Skip to main content
SearchLoginLogin or Signup

Which Source Possesses the Best Data on the Empirical Aspects of Criminal Events? A Theory of Opportunity and Necessary Conditions

Offenders and nonoffenders possess valuable information about crime. But which possesses the best data? This is a complex issue, so I narrow my focus to data on empirical aspects of criminal events. Drawing on the necessary conditions perspective, I theorize that a source’s ...

Published onDec 21, 2018
Which Source Possesses the Best Data on the Empirical Aspects of Criminal Events? A Theory of Opportunity and Necessary Conditions
·
history

You're viewing an older Release (#2) of this Pub.

  • This Release (#2) was created on Jun 09, 2020 ()
  • The latest Release (#3) was created on Jan 26, 2023 ().

Note: This is the postprint of the following paper; publisher version available here.

Jacques, Scott. 2019. Which Source Possesses the Best Data on the Empirical Aspects of Criminal Events? A Theory of Opportunity and Necessary Conditions. Deviant Behavior 40:1543-1552.​

Abstract: Offenders and nonoffenders possess valuable information about crime. But which possesses the best data? This is a complex issue, so I narrow my focus to data on empirical aspects of criminal events. Drawing on the necessary conditions perspective, I theorize that a source’s possession 1) of data varies directly with its involvement in cases; 2) of representative data varies inversely with nonrandom involvement in cases and nonrandom siphoning off from the larger group to which it belongs; and, 3) of accurate data varies inversely with time since involvement in cases. Those general principles suggest that offenders, especially active ones, possess the most data, representative data, and accurate data on empirical aspects of criminal events. I conclude by discussing the implications of those general principles for observation research, sources’ possession of subjective data, and their possession of empirical data on other criminological events, specifically victimization and policing.

Introduction

Criminologists are interested in criminals because they break the law, but also because they possess information about crime (e.g., Becker 1963; Jacobs 1999; Leclerc and Savona 2017; Leclerc and Wortley 2014; Shaw 1930; Sutherland 1937; Whyte 1943; Wright and Decker 1994, 1997). Of course, nonoffenders, such as police, victims, and guardians, have information about crime, too (e.g., Moskos 2008; Reynald 2010). Researchers should tap into the knowledge of both offenders and nonoffenders because each has valuable data. Nonetheless, an interesting question is whether and why offenders or nonoffenders possess more crime data of a higher quality, such as in representativeness and accuracy.[1] Though this issue is of obvious importance to criminological research, it has been insufficiently addressed, with one exception noted below. Thus, in this article, I theorize how offenders compare to nonoffenders in the possession of a particular type of crime data, namely that on the empirical characteristics of criminal events. That is a complicated venture, so it is important to begin by clearly specifying what will be theorized, why, and how. Subsequently, I use a niche of the opportunity perspective – that interested in necessary conditions – to analyze how offenders and nonoffenders compare in the possession of the aforementioned data. Finally, I summarize the analysis and discuss a few of the questions it raises.

Theoretical Focus

No data source knows the most about everything. Rather, some sources know more about certain issues. Herein, the focus is whether offenders or nonoffenders possess more and higher quality data on the empirical facets of criminal events. By “empirical,” I mean things that may be seen, heard, touched, tasted, or smelled. “Criminal events” are strictly defined as acts prohibited by law. When present during a criminal event, offenders and nonoffenders are able to mentally record what occurs, such as seeing a robber pull out a gun and say, “Give me all you have!” Thus, for the sake of theorizing, I will think of people as I do a video camera or audio recorder – as something capable of recording empirical phenomena. (For brevity, I will sometimes refer to empirical data simply as “data” or “information,” and, likewise, criminal events as “crimes” or “offenses.”)

I will not consider data on more subjective phenomena or that pertaining to what happens before or after criminal events. These omissions are not meant to imply that such things are unimportant; clearly, they have great importance (e.g., Presser and Sandberg 2015). Indeed, a theory that explains sources’ possession of such data would be better, at least in generality, than the one I propose. But I restrict my focus to data on empirical aspects of criminal events because it seems wiser to start small and with things that “we” – meaning (most) people – can agree happened based on our senses.

Another focus herein is sources’ potential to provide data, which is a matter of the information they possess. This issue is a question for methodology, or the study of method. Of course, “method” is a catchall term for many behaviors, including various aspects of data collection (e.g., types of sampling, experimenting, instruments) and analysis (e.g., types of qualitative and quantitative analysis). Therefore, methodological studies vary in what exactly they address. As I argue below, methodological texts devoted to criminology have paid insufficient attention to sources’ potential to provide data based on what they possess.

One type of methodological text is that geared toward students. If you look through criminological methods textbooks, what you typically see is a discussion of units of analysis, followed eventually by techniques of sampling, experimentation, recording data (e.g., ways of asking questions and making observations), and analyzing data (e.g., quantitative versus qualitative) (e.g., Maxfield and Babbie 2012). With one exception, namely the “dark figure of crime” (addressed in a following section), what such texts leave largely unaddressed is the quantity and quality of crime data possessed by offenders and nonoffenders. The same shortfall characterizes methodological texts meant for consumption by researchers.

I suspect that the above limitation of methodological texts is attributable to insufficient conceptual clarity, which may be the cause or effect of how they are organized, especially those geared toward researchers. What I see in that literature (e.g., Bernasco 2010; Copes and Miller 2015; Polsky 1969; Miller and Palacios 2015) is the muddling of two distinct processes: what members of a particular population could provide to researchers based on the information they possess, versus the prospects and problems involved in obtaining that data via sampling and data recording techniques (e.g., interviewing, observation). To be clearer about how these differ, I think methodological texts would do best to discuss them in turn: what sources possess, followed by how those data may be accessed. But what such texts do is focus on how to access a given source’s data, during which they sometimes discuss what the source possesses. The problem with this organization is it hides important differences between sources.

To better illuminate the distinction between subjects’ possession of data versus what they share, allow me to succinctly review commonalities found in the literatures on how to obtain data from offenders and police officers (e.g., cf. Bernasco 2010 and Copes et al. 2015 with Fox and Lundman 1974 and Van Maanen 1981). To study them, respectively, first a researcher may have to gain permission from a high ranking correctional or police officer to recruit persons under their authority. Second, the researcher must convince individual offenders or officers to participate; of course, that also is required to study uninstitutionalized criminals (i.e., active ones and desisters). Third, for the offenders and officers who agree to participate, the researcher will want them to speak at length and honestly.

Those steps are important to obtaining data from subjects, and researchers’ success in doing so varies for a variety of reasons. These include factors internal to subjects (e.g., cognitive distortions, such as those due to intoxication), interpersonal in nature (e.g., their trust in the researcher and how questions are asked), and external (e.g., the setting).[2] In light of these influences, researchers have offered a wealth of ideas on how to increase the odds of gaining access and obtaining more information of higher quality. The suggestions are too numerous to review here, but, in short, they amount to manipulating – to the extent possible – the internal, interpersonal, and external forces thought to affect data (e.g., Bernasco 2010; Copes et al. 2015; Fox and Lundman 1974; Van Maanen 1981).

Though such suggestions can be quite useful for getting the most out of subjects, they are not meant to – nor can they – affect subjects’ possession of data. Rather, those steps are a matter of convincing subjects to provide the data they already possess to the researcher. Whether a source will talk, to what extent, how earnestly, and whatnot is conceptually distinct from what the source has to share.

This article’s singular focus is the potential of sources to provide data, which, to reiterate, is a matter of the amount and quality of information they possess.[1] I will not further address what a source will actually provide to researchers or, relatedly, how to get the most out of a source, as those issues have been previously explored at length (e.g., Bernasco 2010; Copes and Miller 2015; Polsky 1969; Miller and Palacios 2015). Nonetheless, theorizing how sources vary in the data they possess has paramount implications for understanding and interpreting the data that they actually provide to researchers when sampled, interviewed, surveyed, and the like. Think of it in mathematical terms: if the amount of data shared by a source with researchers is a count, that count is more meaningful if put as the numerator in an equation, with the data possessed by the source in the denominator; see figure 1.

--FIGURE 1 ABOUT HERE--

Theoretical approach

Another limitation of methodological texts on crime data sources is the absence of a coherent theoretical approach, or any theory at all. For example, some of these works present a series of practically useful insights, but without reference to a particular theoretical orientation (e.g., Polsky 1969). Other texts connect their insights to theory, but do so without a single overarching paradigm. Instead, they draw on one perspective before moving to another, and may also lose sight of theory for some of the ideas (e.g., Copes et al. 2015).

To improve theoretical understanding of the topic at hand, I will draw on a niche of the opportunity perspective, namely that focused on necessary conditions. Its basic premise is that for any given behavior to occur, certain minimal elements must be present (Gottfredson and Hirschi 2003; Jacques 2010; Sparks 1982:29-30). To explain crime, for example, Cohen and Felson (1979) suggest that the necessary conditions for a criminal event are a motivated offender and target. In other words, a situation without both an offender and target cannot result in an offense, irrespective of guardianship (see also Felson and Clarke 1998). Similarly, Cohen, Kluegel, and Land (1981) argue that the definitional properties of offenses affect their rate of occurrence. Auto theft, for example, is constrained by the absence of automobiles (see also Gottfredson and Hirschi 1990).

Thus, the goal of theorists working within the necessary conditions perspective is to specify the minimal elements necessary for something to occur. For this article, then, my theoretical strategy and focus is to determine the minimal elements required for sources to possess 1) data 2) that is representative and 3) accurate. The product of this approach will be a theory stated as general principles that explain how sources compare in their possession of data.

In the following section, I devise those general principles and specify their implications for how offenders and nonoffenders vary in the data they have on empirical facets of criminal events. Because, as seen below, those principles suggest that offenders possess more data of a higher quality, I subsequently continue the analysis by focusing on how and why two broad types of offenders – active and inactive (desisters and inmates) – compare in the information they have on crime. Finally, in the concluding section, I will take tentative steps toward specifying the general principles’ implications for other criminological behaviors, as well as address a couple other questions raised by the analysis.

Theoretical Analysis

Offenders versus Nonoffenders

Both offenders and nonoffenders are present during some offenses. Criminals commit them, of course. Victims are attacked or have their property damaged or stolen. Police and bystanders may watch or intervene. Because offenders and nonoffenders are involved in offenses, both groups possess information about them. For that reason, offenders and nonoffenders may be used as data sources on offenses. However, the major variable affecting the possession of firsthand data is direct involvement in cases, a facet of opportunity. Any data source could be involved in zero to one-hundred percent of cases. Involvement in more cases produces greater opportunity to mentally record what occurred, i.e. obtain “data.” Thus, the more cases in which a source is involved, the more data it should possess. Stated as a general principle: A source’s possession of data increases as does its involvement in cases.

There is a clear implication of that principle for determining whether offenders or nonoffenders possess more data on offending. Though it is uncertain what percent of offenses involving nonoffenders, it must be less than one-hundred because many offenses are “victimless” or unobserved by police, guardians, or anyone else. Yet an offense cannot be committed without an offender, which means that as a collective, or a population, they are involved in one-hundred percent of cases. A related issue is the extent, or span, of participation in any given offense. Potentially, a data source could be involved in zero to one-hundred percent of each case.[3] Because the unit of analysis is empirical knowledge of an offense, the span of each case is from the first to last second of action defined as illegal. In some offenses, offenders and nonoffenders are involved in every moment. For example, a passerby or police officer may intervene upon seeing an offender start to attack someone. When offenders and nonoffenders are present (and conscious) for the entire span of an offense, these parties will possess practically the same amount of data. Yet, clearly, every moment of every offense does not involve a nonoffender. Crime may go undetected for quite some time before victims, police, or guardians learn of it, if they ever do. As a collective, though, offenders partake in every second of every offense; to argue contrary is illogical. In short: Unlike nonoffenders, offenders are involved in the entire span of every offense, so they possess more data on the topic.

That idea relates to the “dark figure” of crime, or the sum of unknown offenses (Biderman and Reiss 1967; Morrison 1897). Offenders and nonoffenders do not report every crime, which results in “dark,” or unaccounted, offenses in data sets and findings derived from them (e.g., Coleman and Moynihan 1996; Polsky 1969). Dark data results, for instance, from the police not learning of every crime, which affects police-based crime data (e.g., UCR), and from victims not realizing they were victimized, which affects victim-based crime data (e.g., NCVS). Also, offenders may not admit what they did to researchers. But whether offenders will report their crimes is different from whether they know about them. As explained above, offenders’ complete involvement in offenses means they possess data on every second of every crime, whereas nonoffenders’ incomplete involvement in offenses precludes total knowledge. Put differently, nonoffenders’ incomplete involvement in offenses means their knowledge of offenses is partially dark, whereas offenders’ total involvement in offenses means that as a collective they could shed light on all of them.

In addition to considering the amount of data possessed by sources, it is important to consider its quality, such as representativeness. Potentially, the information possessed by a source may be zero to one-hundred percent representative of the population of cases (e.g., of all offenses). If a source is not involved in one-hundred percent of cases, it is questionable whether their (non)involvement is only due to chance. If the source’s partial involvement in cases is not random, its data may not be representative of all cases. Put differently, participation in cases opens up the opportunity to obtain information (see above principle), but incomplete participation opens the door to that data not representing the broader whole. The less a source is involved in cases, the greater the risk its data will be unrepresentative, unless their (non)involvement is random. Stated as a general principle: A source’s possession of unrepresentative data increases as does its nonrandom involvement in cases.

As explained above, the total involvement of offenders in offenses means they possess more data on offenses than nonoffenders, as the latter are not involved in every crime, much less the span of each. Moreover, the sources’ different degrees of involvement in criminal events affects the representativeness of their data. Because nonoffenders are not involved in every second of every offense, and this (non)involvement may not be random, the door is open to them becoming involved in some kinds of offenses but not others, which would make their information unrepresentative of all cases. But because offenders are involved in every second of every offense, there is no opportunity to be selectively involved in cases, which prohibits them – as a collective – from obtaining unrepresentative information on offenses.

In addition to the dark figure of crime, which recall refers to offenses missing from data sets, another problem with crime data sets are unjustified counts. This results, for instance, from wrongful arrests as well as citizen complaints made in response to something that is not a crime (e.g., a resident calls 911 about a burglary, but the suspect is the home owner, as happened to Professor Gates of Harvard). Whereas missing information deflates the crime count, unjustified arrests, complaints, and the like increase it. And such unjustified counts pose a problem to representativeness (Sutherland 1940; Black 1970), unless, that is, the labeling of noncriminal events as crime is random. But that process does not appear to be due to chance alone. Rather, members of those groups are more likely to be arrested for crimes they did not commit (e.g., Goffman 2014; Jacobs and Wright 2006), and to have their noncriminal behavior perceived as “suspicious” or criminal and so reported to police (see Reynald, 2012). Thus: Because offenders, unlike nonoffenders, are involved in the entire span of every offense, there is less, if any, opportunity for the data they possess on criminal events to be unrepresentative due to nonrandom involvement in cases.

Active versus Inactive Offenders

              The preceding analysis suggests that offenders possess more data and higher quality data, in terms of representativeness, than nonoffenders on the empirical aspects of criminal events. Of course, “offender” is a broad label (as is “nonoffender”), so, given that, I continue the above analysis to see how two types of offenders – active and inactive – compare in the crime data they possess. I focus on the active/inactive dichotomy for two reasons: one, it cuts across all offender typologies (e.g., those based on crime type, expertise, demographic background); and, two, it has received the most attention in the methodological literature on offender-based research, which, to reiterate, is characterized by the various problems mentioned in the “Theoretical Focus” section (e.g., Bernasco 2010; Copes et al. 2015; Polsky 1969; Wright et al. 1992).

First, there is the issue of how active and inactive offenders compare with respect to the starting general principle: a source’s possession of data increases as does its involvement in cases. It is unknown with certainty how the groups compare in this regard because, to my knowledge, research has not made a definitive say on the matter, and thought experiments could favor either group. In other words, it is unclear whether active or inactive offenders have been involved in more criminal events. For that same reason, the implication of the second general principle – that on representativeness – cannot be determined, either.

How, then, can it be determined whether active or inactive offenders possess more or higher quality data on offenses? The answer reflects the manners in which criminologists classify offenders as “active” or “inactive.” Inactive offenders are those who stopped committing crime by “choice” (broadly conceived) or incapacitation, which implies that active offenders are those who have not terminated lawbreaking for those reasons. The distinction is simple, but, nonetheless, it must be operationalized to be useful in research.

In practice, active and inactive offenders, especially desisters (as compared to inmates, discussed below), often are differentiated by time since last offense. From study to study, for a person to be defined as an active offender, the length of time since their last offense may range from within a week to a year of participation or more. Researchers have flexibility in where, or rather when, to draw this line. The “best place” is not obvious because it involves a trade-off: the upside of longer periods since last offense is an increase in the population size to draw on, which can help to increase the sample size; but the downside of longer periods is they permit participation by persons with more memory distortion (Copes et al. 2015; for forgetting generally, see Wixted 2004). This problem points to the importance of data being accurate, which, like representativeness, is a facet of data quality. Potentially, what a source remembers happening, i.e. the data they possess, may be anywhere from zero to one-hundred percent accurate, meaning in line with what actually happened, empirically speaking. This points to a general principle: A source’s possession of accurate data increases as time since involvement in cases decreases.

If active offenders and desisters are differentiated according to time since last offense, the above principle has clear implications for which one should possess the most accurate data on criminal events. In any given study that divides up offenders as active or desister on the basis of time, active offenders always are defined, of course, as more recently offending than desisters. Since memory distortion increases with time, the implication of that categorization scheme is that active offenders possess more accurate information on offenses than do desisters. Put succinctly: Because active offenders are more recently involved in offenses than desisters, the former possess more accurate data on the topic.

Now, let us examine how differentiating active and inactive offenders on the basis of incarceration affects the value of their information and, thus, their potential usefulness in research. First, it should be reiterated that it is unknown whether active offenders or inactive offenders, including inmates, have been involved in more offenses and, therefore, which group possesses more information on offenses. Plus, it should be made clear that active offenders are not always more recently involved in crime than inmates. Someone detained in jail, for instance, may have committed an offense only a few days ago, whereas a criminal may be defined as active despite last committing an offense about a year ago. Because active offenders and inmates do not differ by definition in time since last offense (as do active offenders and desisters, above), it is uncertain how time should affect the data they respectively possess.

However, what is practically certain is that the inmate population is unrepresentative of the active offender population. The unrepresentativeness is due to inmates being nonrandomly drawn into the criminal justice system. The odds of apprehension and incarceration depend on offenders’ expertise and deterrability, for instance (Jacobs 2010; Nee 2015; Wright and Decker 1994). Also, it is well established that jails and prisons are disproportionately filled with socially disadvantaged souls (Goffman 2014). This is because advantaged offenders are less likely to be complained about by citizens, investigated or caught by police, prosecuted in court, or severely punished. The nonrandom reality of control means inmates’ offense data will disproportionately represent the experiences of offenders who are less skilled, less deterrable, less advantaged, or otherwise different from active offenders who escape apprehension. Stated as a general principle: A source’s possession of unrepresentative data increases as it is more frequently and nonrandomly siphoned off from the broader group to which it belongs. The nonrandom process by which active offenders become inmates means the latter will possess less representative data on offenses. In sum: Because inmates are not randomly drawn from the population of active offenders, the former possess less representative data on offenses than do the latter.                                 

Discussion

Offenders and nonoffenders are a mainstay of the criminological enterprise because of their role in crime, of course, but also because they are a source of data. The purpose of this article has been to determine which source possesses the most and highest quality data on the empirical aspects of criminal events. Drawing on the opportunity perspective, general principles were deduced: A source’s possession 1) of data varies directly with its involvement in cases; 2) of representative data varies inversely with nonrandom involvement in cases and nonrandom siphoning off from the larger group to which it belongs; and, 3) of accurate data varies inversely with time since involvement in cases. In light of those principles, it seems that offenders, especially active ones, possess the most data, representative data, and accurate data on empirical facets of criminal events.

The above analysis may raise many questions. Because it is not possible to answer all of them, at least not at length, I conclude by discussing what I believe are the three most important issues left unaddressed to here. First, what are the implications of the general principles for observation research? In some sense, researchers who make observations of crime and control are a data source unto themselves (see, e.g., Becker 1963; Duneier 1999; Goffman 2014; Whyte 1943). Traditionally, researchers have made observations in situ, but, nowadays, they also can do so by watching CCTV footage (e.g., Piza and Sytsma 2016; Moeller 2017). In either case, for a criminologist’s observations to be worth anything, they have to be of something and, often, of someone directly involved in crime or control. Thus, observation research has two data sources: the researcher who makes observations, and the people who are observed. Put succinctly, the aforementioned sources compare thusly: Potentially, criminologists could obtain more data, including more representative data, on offenses by observing offenders than nonoffenders because only the former are involved in the entire span of every crime; and, researchers cannot obtain any observation data on offenses from inactive offenders because they no longer commit crime.

Second, what are the implications of the general principles for subjective data on criminal events? Such information may include the emotions and motivations of offenders and nonoffenders, including victims and agents of control (e.g., Athens 1992; Katz 1988). Recall that I focus on empirical data in this article because it seems prudent to “start small.” What I meant by that, in part, is that issues surrounding subjective data are more complex.[4] For instance, and this strikes me as far less true of empirical data, criminologists do not agree – not that they should – about which type of subjective data is “best.” Of course, the differences reflect their respective orientations toward theory. Some criminologists prefer data on what offenders were thinking and feeling at the time of offenses (e.g., Leclerc and Savona 2017; Leclerc and Wortley 2014; Wright and Decker 1994, 1997), whereas other criminologists want data on how offenders’ assessments of and emotions about their lawbreaking evolve over time (e.g., Presser and Sandberg 2015). If thought of as types of wine, some motivations are ready to study shortly after harvest, such as Beaujolais nouveau, whereas others are best if aged in the mind like a well hung claret (for a further discussion of these issues as pertain to data quality, see Copes et al. 2015).

Given those complexities, the process of theorizing the possession of subjective data on criminal events is an especially difficult venture. Moving forward, however, I hope researchers will seek to do so. Toward that end, I offer some tentative ideas on the implications of the general principles for such data. The first principle suggests that a source’s possession of subjective data should vary directly with its involvement in cases. Thus, offenders should have more subjective data on criminal events than do nonoffenders. The implication of the second general principle is that a source should hold less representative subjective data if its involvement in criminal events is neither total nor nonrandom. However, it is unclear, at least to me, what exactly that means because, unlike empirical data, subjective data from any source (e.g., criminals, victims, control agents) strike me as only “representative” of criminal events from the perspective of that particular source. Finally, the third general principle suggests that a source should possess less accurate subjective data on a criminal event as time since then increases. If a researcher’s focus is on what a person subjectively experienced during an offense, that idea makes sense, but it will be of little consequence to researchers interested in how a person’s subjective understanding varies over time (e.g., Copes et al. 2015; Presser and Sandberg, 2015).

Throughout this article, and here for a final time, it has been stated that there is more than one useful data source on offending, which is true for most any other topic, as well. So while (active) offenders may possess the most and highest quality data on empirical features of offenses, this should not imply that nonoffenders or inactive offenders should be excluded from studies. Yet that argument is uncontroversial for the very reason it is obvious. What is not well understood is whether, how, and why various sources differ in the data they possess.

Third, then, what are the implications of the general principles for research on other criminological behaviors? The answer should be the product of a detailed analysis, but, by way of illustration, consider the following. First, note that though offenders were found to possess the most and highest quality data on offenses, it should not be assumed that victims will possess the most data on victimization, police on policing, and so on. Determining which source possesses the most data on any given behavior will depend on how it is defined. For example, if victimization is defined as subjection to an act prohibited by law, the general principles suggest it may be better to engage in “offenders on victimization” research than “victims on victimization” research because offenders observe every moment of every victimization event, whereas some victims only see the aftermath (e.g., discover theft or destruction of their property after the crime is committed). However, if the focus is interpersonal victimization, victims and offenders may possess the same amount of empirical data on such events because they require both parties to be present during their commission.

Let us further consider the general principles’ implications for the study of policing. Again, determining which source will have the most and highest quality policing data will depend on how the unit of analysis is defined. If the focus was made narrow by only considering arrest, for instance, then the interactional nature of this behavior would imply two roles possess the most data: police officers and arrestees (though note that not all arrestees are actually offense perpetrators, i.e. offenders). But the answer will change if policing is defined more broadly, such as any act undertaken by an officer in an official capacity. While offenders, victims, and guardians, among others, are involved in cases of policing, only officers are involved in every moment of it and, therefore, only they possess empirical data on one-hundred percent of policing acts. Moreover, that means police should possess more representative data on policing than do other sources. Whether police officers, offenders, or any other criminological actor will share what they know is a different issue, albeit one that can only be fully understood by determining what they have to tell.

References

Agar M. 1973. Ripping and Running: A Formal Ethnography of Urban Heroin Addicts. New York: Seminar Press.

Akerstrom, Malin. 1985. Crooks and Squares: Lifestyles of Thieves and Addicts in Comparison to Conventional People. New Brunswick, NJ: Transaction.

Anderson, Elijah. 1999. Decency, Violence, and the Moral Life of the Inner City. New York: W.W. Norton & Company.

Athens, Lonnie H. 1992. The Creation of Dangerous Violent Criminals. Chicago: University of Illinois Press.

Becker, Howard S. 1963. Outsiders: Studies in the Sociology of Deviance. New York: Free Press.

Bernasco, Wim (ed.). 2010. Offenders on Offending: Learning About Crime From Criminals. Cullompton, UK: Willan Publishing.

Biderman, Albert D., and Albert J. Reiss. 1967. “On Exploring the ‘Dark Figure’ of Crime.” The Annals of the American Academy of Political and Social Science 374:1-15.

Black, Donald. 1970. “Production of Crime Rates.” American Sociological Review 35:733-748.

Coleman, Clive, and Jenny Moynihan. 1996. Understanding Crime Data: Haunted by the Dark Figure. Philadelphia: Open University Press.

Copes, Heith, Scott Jacques, Andrew Hochstetler, and Timothy Dickinson. 2015. “Interviewing Offenders: The Active vs. Inmate Debate.” Pp. 157-172 in Routledge Handbook of Qualitative Criminology, edited by Heith Copes and J. Mitchell Miller. London: Routledge.

Duneier, Mitchell. 1999. Sidewalk. New York: Farrar, Straus and Giroux.

Felson, Marcus. 1998. Crime and Everyday Life, 2nd ed. Thousand Oaks, CA: Pine Forge Press.

Felson, Marcus, and Ronald Clarke. 1998. Opportunity Makes the Thief. Crime Prevention and Detection Series, paper 98. London: Home Office.

Fox, James C., and Richard J. Lundman. 1974. “Problems and Strategies in Gaining Research Access in Police Organizations.” Criminology 12:52-69.

Geertz, Clifford. 1983. Local Knowledge: Further Essays in Interpretative Anthropology. New York: Basic Books.

Goffman, Alice. 2014. On the Run: Fugitive Life in an American City. Chicago: University of Chicago Press.

Jacobs, Bruce. 1999. Dealing Crack: The Social World of Streetcorner Selling. Boston: Northeastern University Press.

_____. 2010. “Deterrence and Deterrability.” Criminology 48: 417-441.

Jacobs, Bruce A., and Richard Wright. 2010. Street Justice. New York: Cambridge University Press.

Jacques, Scott. 2010. “The Necessary Conditions for Retaliation: Toward a Theory of Non-Violent and Violent Forms in Drug Markets.” Justice Quarterly 27:186-205.

Jacques, Scott, and Richard Wright. 2008. “Intimacy with Outlaws: The Role of Relational Distance in Recruiting, Paying, and Interviewing Underworld Research Participants.” Journal of Research in Crime and Delinquency 45:22-38.

Katz, Jack. 1998. Seduction of Crime. New York: Basic Books.

Kuhn, Thomas S. 1977. The Essential Tension: Selected Studies in Scientific Tradition and Change. Chicago: University of Chicago Press.

Leclerc, Benoit, and Ernesto U. Savona. 2017. Crime Prevention in the 21st Century: Insightful Approaches for Crime Prevention Initiatives. New York: Springer.

Leclerc, Benoit, and Richard Wortley (eds.). 2014. Cognition and Crime: Offender Decision Making and Script Analyses. New York: Routledge.

Maxfield, Michael G., and Earl R. Babbie. 2012. Basics of Research Methods, 3rd edition. Belmont, CA: Wadsworth.

Miller, Jody, and Wilson R. Palacios (eds.). 2015. Qualitative Research in Criminology. Advances in Criminological Theory, Vol. 20. New Brunswick, NJ: Transaction Publishers.

Moeller, Kim. 2017. “Video-recorded Retail Cannabis Trades in a Low-Risk Marketplace: Trade Value and Temporal Patterns.” Journal of Research in Crime and Delinquency.

Morrison, William Douglas. 1897. “The Interpretation of Criminal Statistics.” Journal of the Royal Statistical Society 60:1-24.

Moskos, Peter. 2008. Cop in the Hood: My Year Policing Baltimore’s Eastern District. Princeton, NJ: Princeton University Press.

Nee, Claire. 2015. “Understanding Expertise in Burglars: From Pre-Conscious Scanning to Action and Beyond.” Aggression and Violent Behavior 20:53-61.

Piza, Eric L. and Victoria A. Sytsma. 2016. “Exploring the Defensive Actions of Drug Sellers in Open-air Markets—A Systematic Social Observation.” Journal of Research in Crime and Delinquency 53:36-65.

Polsky, Ned. 1969. Hustlers, Beats, and Others. New York: Doubleday.

Pratt, Travis C., Francis T. Cullen, Kristie R. Blevins, Leah E. Daigle, and Tamara D. Madensen. 2006. “The Empirical Status of Deterrence Theory: A Meta-Analysis.” Taking Stock: The Status of Criminological Theory 15:367-396.

Presser, Lois, and Sveinung Sandberg (eds.). 2015. Narrative Criminology: Understanding Stories of Crime. New York: New York University Press. 

Reynald, Danielle. 2012. “Guardians on Guardianship: Factors Affecting the Willingness to Supervise, the Ability to Detect Potential Offenders, and Willingness to Intervene.” Journal of Research in Crime & Delinquency 47:358-390.

Shaw, Clifford R. 1930. The Jack-Roller: A Delinquent Boy’s Own Story. Chicago: University of Chicago Press.

Sutherland, Edwin H. 1937. The Professional Thief. Chicago: University of Chicago Press.

_____. 1940. “White-Collar Criminality.” American Sociological Review 5:1-12.

Van Maanen, John. 1981. “The Informant Game: Selected Aspects of Ethnographic Research in Police Organizations.” Urban Life 9:469-494.   

Weis, Joseph G. 1986. “Issues in the Measurement of Criminal Careers.” Pp. 1-51 in Criminal Careers and “Career Criminals”, edited by Alfred Blumstein, Jacqueline Cohen, Jeffrey A. Roth, and Christy A. Visher. Washington, DC: National Academy Press.

Whyte, William Foote. 1943. Street Corner Society: The Social Structure of an Italian Slum. Chicago: University of Chicago Press.

Wixted, John T. 2004. “The Psychology and Neuroscience of Forgetting.” Annual Review of Psychology 55:235-269.

Wright Richard, Scott Decker, Alison K. Redfern, and Dietrich L. Smith. 1992. “A Snowball’s Chance in Hell: Fieldwork with Active Residential Burglars.” Journal of Research in Crime & Delinquency 29:148-161.

Wright, Richard, and Scott Decker. 1994. Burglars on the Job: Streetlife and Residential Break-Ins. Boston: Northeastern University Press.

_____. 1997. Armed Robbers in Action: Stickups and Street Culture. Boston: Northeastern University Press.

Figure 

Figure 1. Visualization of How Data Possessed by Source is Distinct yet Relevant to Data Obtained by Researcher from Source

Notes

[1] Some readers may think it is obvious that offenders know more than nonoffenders, and, thus, a detailed examination of how they compare as such is unnecessary. But I think that perception is incorrect for a few reasons. One, even “common sense” needs theoretical explanation because otherwise the mechanisms remain dark (see Geertz, 1983). Two, there are plenty of “obvious” ideas that have proved difficult to verify; the effect, or lack thereof, of formal deterrence measures on crime is a case in point (Pratt et al., 2006). Three, not everyone would agree that criminals obviously possess more and higher quality data on crime; for example, police officers, prosecutors, and judges may think they know more than criminals.

[2] For the sake of illustrating how internal, interpersonal, and external factors may affect data, consider some examples from a study I did of active robbers. One internal factor is intoxication, which usually is seen as hurting the amount and quality of data obtained (Akerstrom 1985; Copes et al. 2015; Weis 1986). This became evident to me after interviewing, sort of, a robber at a bar. Upon our arrival there, I bought him a pint, which he chugged down. He asked for another, which he also chugged. Then he went to the bathroom and, afterwards, walked out the front door, rather than finish the interview. Perhaps the most important interpersonal factor is how closely tied a researcher is to potential subjects, including their overlapping ties, because this is thought to enhance trust (Agar 1973; Bernasco 2010; Jacques and Wright 2008). Though it took me almost no time to locate and interview the aforementioned robber, I had no success at recruiting active robbers over the following two months. This changed when I had a chance encounter with someone whom I had interviewed a couple years back for another study. Not only did he recruit 28 robbers into the study, but he would sit with me during interviews and call out participants’ “bullshit” to help get their answers closer to the truth. External factors are significant, too. Where offenders are interviewed, such as in- or outside prison, is thought to affect how much and what they say (Bernasco 2010; Copes et al. 2015; Polsky 1969; Wright et al. 1992). This is seen in how I had a very difficult time getting active robbers to sit still for even an hour long interview, but my colleague was able to interview inmate robbers for hours on end. However, and unlike the active robbers, the inmates were largely unwilling to discuss crimes for which they had not been apprehended.

[3] How to define the span of a case is up to the analyst, but it could start and end with interaction in a situation, the life of an individual, the calendar year of a community, or any other frame.

[4] I also meant that it is better to start with one sort of data (the empirical type) than two types (e.g., empirical and subjective data) because starting “too big” is more likely to produce invalid statements by complicating the matter.

Comments
0
comment
No comments here
Why not start the discussion?