The past quarter century has witnessed the emergence of a rich methodological literature devoted to various ways of tapping into the offender’s perspective on crime. Whatever its virtues, that literature has remained almost wholly atheoretical. We recently introduced a ...
Note: This is the postprint of the following paper; publisher version available here.
Jacques, Scott, and Richard Wright. 2010. Apprehending Criminals: The Impact of Law on Offender-Based Research. Chapter 3 in Offenders on Offending: Learning About Crime from Criminals, ed. Wim Bernasco. Cullompton, UK: Willan Publishing.
Abstract: The past quarter century has witnessed the emergence of a rich methodological literature devoted to various ways of tapping into the offender’s perspective on crime. Whatever its virtues, that literature has remained almost wholly atheoretical. We recently introduced a preliminary theory of research grounded in the perspective of pure sociology. In this chapter we seek to extend that theory by examining how law and normative status affect offender-based research. We argue that as more law is applied to actors (i.e., as the normative status of persons or groups declines), the probability that those actors are recruited for offender-based research increases, the amount of remuneration provided to them for participation decreases, and the quality of data obtained from them decreases. We conclude by offering theoretically situated, practical advice about the ways in which criminologists might maximize data while minimizing costs associated with recruitment and remuneration.
Criminologists require data collected via fieldwork, interviews, and experiments with offenders in order to understand fully the factors that influence lawbreaking (see, e.g., Brookman et al., 2007; Carr, 2005; Copes, Forsyth, and Brunson 2007; Curtis, 2003; Jacobs and Wright, 2006; Jacques, forthcoming; Jacques and Wright, 2008a, 2008b; Morselli, 2001; Nee and Meenaghan, 2006; Shover, 1996; St. Jean, 2007; Venkatesh, 2006; Zhang, Chin, and Miller, 2007). Research has shown conclusively that criminals know things about crime that others do not (see, e.g., Wright, Logie, and Decker, 1995; Logie, Wright, and Decker, 1992), and this makes them a valuable source of information about lawbreaking. The challenge for methodologists and criminologists is to determine how best to access that information. What sort of research strategies yield the most valid and plentiful data at the lowest cost?
The past quarter century has witnessed the emergence of a rich methodological literature devoted to various ways of tapping into the offender’s perspective on crime (see, among many others, Adler, 1990; Dunlap and Johnson, 1999; Ferrell and Hamm, 1998; Glassner and Carpenter, 1985; Jacobs, 1998, 2006; Mieczkowski, 1988; Williams et al., 1992; Venkatesh, 2008; Wright et al., 1992). Whatever its technical merits, this literature has remained strikingly atheoretical. The advancement of methodology and criminology are slowed by atheoretical ideas because they are difficult to falsify through empirical research (see Popper, 2002).
Without a theory of offender-based research, the task of finding the most economical strategy for accessing information held by criminals is, at best, foggy and inefficient or, at worst, practically impossible. This is true because if recommendations for improving research are not deduced from a falsifiable theory, then those recommendations cannot be shown to be wrong with empirical data. And if recommendations cannot be falsified, then bad advice may linger longer than it should and good advice may not be adopted at the rate or magnitude it deserves.
With the limits of atheoretical method in mind, we recently introduced a preliminary theory of research (Jacques and Wright, 2008c). In this chapter, we seek to extend our theory by examining how law and normative status affect offender-based research. In turn, we review the current state of our conceptualizing and theorizing on offender-based research, define the concepts of law and normative status, hypothesize the impact of those variables on offender-based research, and conclude by offering theoretically situated, practical advice about the ways in which criminologists might maximize data while minimizing costs associated with recruitment and remuneration.
There are at least two first steps in developing a theory of offender-based research. The first step is to specify what is being explained. What is “offender-based research”? The next step is to make predictions regarding how offender-based research is affected by any given variable. What causes “offender-based research”?
Research, or method, is a social behavior defined as the process of recording or analyzing information. Research is a quantitative variable: it increases with every additional datum collected or analyzed. Because research can be quantified, measured, and empirically observed, it is theoretically explainable. In other words, we can predict what situations, persons, and groups are most likely to be involved in surveys, interviews, experiments, observations, or analyses.
Offender-based research is defined here as the process of obtaining information about crime through conversation with or observation of persons who have personally engaged in crime (at least once). More narrowly, there are at least three distinct parts of collecting information from offenders: recruitment; remuneration; and data collection (Jacques and Wright, 2008c; also see Dunlap and Johnson, 1999; Wright et al., 1992).
Recruitment is defined as the process of interacting with criminals and convincing them to provide data. Remuneration is defined as payment for participation in research. Data collection, or data quality, is defined as the process whereby knowledge about crime is ascertained by criminologists through conversation with or observation of criminals.
Defined thusly, offender-based research is a social, empirical, and quantifiable variable measured by (1) the number of recruitments or the amount of effort (e.g., time) spent on recruitment, (2) the number of remuneration payments or the amount of remuneration provided, and (3) the number of data collection sessions (e.g., interviews) or the amount of valid data collected. More recruitment of, remuneration of, or data collection with offenders is equivalent to more offender-based research.
Although offender-based research is first and foremost a means of knowledge production, it is also a quantitative variable and therefore it should be theoretically explainable. This raises the question: What factors affect the amount of recruitment, remuneration, and data quality in offender-based research across situations, people, or groups? Although many theoretical perspectives may have the ability to explain research, our theory draws on the orienting paradigm known as pure sociology (see, especially, Black, 1976, 1995, 1998, 2000; Cooney, 2006).
The goal of pure sociology is to understand how the quality (type) and quantity (amount) of social behavior is affected by social structure. Social structure is defined as the relative social status of and also social distance between every actor in a particular situation.
Actors, meaning individuals or groups, vary in social status, meaning their position in a social hierarchy. There at least five forms of social status: vertical status, measured by wealth; radial status, measured by community involvement and production; corporate status, measured by memberships and organization; symbolic status, measured by knowledge and conventionality; and, normative status, measured by freedom from social control. Higher social status is equivalent to more wealth, community involvement, production, memberships, organization, knowledge, conventionality, and freedom.
Actors also vary in social distance. Whereas social status is the characteristic of one actor, social distance may be thought of as the characteristic of (at least) two actors. There are at least three forms of social distance: relational distance, measured by the amount of interaction between two actors; corporate distance, measured by the number of common memberships; and, social distance, measured by the number of cultural similarities. Closer social distance is equivalent to more prior interaction between actors, more common memberships held by them, and more similarities in their ideas and forms of expression.
Pure sociologists attempt to specify how variability in the social status of and social distance between actors involved in a situation affects their social behavior. There are many different kinds of social behavior. Broadly speaking, however, pure sociology is used to explain the social behaviors of resource transfer and accumulation, interaction, organization, culture, and social control. In short, the goal is to determine how variability in social structure causes variability in the kind and amount of social behavior that occurs.
For instance, Black theorizes (1976) that – all else equal – crimes by lower-status persons against higher-status persons are more likely be punished by the government than are similar crimes by higher-status persons against lower-status persons. Related to this idea, Black (1983) also argues that retaliation – or justice without the government – is more likely to occur as access to law decreases, and so the social structures least likely to lead to law are most likely to result in vengeance. The perspective of pure sociology has also been used to explain a number of other behaviors, including avoidance (Baumgartner, 1988), retaliation (Black, 1983, 1998; Cooney, 1998), ideas (Black, 2000), welfare (Michalski, 2003), art (Black, 1998: 168-9), predation (Cooney, 2006), genocide (Campbell, 2009), and victimization related to active offender research (Jacques and Wright, forthcoming).
As a theoretical perspective, the value of pure sociology is it generality, simplicity, testability, and originality (for details, see Black, 1995). The paradigm is original because it intentionally disregards what other sociological perspectives find key: motivations and emotions. By disregarding those factors, purely sociological theories become more testable because motivations and emotions are subjective and so beyond direct empirical measurement. Although motivations and emotions often are construed to explain “why” behavior happens, the concepts of motivation and emotion are, at best, unobservable or, at worst, nothing real at all; thus their inclusion in theoretical models makes them less simple and, again, less testable than theories that eschew such explanatory variables. Perhaps in the end, however, the ultimate question is whether any given theory is valid. Although empirical data and statistics can be used to test pure sociological theories and directly falsify them, theories concerned with motivations and emotions can never be directly falsified because those factors are subjective. Therefore, the validity of such theories will forever remain opaque.
An important point to clarify is that even if a purely sociological theory – such as Black’s (1976, 1983) well-known theory of law or theory of self-help – was falsified through empirical testing, this does not mean that the purely sociological perspective is falsified or invalid. A perspective, or paradigm, is a way of conceptualizing the world and a strategy for explaining it (e.g., social structure). On the other hand, a theory is a statement that specifies how concepts are causally related to each other. Thus, a purely sociological theory may be incorrect (falsified) if it does not significantly predict how sociological factors affect each other, but this does not necessarily mean that those sociological factors do not exist or affect each other in different ways. For example, even if more intimate relationships do not in fact reduce retaliation (which is a theory, see Black, 1983), this does not necessarily mean that there are no real things such as “intimacy” and “retaliation” (which are concepts) or that they have no causal relationship.
We have begun to develop a preliminary “theory of method”, or “theory of research”, nested in the paradigm of pure sociology (Jacques and Wright, 2008c, forthcoming). Although the theory we are developing may eventually be falsified, our ultimate goal is to stimulate a debate in science, especially criminology, which leads to greater theoretical understanding of research. As discussed above, a valid theory of research would be useful because its implications could be adopted to improve research and, in turn, the findings that emerge from it.
In our initial attempt to explain research, we restricted our theoretical attention to active offender research and the explanatory variable known as relational distance, or intimacy. By definition, the relational distance between a researcher, liaison (e.g., a recruiter), and offender decreases as the amount of prior contact between them increases (see Black, 1976). Two persons, for example, who drink alcohol or smoke marijuana together are closer in relational distance to each other than are two people who do not. Likewise, two persons who have a mutual friend, acquaintance, or colleague are closer in relational distance than two people without such a contact. We addressed the question: How does relational distance affect the process of obtaining information from unincarcerated criminals?
Our own experiences in recruiting, paying, and interviewing drug dealers (Jacques and Wright, 2008c) alongside those of other criminologists (Bourgois, 2003; Hoffer, 2006; Mieczkowski, 1988) led to the following hypotheses: As the relational distance between a researcher, liaison, and offender decreases, (1) the likelihood that the offender will be recruited increases, (2) the amount of remuneration provided to the offender decreases, and (3) the quality of data collected increases. For example, our theory predicts that a researcher’s criminal-friend or family member is more likely than a stranger to agree to participate in research, to do so for little or no compensation, and to be honest.
We concluded our initial attempt at theorizing method by observing that relational distance is not the sole factor influencing method. Among other possibilities, we speculated that the quantity of law applied to offenders represents an especially strong candidate in this regard (Jacques and Wright 2008c:34). In this chapter, we seek to extend our preliminary theory of research by examining how law and normative status affect recruitment, remuneration, and data quality in offender-based research.
Social control refers to “the normative aspect of social life. It defines and responds to deviant behavior… Social control is found wherever and whenever people hold each other to standards [whether] on the street, in prison, [or] at home” (Black, 1976:105). Social control is found among many living organisms including, for example, insects (Wenseleers and Ratnieks, 2006), non-human primates (de Waal, 1989), archaic tribes (see Boehm, 2000), modern suburbanites (see Baumgartner, 1988), and gang members (see Decker and Van Winkle, 1996). The concept of social control encompasses many kinds of behaviors (see Black, 1998; Horwitz, 1990; Jacques and Wright, 2008b), such as law, retaliation, avoidance, negotiation, humiliation, gossip, protest, apology, and suicide.
Law is “governmental social control” (Black, 1976:2). Put differently, law “is…the normative life of a state and its citizens, such as legislation, litigation, and adjudication” (p. 2). We can theorize law because it is a quantitative variable – some situations involve more law than others. Law increases with every additional prohibition, arrest, indictment, prosecution, fine, or imprisonment.
Actors vary in the amount of law that has been applied to them in the past. Law “divides people into those who are respectable and those who are not” (Black 1976:105, emphasis added). Normative status, or what is also called respectability, is a “record” of an actor’s subjection to social control (Black, 1976, 1998). The more law that has been applied to an actor in the past, the lower that actor’s formal normative status. For example, a person who has been arrested, indicted, prosecuted, fined, or imprisoned is lower in normative status than someone who has not been subjected to those kinds of formal social control.
Law and normative status are important variables to explain, but they are also explanatory factors that influence social life. Theory and research suggest, for instance, that actors who are poor, unemployed, unmarried, and uneducated not only are more likely to be punished with formal control, but also that actors who have been punished by formal control are subsequently more likely to be poor, unemployed, unmarried, and uneducated (Black, 1976, 1998; Cooney, 1998; Western, 2006; Wilson, 1987, 1996). The point here is that a number of factors determine the variable amounts of law applied across situations, persons, and groups, and that this variability in the application of law has consequences for subsequent behavior.
We now propose a theory comprised of a series of propositions that use the explanatory variable of formal normative status to predict variability in offender-based research. To empirically illustrate the theory and suggest how to test it, the propositions are directly followed by hypotheses – or empirical predictions – about the way in which normative status, as measured by institutionalization, affects recruitment, remuneration, and data quality.
There are many measures of formal normative status, such as the number of times a person has engaged in crime or been arrested, prosecuted, or punished, or the amount of punishment received in any particular case (Black, 1976). For purposes of clarity and brevity, however, this chapter focuses on normative status as measured by government institutionalization, such as confinement in a jail or prison. All else equal, a person who is confined in a government institution has lower normative status than a person who is free. In other words, criminals who are free have higher normative status than those who are institutionalized, all else equal (e.g., their arrest record or the amount of time they are institutionalized).
To be clear, the purpose of this theoretical exercise is to stimulate a discussion regarding which empirically measurable factors have a statistically significant and substantial impact on the various parts of offender-based research.
Recruitment is the process by which a researcher goes about locating criminals and convincing them to cooperate in data collection. Recruitment is a quantitative variable measurable by the amount of social interaction devoted to convincing actors to participate in research. Recruitment can last minutes, days, or longer, and the number of actors recruited could be one, one hundred, or more.
Most offender-based research relies on criteria-based sampling to locate criminals, meaning only those individuals who possess the social, psychological, or biological characteristics relevant to research are recruited. The recruitment of offenders for research can take place within the walls of a government institution or in the free world. In both settings, researchers often depend on a liaison, or “broker,” to recruit criminals for participation. In government institutions, a researcher relies on what could be called a “gatekeeper,” that is, a person who works for a government institution and has the “right” to grant researchers access to persons residing therein. In the free world, a researcher depends on what is commonly called a “recruiter,” who is in a position to connect criminals to researchers because of her or his personal and professional ties to both groups.
Just as there are two locales in which criminals can be recruited, so too there are two “kinds” of criminals: institutionalized and “free” (i.e., not institutionalized). As relates to law and normative status, institutionalized criminals by definition are lower in normative status than free criminals because they are subject to more law (holding constant other past subjections to law). This conceptual distinction raises the empirical and theoretical question: Does normative status, as measured by institutionalization, affect recruitment in offender-based research?
We speculate that it does and offer the following proposition: Recruitment to offender-based research increases as the normative status of a criminal decreases. In other words, a criminal is more likely to be recruited for research as more law is applied to that criminal (cf. Black, 2000: 350). We hypothesize, then, that offender-based research has more recruited institutionalized criminals than free ones (see figure 1).
--FIGURE 1 ABOUT HERE--
The proposition and hypothesis predict that throughout the history of offender-based research, there have been more recruitments of institutionalized than free criminals. For studies that collect data from institutionalized criminals or free criminals (but not both), the above proposition predicts institutionalized criminals are more likely to be recruited. And for studies that collect data from both kinds of offenders, the proposition predicts a larger number of recruited institutionalized offenders per study and on average.
Remuneration refers to objects and services given to a criminal in return for participation in research. Remuneration is perhaps the most straightforward, generalizable, and successful strategy for convincing persons to provide data (see Dunlap and Johnson, 1999). Like recruitment, remuneration is a quantitative variable, and is measurable by the amount of resources (objects or services) provided to a criminal as compensation for taking part in research. For instance, a researcher can compensate a participant with no money, $50, or $100.
Does normative status, as measured by institutionalization, affect remuneration in offender-based research? We believe that it does and suggest the following proposition: Remuneration for participation in offender-based research increases as the normative status of a criminal increases. Stated differently, criminals will receive greater remuneration as less law is applied to them. As relates to institutionalized versus free criminals, then, we hypothesize that the latter group typically receives more compensation than the former for participation (see figure 2).
--FIGURE 2 ABOUT HERE--
The above proposition and hypothesis predict that, on average, institutionalized participants have been remunerated less than free-ranging ones. For studies that collect data from institutionalized criminals or free criminals, the above proposition predicts a larger average remuneration fee in studies of free criminals. For studies that collect data from both kinds of criminals, the proposition predicts larger remuneration for the free-ranging participants, both on average and per study.
The time, effort, and resources devoted to recruitment and remuneration are squandered if the resultant data are sparse or false. Data quality is defined as the amount of valid information collected from an offender, and is a quantitative variable measurable by the amount of data produced and the truthfulness of that information. For example, some offenders never stop talking but others never start, and some interviewees never tell the truth, others tell half of it, and still others are almost wholly forthright and honest.
Does normative status, as measured by institutionalization, affect the quality of data produced in offender-based research? We hypothesize that it does and offer the following proposition: Data quality in offender-based research increases as the normative status of a criminal increases. What this means is that a criminal provides more plentiful and valid data as less law is applied to that criminal. If we compare free criminals to those who are institutionalized, we hypothesize that institutionalized offenders provide less valid data than do free-ranging ones (see figure 3).
--FIGURE 3 ABOUT HERE--
The proposition predicts that, on average, institutionalized research participants have provided lower quality data than have free participants. For studies that collect data from institutionalized criminals or free-ranging ones, the above proposition predicts that per participant, studies of free offenders produce higher quality data. For studies that collect data from both kinds of criminals, the proposition predicts that per participant, less valid and truthful data are provided by the institutionalized participants than by the free-ranging ones.
Above, we propose a preliminary theory of how research with offenders “behaves.” Actors’ normative status increases as less law is applied to their behavior, and we have addressed the question: How do law and normative status affect recruitment, remuneration, and the quality of data produced in offender-based research? We theorize that as criminals become lower in normative status then they become more likely to be recruited for research, less likely to receive remuneration if they are recruited, and less likely to provide valid and plentiful data. Empirically, speaking, this theory suggests the following hypotheses: compared to free criminals, institutionalized offenders are more likely to be recruited for research, to receive less remuneration, and to provide less valid information.
It is important to note, however, that the purpose of this chapter is to stimulate an academic debate surrounding what factors influence the quality and quantity of research, especially of the offender-based variety. With that said, it is obvious that there is much more to understanding variability in recruitment, remuneration, and data quality in research than can be explained solely as a function of formal normative status and government institutionalization.
Therefore, substantial work remains to be done by theorists and researchers before methodologists, including criminologists, will have produced a clear picture of what factors influence variability in research. At present, advancing a theoretical understanding of method requires theorists to develop explanations that can be falsified empirically, and researchers to put those theories to the test with empirical data and statistical analyses.
Science advances by falsifying or supporting theory through empirical testing (Popper, 2002). As mentioned above, the two first steps in developing a theory of offender-based research are to specify what it is and then to make predictions about what causes it. The next step in developing a theory of method is testing it with empirical data and then refining or discarding it based on those findings, if necessary.
At present, we – as a field – have practically no readily available empirical information bearing on how offender-based research behaves. Of course, it would have been possible for us to provide anecdotal evidence for our proposed hypotheses (see, e.g., the history of burglary research provided by Nee, this book), but doing so would be haphazard. We urge researchers to compile a “purpose built” data set and use statistical analyses to determine what empirical factors do or do not influence offender-based research.
Throughout this chapter we have alluded to how our theory might be tested: count and statistically analyze whether, in comparison to free offenders, institutionalized offenders are (1) more likely to be recruited for research, (2) less likely to be remunerated for participating, and (3) more likely to provide lower quality data. There are many conceivable research designs through which empirical data bearing on these hypotheses might be obtained.
Although we do want to stymie researchers’ creativity, allow us to outline two possible methods for testing our theory of method:
Survey criminologists to collect information on the following:
The number of institutionalized criminals and free criminals recruited for research interviews in a specific time span, such as over the past calendar year.
The theory is supported if, controlling for all other influences (including their relative population size), institutionalized criminals are, statistically speaking, significantly more likely than free criminals to be recruited for research.
The theory is falsified if institutionalized criminals are less likely than free criminals to be recruited or if there is no difference.
For each interview, the amount of remuneration offered to the participant concerned.
The theory is supported if, controlling for all other influences (including the quality of the data provided), institutionalized criminals receive significantly less remuneration than free criminals for participation.
The theory is falsified if institutionalized criminals receive more remuneration than free criminals or if there is no difference.
For each interview, its length.
The theory is supported if, controlling for all other influences (including the amount of remuneration), interviews with institutionalized criminals are significantly shorter than interviews with free criminals.
The theory is falsified if interviews with institutionalized offenders are longer or if there is no difference.
For each interview, the number of questions asked by the researcher(s) in relation to the number answered by the participant(s).
The theory is supported if, controlling for all other influences (including the amount of remuneration and time length), interviews with institutionalized offenders have a significantly lower rate of response than interviews with free criminals.
The theory is falsified if interviews with institutionalized offenders have a higher rate of response or there is no difference.
Such a research design would provide empirical evidence on how institutionalization affects recruitment, remuneration, and data quantity in the real world of offender-based research.
What that research design would not tell us, however, is whether institutionalization affects data validity. Given the problems inherent in measuring validity (especially in a field such as criminology), an experimental design is probably the best option for determining the validity of our theory of method. One conceivable experiment is outlined below:
To control for the influence of recruitment, remuneration, and data quantity:
Use purposive sampling to recruit one armed robber who is institutionalized and one who is free, and then use snowball sampling to recruit a sample of institutionalized and free armed robbers (e.g., a total of 30 for each group).
Participants will not receive remuneration directly, but $100 in cash will be delivered to a non-institutionalized person of their choosing following completion of the interview.
Promise all participants that the interview will take approximately 5 minutes to complete and consists of four questions. To avoid or minimize travel costs and time for the participant, the interview can take place at any time and place of their choosing (given necessary constraints, such as being inside the institution).
To determine the relationship between institutionalization and data validity, ask each participant the following questions:
What is your birth name?
In which city were you born?
What is your date of birth?
What is your mother’s year of birth?
Using the answers to those questions, attempt to obtain a birth certificate for each participant. If an adequate number of birth certificates can be obtained for both institutionalized and free armed robbers and they do not significantly differ in this regard, then a test of validity may be made:
The theory is supported if, controlling for all other influences, institutionalized offenders provide a significantly less valid – meaning less accurate – answer regarding their mother’s year of birth than do free criminals.
The theory is falsified if institutionalized criminals provide a significantly more accurate answer regarding their mother’s year birth than do free criminals or there is no difference.
Such an experiment is just an example of how to test whether institutionalization affects the validity of findings in offender-based research. Whatever research design is used to test our theory of validity, the key will be to ask questions that have independently verifiable answers, such as those found on birth certificates or in official (and accurate) records.
Before tests take place, however, we recommend the development of theories – from various perspectives – to guide the process.
Testing requires controlling for extraneous factors. Therefore, when testing theories of method it is important to control for the predictions of alternative perspectives. Unfortunately, and as discussed above, there are practically no other theories of offender-based research. As far as we are concerned, it is better to have empirically untested (yet falsifiable) theories of offender-based research than it is to have no such theories at all. Before the study of offender-based research can progress with tests, we must first create theories to test. This is not a venture that can be accomplished overnight. Yet until accomplished, research on offender-based research will be crippled for lack of the orienting force of theory. It is analogous to searching in the dark.
What we are saying here is that the field is best off if we first produce theories of offender-based research and then test them with empirical data and statistical analyses. We can no longer wait for theory development, but holding off on formal testing is perhaps prudent at this time because theory on offender-based research is virtually non-existent. Economically speaking, it is less costly to develop and then test theories than it is to collect data, mine them to form a theory, and then test that theory with other data. This is especially true when you consider that the invalidation of theories – so long as they are falsifiable – provides knowledge by telling us what factors do not influence a particular behavior, such as research.
The necessity of theory development before testing becomes clear when considering the implications of perspectives for the factors being explained. Perspectives are abstract ways of understanding the empirical world around us. Whereas theories differ in their predictions, perspectives differ in their conceptualizations and strategies for forming predictions. The way a perspective conceptualizes the world will affect how we define offender-based research and its component parts. For instance, a purely sociological theory of research defines “data quality” as an entirely objective measure: the amount of information (whether obtained through sight, taste, small, touch, or hearing) that is true, meaning congruent with empirical, social events as they actually happened. Because (1) pure sociology is only concerned with objective, social variables, and (2) information on those variables is obtained through empirical observations (including communication with offenders), then data quality is entirely objective according to this perspective – it is the amount of information that accurately describes what actually happened (validity).
On the other hand, there are other perspectives, such as ethnography, that view “validity” as subjective and not empirical. Thus, a theory of validity can never itself be valid because there is no such thing as validity (see footnote 3, above). If a theoretical perspective takes the view that there is no such empirical thing as validity, then a theory of offender-based research grounded in that perspective will, perhaps, have a qualitatively different set of behaviors to explain than the ones explained in this chapter (recruitment, remuneration, and data quantity and validity).
Theses conceptual differences make clear that a theoretical understanding of offender-based research must first be preceded by definitions. Not all perspectives will have the same conception of offender-based research. Methodologists and criminologists should view the behavior of offender-based research through the “lens” of various perspectives in order to define it so that theories grounded in those perspectives can be developed. Not only different theories – but also different conceptions – of research may serve to improve our scientific understanding of how data is obtained.
The theory proposed in this chapter is preliminary and builds on our earlier work (Jacques and Wright, 2008). When combined with our initial work on the impact of relational distance on offender-based research (Jacques and Wright, 2008c), our current theory is comprised of six propositions that explain three behaviors: (1) Recruitment to a study increases as criminals lose normative status and become closer in relational distance to researchers. (2) Remuneration for participation increases as criminals gain normative status and increase in relational distance from researchers. And, (3) data quality increases as criminals gain normative status and reduce relational distance from researchers.
There is much more to social structure than normative status or relational distance, and so future work should specify what other factors may affect the quality and quantity of method. For instance, variability in knowledge – which is an aspect of symbolic status – stands out as being particularly promising. After all, the goal of science is to produce knowledge.
Knowledge about crime is what criminologists are attempting to generate, and offender-based research is one way of doing so. Knowledge is a quantitative variable, and is measurable by the total and relative number of ideas a person possesses (Black, 1976). For instance, some people know more things than others (e.g., the difference between a professor and a college student), and some things are more widely known than others (e.g., more people know how to drive a car than fly a plane).
It seems to us that research may behave as a function of knowledge. For example, one conceivable hypothesis is that recruitment, remuneration, and data quality increase as the knowledge of a criminal increases. In other words, the more a criminal knows about a certain behavior (e.g., robbery or murder), and the less others know about that behavior, the more likely that criminal is to be recruited for research, to be paid more for doing so, and to provide a substantial amount of truthful information. This hypothesis predicts, for instance, that a murderer may receive greater remuneration for participation in research than a robber for the reason murderers and murders are less common than robbers and robbery.
The idea that knowledge affects method is important because, among other things, the effect of normative status on method may otherwise be distorted when analyzed empirically. If knowledge is not simultaneously considered, for example, it might appear that the relationship between normative status and remuneration is U-curved rather than linear (see figure 4).
--FIGURE 4 ABOUT HERE--
It seems eminently plausible that convicted criminals typically are paid less than non-convicted persons for doing the exactly same work. This pattern in life, we suggest above, also applies to remunerating criminals in offender-based research. At the same time, however, we might ask: Why would a person who has never committed a crime be paid for providing information about how to commit crimes? The proposition above suggests that because non-criminals know less about crime – and not because they have more normative status – they are relatively unlikely to be recruited and remunerated because they are relatively unlikely to provide useful data.
All of the above is simply to say that the purely sociological theory of research is in its infancy and requires further elaboration for the total effect of social structure to be understood.
Related to this point, it is important to recognize that there are many different ways of operationalizing the concept of “normative status.” This chapter has focused on how institutionalized offenders are more likely than free criminals to be recruited for research, less likely to receive remuneration for taking part, and less likely to provide high quality data. However, our theory of normative status and offender-based research also predicts, for instance, that a person who has committed 100 robberies is more likely to be recruited for research than is a person who has committed just one robbery; a criminal who has been arrested many times should receive less remuneration for participation than does an offender who never been arrested; and, a murderer is less likely to provide a plentiful amount of valid information about her or his crime than is a burglar. The point here is that different operationalizations of normative status may lead to different albeit similar predictions, and that the full range of these predictions should be specified and examined in future work.
Although our theory is grounded in the paradigm of pure sociology, there are other theoretical perspectives that could be and should be employed to explain method. To be sure, method is a complex behavior that is likely affected by a broad range of factors, such as rationality, learning, and opportunity, among others. Methodologists and criminologists should explore the potential of all theoretical perspectives for understanding and improving the process that produces much of the data used in criminological research. We must begin to take atheoretical ideas about how method behaves and turn them into falsifiable predictions that can be tested with empirical research.
Copes and Hochstetler (this book), for instance, suggests a number of hypotheses (some more falsifiable than others) that may explain variation in data quality. Their review suggests that the quantity and validity of data obtained from offenders may increases as (1) therapeutic benefits for participation increase, (2) perceived restitution for participation increases, (3) free time for an offender increases (i.e., as boredom increases), (4) admiration held by an offender for a researcher increases, (5) costs for participation decrease or benefits increase, (6) trust for a researcher held by an offender increases, (7) social distance between a researcher and government officials increases, (8) intoxication of an offender decreases, and (9) the amount of time between an offender’s action and researcher’s questions about the action increases.
Other contributions to this book may move the field towards a comprehensive theory of offender-based research. For example, Shover and Hunter (this book) provide good reasons for why the validity of data provided by white-collar criminals may vary substantially from the data provided by lower-class criminals. Zhang (this book) is in some respects suggesting that as the amount of guanxi between a researcher and participant increases, then so too should the validity and quantity of information obtained from that participant. And Miller’s (this book) insights provide a starting point for a gendered theory of offender-based research (but also see Weinreb, 2006). These theoretical alternatives should be fully developed.
With that in mind, we remind ourselves that an important challenge for criminologists is to determine how best to access the undoubtedly valuable information possessed by criminals. What sort of methodological strategies yield the most valid and plentiful data at the lowest cost? The idea that method can be theorized is important because explanations of method can be used to generate practical strategies for improving it, whether by increasing recruitment success, reducing remuneration payments, or enhancing data quality.
Allow us to conclude, then, by suggesting three pieces of practical advice deduced from the theory developed in this chapter:
Criminals can be more easily recruited from institutions than from the free world.
Criminals can be paid less for providing data in an institution than in the free world.
Criminals provide more valid data when they are free than when they are institutionalized.
These practical suggestions tell us that while it may be easier and cheaper to interview institutionalized criminals, on the whole they may be less valuable informants than active criminals.
Whatever the validity of the above assertions, the key takeaway point is that because this chapter makes falsifiable predictions regarding the effect of law on offender-based research, the practical advice suggested herein can be discarded or adapted according to the findings of empirical research that tests the theory. Before testing our theory, however, we hope that competitors will emerge who provide their own falsifiable theories that shed additional light on the behavior of offender-based research.
Adler, P. 1990. “Ethnographic Research on Hidden Populations: Penetrating the Drug World.” NIDA Monograph 98:96-111.
------. 1993. Wheeling and Dealing: An Ethnography of an Upper-Level Drug Dealing and Smuggling Community, second edition. New York: Colombia University Press.
Baumgartner, M. P. 1988. The Moral Order of a Suburb. New York: Oxford University Press.
Black, D. 1976. The Behavior of Law. New York: Academic Press.
------. 1983. “Crime as Social Control.” American Sociological Review 48:34-45
------. 1995. “The Epistemology of Pure Sociology.” Law and Social Inquiry 20:829-870.
------. 1998. The Social Structure of Right and Wrong, revised edition. San Diego, CA: Academic Press.
------. 2000. “Dreams of Pure Sociology.” Sociological Theory 18(3):343-367.
Boehm, C. 2000. “Conflict and the Evolution of Social Control.” In Evolutionary Origins of Morality: Cross-Disciplinary Perspectives, ed. Leonard D. Katz. Exeter, UK: Imprint Academic, Ltd.
Bourgois, P. 2003. In Search of Respect: Selling Crack in El Barrio, second edition. New York: Cambridge University Press.
Brookman, F., C. Mullins, T. Bennett, and R. Wright. 2007. “Gender, Motivation and the Accomplishment of Street Robbery in the United Kingdom.” British Journal of Criminology 47(6):861-884
Campbell. B. 2009. “Genocide as Social Control.” Sociological Theory 27:150-172.
Carr, P. J. 2005. Clean Streets: Controlling Crime, Maintaining Order, and Building Community Activism. New York: New York University Press.
Copes, H., Forsyth, C. J. and Brunson, R. K. 2007. “Rock Rentals: The Social Organization and Interpersonal Dynamics of Crack-for-Cars Transactions in Louisiana, USA.” British Journal of Criminology 47(6):885-899.
Copes, H., and Hochstetler, A. This book. “Interviewing the Incarcerated: Pitfalls and Promises.” In Offenders on Offending: Learning About Crime from Criminals, ed. Wim Bernasco. Cullompton, UK: Willan Publishing.
Cooney, M. 1998. Warriors and Peacemakers: How Third Parties Shape Violence. New York: New York University Press.
------. 2006. “The Criminological Potential of Pure Sociology.” Crime, Law, and Social Change 46:51-63.
Curtis, R. 2003. “Crack, cocaine and heroin: Drug eras in Williamsburg, Brooklyn.” Addiction Research and Theory 11(5):349-366.
Decker, S., and Winkle, B. V. 1996. Life in the Gang: Family, Friends, and Violence. New York: Cambridge University Press.
Dunlap, E., and Johnson, B. D. 1999. “Gaining Access to Hidden Populations: Strategies for Gaining Cooperation of Drug Sellers/Dealers and Their Families in Ethnographic Research.” Drugs and Society 14:127-149.
Ferrell, J., and Hamm, M. S. 1998. Ethnography at the Edge: Crime, Deviance, and Field Research. Boston, MA: Northeastern University Press.
Glassner, B., and Carpenter, C. 1985. The Feasibility of an Ethnographic Study of Property Crime: A Report Prepared for the National Institute of Justice. Washington, DC: National Institute of Justice. Mimeo.
Hoffer, L. D. 2006. Junkie Business: The Evolution and Operation of a Heroin Dealing Network. Belmont, CA: Thomson Wadsworth.
Horwitz, A. V. 1990. The Logic of Social Control. New York: Plenum Press.
Jacobs, B. A. 1998. “Researching Crack Dealers: Dilemmas and Contradictions.” Pp. 160-177 in Ethnography at the Edge: Crime, Deviance, and Field Research, edited by J. Ferrell and M. S. Hamm. Boston, MA: Northeastern University Press.
------. 2006. “The Case for Dangerous Fieldwork.” Pp. 157-168 in The Sage Handbook of Fieldwork, edited by D. Hobbs and R. Wright. Thousand Oaks, CA: Sage Publications.
Jacobs, B., and Wright, R. 2006. Street Justice: Retaliation in the Criminal Underworld. New York: Cambridge University Press.
Jacques, S. Forthcoming. “The Necessary Conditions for Retaliation: Toward a Theory of Non-Violent and Violent Forms in Drug Markets.” Justice Quarterly.
Jacques, S., and Wright, R. 2008a. “The Relevance of Peace to Studies of Drug Market Violence.” Criminology 46(1):221-253.
------. 2008b. “The Victimization—Termination Link.” Criminology 46(4), forthcoming.
------. 2008c. “Intimacy with Outlaws: The Role of Relational Distance in Recruiting, Paying, and Interviewing Underworld Research Participants. Journal of Research in Crime and Delinquency 45(1):22-38.
------. Forthcoming. “Dangerous Intimacy: Toward a Theory Violent Victimization in Active Offender Research.” Journal of Criminal Justice Education.
Jansen, A. C. M. 1991. Cannabis in Amsterdam: A Geography of Hashish and Marihuana. Amsterdam, NL: Coutinho.
Korf, D. J. 1995. Dutch Treat: Formal Control and Illicit Drug Use in the Netherlands. Amsterdam, NL: Thesis Publishers.
Logie, R., R. Wright, and S. Decker. 1992. “Recognition Memory Performance and Residential Burglary.” Applied Cognitive Psychology 6:109-123.
Michalski, J. H. 2003. “Financial Altruism or Unilateral Resource Exchanges? Toward a Pure Sociology of Welfare.” Sociological Theory 21(4):341-358.
Mieczkowski, T. 1988. “Studying heroin retailers: A research note.” Criminal Justice Review 13:39-44.
Miller, J. This book. “Negotiating Gender When Interviewing Offenders on Offending.” In Offenders on Offending: Learning About Crime from Criminals, ed. Wim Bernasco. Cullompton, UK: Willan Publishing.
Morselli, C. 2001. “Structuring Mr. Nice: Entrepreneurial opportunities and brokerage positioning in the cannabis trade.” Crime, Law, & Social Change 35(3):203-244.
Nee, C. This book. “Research on Residential Burglary: Ways of Improving Validity and Participants’ Recall When Gathering Data.” In Offenders on Offending: Learning About Crime from Criminals, ed. Wim Bernasco. Cullompton, UK: Willan Publishing
Nee, C., and Meenaghan, A. 2006. “Expert Decision Making in Burglars.” British Journal of Criminology 46:935-949.
Phillips, S. 2009. “Legal disparities in the capital of capital punishment.” Journal of Criminal Law and Criminology 99, forthcoming.
Popper, K. 2002. The Logic of Scientific Discovery. New York: Routledge Classics.
Shover, N. 1996. Great Pretenders: Pursuits and Careers of Persistent Thieves. Boulder, CO: Westview Press.
Shover, N., and Hunter, B. This book. “Blue-collar, White-collar: Crimes and Mistakes.” In Offenders on Offending: Learning About Crime from Criminals, ed. Wim Bernasco. Cullompton, UK: Willan Publishing.
St. Jean, P. K. B. 2007. Pockets of Crime: Broken Windows, Collective Efficacy, and the Criminal Point of View. Chicago, IL: University of Chicago Press.
Venkatesh, S. A. 2003. “Review of Ethnography at the Edge: Crime, Deviance, and Field Research edited by Jeff Ferrell and Mark S. Hamm.” American Journal of Sociology 105(1):284-6.
------. 2006. Off the Books: The Underground Economy of the Urban Poor. Cambridge, MA: Harvard University Press.
------. 2008. Gang Leader For A Day: A Rouge Sociologist Takes to the Streets. New York: Penguin Press.
de Waal, F. 1989. Peacemaking among Primates. Cambridge, MA: Harvard University Press.
Weinreb, A. 2006. “Limitations of Stranger-Interviewers in Rural Kenya.” American Sociological Review 71:1014-39.
Western, B. 2006. Punishment and Inequality in America. New York: Russell Sage Foundation.
Wenseelers, T., and Ratnieks, F. L. W. 2006. Enforced Altruism in Insect Societies. Nature 444(7115):50.
Williams, T., Dunlap, E., Johnson, B. D., and Hamid, A. 1992. “Personal Safety in Dangerous Places.” Journal of Contemporary Ethnography 21(3):343-74.
Wilson, W. J. 1987. The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. Chicago, IL: University of Chicago Press.
------. 1996. When Work Disappears: The World of the New Urban Poor. New York: Vintage.
Wright, R., Decker, S., Redfern, A., and Smith, D. 1992. “A Snowball’s Chance in Hell: Doing Fieldwork with Active Residential Burglars.” Journal of Research in Crime and Delinquency 29:148-161.
Zhang, S. This book. “Talking to Snakeheads: Methodological Considerations for Research on Chinese Human Smuggling.” In Offenders on Offending: Learning About Crime from Criminals, ed. Wim Bernasco. Cullompton, UK: Willan Publishing.
Zhang, S. X., Chin, K., and Miller, J. 2007. “Women’s Participation in Chinese Transnational Human Smuggling: A Gendered Market Perspective.” Criminology 45(3):699-733.
 Even if a theory of offender-based research is eventually disproved through empirical research, both the theory and the research benefit methodology and criminology by revealing what factors do not produce more valid and plentiful data at a lower cost.
 It is possible to develop purely sociological theories in several ways. One way is to look at the available data on a behavior and attempt to identify its patterning with social status and social distance. If patterns are found, then they might provide a good start toward a more comprehensive and nuanced theoretical understanding. This is what we might call a “mining” approach to theory development. Another way, however, is to specify a theory – meaning make predictions about how variability in social structure affects behavior – with nothing but intuition, which may or may not be built on years of experience with the topic being theorized. This is what we might call an “intuitive” approach to theory development. The predictions produced by both forms of theory development, mining and intuitive, will ultimately require formal testing to ascertain their worth. How theories are “made” is distinct from – and less important than – whether theories are “good”, meaning relatively valid, simple, general, testable, and original (see Black, 1995).
 Note, however, that if there are no such real things as “intimacy” and “relational distance” then a theory that uses them as independent or dependent variables will be necessarily invalid, at least in part.
 On the other hand, an actor’s informal normative status decreases as more informal social control (e.g., vengeance or avoidance) is applied to that actor.
 It will not be possible to test the theory based on the answers to the first three questions because institutions often have such information readily available for each offender and, knowing this, such persons are, we theorize, more likely to provide valid information on those topics than free offenders (for details see the section, “Toward a Purely Sociological Theory of Research”, below). This is why it is necessary to first determine whether the two groups are equally likely to provide enough information to obtain a birth certificate; again, if they differ in this regard than the following test cannot precede.
 Which themselves are social behaviors.
 Better yet would be to have a coherent theoretical perspective to organize those predictions.
 Regardless of what factors are found to affect method, it should be possible to manipulate them in order to achieve our desired goals. The findings of science are themselves objective factors that influence the world. For example, criminologists have found that governments sometimes discriminates against minorities (Phillips, 2009), and this “finding” is itself a “social fact” – it exists in the real world; it is read, heard, and spoken about. This social fact, in turn, has changed the real world to the degree that it has reduced such forms of government discrimination against minorities. In other words, the goal of science is to produce findings about the real world, and those findings become real things that themselves affect that world.