological and Definitional Problems (1998), authors Mikhail Simkin and Vwani Roychowdhury of Stochastic Modeling of a Serial Killer (2014), Marissa Harrison, author of Female Serial Killers in the US: Means, Motives, and Makings and Virginia Beard, author of Death-Related Crime: Applying Bryant's Conceptual Paradigm of Thanatological Crime to Serial Homicide. To encourage further empirical studies, Aamodt and Yaksic provided an opportunity to bolster future work in this arena and increase sample sizes by supplying each with the entire Radford/FGCU-SHEISC Serial Killer Database (Aamodt, 2014). 

In the past, serial homicide databases were sparsely populated as academicians collecting information made it available only to their own research team (Ramsland, 2010). To protect self-interests, most operated in ‘information silos’, hoarding data and secreting source material. Procurement of these files now comes at exorbitant financial cost (Canter, 2011; Godwin, 2012) as records were often kept in paper format (Egger, 2014; Geberth, 2010) or within now defunct or corrupted computer systems (Waters, 2013). As a result, some materials no longer exist (Leyton, 2013). Byproducts of these actions led to the creation of different definitions, the use of assorted sources, the surveillance of non-comparative variables and the inability to validate information.

The variance of emphasis on the nuances of serial murder has led to the use of multiple definitions (Morton and McNamara, 2005) as criminologists have historically been restrictive in their interpretations of what constitutes a serial murderer (Ostrosky-Solis et al., 2008). Wright et al. (2009) state that much of the scholarly research has been hindered by the definition of each type of killer, differing on requirements such as the number of murders, the types of motivations and the temporal aspect (Morton and Hilts, 2008). Starting from a more inclusive definition may hold advantages, according to Osborne and Salfati (2015), but, by excluding certain types of killers, analysts have reduced the pool of research subjects (Cluff et al., 1997). The Radford/FGCU-SHEISC Serial Killer Database (Aamodt, 2014) employs the broad two victim definition provided by the Federal Bureau of Investigation (FBI) as it encompasses the full array of serial murderers by not referencing underlying motivation, behavior or psychological characteristics (Brantley and Kosky, 2005; Hargrove, 2010). Radford data include financial, professional contract, revenge, gang, organized crime, robbery/homicide and witness elimination offenders among the more ‘traditional’ serial sexual murderers. Research efforts have been damaged in recent years due to a hyperfocus on the serial sexual killer as the terms sexual homicide and serial murder have become synonymous (Petee and Jarvis, 2000).

Criminologists protest the inclusion of gang and professional murderers in serial offender databases since they occur alongside functional (Ferguson et al., 2003) violence and have rational motives (Jenkins, 2002). This viewpoint reinforces that offenders must utilize violence unconventionally to qualify as a serial murderer. Many, including the killers themselves (Valencia et al., 2013), believe that self-motivated serial murderers do not belong juxtaposed to ‘hitmen’ since they are paid to kill and do not choose their own targets (Hickey, 2013). Conversely, these murderers display the intent to kill anyone, at any time (Vronsky, 2013) and are seemingly gratified by appeasing their greed. Lester and White (2014), as well as agents of the FBI (Morton and McNamara, 2005), acknowledged that organized crime, contract and drug/gang killings are serial murders that can be motivated by revenge, loyalty or profit.

Evident is the risk of diluting a sample and compromising the integrity of data by conflating various types of multiple murderers together (Giannangelo, 2013). However, excluding killers that cross between subcategories hinders the advancement of serial homicide research (Clemente, 2013). Hickey (2015) notes that broad definitions may not be meaningful, but excluding financial motives and those who kill acquaintances is based on speculation rather than on any empirical foundation. Aamodt and Yaksic’s application of the Modified Delphi Technique (Custer et al., 1999) on an instituted panel of experts demonstrate a wide range of opinions on the topic (Aamodt and Yaksic, 2015a). Although these disagreements highlight the difficulty of maintaining collaborative relationships, the group posited that the broad categorical term “Multiple Event Murderer” could be used as a means to study the multitude of subsets of offenders either together or separately, defined henceforth as: Any person(s) that cause(s) the death of another through illegal means more than once in at least two locations. 

The benefit of cataloging multiple murderers using one overarching classification scheme is that it allows individual researchers to adapt as needed and avoid the shortfalls of the existing definitions (Adjorlolo and Chan, 2014). Rather than engaging in the process of retrofitting offenders into categories to suit specific study parameters, whole subsets of offenders can be included or excluded from analysis, depending on the author’s pre-defined criteria. The detriment of this approach arises when psychologically dissimilar populations are compared between studies (Aamodt and Yaksic, 2015a). Researchers must be cognizant that serial murderers cannot be fit into one single behavioral profile (Hickey, 2015) and, based upon the uniqueness and variances of human behavior, be aware of the limitations in typing the serial murderer (Morton and McNamara, 2005).

Fox, Levin (2011) and Yaksic’s meticulous tracking of adjudicated serial murder series over several decades revealed that the phenomenon is in decline (Beam, 2011). Yaksic’s (2013a) reasoning behind the decrease was initially met with trepidation but several of the proposed theses have since been accepted (Fox and Levin, 2014; Hickey, 2015) as probable explanations for the downward trend. It is presumed that advances in technology have made it easier for law enforcement to consider that a serial murderer may be operating in their area (Kaste, 2015). Jurisdictional conflicts have subsided in favor of increased collaboration with the FBI and surrounding departments while the public is consulted for assistance (Backus, 2015a) more frequently. Cell phones, always connected social media accounts and the dawn of the surveillance age add other measures of risk to an offender’s decision to victimize others. The Internet provides would-be offenders the opportunity to placate themselves without exploiting unwilling participants. Greater utilization of the underground sex trade and the likelihood of offenders warehousing abductees contributes to a decreased need to eliminate complaining witnesses with such regularity. Efforts to educate the public about these offenders led to increased awareness that odd behaviors, stalking offenses, paraphilias and violent tendencies toward animals or others in youth are part of a larger group of warning signs. A greater distrust of strangers led to the abolishment of hitchhiking and parents reluctant to allow children to play freely without supervision (Aamodt and Surrette, 2013), diminishing potential victim pools. Harsher punishments and less use of parole ensures that would-be serial murderers are incarcerated for longer. Serial murder is not viewed as the shortcut to celebrity status it once was since news coverage of these events has lessened over the years. Many would-be serial murderers are captured after their first murder due to incompetence, before they have the opportunity to amass larger victim counts (Aamodt and Yaksic, 2015b). Future generations may witness an even greater decrease in victimization as long haul truckers are replaced by automated vehicles (Davies, 2015). The decline in serial homicide calls into question the image of the infallible, successful killer these offenders were once thought to be. Societies’ past ignorance of their means and motives allowed serial murderers freedoms they can no longer enjoy. While the desire to become a serial murderer has not dissipated among offenders (Aamodt and Yaksic, 2015b), these factors may have permanently displaced some offenders, forced others into altering their modus operandi or into early retirement.

Still others may have begun adopting tactics commonly associated with the spree murderer (Aamodt and Yaksic, 2015b). Long after Muhammad and Malvo (Koerner, 2002), the debate about how to typify spree murderers continues due to those that carry out their crimes in ‘spree-like’ frequency but, as agents of the FBI have stated (Morton and Hilts, 2008), the motives and tactics of a serial murderer (Earl, 2013). Researchers have been reluctant to juxtapose these offenders or admit witnessing the convergence of the spree and serial murderer in recent years. It is disputable that spree and serial murderers should continue to be stratified separately since the ‘cooling off period’, used to make distinctions between potentially similar serial offenders, is now characterized as a historical artifact (Douglas et al., 2013). Osborn and Salfati (2015) concluded that spree and serial homicide may not be distinguishable, including in their data eight instances of one day time intervals and one series that lasted only three days.

Nearly a decade ago, Salfati et al. (2006) concluded the only study comparing serial and spree murderers, finding that spree offenders were generally 29 years old, killed on average six victims with firearms, most often in one day at anywhere from two to six locations. Seventy-three percent of these perpetrators were killed before capture. Aamodt and Yaksic (2015b) found that spree offenders in their data were also typically 29 years old, but killed on average three victims with firearms most often over a one week period at three locations. The majority of offenders in the Aamodt and Yaksic (2015b) inquiry were arrested rather than being killed in police shootouts after embarking on a suicide mission. Although spree offenders did share similarities with serial murderers, it is not yet known which psychological traits are mutual or if patterns in precipitating factors can be established. Aamodt and Yaksic (2015b) concur with Salfati et al. (2006) that more data is needed to determine if spree and serial murderers should continue to be stratified separately.

After incorporating spree murderers and instances of urban violence into their serial homicide dataset, Hickey (2015) and Yaksic found evidence of a change in victimization rates. Fewer cases of only female victims are co-occurring with an increase in males being targeted. Fewer strangers and more family members are killed with a slight drop in the number of prostitutes being singled out. There are fewer numbers of victims per offender and fewer cases involving more than one state. Fewer cases of strangulation are recorded alongside more cases involving shooting as the sole method of killing.

Still, Homant and Kennedy (2014) stipulate that a serial murderer’s subsequent killings should be part of a separate sequence of behaviors. Those committed by spree murderers are said to be part of one continuous event, often the byproduct of situational violence (Daniels, 2015). Not only is the sequence distinction less useful if these offenders are viewed as Multiple-Event Murderers (Aamodt and Yaksic, 2015a), but serial murderers can and have been fueled by similar precipitating incidents.

Although Beasley (2004) noted feelings of restlessness and impulsivity among serial murderers, they have been characterized as cold and calculating methodical planners that commit unprovoked, patterned and predatory sexual attacks on strangers. Spree killers are typed as temperamental, impulsive and bombastic with initial murders being haphazard retaliatory reactions directed at acquaintances (Salfati, 2006) whose subsequent murders often help them obtain necessities (Daniels, 2015). Perhaps, however, there are circumstances and environmental factors that may preclude multiple murderers from acting in the manner that they would ideally choose. Some offenders are given the luxury of time between offenses, not because of superior knowledge, talent or skill, but due to variables beyond their control including the degree of witness involvement, varying levels of police pressure and even luck. If a spree-type offender eludes police, they may bide their time, reenter society, and possibly continue killing again in the future, unabated. Similarly, serial murderers can exhibit ‘run and gun’ behaviors by the end of their series (Smith, 2009; Sparacello, 2011). By ignoring the stratification question, researchers overlook the evolving nature of serial homicide and disregard how serial murderers are molded by social conditions, cultural changes and external pressures (Warf and Wardell, 2002).

Early qualitative research efforts were ambitiously spearheaded by agents of the FBI who conferenced with twenty-five serial murderers in the early 1980s in an attempt to understand the circumstances behind their existence by immersing themselves in their life histories (Ressler and Burgess, 1985; Beasley, 2004). Since then, researchers have instituted lore by either intentionally or unwittingly overemphasizing anecdotes gathered from these interviews. Stereotypes emerged over the past forty years, forming due to the scarcity of systematic studies (Arndt et al., 2004; Jenkins, 1994), the desire to monetize the concept and the tendency for some to consciously invest in catering to those that actively embrace the more gruesome aspects of serial murderers and their deeds (Yaksic, 2014a). As such, researchers have spent the greater part of the last decade combating antiquated viewpoints and dismantling myths from past eras (Fox and Levin, 1999; Sterbenz, 2015; Yaksic, 2013b). Consequently, there were very few comparative studies, virtually no biopsychosocial studies, an absence of more sophisticated statistical analysis and a repetitive use of small, nonrandom samples using retrospective data at the beginning of the millennium (Meloy, 2000). Dowden (2005) suggests that researchers refocus their efforts as inquiries made by those working outside the field of serial homicide research are often the most potent. 

Serial murder has been transformed into a profitable construct, often exploited to satisfy our curiosity (McNamara and Morton, 2004; Beasley 2004). Falsehoods are often promulgated by unqualified individuals, self-ascribed forensic criminologists, due to the ease with which information can now be generated and disseminated for financial incentives (Yaksic, 2014a). There are those that imprudently wield criminal profiling as a weapon in a game of good versus evil where the good guys battle and triumph over the bad guys (Mains, 2015a). This infantile worldview may produce good entertainment (Mains, 2015b) but also inevitably induces a false sense of knowing these offenders. As such, serial murderers have been portrayed as prolific evil geniuses, inhuman psychopathic monsters and dysfunctional white male loners, unusual in appearance and incapable of maintaining long term relationships. They are called cunning predators that ceaselessly hunt dozens of strangers to fulfill a desire for bloodlust. They supposedly possess the ability to not only evade the police but expertly engage them through the media and frequent their hangouts to learn about cases, all while yearning to be caught.

Leading research shows that serial murderers do not consistently behave this way (Fox and Levin, 1999; Hickey, 2015). The current authors have each separately imparted their findings on the issue of race and the serial murderer. Yaksic (2006) contentiously stated that every other serial killer since 1995 has been African American. Others (Hickey, 2015) have since validated this proposal, finding that it has held true almost a decade after being posited. Aamodt (2008) found that the commonly held demographic profile correctly matches only eighteen percent of serial murderers. As Beasley (2004) notes, the prevailing thought that serial murderers sadistically kill for sexual gratification, engage in animal torture, are physically or sexually abused as children, become more evidence conscious over time and allow media coverage to alter their criminal intentions must be challenged. Results from Bateman and Salfati (2007) indicate that serial murderers are not consistently performing the same crime scene behaviors throughout their series. Coupled with the findings of Schlesinger et al. (2010), these revelations test the notions that serial murderers consistently take souvenirs or leave signatures, escalate in their violence as they continue killing, improve their methods and change their strategies over their careers. Violent acts performed on a body also should not automatically signal the presence of a serial murderer.

Serial murder is thought to be an affliction that overtakes an offender’s life but they are not hostages to this cycle; they can control their desire to kill. Offenders are not always in prison, the military, at college, or in a mental health facility during time intervals (Morton et al., 2015) and rarely begin to unravel at the end of their series. Many have criminal records that reflect histories mired in anti-social behavior but some do display remorse (Yaksic, 2012) and can be affected by what they do (Levin and Fox, 2008), sometimes coping with their actions by abusing substances. Murders can be committed serially due to subtle factors such as the offender’s perception that a victim lied, cheated, insulted, or hurried them (Quinet, 2011). Victims are often selected not as substitution for someone they know but based on availability, vulnerability, and desirability (Morton et al., 2015). It is also obsolete to classify serial murderers as either organized or disorganized as this investigational dichotomy has not been used in practice during much of the past decade (Morton et al., 2015).

Abiding by tropes can influence the course of serial homicide investigations. A private investigator from Sarasota, Florida used only the physical attributes of eighteen potential victims to connect the disappearances simply because they “look just like sisters”, drawing a comparison to the archaic archetype of Ted Bundy (Gleiter, 2014). Although firearms are the highest proportion of kill method in the Radford data (Aamodt, 2014), an investigator characterized a recently apprehended serial murderer’s use of a firearm in each of four homicides as “rare” (Alexander, 2014). Most still consider as fact the old adage that serial murderers desire a ‘hands-on’ kill and must strangle their victims to feel the life leave their bodies.

To avoid further pitfalls, Beasley (2004) recommends supplementing interviews, like those conducted by Pino (2005) and Reavis (2011), with information from large scale databases in order to obtain a more informed view of the serial murderer. ‘Big Data’ provides many benefits to almost all sectors of society (Davenport, TH., and Dyché, J., 2013; Venton, 2015) but has yet to be utilized to effect this microcosm of the criminal justice system. While researchers have now officially convened in an effort to attain this aim (Boyne, 2014), several limitations have been encountered.

All facets of serial homicide research are fraught with quantification issues due to the use of outmoded data collection methods. No official mechanism exists for capturing instances of serial homicide as the FBI’s Uniform Crime Report (UCR) and Supplementary Homicide Report (SHR) lack a specific entry field. Researchers inclined to count the frequency of events and make determinations about their instance in society cannot track serial homicide easily as there is a great amount of disagreement as to what exactly constitutes a serial murder series. Categorization efforts have been ongoing since the inception of this phenomenon but researchers are no closer to consensus (Ostrosky-Solis et al., 2008) on classification today than they were forty years ago. Attempting to include motive into the definition of serial murder results in some degree of subjectivity (Ferguson et al., 2003) as it must be inferred from observable behavior (Kraemer et al., 2004) and depends on history, context and expectation; constructs that are not only impossible to quantify but unique to each individual. The news media oftentimes does a poor job of accurately labeling a serial offender based on motive, as evident in the recent trials of Aaron Hernandez (Jones, 2014) and Dzhokhar Tsarnaev (Zalkind, 2014). Those regularly deemed spree killers can also erroneously be given the serial killer designation (Winton et al., 2014). For these reasons, the addition of serial homicide offender records to databases can be an unnecessarily subjective process.

The emotional ‘cooling off period’ is a confounding concept that has been used to differentiate between multiple murderers. While experts disagree on a standard ‘cooling off’ length, many continue to utilize this temporal element as part of their exclusion criteria. Osborne and Salfati (2015) note that there is a lack of research to develop an understanding of ‘cooling off’ and that it should be discarded and re-conceptualized into time intervals. Since these intervals could be less than one day, the continued use of the temporal distinction has restricted the utility of the current serial murder definition (Adjorlolo and Chan, 2014). Although serial murderers may return to their usual way of life in the time breaks between homicides, researchers have little understanding of the behaviors that they partake in during these interruptions. It is impossible to discern the degree to which serial murderers remain entrenched in their killing lifestyle. While it may appear that they are taking care of their children or going to the store, they may actually be dedicating effort to planning future crimes, ruminating about past ones or managing the impression made on others (Levin and Fox, 2008). An offender’s dormant period is thought to be psychologically beneficial to them but implies that they engage in violence as part of a stress-relief regimen to resolve a buildup of internal conflict (Corzine, 2014). This viewpoint wrongfully insinuates that serial murder is driven only by aggression and agitation and that offenders work themselves into a frenzy, or ‘heat up’ before each kill (McClellan, 2014). This outlook, as well as the use of the word ‘emotional’ to describe this period (Aamodt and Yaksic 2015b), conflicts with the perspective that serial murder is a predatory crime whose offenders are well-controlled, cold and calculating (Myers, 2014).

Unlike in other fields of inquiry, attempts to collect primary source information are hampered by extraneous factors. Researchers must contend with limited access to subjects due to their incarceration, death or refusal to participate in research studies. When access to subjects is granted, offenders often alter their version of events to better suit their needs or cannot accurately recall how events transpired. Although some serial murderers are consumed by their killing lifestyle (Fox and Levin, 2011), most often lack enough insight into themselves to provide useful and actionable information. Since opportunities for interviews are sparse, statements given by murderers are overblown, generalized and given more weight than they deserve. Building rapport with a subject can be a time consuming process, often lasting months or even years. Regardless, primary source interviews (Pino, 2005; Reavis, 2011) should supplement the information culled from secondary sources so that offenders’ actions are not minimized after being reduced to mere data points in an Excel worksheet. Sorting offenders by only demographic variables can lead to a loss of narrative components and an overreliance on selected anecdata. 

Available data can also be rather difficult to aggregate and collate. Titlow (2014) cataloged some of the hardships the current authors encountered during the process of data transformation while merging each of the databases received through the SHEISC initiative. Although a more real-time, ‘root cause analysis’ of offenders’ methods and motivations would greatly enhance law enforcement’s response to these matters, gathering primary source information on each of the 3,949 serial homicide offenders listed in the Radford/FGCU-SHEISC Serial Killer Database (Aamodt, 2014) would take more than a lifetime, even if they were universally available. While potentially able to amass far larger victim counts than their American counterparts, international killers are woefully understudied due to substandard record keeping and unestablished judicial systems in foreign countries as well as cultures that do not obsess about the criminal mind.

Linking and attributing homicides to one killer can be a challenging process, spanning many years and encompassing the efforts of hundreds of individuals (LePard, 2015; Lohr, 2015). Taskforces are charged with identifying the perpetrator after a large expanse of time since the initial murder occurred (Backus, 2015b). The length of legal proceedings contributes to the dearth of primary source research materials, such as an offender’s journal, as these items are warehoused in evidence lockers and not revealed until a trial (Zapotosky, 2015). As is the case with serial murder suspects Lonnie Franklin (Gerber, 2015) and Felix Vail (Yaksic, 2014b), bringing offenders to trial can take half a decade. Due to the threat of lawsuits from individuals that are wrongly labeled with the ‘serial killer’ tag, researchers often wait until their conviction to add them to datasets. This artificial lag greatly diminishes an offender’s impact on the data because researchers customarily generate statistical output based on the data that becomes available during the new edition cycle of their academic textbooks. Excluded from analysis would be those offenders discovered during the resolution of a cold case (Hoffer, 2015), for example, since they occurred outside of the parameters of the author’s timeframe. The drawback of such a stringent focus on cases surrounding the textbook’s publication period is the inability to conduct any type of systematic comparison of offenders from previous eras to these more recent killers.

Complicating matters are those individuals that have no involvement in a series of murders but claim responsibility for them and those that boast about murders beyond the scope of their series. Each scenario makes adjudicating these crimes difficult. Still, others refuse to accept responsibility for the totality of the killings they did cause. Some individuals listed in The Innocence Project (Scheck and Neufeld, 2015) and The National Registry of Exonerations (2015) were released after the true perpetrator was revealed to be a serial murderer (Coker, 2015). Over the years, a few series were even determined to likely be nonexistent – the Smiley Face killings of college aged men in the Midwest and Northeast (Drake et al., 2010), the canal ‘Pusher’ case in Manchester, England (Slater, 2015) and the Phantom of Heilbronn, supposedly operating in Austria, France and Germany from 1993 to 2009 ().

Due to Institutional Review Board restrictions and confidentiality agreements (Salfati, 2011), the most detailed records of the crime, those of law enforcement, usually are not available to researchers (Kraemer et al., 2004). Also routinely out of reach is comprehensive documentation related to a case, including investigative, autopsy, forensic and evidence analysis reports; crime scene and autopsy photographs, diagrams, sketches, and maps; victimology information; offender background; and any confessions or admissions by the offender (Beasley, 2004). Even if available, these documents cannot account for missed offenses that the offender committed that are unknown to police (Osborne and Salfati, 2015). Information gleaned from law enforcement sources is also processed through several layers of individuals before it ends up as a data point in a database and is therefore exposed to all the biases and distortions that accompany human interpretation of facts. These records can contain misinformation as investigators’ levels of experience with and understanding of such offenses may vary widely (Morton and McNamara, 2005), leading them to only record information fitting their notions of what will be important for the investigation (Bateman and Salfati, 2007). An unavoidable limitation of all serial homicide datasets is that what is not known about victims cannot be known (Quinet, 2011) as they are the primary, and perhaps the sole, witness. Due to the outcome of the crime, he or she is no longer able to provide the necessary evidence (Bateman and Salfati, 2007). 

Because of these reasons, most researchers are forced to collect information from secondary sources, such as true crime books and news media reports (Morton and McNamara, 2005). Reliance on this approach is problematic because the veracity of claims made in true crime books are often unverifiable and reports often contain varying degrees of information pertinent to investigating this phenomenon (Morton and McNamara, 2005), which oftentimes contributes to missing data (Bateman and Salfati, 2007). For example, universally incorporating offenders that commit two homicides is arduous since news outlets infrequently refer to these offenders as serial murderers. Instead, these offenders are muddled with double-murderers (Hoffer, 2015), killers whose two homicides occur in the same incident, at the same time. This oversight has led to several records being overlooked during searches.

Estimating the true prevalence of serial homicide is complicated by the complex nature of the data. Due to the serial murder entertainment industry, superfluous content often overtakes search results during research, leading to an oversampling of the most bizarre cases, at the expense of more mundane and more common offenders (Beasley, 2004). Researchers do not factor unsolved homicides into their analysis since these victims cannot be definitely attributed to the actions of serial murderers. Also, as unapprehended serial murderers are probably hidden within the nations numerous DNA backlogs (Nelson, 2010), some researchers deem it unethical to produce a number that would adequately capture the serial murderer’s occurrence in society. It is also impossible to know the identities of burgeoning serial murderers. Those that maintain the intent to commit a subsequent offense after being apprehended for their first homicide are not included in the data as no academic study has been dedicated to this offender population. The instance of false positives, those that express the desire to kill serially but are not responsible for any homicides, is unknown.

An interdisciplinary team of researchers, journalists and data scientists have been assembled in an effort to address the limitations outlined in this report (Boyne, 2014). To overcome these barriers, we must begin to adapt processes proven useful in other industries. Virtual breakthrough series collaboratives (Zubkoff, 2014), encouraging the free exchange of ideas and network building, inspired the creation of the SHEISC in 2010 which resulted in the capture of serial murder suspect Felix Vail (Yaksic, 2014b). IBM’s Watson supercomputer is primed for use in the criminal justice system (Wyllie, 2011) but its potential impact on serial homicide research and investigations is unknown at this time.

Digital information dashboards, data displays and electronic surveillance tools are used in healthcare systems to track performance on key indicators, allowing for ‘drill down’ capabilities to reveal information hidden amongst lower level data (Chen et al., 2014). These displays would serve the criminal justice system by subsuming data from various sources into one location to provide a semi real-time overview of probable serial murder activity while compiling data for scientists to analyze.

After completing work applying an algorithm on unsolved homicide cases using the statistical technique of cluster analysis (Hargrove et al., 2011), Thomas Hargrove founded the Murder Accountability Project (Hargrove, 2015) to ensure that police increase their reporting of homicide occurrences to the FBI using the UCR and SHR mechanisms. Dallas Drake of the Center for Homicide Research has begun work on the Wyoming Homicide Database Project (Drake, 2015), a pilot to test the feasibility of obtaining information on solved homicides directly from law enforcement institutions. Data scientist Peter Brendt is at work applying natural language processing techniques to large swaths of data in an effort to generate a type of ‘weather map’ detailing probable serial murder activity among listings of unidentified bodies and missing persons (Brendt, 2015). Aamodt and Yaksic continue to populate the Radford/FGCU-SHEISC Serial Killer Database (Aamodt, 2014) with new cases. Relationships should be fostered with local law enforcement crime analysts and statewide fusion centers under a joint understanding that utilizing data to address this longstanding criminal justice issue is of critical importance. Linking these efforts together under one umbrella may have distinct advantages, as yet both unexplored and unexploited.


Aamodt, M. (2008). Reducing Misconceptions and False Beliefs in Police and Criminal Psychology. Criminal Justice and Behavior. 35 (10). Available at: Accessed: January 14, 2015.

Aamodt, M. (2014). Serial killer statistics. Available at: Accessed: March 10, 2015.

Aamodt, M. and Surrette, MA. (2013). Is the Decline in Serial Killing Partially Explained by a Reduction in "Free Range Kids"? Presentation at Society for Police and Criminal Psychology, Ottawa, Canada.

Aamodt, M. and Yaksic, E. (2015a). Results from the application of the Modified Delphi Technique. Author’s Files.

Aamodt, M. and Yaksic, E. (2015b). Serial Murder: Separating Fact from Fiction. Webinar. Hosted by the Justice Clearinghouse. Available at: Accessed April 7, 2015.

Adjorlolo, S., and Chan, HC. (2014). The controversy of defining serial murder: Revisited. Aggression and Violent Behavior. 19:486-491.

Alexander, B. (2014). Atlanta’s Alleged Serial Killer. WXIA‑TV. Available at: Accessed: April 5, 2015.

Allely, C., Minnis, H., Thompson, L., Wilson, P., and Gillberg, C. (2014). Neurodevelopmental and Psychosocial Risk Factors in Serial Killers and Mass Murderers. Aggression and Violent Behavior. 19(3): 288-301. Accessed February 9, 2015.

Arndt, WB., Hietpas, T., Kim, J. (2004). Critical Characteristics of Male Serial Murderers. American Journal of Criminal Justice. 29(1):117-131.

Backus, L. (2015a). More work ahead in New Britain serial killer investigation. The Bristol Press. Available at: Accessed: May 9, 2015.

Backus, L. (2015b). Woman visits spot where remains of sister, 2 other women found. New Britain Herald. Available at: Accessed: April 24, 2015.

Bateman, AL., and Salfati, G. (2007). An examination of behavioral consistency using individual behaviors or groups of behaviors in serial homicide. Behav Sci Law. 25(4):527-44.

Beam, C. (2011). Blood loss: the Decline of the Serial Killer. Slate. Available at: Accessed: December 17, 2014.

Beasley, JO. (2004). Serial Murder in America: Case Studies of Seven Offenders. Behav Sci Law 22: 395–414. Available at: Accessed: March 28, 2015.

Boyne, ES. (2014). Serial Homicide Collaborative Brings Research Data Together. (Fall/Winter 2014). Criminal Justice Update (CJ Update). An Online Newsletter for Criminal Justice Educators. From Routledge and Anderson. 43(1):2. Available at: Accessed: February 5, 2015.

Brantley, AC., and Kosky, RH. (2005). Serial Murder in the Netherlands: A Look at Motivation, Behavior and Characteristics. FBI Law Enforcement Bulletin. 74(1):26-32. Available at: Accessed: April 10, 2015.

Brendt, Peter. (February, 2015). Personal Communication.

Canter, David. (April, 2011). Personal communication.

Chen, Q., Shin, MH., Chan, JA., Sullivan, JL., Borzecki, AM., Shwartz, M., Rivard, PE., Hatoun, J., Rosen, AK. (2014). Partnering With VA Stakeholders to Develop a Comprehensive Patient Safety Data Display: Lessons Learned From the Field. American Journal of Medical Quality. Available at: Accessed: December 1, 2014.

Clemente, J. (2013, June 18). Re: Former hit man John V. Martorano says he’s no serial killer‏. Message posted to Multidisciplinary Collaborative on Sexual Crime and Violence electronic mailing list, archived at

Coker, M. (2015). Eder Giovanni Herrera, Once Held in Slayings of His Mom and Brother, Wins $700K Settlement. OC Weekly. Available at: Accessed: April 30, 2015.

Cluff, J., Hunter, A., Hinch, R. (1997). Feminist Perspectives on Serial Murder. Homicide Studies. 1(3): 291-308.

Corzine, J. (2014, April 15). Re: Cooling-Off Period‏. Message posted to Multidisciplinary Collaborative on Sexual Crime and Violence electronic mailing list, archived at

Custer, RL., Scarcella, JA., Stewart, BR. (1999). The Modified Delphi Technique - A Rotational Modification. Journal of Vocational and Technical Education. 15(2). Available at: Accessed: November 17, 2014.

Daniels, S. (2015). Do We Need a New Category of Multiple Murder? A Lesson in Typing Homicide. Unpublished manuscript.

Davenport, TH., and Dyché, J. (2013) Big Data in Big Companies. SAS Institute. Available at: Accessed: March 24, 2015.

Davies, A. (2015). The World’s First Self-Driving Semi-Truck Hits the Road. Wired. Available at: Accessed: May 7, 2015.

Douglas, J., Burgess, AW., Burgess, AG., Ressler, RK. (2013). Crime Classification Manual: A Standard System for Investigating and Classifying Violent Crime. Hoboken, NJ. Wiley.

Dowden, C. (2005). Research on Multiple Murder: Where Are We in the State of the Art? Journal of Police and Criminal Psychology. 20(2):8-18.

Drake, D., Bebernes, CM., Wheeler, BJ., Stelflug, EZ., Espenson, PD., Van Auken, AM., Olson, KR., Panaretos, AM., and Anderson, BJ. (2010). Drowning the Smiley Face Murder Theory. Research Brief. Center for Homicide Research. Available at: Accessed: October 30, 2014.

Drake, D. (2015). Wyoming Homicide Database Project. Center for Homicide Research. Available at: Accessed: April 24, 2015.

Earl, D. (2013). Nikko Jenkins: ‘I need mental health treatment’. KETV. Available at: Accessed: April 5, 2015.

Egger, Steven. (May, 2014). Personal communication.

Ferguson, CJ., White, DE., Cherry, S., Lorenz, M., and Bhimani, Z. (2003). Defining and classifying serial murder in the context of perpetrator motivation. Journal of Criminal Justice, 31(3):287–292.

Fox, J. and Levin, J. (1999). Serial Murder: Popular Myths and Empirical Realities (From Homicide: A Sourcebook of Social Research, P 165-175. M. Dwayne Smith and Margaret A. Zahn, eds.)

Fox, J., and Levin, J. (2011). Extreme killing: Understanding serial and mass murder. Los Angeles, CA: Sage Publications.

Fox, J., and Levin, J. (2014). A surprising truth about serial killings. Available at: Accessed: December 9, 2015.

Geberth, Vernon J. (October, 2010). Personal communication.

Gerber, M. (2015). Grim Sleeper serial killings: Trial date set after 4 1/2 years of delays. Los Angeles Times. Available at: Accessed: April 24, 2015.

Giannangelo, S. (2013, June 18). Re: Former hit man John V. Martorano says he’s no serial killer‏. Message posted to Multidisciplinary Collaborative on Sexual Crime and Violence electronic mailing list, archived at

Gleiter, D. (2014). Investigator notes similarities between Kortne Stouffer, other missing women; DA doubts there's link. The Patriot-News. Available at: Accessed: April 5, 2015.

Godwin, Maurice. (September, 2012). Personal communication.

Hargrove, T. (2010, April 22). Interview of Mark Hilts, Unit Chief of the Federal Bureau of Investigation’s Behavioral Analysis Unit 2. Scripps Howard News Service. Author’s Files.

Hargrove, T., Bowman, L., Wolf, I., Lucas, E., and Bartz, J. (2011). Murder Mysteries. Available at: Accessed: September 8, 2015.

Hargrove, T. (2015). Murder Accountability Project. Available at: Accessed: April 30, 2015.

Hickey, E. (2013, June 18). Re: Former hit man John V. Martorano says he’s no serial killer‏. Message posted to Multidisciplinary Collaborative on Sexual Crime and Violence electronic mailing list, archived at

Hickey, E. (2014). The Evolution of Serial Murder as a Social Phenomenon in American Society: An Update. Academy of Criminal Justice Sciences. XXXVIIII(3):12-16. Available at: Accessed: January 12, 2015.

Hickey, E. (2015). Serial Murderers and Their Victims. (7th ed.) Belmont, CA: Wadsworth, Cengage Learning.

Germany's Phantom Serial Killer: A DNA Blunder. Time. Available at:,8599,1888126,00.html. Accessed: April 24, 2015.

Hinch, R., and Hepburn, C. (1998). Researching Serial Murder: Methodological and Definitional Problems. Electronic Journal of Sociology. 3(2):1–11. Available at: Accessed: March 19, 2015.

Hoffer, S. (2015). Phoenix Police Make Arrest in 1990s Double-Murder Cold Case‏. The Huffington Post. Available at: Accessed: April 1, 2015.

Homant, RJ., and Kennedy, DB. (2014). Understanding serial sexual murder: A biopsychosocial approach. In W. Petherick (Ed.), Profiling and serial crime (341–372) (3rd ed.). Boston: Anderson Publishing, Ltd.

Jenkins, P. (1994). Using murder: The social construction of serial homicide. New York: A. de Gruyter.

Jenkins, P. (2002). Catch Me Before I Kill More: Seriality as Modern Monstrosity. Cultural Analysis 3:1-17. Available at: Accessed: April 24, 2015.

Jones, LH. (2014). Aaron Hernandez's behavior more 'gangster' than serial killer. USA Today. Available at:

Kaste, M. (2015). Open Cases: Why One-Third of Murders in America Go Unresolved. Available at: Accessed: March 30, 2015.

Koerner, B. (2002). Serial Killers and Spree Killers: What’s the Difference? Slate. Available at: Accessed: April 24, 2015.

Kraemer, GW., Lord, WD., and Heilbrun, K. (‎2004). Comparing Single and Serial Homicide Offenses. Behavioral Science and the Law. 22(3):325–343.

LePard, D., Demers, S., Langan, C., and Rossmo, DK. (2015). Challenges in Serial Murder Investigations Involving Missing Persons. Police Practice and Research. 16(4):328-340.

Lester, D. and White, J. (2014). A Study of African American Serial Killers. Journal of Ethnicity in Criminal Justice. 12:1-9.

Levin, J., and Fox, J. (2008). Normalcy in Behavioral Characteristics of the Sadistic Serial Killer. Available at: Accessed: April 2, 2015.

Leyton, E. (1996). Second Thoughts on Theoretical Approaches to Multiple Murder," in Thomas O'Reilly- Fleming (Ed.) Serial and Mass Murder: Theory, Research and Policy. Toronto: Canadian Scholars' Press: 37-51.

Leyton, Elliott. (September, 2013). Personal communication.

Lohr, D. (2015). CHILLICOTHE'S MISSING WOMEN: 'These Girls Knew Each Other' ... And Now They're Gone. The Huffington Post. Available at: Accessed: May 27, 2015.

Mains, KL. Ken’s Corner. (2015a). Available at: Accessed June 2, 2015.

Mains, KL. (2015b). Signing my talent contract. Available at: Accessed June 2, 2015.

McClellan, J. (2014, April 15). Re: Cooling-Off Period‏. Message posted to Multidisciplinary Collaborative on Sexual Crime and Violence electronic mailing list, archived at

McNamara, JJ., and Morton, RJ. (2004). Frequency of Serial Sexual Homicide Victimization in Virginia for a Ten-Year Period. J Forensic Sci. 49(3):529-33.

Meloy, JR. (2000). The Nature and Dynamics of Sexual Homicide: An Integrative Review. Aggression and Violent Behavior. 5(1):1-22. Available at: Accessed: November 3, 2014.

Morton, RJ., and McNamara, JJ. (2005). Serial murder. In J. Payne-James (Ed.), Encyclopedia of forensic and legal medicine (pp. 4753). Oxford: Elsevier.

Morton, RJ. and Hilts, MA. (2008). Serial Murder: Multi-Disciplinary Perspectives for Investigators. National Center for the Analysis of Violent Crime. Washington, D.C.: Federal Bureau of Investigation. Available at: Accessed: April 5, 2015.

Morton, RJ., Tillman, JM., Gaines, SJ. (2015). Serial Murder: Pathways for Investigations. National Center for the Analysis of Violent Crime. Washington, D.C.: Federal Bureau of Investigation. Available at: Accessed: April 5, 2015.

Myers, W. (2014, April 16). Re: Cooling-Off Period‏. Message posted to Multidisciplinary Collaborative on Sexual Crime and Violence electronic mailing list, archived at

National Registry of Exonerations, The. (2015). Project of the University of Michigan Law School. Available at: Accessed: January 24, 2015.

Nelson, M. (2010). Making Sense of DNA Backlogs - Myths vs. Reality. Office of Justice Programs. National Institute of Justice Journal No. 266. Available at: Accessed: April 30, 2015.

Osborne, JR. and Salfati, G. (2015). Re-Conceptualizing “Cooling-Off Periods” in Serial Homicide. Homicide Studies. 19(2):188-205.

Ostrosky-Solis, F., VélezGarcía, A., SantanaVargas, D., Pérez, M., and Ardila, A. (2008). A middleaged female serial killer. Journal of Forensic Sciences, 53(5):1223–1230.

Petee, TA., and Jarvis, J. (2000). Analyzing Violent Serial Offending. Homicide Studies. 4(3):211-218.

Pino, NW. (2005). Serial Offending and the Criminal Events Perspective. Homicide Studies. 9(2):109-148.

Quinet, K. (2011). Prostitutes as Victims of Serial Homicide Trends and Case Characteristics 1970-2009. Homicide Studies. 15(1), 74-100.

Ramsland, Katherine. (August, 2010). Personal communication.

Reavis, JA. (‎2011).  Serial Murder of Four Victims, of Both Genders and Different Ethnicities, by an Ordained Baptist Minister. Case Reports in Psychiatry. Available at: Accessed: March 14, 2015.

Ressler, RK., and Burgess, AW. (1985). The Men Who Murdered. FBI Law Enforcement Bulletin. 54(8) 2–31. Available at: Accessed: February 27, 2015.

Salfati, G., Petee, T., and Field, JL. (2006). Serial and spree homicide: Different crimes or all serial crimes? Presentation at American Society of Criminology Conference, Los Angeles, USA.

Salfati, Gabrielle. (November, 2011). Personal communication.

Scheck, B., and Neufeld, P. The Innocence Project. (2015). Available at: Accessed: January 24, 2015.

Schlesinger, LB., Kassen, M., Mesa, VB., Pinizzotto, AJ. (2010). Ritual and Signature in Serial Sexual Homicide. J Am Acad Psychiatry Law. 38(2):239-46.

Simkin, M., and Roychowdhury, V. (2014) Stochastic Modeling of a Serial Killer. Journal of Theoretical Biology. 355:111-116. Available at: Accessed: January 1, 2015.

Skrapec, CA. (2001). Defining serial murder: A call for a return to the original Lustmörd. Journal of Police and Criminal Psychology. 16(2):1024.

Slater, C. (2015). Canal 'serial killer': Police dismiss 'myth' of pusher stalking Manchester's waterways after meeting professor. Manchester Evening News. Available at: Accessed: April 30, 2015.

Smith, V. (2009). FBI calls serial killer suspect in cold case. Times West Virginian. Available at: Accessed: September 15, 2014.

Sparacello, M. (2011). 22 phones in bridge jumper's car, house suggest more victims, Jefferson sheriff says. The Times-Picayune. Available at: Accessed: September 15, 2014.

Sterbenz, C. (2015). Here's a surprising look at the average serial killer. Business Insider. Available at: Accessed: May 16, 2015.

Titlow, J. (2014). The Productivity Snag That Keeps Data Scientists Up At Night. Fast Company magazine. Available at: Accessed: April 24, 2015.

Valencia, MJ., Murphy, S.,  and Ellement, JR. (2013). Former hit man John V. Martorano says he’s no serial killer — and he tried to be a ‘nice guy’. Boston Globe. Available at: Accessed: July 8, 2014.

Venton, D. (2015). One Doctor’s Quest to Save the World with Data. Wired. Available at: Accessed: April 28, 2015.

Vronsky, P. (2013, June 18). Re: Former hit man John V. Martorano says he’s no serial killer‏. Message posted to Multidisciplinary Collaborative on Sexual Crime and Violence electronic mailing list, archived at

Warf, B., and Waddell, C. (2002). Heinous spaces, perfidious places: The sinister landscapes of serial killers. Social and Cultural Geography. 3(3):323-345.

Waters, Nicole L. (November, 2013). Personal communication.

Winton, R., Mather, K., and Stevens, M. (2014). Suspect in five-day shooting rampage is a 'serial killer,' police say. Los Angeles Times. Available at: Accessed: March 1, 2015.

Wright, KA., Pratt, TC., and DeLisi M. (2009). Multiple Homicide Offenders Arbitrary Cut-Off Points and Selection Bias. Homicide Studies. 13(2):193-199.

Wyllie, D. (2011). IBM supercomputers help law enforcement gather, analyze, and manage crime data. PoliceOne. Available at: Accessed: January 31, 2015.

Yaksic, E. (2006). Can a Demographic Make you Psychopathic? Northeastern University Research Expo. Poster Presentation. Available at: Accessed: September 3, 2014.

Yaksic, E. (2012, July 9). Re: Serial Killers and Significant Others‏. Message posted to Multidisciplinary Collaborative on Sexual Crime and Violence electronic mailing list, archived at

Yaksic, E. (2013a). An Attempt to Explain the Decline in Instances of Serial Homicide. Accessed: October 4, 2014.

Yaksic, E. (2013b). Breaking the stereotype of a serial killer. Op-Ed. Clarion Ledger. Available at: Accessed: April 16, 2015.

Yaksic, E. (2014a). Review of Why We Love Serial Killers. Available at:

Yaksic, E. (2014b). A Guest Post: The Felix Vail Case. Available at: Accessed: December 3, 2015.

Zalkind, S. (2014). The Murders Before the Marathon. Boston Magazine. Available at:

Zapotosky, M. (2015). Judge moves murder trial to Fairfax. Washington Post. Available at: Accessed: April 24, 2015.

Zubkoff L., Neily J., Mills PD., Borzecki A., Shin M., Lynn MM., Gunnar W., Rosen A. (2014). Using a virtual breakthrough series collaborative to reduce postoperative respiratory failure in 16 Veterans Health Administration hospitals. Jt Comm J Qual Patient Saf. 40(1):11-20. Available at: Accessed: April 5, 2015.