Abstract
It is frequently reported by the media and public officials that 100,000 registered sex offenders (RSOs) in the United States are “missing.” This policy note first describes the origin of this figure, which was initially derived from a 2003 informal survey of state registries conducted by a grassroots advocacy organization. Then, we explore the definitional ambiguities that complicate the process of calculating the national number of fugitive sex offenders. Finally, we present emerging research efforts to develop reliable estimates of the number and proportion of RSOs officially recorded by states as absconded, whereabouts unknown, or noncompliant with registration requirements. While such data remain limited, we find little evidence to support that 100,000 sex offenders are “missing,” using even the most inclusive definitions. Implications for policy and practice are discussed.
Like traditional folklore, urban legends tell one kind of truth. They are a unique, self conscious reflection of major concerns of individuals in the societies in which they circulate. . . . the lack of verification in no way diminishes the appeal urban legends have for us.
The Story of a Number
The number is 100,000. Depending on the account, it has been used to represent the number of sex offenders in the United States who are “missing,” “noncompliant,” “fugitive,” “evading authorities,” or “absconded.” Since early 2003, this attention-grabbing statistic has occupied a prominent place in the discourse over our nation’s sex offender policy. This figure (or in recent cases, an upwardly adjusted variant of it) has appeared in dozens of news articles, testimony from advocates and public officials, Congressional and government agency reports, press releases, and federal appropriations bills. The number is commonly cited without attribution as an unassailable fact—a self-evident fixture in the dialogue over how best to protect society from dangerous sexual predators.
The routine use of the “100,000 missing” figure by the media and public officials, and the context in which it is typically invoked, suggest a number of implicit assumptions: (a) that the figure is accurate and derived from a valid data source, (b) that it is based on a reasonably uniform and commonly accepted definition, (c) that registration noncompliance is an indication of an intent to evade authorities for purposes of committing new sexual offenses and is therefore synonymous with elevated risk, and (d) that the figure is attributable to the actions or inactions of individual registrants rather than to data management or other administrative system limitations. This article critically examines and evaluates these assumptions and then provides new data estimating the number of noncompliant sex offenders in the United States. We conclude with implications for policy and practice.
Where Did the Number Come From?
According to the National Center for Missing and Exploited Children (NCMEC), there are currently approximately 730,000 registered sex offenders (RSOs) in the United States (NCMEC, 2010). Scholarly research on the national sex offender population has been hindered by the fact that state sex offender registries are independently operated, creating significant challenges in generating a national database for empirical inquiry. Furthermore, because the sex offender registry population fluctuates daily as new offenders are added and status changes occur for those already listed, any registry research represents only one snapshot in time. Though some state government reports have provided information about the status of offenders on their registries, only one known scholarly article has been published to date investigating the characteristics of the nationwide sex offender population and estimating of the number of missing sex offenders (Ackerman, Harris, Levenson, & Zgoba, 2011).
Although the 100,000 figure has been attributed to the NCMEC or the US Marshals Service (USMS), news archives suggest that the original source was a grassroots advocacy organization called Parents for Megan’s Law (PFML). Prompted by an Associated Press investigation revealing that California had lost track of at least 33,000 sex offenders, PFML initiated a survey of state registries in 2003 and reported that, “on average,” states were unable to account for 24% of sex offenders required to register. Their website claims that they “found that nearly 25% of the nations’ [sic] registered sex offenders were not complying with state registration requirements . . . over 100,000 sex offenders were ‘missing’ from registries across the country” (Parents for Megan’s Law, 2010). Like dormitory gossip or the childhood game Whisper Down the Lane, this statement galvanized the urban legend that over years of telling took on a life of its own.
PFML has not made their methodology or detailed results of their survey publicly available and they did not respond to multiple inquiries from the current authors. A 2003 story by Associated Press reporter Kim Curtis, however, revealed that although registry managers in all states responded to PFML’s telephone survey, only 32 were able to provide failure-to-register rates (Curtis, 2003a). The numbers were based on the word of officials in each state; no raw data were provided to PFML for analyses. Many registry officials said they had never audited their sex offender registries and could provide only rough estimates of their accuracy. The survey results reported that California, Massachusetts, Oklahoma, and Tennessee had the highest rates of noncompliance, all at around 50%. A Tennessee official disputed this, and Oklahoma’s Corrections Department spokesman, Brian Johnson, said the figure was just his best guess. “I don’t have any specific actual information in terms of the level of noncompliance,” he told the Associated Press. “We’ve not done a study of that that I’m aware of.” Florida was among the states with the best compliance, successfully tracking about 95% of their RSOs (Curtis, 2003a).
An earlier Associated Press analysis of the California registry, based on actual registry data, suggested that there were indeed systemic problems with that state’s registry and that the records of as many as 33,296 sex offenders (44% of the 76,350 who were registered at the time) were out of date (Curtis, 2003b). Poor sex offender tracking made news in other states as well. For example, a Boston Herald investigation in November 2003 found that approximately 49% of the 18,120 sex offenders in Massachusetts were not registered at all with either the state or with local police departments. More than 1,000 sex offenders were living in Boston but the addresses of nearly half were unknown to police. The Herald found that the job of tracking sex offenders released from prison and verifying the accuracy of their addresses had fallen through the cracks (Mullvihill, Wisniewski, Meyers, & Wells, 2003).
Looking back on the news coverage, it is easy to see how a fairly cohesive narrative emerged. Extrapolating from their state survey, PFML estimated that about 25% of the nation’s sex offenders (about 100,000 at the time) did not have a verified current address. Ever since that alarming statistic was quoted in the media, it has permeated the public discourse of what is “known” about sex offenders. The fallacy, however, might be that the problem of “missing” sex offenders was more a reflection of failures in the operation and administration of the registries than of noncompliance by individual RSOs. In other words, challenges related to record keeping, personnel demands, and data management during the formative years of sex offender registration laws were likely conflated with the issue of offender noncompliance.
Definitional Ambiguity
The actual numbers aside, reviewing the history of the issue sheds light on important definitional ambiguities inherent in discussions of “missing” sex offenders. There is a clear lack of uniformity in how certain terms and designations are applied across jurisdictions and how these concepts are interpreted and communicated. States use different language in their registries to describe sex offenders whose addresses might not be verified; examples include terminology that may denote a lack of permanent address (e.g., “homeless” or “transient”), technical failure to comply with registration requirements (“in violation” or “noncompliant”), and official designations of fugitive status ( “absconded,” “address unknown,” or ”whereabouts unknown”). Most states appear to track and report any form of registry noncompliance—not specifically fugitive or absconder status.
A 2011 investigation by Politifact (a fact-checking news outlet) attempted to ascertain the number of missing sex offenders in several New England states (Lord, 2011). They found that 4% of sex offenders in Rhode Island were “unaccounted for,” 7% in Connecticut were “not in compliance,” 1.5% of New York sex offenders were identified as “location unknown,” and nearly 3% of Massachusetts offenders were designated as “violators.” According to this report, official figures from California in March 2011 indicated that about 24% of sex offenders from that state were “in violation” of the registration law (Lord, 2011). In 2009, the California Sex Offender Management Board noted in a report that 18% of their registrants were defined as “in violation” of some sort (California Sex Offender Management Board, 2010).
We can see that states show a vast range of noncompliance rates. Five states comprise nearly half the registered sex offenders in the United States (California, Florida, Texas, Michigan, and New York). Larger states might have more challenges tracking their offenders, and California has consistently been an outlier with higher reports of noncompliance than other states. PFML’s strategy of extrapolating a national estimate from a state such as California is a flawed methodology; national averages cannot be inferred from single states. We further suggest that a more refined definition of “missing” sex offenders might produce more accurate rates of registrants whose whereabouts are truly unknown.
References to the 100,000 missing sex offenders have dominated policy discussions since 2003 and in fact were instrumental in persuading Congress to include certain parameters in the passage of the Adam Walsh Sex Offender Registration and Notification Act (2006). Consider the compelling testimony of the director of the NCMEC during hearings in 2005
In 31 States, the penalty for failure to register is just a misdemeanor. In three States, offenders have more than 10 days to notify authorities when they change their address. We suspect that those who represent the greatest threat to children are also the least likely to be compliant. There are at least 100,000 non-compliant offenders; people like the killer of Jessica Lunsford, who was not where he was supposed to be and whose presence was unknown to police or Jessica’s family . . . . (109th Congress, 2005, p. 15)
In response to this and other testimony, the Adam Walsh Act required RSOs to notify authorities of an address change within three business days and made failing to do so a felony offense punishable by a prison term of 1 to 10 years (Adam Walsh Sex Offender Registration and Notification Act, 2006).
As well, definitional ambiguity articulating the number of “missing” sex offenders may lead to certain assumptions that drive funding allocations. In a 2009-2010 appropriations bill, Congress placed the number of missing sex offenders at 135,000 as part of the rationale for a substantial increase in the U.S. Marshal Service budget (111th Congress, 2009). Although the bill does not cite a path to this figure, the most likely explanation is that it represented an inflation of the original 100,000 figure, accounting for growth in the number of RSOs nationwide. Alluding to the vast magnitude of the problem of fugitive sex offenders, the bill commended the Marshals Service and appropriated an additional US$10,000,000 above their US$50,985,000 request to expand Adam Walsh Act enforcement.
Registration Noncompliance and Risk for Sexual Recidivism
Implicit in the interchangeable use of the terms described earlier (e.g., “fugitive,” “absconder,” “missing”) is the assumption that all sex offenders who are noncompliant with registration are deliberately eluding authorities and that they represent the highest risk of sexual offending. This belief is reflected throughout official testimony, including the following statement from a 2011 hearing to reauthorize the Adam Walsh Act:
. . . the mobility of offenders and inconsistencies among state registration laws have resulted in as many as 100,000 missing sex offenders. Law enforcement does not know where these missing sex offenders are, yet they are living in our communities . . . The offenders who take advantage of these loopholes are attempting to evade their registration duties—which could present a threat to the safety of our communities. (Reauthorization of the Adam Walsh Act, 2011, pp. 5-6).
In another example, a press release announcing the introduction of a bill called the Finding Fugitive Sex Offender Act of 2011 stated, “Sex offenders often fail to register precisely so they can evade detection and in many cases, find new victims . . . Marshals must . . . be able to find missing predators more easily and greatly curb the threat of future offenses” (Blumenthal, 2011). The former director of the U.S. Sex Offender Sentencing, Monitoring, Apprehending, Registering, and Tracking (SMART) Office warned in a USA Today story, “The people you need to be worried about most are the ones who aren’t registering at all” (Koch, 2007, p. 1).
These assertions are problematic on at least two levels. First, while research on sex offender registration compliance is in its infancy, studies that have been conducted do not demonstrate a direct linkage between failure to comply with registration requirements and sexual recidivism risk. Duwe and Donnay (2010), noting that failure to register (FTR) has become the most common recidivism offense for sex offenders released from prisons in Minnesota, examined outcomes of 1,561 released high-risk sex offenders. Of these, about 11% had been convicted of failing to register, but FTR was not predictive of either sexual or general recidivism. The authors concluded that registration noncompliance did not appear to increase the risk of sexual reoffending (Duwe & Donnay, 2010). A study conducted in South Carolina involving 2,970 RSOs (Levenson, Letourneau, Armstrong, & Zgoba, 2010) found no statistically significant differences between sexual recidivism rates of those who failed to register (11%) and compliant registrants (9%). FTR did not predict sexual recidivism, and the authors suggested that FTR and sexual offending tap separate constructs, with FTR related to rule-breaking behavior patterns while sexual offending is driven by sexual deviance. A similar study in New Jersey revealed that failure to register was not a significant predictor of sexual recidivism. FTR offenders were more likely to have adult female victims, debunking the myth that pedophiles abscond in order to prey on children (Zgoba & Levenson, in press).
Beyond these findings, research on general criminal offenders has found that sex offenders are among those least likely to abscond (Williams, McShane, & Dolny, 2000), that prior offense severity does not predict absconding, and that that absconders are not necessarily a high-risk criminal group (Mayzer, Gray, & Maxwell, 2004; Schwaner, McGaughey, & Tewksbury, 1998; Williams et al., 2000). Furthermore, it appears that some parole violators are motivated to flee due to a perceived inability to comply with an overwhelming, complex, and rigid set of rules (Schwaner et al., 1998) rather than to commit new crimes.
The second major fallacy involved in the presumption that noncompliance is associated with risk is the lack of recognition that failure to register may be linked to a wide range of circumstances ranging from RSO oversight or carelessness to deliberate evasion of authorities. It is improbable that all sex offenders with invalid addresses are willful violators; rather, most noncompliant sex offenders do not appear to have absconded (Duwe & Donnay, 2010; Levenson et al., 2010). Reports from law enforcement agencies that have embarked on initiatives targeting noncompliant registrants suggest that the substantial majority of those who fail to adhere to registration requirements are easily located through routine local law enforcement actions. Some may appear to be missing due to inadequate or incomplete address information, data entry errors, lag times in administrative updating, unauthorized travel, or homelessness (Harris & Pattavina, 2009). Confused by complex registration laws, some sex offenders might carelessly neglect to update their registration as required but continue to report to parole agents and remain in their known locations despite their lapse in compliance.
Among those who do indeed willfully fail to register, it cannot be assumed that all of these individuals do so to evade authorities for purposes of committing new sexual offenses. It is well documented that many RSOs experience unemployment, housing disruption, harassment, and social alienation as a result of sex offender registration and notification (SORN) laws (Levenson & Cotter, 2005; Levenson, D’Amora, & Hern, 2007; Mercado, Alvarez, & Levenson, 2008; Tewksbury, 2005; Tewksbury & Lees, 2006; Zevitz & Farkas, 2000). It would therefore be expected that some sex offenders fail to comply in an effort to avoid the collateral consequences of SORN.
Offender Noncompliance or Administrative Failure?
One final noteworthy point regarding the “100,000 missing” is that offender noncompliance may be conflated with inherent problems involving the administration of the registries. It is likely that the number of sex offenders whose whereabouts were unknown was higher in the early years of the new millennium than it is currently and that much of the “missing sex offender” problem was due to the comparatively primitive state of SORN systems at that point in time. States were mandated to come into compliance by 1999 with the Jacob Wetterling Act and Megan’s Law. The PROTECT Act, an additional Wetterling amendment, required states to develop and maintain online registries by 2003. Law enforcement officers and probation agents expressed concerns early on about increased personnel demands as states struggled to implement new registration and notification procedures (Matson & Lieb, 1996; Zevitz & Farkas, 2000). It is not surprising that tracking and monitoring of sex offenders in the early days of the law was far from perfect. Since that time, enhancements in technology and automated tracking procedures have led to improvements in monitoring and verifying the locations of RSOs.
Although states have improved their procedures over the past decade, some states historically have been more or less successful than others. There is little evidence to suggest that either registry noncompliance or “missing” status is evenly distributed across jurisdictions. For reasons that are unclear, for instance, California has consistently demonstrated a higher rate of noncompliance than other states and Florida has historically been one of the most successful states in tracking offenders. Some states may have more sophisticated technology or more substantial personnel resources to verify addresses.
How Many Sex Offenders in America are Truly Missing?
To answer this question, we rely on two recent studies investigating the characteristics of U.S. sexual offenders. The methodologies included efforts to ascertain how many were documented by states as truly missing or absconded. The studies are summarized very briefly here and interested readers can refer to the full reports for a more detailed understanding of the methods and results.
Official designations of “missing” sex offenders, using data downloaded from public Internet registries
The first study analyzed data on 445,127 RSOs obtained in 2010 from the Internet registries of 49 states, Washington, D.C., Puerto Rico, and Guam (Ackerman et al., 2011). Using data downloaded directly from public registries (about two thirds of the total RSO population), the authors were able to identify 17,688 RSOs who were designated by states to be transient, homeless, absconded, noncompliant, or whose address or whereabouts were otherwise unknown. Nationwide, a total of 5,349 offenders were officially listed as absconded, 1,264 were listed as missing or unable to be located, and 4,152 were listed as having failed to comply with registration requirements. Using these categories, approximately 10,000 sex offenders (2.4%) might be considered “truly missing.” Utilizing the more inclusive figure of 17,688 offenders (including those designated homeless or transient), approximately 4% of the sample might not have a valid address. No evidence was found to support the frequently repeated notion that 100,000 of the nation’s sex offenders are missing or unaccounted for.
A second study (Harris, Levenson, & Ackerman, under review) analyzed data obtained in 2010 via email and telephone surveys of state registry managers. Of 42 responding jurisdictions, 29 provided counts of missing and absconded RSOs, reporting a total of 28,678 cases with such designations. Notably, more than half of these cases were derived from one state (California) and nearly three quarters from just four states. Rates varied significantly across jurisdictions, ranging from less than 1% to just above 13%, with a median rate of 2.7%. Though states were asked to report in this category only those RSOs officially designated as missing or absconded, follow-up analyses indicated that many states (including those with the highest rates such as California and Wisconsin) included any form of registry noncompliance in their missing/absconded status. States using more refined definitions (e.g., unsuccessful attempts to locate the offender) produced more moderate estimates.
The study identified significant variation in how state systems defined and captured these data. Citing both definitional variation and the skewed distribution of those officially designated as missing or absconded, the study’s authors cautioned against extrapolating figures based on a limited sample to derive a national estimate. They concluded, however, that there is little evidence to support the assertion that more than 100,000 RSOs are missing or absconded, even using the most liberal definitional criteria.
Summary and Policy Implications
Our review concludes that despite its influence and significant staying power, the continued assertion that 100,000 of the nation’s sex offenders are unaccounted for has not been substantiated. Recent research suggests that the figure is highly inflated; using data obtained directly from online registries and state registry managers, it is estimated that approximately 2.5% of the nation’s RSOs might be truly missing but that the proportion increases to about 4% when including homeless and transient offenders with unverified addresses. Moreover, continued reference to 100,000 missing sex offenders carries with it numerous problematic assumptions, including the expectation of increased risk and the presumption of intentional evasion. Finally, it appears that rates of registration failure vary widely across the states, suggesting that the problem may be, in large part, a reflection of idiosyncratic data challenges and inadequate law enforcement resources in particular jurisdictions rather than a more generalized crisis of sex offender absconding.
In addition, the definitional ambiguity across states creates significant challenges for both policy makers and researchers. The Adam Walsh Act intended to facilitate uniformity in classification systems, Internet notification procedures, data elements, and parameters of registration. It does little, however, to address the considerable definitional differences across state registration systems. Registries are intended primarily for law enforcement tracking and public awareness purposes, but they also provide important information for those engaged in research and policy development. With that in mind, federal policy makers seeking greater standardization might consider the development of common terminology and definitions that could be implemented as states pursue Adam Walsh Act compliance. For instance, if states were required to differentiate “noncompliant” or “unverified” offenders from those who have “absconded,” the number of truly “missing” sex offenders at any given time might be more readily estimated. There should be a specific designation for offenders who remain unaccounted for after multiple failed attempts to locate them or when it has been confirmed that the offender is no longer living at the registered address and is no longer checking in with probation, parole, treatment providers, or law enforcement.
Regardless of how artfully the statistic is framed, the implicit subtext—that there are more than 100,000 dangerous sex offenders lurking in the shadows and deliberately evading detection so that they can prey on children—exerts a profound impact on policy development and on federal funding allocations. These decisions are influenced in part by public perception. As Brunvand cautioned, “The mass media themselves participate in the dissemination and apparent validation of urban legends . . . adding to their plausibility” (Brunvand, 1981, p. xii). If the number of noncompliant, missing, or fugitive sex offenders is to continue to wield an impact on allocation of taxpayer dollars and the broader sex offender policy discourse, we need better and far more transparent data.
Ultimately, devising informed and effective responses to registration lapses requires a more refined understanding of RSO noncompliance and its circumstances. With more accurate data, we can more effectively identify and respond to potentially dangerous predators who truly abscond to avoid detection. Reliance on a single and outdated statistic—one that proves unsubstantiated when examining current data—does little to advance understanding of the needs and limitations of the nation’s sex offender registries and obscures important definitional and practical distinctions that are vital to the efficient distribution of resources across levels of government.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.