Criminal Justice Other

Research Methods in Criminal Justice: Research Proposal

The present research proposal provides essential information about the upcoming research paper titled “Are Juveniles from Under-Privileged Backgrounds More Likely to Accept Plea Bargains?”. All six parts covered by the proposal include the topic-related information and research details. Since this work aims to prepare a researcher for the thorough investigation, the presentation of all details associated with theoretical and practical aspects of the topic assumes great significance.

Introduction to the Topic


The importance of the selected topic for the research investigation is determined by the currently observed increase of juvenile crime rate. For example, according to the evidence, “since 1995, juvenile crime levels in many countries in Eastern Europe and the Commonwealth of Independent States have increased by more than 30 per cent” (Juvenile Delinquency 2003, p. 189). This way, today, juvenile crimes have become the real-life problem for many countries. However, exactly the USA is associated with high rates of juvenile crimes, since in this country, “juvenile offenders are the third largest category of criminals” (Shepard, 2009, p. 192). Overall, the mentioned trend is rooted in more global problem, social delinquency. Being mostly embodied in juvenile crimes (mostly, from under-privileged backgrounds), social delinquency has become the reality of the modern 21st-century society characterized by general disorganization and imbalance that affects almost all aspects (mainly, economic, psychological, etc.) of human lives (Siegel & Welsh, 2008). At the background of the mentioned facts, the in-depth investigation of the topic of juvenile delinquency seems essential in order to understand all details related to this phenomenon.

Another trend that increases the importance of the investigation of the chosen topic is associated with the current efforts of policymakers to find an effective way to struggle with social delinquency and juvenile crimes, in particular. In this context, plea bargain as a successful prevention and punishment measure assumes great popularity in both adult and juvenile criminal justice systems of many countries (including the USA). Hence, it is not surprising that “plea bargaining occurs in juvenile court delinquency proceedings”, where “minors plead guilty in exchange for lighter punishment” (Myers & American Professional Society on the Abuse of Children, 2010, p. 54). Hence, the growing popularity of plea bargains in national juvenile criminal justice systems cannot be underestimated; in this context, the detailed investigation of the peculiarities of this legal measure in case of under-privileged juveniles is relevant.

Research Problem

The research problem of the topic relates to the likelihood of juveniles from under-privileged backgrounds to accept plea bargains. It is worth admitting that the criminal justice specialists’ opinions about this likelihood differ. On the one hand, some researchers (Edelman, 2007; Devers, 2011) suggest that juveniles from disadvantageous backgrounds are less likely to accept a plea bargain. On the other hand, other researchers (Kelleher, 2000; Siegel & Welsh, 2008, etc.) imply that juveniles’ backgrounds may not necessarily influence a court decision, so a plea bargain may be equally accepted by adolescent criminals from both advantageous and disadvantageous backgrounds.

Juveniles from under-privileged backgrounds may be less likely to accept plea bargains than those who have a more privileged status do. Edelman’s (2007) findings indicated that white adolescents from the family with an advantageous socio-economic status are less likely to be severely punished. However, if juveniles belong to non-white races, and proceed from a family with a disadvantageous socio-economic status, they are more likely to be imprisoned. The results demonstrated that since courts are notably biased to delinquents’ socio-economic and racial backgrounds, a plea bargain can hardly be accepted by adolescent criminals from under-privileged backgrounds (Edelman, 2007). Devers (2011) revealed that race and socio-economic status refer to the extralegal characteristics that affect court decisions. These characteristics “make it more likely that a defendant will not receive a reduced plea charge” (Devers, 2011, p. 3). For example, the US criminal justice experience demonstrates that “blacks are less likely to receive a reduced charge compared with whites” (Devers, 2011, p. 3). Overall, the researchers’ findings presented above suggest that the relationship between disadvantageous backgrounds and non-acceptance of a plea bargain is obvious.

At the same time, Kelleher (2000) noted that sometimes, acceptance or non-acceptance of a plea bargain may not be influenced by juveniles’ backgrounds. The author even mentioned that according to numerous court cases of adolescent crimes, plea bargains were not accepted in cases of juveniles from close families or privileged backgrounds (Kelleher, 2000).  Besides, Siegel and Welsh’s (2008) investigation demonstrated that officially, court practices and decisions remain impartial to delinquents’ backgrounds, so a plea bargain may relieve the punishment in case of juveniles of both white and non-white race, socio-economically privileged and under-privileged families. Taking into consideration the existence of opposite opinions about the connection of adolescent criminals’ backgrounds and plea bargain acceptance, one may agree that the upcoming investigation should be conducted in order to clarify the likelihood of juveniles from under-privileged families to accept plea bargains.

Research Questions

According to the research problem presented above, the influence of adolescent criminals’ origin/status on the acceptance of a plea bargain remains ambiguous. Hence, the upcoming investigation will be conducted in order to clarify the likelihood of juveniles from under-privileged backgrounds to accept plea bargains. The following research questions will guide the research:

RQ1: How do juveniles’ backgrounds affect plea bargain acceptance?

RQ2: Which adolescent criminals’ extralegal characteristics create favourable conditions for the acceptance of a plea bargain?

RQ3: Are juveniles from under-privileged backgrounds or those from advantageous families more likely to accept a plea bargain?


The investigation of the chosen topic will have both theoretical and practical implications that cannot be underestimated. First, the results of the study will fill the gap that exists in the research literature about the relationship between juveniles’ backgrounds and plea bargain acceptance, and will reveal other aspects of the topic that can be researched in future. Second, the empirical data and evidence-based information provided in the investigation will help to design a policy aimed to promote unbiased approach to adolescent criminals’ punishment, and to apply it to real-life settings (on the territory of the USA and other countries) that have a certain relation to juvenile criminal justice. Overall, the applicability of research findings to theoretical and practical areas points to the significance of the upcoming investigation.

Literature Review

The research works of numerous authors (including Shepherd, 2008; Viljoen, Klaver, & Roesch, 2005; Carmen & Trulson, 2005, etc.) contributed to the topic of the upcoming investigation. The key findings about the peculiarities of a plea bargain in juveniles’ cases deserve particular attention.  Shepard (2009) admitted that juveniles who have been involved in criminal activity should be punished according to the standard juvenile criminal justice systems. As numerous court trials focused on the cases of juvenile crimes demonstrated, a judge may decide either to make an arrest of an adolescent criminal or to accept a plea bargain. According to Siegel and Welsh (2008), a “plea bargain is an agreement between the prosecution and the defence by which the juvenile agrees to plead guilty for certain considerations” (p. 290). This way, acceptance of a plea bargain helps a juvenile to avoid a severe punishment. For this reason, it is not surprising that a plea bargain has gained popularity in juvenile cases among western democratic countries; for example, the US statistical data demonstrated that about 90% of juvenile cases result in a plea bargain (Carmen & Trulson, 2005, p. 254).

The recent research evidence suggested that juveniles’ extralegal characteristics (such as educational background, socio-economic status, race, etc.) have a great impact on the acceptance of a plea bargain (Shepherd, 2008). Carmen and Trulson (2005) mentioned that a plea bargain supposes that “a juvenile agrees to plead guilty to an offence in exchange for a lower charge, a lower sentence, or other considerations” (p. 268). However, the existence of this right does not suppose that a plea is satisfied in all cases. In a real life, juveniles from advantageous backgrounds have more chances to have their plea bargains accepted. Viljoen et al. (2005) explained that in cases of plea bargaining, juvenile’s avoidance of severe punishment requires a good paid lawyer whose wise legal guidance leads to desirable consequences. The reality demonstrates that white and educated adolescent criminals from the families with a relatively high socio-economic status are more likely to afford such lawyer. This way, in cases of juveniles from under-privileged backgrounds, a plea bargain is less likely to be accepted (Viljoen et al., 2005).

Key Concepts and Their Measurement

The topic of the upcoming investigation embraces several variable and non-variable key concepts worth mentioning. Overall, according to the topic of the upcoming research, juvenile delinquency, plea bargaining, and juveniles’ backgrounds represent the key concepts. Juvenile delinquency and plea bargaining refer to invariables; at the same time, the study will include some variables, and juveniles’ backgrounds (associated with extralegal characteristics affecting a court decision) represent the key one. To be more specific, the concept of juveniles’ background can be measured with the help of such variables as age, race, gender, and a socio-economic status.

Such concepts as juvenile delinquency and plea bargaining are invariables in the context of the upcoming study, since their essence is constant, and cannot be altered, and therefore – measured. Juvenile delinquency can be treated as adolescent activity of criminal and non-criminal nature; however, in the context of the upcoming investigation and for the sake of the achievement of research purposes, juvenile delinquency involves criminal behaviour. From this perspective, juvenile delinquency can be treated as “behaviour that is a violation of the criminal code and committed by a youth who has not reached adult age” (Regoli, Hewitt, & Matt DeLisi, 2009, p. 19). Shoemaker (2009) specified that the maximum age for juvenile court jurisdiction may vary from 16 to 18 depending on a country and its criminal justice system (Shoemaker, 2009).

Juvenile delinquency usually includes a wide range of criminal activities: murder, theft, rape, assault, robbery, and sale of illegal substance, etc. (Shepard, 2009). Another invariable concept involved in the upcoming study is plea bargaining. It can be defined as an adolescent criminal’s agreement to plead guilty in return for the promise of benefit. Plea bargaining is a concept widely used in juvenile criminal cases; this agreement “allows the parties to agree on the outcome and settle the pending charge” or simply “walk off with a plea” (Sirohi, 2009, p. 2152). The result of plea bargaining is a lenient sentencing or other types of less severe punishment. Overall, one may admit that the mentioned two concepts will remain unchanged regardless of juvenile’s backgrounds.

Juveniles’ backgrounds present the key variable concept included in the upcoming study. This concept includes adolescent criminal’s age, race, gender, and a socio-economic status. Exactly these extralegal characteristics of adolescent criminals will help to answer the research questions, and to achieve the ultimate purpose of the study. According to the evidence, owing to globalized and changing society, today’s juveniles may proceed from diverse backgrounds (McShane & Williams, 2003). In other words, they may have different age, race, gender, and a socio-economic status. A standard juvenile criminal justice system supposes that an age of a juvenile may vary from 12 to 18 (as it was mentioned, sometimes, it is maximum 16) (Devers, 2011). Race of an adolescent criminal may include white and non-white individuals who can be represented by Mexicans, Afro-Americans, Asians, etc. (Devers, 2011). Naturally, the gender of a juvenile can be either male or female (Devers, 2011). A socio-economic status is probably the most complex construct of juvenile’s backgrounds, since it is multidimensional. To be more specific, a socio-economic status includes three factors that have a relation to a juvenile’s family: educational, parents’ occupations, and financial income (Bornstein & Bradley, 2003). One may agree that an adolescent criminal may have various educational achievements, and his or her parents may occupy different occupational statuses, and possess diverse financial income.

Since juveniles’ backgrounds are variables they can be measured in a corresponding way. It is worth mentioning that variables may represent different levels of measurement. In terms of these levels, a juvenile’s age refers to ratio measure, race and gender – to nominal ones, while socio-economic status represents an ordinal measure (Hagan, 2009). Nominal measures are those representing discrete data. Ordinal variables are those that can be logically rank ordered. Finally, ratio measures represent the variables based on a true zero point (Hagan, 2009). Overall, one may admit that these different levels of measurement may reveal what sorts of information a researcher can gain from the scores assigned to the values of a variable. Another possible way to measure such variables as juvenile’s backgrounds is to apply Likert scale rating (mainly, in case of an adolescent criminal’s socio-economic status). The scores of this psychometric scale will allow evaluating juvenile’s education, his or her family’s occupational statuses and financial incomes in a corresponding way (Maxfield & Babbie, 2011).

Research Design

In the context of the investigations of criminal justice issues, the research design should be selected properly; otherwise, the research goals will not be accomplished. Taking into consideration the nature of the chosen topic of the upcoming study, quasi-experimental research design seems appropriate.  According to the evidence, quasi-experimental research design “is similar to controlled experimentation”, since both controlled experiments and quasi-experiments aim to answer questions about causation (Ellis, Hartley, & Walsh, 2010, p. 338). To be more specific, quasi-experimental research design is applied to an empirical study used to estimate the causal impact of an intervention on the selected target population (Hagan, 2009). This research design is used when randomization is not possible. In the context of the upcoming research, randomization seems inappropriate because the researcher will be interested only in a specific population group. Hence, a quasi-experimental research design is worth using to investigate the chosen topic.

The selected quasi-experimental research design supposes the implementation of non-randomized purposive sampling. First, the participants will not be chosen at random, so non-randomized sampling will be used; randomization will be irrelevant, since the researcher will know who exactly should be included in the sample (juveniles from under-privileged backgrounds) (Hagan, 2009). Second, the selected sample will be purposive (or judgment), since the selection of the specific population group fit a set purpose of the study (Maxfield & Babbie, 2011). Overall, the proposed sampling design suggests selecting a non-randomized purposive sample. This design seems appropriate because it is aligned with the ultimate research goals.

Measurement of variables will be realized in correspondence with the quasi-experimental design. Two groups of the target population will take part in the study, with the experimental group possessing the under-privileged background characteristics: presence of sufficient/insufficient educational background, belonging to the family with average/poor financial income, and possession of families with satisfactory/unsatisfactory occupational status.  The control group will consist of people conventionally considered to belong to normal, usual backgrounds with such characteristics as sufficient education, normal family income, possessing of a house, and a normal occupational status of family members. Overall, pre-intervention and post-intervention measurements will be realized; in this context, the intervention supposes criminal prosecution and plea bargaining. It is also worth mentioning that a questionnaire will serve a measurement method implemented in the pre-intervention and post-intervention stages.

The proposed research design and approach suppose the implementation of a corresponding data collection strategy. Secondary data will be collected with the help of document analysis considered an effective method for gathering detailed information about the topic in which a researcher is interested (Hagan, 2009). In the context of the upcoming study, the archival information about juvenile cases and adolescent criminals’ backgrounds along with the topic-related data received from previous investigations in the same field will be collected.  The primary data will be received with the help of questionnaire specially designed for the research participants, and legal document analysis on the gravity of offense and plea bargains (Maxfield & Babbie, 2011). The questionnaire will contain open- and closed-ended questions the answers to which will reveal the necessary peculiarities of adolescent criminals’ extralegal characteristics and juvenile court case with plea bargaining. This way, exactly a mixed-method approach will be implemented in order to maximize and optimize the data collection and analysis processes, and to make results and findings more thorough and in-depth.

Overall, it is worth mentioning that the mentioned data collection methods (document analysis and questionnaire) will help to receive both qualitative and quantitative information that will ultimately influence the findings and results of the research. According to the nature of the topic and the selected research design, the application of a mixed method approach is appropriate. This approach is used when a researcher wants to collect and analyze both qualitative and quantitative data (Hagan, 2009). Within the framework of the upcoming investigation, both types of data related to juveniles with under-privileged backgrounds will be collected and analyzed.

The usage of appropriate coding instruments (mainly, five-point Likert scale and coding frame) seems also important for the analysis of the collected data. The five-point Likert scale will help to receive evaluations of juveniles’ backgrounds from the participants (Ellis et al., 2010). This scale has different measures of intensity, where 1 stands for “strongly agree”, 2 – for “agree”,3 – for “neutral”,4 – for “disagree”, and 5 – for  “strongly disagree” (№5). These five evaluations will help to understand what kind of juveniles’ education, their parents’ occupations, and financial income influence the acceptance or non-acceptance of a plea bargain. Besides, the so-called “coding frames” will be used for the questionnaire (Oppenheim, 2000, p. 267).  Coding frames present the effective coding instrument that show how verbal information has been converted into numeric data for purposes of the analysis; in other words, by providing the link between the verbal data and the numeric data, this instrument provides the explanation what the numeric data mean. The development of coding frames will provide the rules for data entry; for example, the possible answers to closed-ended questions can be coded with the help of the letters “a”, “b”, etc. Overall, the mentioned coding instruments will speed up and optimize the data analysis process.

Some strengths and weaknesses of the selected data collection methods along with the ways of improving the quality of the research findings are worth mentioning, as well. Document analysis allows receiving ready-made materials related to the topic, and collecting data about past events (mainly, juvenile cases) in a cost-effective manner. At the same time, secondary sources may contain irrelevant data and outdated information (Bayens, & Roberson, 2010). The credibility of the collected secondary information can be increased through the implementation of high-quality document analysis embodied in the critical presentation of the data and avoidance of misrepresentations (Hagan, 2009). The strengths of questionnaire lie in a relatively high level of objectivity (in comparison with interviews) and a quick manner of data collection from a large participant group, while its weaknesses are associated with participants’ misinterpretation of the information and consumption of much time to analyze open-ended questions (Oppenheim, 2000, p. 267). For increasing the questionnaire’s reliability, an appropriate sample size (100 people) will be selected, and the peer review of the analyzed questionnaire will be implemented by involving special experts (Hagan, 2009).

Analytic Plan

The plan of data analysis for the planned study stems from the hypothesis chosen for testing: the hypothesized relationship is that between socio-economic background and readiness to accept a plea bargain. It is hypothesized that the lower the socio-economic status of a person is, the more ready he or she is to accept the offered plea bargains. First, the concepts that are to be measured should be operationalized (Hagan, 2010). The under-privileged background of juvenile delinquents is to be determined with the help of assessing their educational status, employment and/or employment status of their parents, the family income and housing, etc. Thus, the juveniles with average, middle-class backgrounds will be selected to the control group, and the individuals with under-privileged backgrounds will be selected to the experimental group. In the context of pre-intervention measurement, their criminal records and the number of registered criminal offenses, as well as accepted plea bargains will be registered for both groups.

The intervention will represent the involvement in a criminal offense, which is not under the control of a researcher, but will be traced through official police records. Hence, only those who will be detained and arrested by the police within 6 months will be included into the study sample. In 6 months, after the samples are formed from those who were accused of a crime, the post-intervention measurement will be done with the help of a similar survey to register the impressions and ideas of individuals about plea bargaining. As for quantitative data, the plea bargaining cases for both groups will be elicited from police records, and a frequency distribution table will be made to illustrate the acceptance of plea bargains by gravity of socio-economic background of juvenile delinquents. The frequency distribution will also be formulated according to the offense type to see for which offenses plea bargains are more often accepted.

Implications of Possible Findings

The present study has many theoretical and practical implications because it aims at investigating the relationship between juveniles’ backgrounds and their willingness to accept plea bargains. The present relationship is of much research interest because plea bargaining involves not only the issue of reducing the gravity of legal punishment, but also the false assumption that some adolescents hold that they will still be accused of the crime, so they may accept plea bargains for the crimes they did not commit. The fear of adolescents from under-privileged backgrounds is much more reasonable because they do not have funds to afford a lawyer, so many crimes in which juveniles from privileged and under-privileged backgrounds are involved end in the former serving a long term in prison, and the latter receiving only a probation sentence or a mild prison sentence. The present legal injustice does not stem from an intentional corruption of the legal system’s approach to juveniles with different legal backgrounds – it is of a much more complex nature of neglect towards those whose lawyers do not bother to undertake a detailed analysis of the case, point to the inconsistencies, and win the case.

It is necessary to keep in mind that the present study is not aimed at the reduction of crime rates; however, it is still related to a very significant aspect of criminal justice research – ensuring the social justice in punishment for crimes committed and not committed by certain individuals. Hence, adolescents from the under-privileged backgrounds are often desperate about being accused for criminal offenses and feel that they will be put to prison if they do not confess, and do not accept a plea bargain on the crime they did not commit. Such a situation, especially under the condition of the growing adolescent crime rates, is a very serious aspect of criminal justice requiring urgent attention. As Hagan (2010) repeatedly noted, justice is not about putting as many people to prison as possible, but about punishing the criminals and protecting the innocent. Deriving the rationale for the present study from this argument, it is essential to create a clear image of why and how individuals from under-privileged backgrounds accept plea bargains, which will help to reconsider the system of plea bargaining in the juvenile justice courts. The present effort is needed to help people with no sufficient financial resources to get out of the situations in which they are urged to confess to crimes for the sake of getting a miler sentence for something they did not do instead of serving a full-scale prison sentence for that.

A significant implication of the present study is also about understanding that juveniles are predominantly under the care of their parents, and the financial status of parents is a much more serious determinant of acceptance of plea bargains than the educational level or race of a juvenile. The reason for this is that parents able to pay huge sums for superb lawyers can free their children from a prison sentence, while crimes committed in groups still require somebody to be punished for a record. Hence, to uncover such hidden relationships, the socio-economic status of the juveniles’ families in the sample is also included in the framework of analysis to understand how some juveniles manage to avoid prison sentences, while others accept plea bargains for crimes they did not commit. Understanding of these hidden processes in the juvenile court system is potentially able to initiate a large-scale, fundamental change in the mechanism of administering justice for juvenile delinquents of all backgrounds.

Resource Needs

The research will not require significant financial investment because the majority of quantitative data used for elicitation of findings is of secondary nature – the police records and plea bargaining records of juvenile delinquents from the study sample will be borrowed from the state police authorities for production of the analysis. Prior agreement about the use of police records will be sought by the researcher from his educational institution and from the police authority to ensure smooth data provision and ready access to research data. Therefore, the main staffing issue is that of conducting open-ended surveys with the sample participants. Since the sample size is rather large, at least at the pre-intervention stage, the researcher will require at least 10 research assistants who will be able to collect the pre-intervention and post-intervention data. The statistical analysis of frequency distribution, as well as content analysis and Likert-scale analysis of surveys will be performed by the researcher personally for the sake of maintaining the confidentiality of subjects. However, peer review and peer checks are presupposed for the data analysis drafts, as well, to reduce the probability of error or researcher bias.

Hence, taking into account the simple structure of research, the cost implications include only staffing for data collection, which make take up to $1,000. In case the researcher is involved in a commercial basis and not in the framework of educational research effort, the salary for the researcher should also be included (6 months of intervention interval, plus 3 months from writing the paper – one month before intervention and two months after the intervention), which takes up to 9-month salary.

As for the research timetable, it is planned as follows:

1 month – confirmation of the research proposal, preliminary research and formulation of objectives, writing chapters 1-3 of research, composing and confirming questionnaires

6 months – the intervention (observation of the juveniles with normal and under-privileged background regarding criminal prosecutions and acceptance of plea bargains)

2 months – post-intervention measurement (for only those who involved in criminal prosecutions); analysis of data; writing chapters 4 and 5. Presenting the research results.


Bayens, G. J., & Roberson, C. (2010). Criminal Justice Research Methods: Theory and Practice (2nd ed.). Philadelphia, PA: CRC Press

Carmen, R. V., & Trulson, C. R. (2005). Juvenile Justice: The System, Process, And Law. Belmont, CA: Cengage Learning.

Devers, L. (2011). Plea and charge bargaining. Retrieved from

Edelman, M. W. (2007). Let’s put the ‘justice’ back in our juvenile justice system. Baltimore Afro-American (8 Dec 2007), p. A8.

Ellis, L., Hartley, R. D., & Walsh, A. (2010). Research Methods in Criminal Justice and Criminology: An Interdisciplinary Approach. Lanham, MA: Rowman & Littlefield.

Hagan, F. E. (2009). Research Methods in Criminal Justice and Criminology (8th ed.). Upper Saddle River, NJ: Prentice Hall.

Juvenile Delinquency (2003). World Youth Report. Retrieved from

Maxfield, M. G., & Babbie, E. R. (2011). Basics of Research Methods for Criminal Justice and Criminology. Belmont, CA: Cengage Learning.

McShane, M. D., & Williams, F. P. (2003). Encyclopedia of Juvenile Justice. Thousand Oaks, CA: SAGE.

Myers, J. E., & American Professional Society on the Abuse of Children. (2010). The APSAC Handbook on Child Maltreatment. Thousand Oaks, CA: SAGE.

Oppenheim, A. N. (2000). Questionnaire Design, Interviewing and Attitude Measurement. London, UK: Continuum International Publishing Group.

Regoli, R., Hewitt, J., & Matt DeLisi, M. (2009). Delinquency in Society. Sudbury, MA: Jones & Bartlett Learning.

Shepherd, R. E. (2008). Plea Bargaining in Juvenile Court. Criminal Justice, 23(3), 61-63.

Shepard, J. M. (2009). Sociology. Belmont, CA: Cengage Learning.

Shoemaker, D. J. (2009). Juvenile Delinquency. Lanham, MA: Rowman & Littlefield.

Siegel, L. J., & Welsh, B. C. (2008). Juvenile Delinquency: Theory, Practice, and Law. Belmont, CA: Cengage Learning.

Sirohi, S. (2009). Plea bargaining: a new horizon in Indian criminal justice system. Pratiyogita Darpan, 36(6), 2151-2153.

Viljoen, J. L., Klaver, J., & Roesch, R. (2005). Legal Decisions of Preadolescent and Adolescent Defendants: predictors of Confessions, Pleas, Communication with Attorneys, and Appeals. Law and Human Behavior, 29(3), 253-277.