Research methods are procedures for obtaining information on individual and/or aggregate phenomena for the purpose of (1) creating a general explanation or theory to explain a phenomenon; (2) testing the applicability of an existing theory to a subgroup of the population; or (3) testing the effectiveness of an existing social policy or program. Topics (1) and (2) are critical to the dialectic of scholarly knowledge in criminology and criminal justice. Somewhat unique to the field of criminal justice, however, is a heavier emphasis on (3) as a product of research. The phenomena of primary interest to criminologists include juvenile delinquency, adult criminality, and victimization, at both the individual and aggregate levels. The interests of criminal justice researchers appear more eclectic, only a few of which include police practices and effectiveness, the dynamics of criminal case processing, sentencing discrimination, inmate violence, and correctional pro gram effectiveness.
Research Methods in Criminology and Criminal Justice
- Citation and Content Analysis
- Crime Classification Systems
- Crime Mapping
- Crime Reports and Statistics
- Criminal Justice Program Evaluation
- DAWN and ADAM
- Edge Ethnography
- Experimental Criminology
- Fieldwork in Criminology
- Quantitative Criminology
The methods employed in criminological and criminal justice research are identical to those in the behavioral and social sciences in general. A critical assumption underlying the use of these procedures involves the belief in an objective reality, or a world that different people perceive in similar fashion. Related to this assumption is that such a reality can be studied objectively. The perspective that individual and social processes can be studied dispassionately or scientifically is referred to as positivism.
Not all criminologists share the positivist perspective. For example, any effort to derive a social psychological theory of criminality relies on the idea that social processes operate uniformly across most (if not all) individuals. One might argue, however, that such uniformity does not exist, due to individual differences in perceptions of these processes. Ethnomethodology involves the perspective that all ‘‘realities’’ are socially constructed. From this perspective, individuals perceive their world in terms of how it makes sense to them, thus introducing different perceptions of reality that may not be reconcilable.
The pieces of information that are gathered and examined during the course of research are referred to as data, which may be either qualitative or quantitative in form. Both forms of information may be gathered through observations of the phenomena under study, and quantitative information may also be compiled through survey research or a review of archival data. Qualitative observations are recorded by researchers as verbal statements that describe particular processes and outcomes, whereas quantitative observations consist of pieces of information recorded in numerical form. Both qualitative and quantitative methods are useful for theory development and testing, although a heavier emphasis in criminology and criminal justice appears to be placed on qualitative research for theory development versus quantitative research for theory/hypothesis testing and program evaluation. Many investigators use both approaches in a single study, however, because findings from each serve as a check on the other.
Ethnography is used to refer to a qualitative study of a social group or (sub)culture in which a researcher compiles a detailed description of processes and outcomes related to the phenomenon of interest. An example of ethnography would be a study of prison inmate social systems and adaptation to incarceration in a particular prison (such as the classic studies conducted by Clemmer, Sykes, Carroll, Jacobs, and Irwin). A penologist might make observations about the types of inmates that exist in that prison and how they interact with each other in order to understand, for example, why some inmates adapt to incarceration more easily than others. This information could then be used to create a general theory of inmate behavior that extends beyond the specific prison to all similar inmate populations. Critical to the success of such an endeavor is the researcher’s objectivity in making and recording his or her observations regarding inmate behaviors.
In contrast to qualitative research, a quantitative study involves gathering information and attaching numerical values to each piece. Some types of information already have numbers attached to them (e.g., a person’s age in years), whereas other types are assigned numerical values by the researcher (e.g., the sex of an individual, where every male in a sample is coded as ‘‘0’’ and every female in the sample is coded as ‘‘1’’). When a researcher attaches his or her own numerical values, these values are determined by the researcher and must be defined for someone who is trying to understand the study. These scales or variables are then analyzed with statistics in order to make sense of the information for subsequent interpretation. Statistics, therefore, are also pieces of information, the difference being that the statistical information is a more general summary of the information gathered by a researcher. Numbers are assigned to pieces of information only when a researcher intends to apply statistics in order to produce new information that cannot be obtained through verbiage.
Unlike qualitative research, where a researcher remains ‘‘open’’ to new information, the types of information gathered from a quantitative study are determined before data collection begins. This is one reason why quantitative research is used primarily for theory/hypothesis testing, because such research involves collecting information that has already been described in a specifically worded hypothesis derived from a testable theory. Quantitative research can be used for theory development when the theory of interest focuses on the causal order of events and behaviors rather than the substance of those events/behaviors. Even then, however, the application is usually limited to reducing the number of possible orders rather than pinpointing the exact causal model.
More steps are typically involved in quantitative research designed to test a theory/ hypothesis compared to qualitative exploratory research for the purpose of theory construction. The research design of such a quantitative study always falls into one of three broad types: experimental, quasi experimental, and non experimental or correlational. These groupings reflect differences in methodological rigor, or the ability of a study to establish the causal order of events (which is relatively rare in criminological and criminal justice research). The specific steps involved in this application of quantitative research include the following:
- Begin with a theoretical model (paradigm) of interest, which, in criminology, often involves a general perspective of a social, political, and/or economic process. For example, a ‘‘conflict paradigm’’ involves the perspective that many social problems such as discrimination, poverty, environ mental pollution, and crime in a capitalist society are consequences of economic (and thus power) ‘‘conflicts’’ between groups.
- The theoretical paradigm selected at step 1 is applied to a particular aspect of society. For example, a conflict criminologist is only concerned with the part of the conflict perspective that explains crime in a capitalist society.
- Theories involve theoretical, or abstract, concepts (e.g., ‘‘economic power’’ and ‘‘crime’’). In order to test a theory, one must be able to transform the theoretical concepts into operational definitions that are directly observable and measurable (e.g., ‘‘economic power’’ may be operationalized as gross annual household income). These definitions are then placed into a hypothesis, or a proposition that describes the predicted (hypothesized) relationship between the variables (e.g., persons with lower household incomes are more likely to be arrested). Any test of a theory actually involves a test of a specific hypothesis stemming from a general theory, and so the specific nature of any hypothesis means that a theory can never be tested directly. It is always possible that the measures tested do not accurately reflect the ‘‘true’’ theoretical concept. This is why such measures are constantly being refined.
- A researcher then plans the data collection that is required for the hypothesis test(s), involving the determination/selection of the (a) target population, or the population to which the results will be generalized, (b) units of analysis reflected in each hypothesis (individuals, organizations, cities, counties, etc.), (c) time dimension to be reflected in the data (e.g., one point in time versus two or more points in time), (d) research design (based on the hypothesis and the level of rigor desired, such as matched pairs, factorial, pretest posttest, time series, etc.), (e) sample that represents the target population (using one of a number of probability sampling techniques such as simple random sampling, systematic random sampling, sampling proportionate to size, etc.), (f) data collection instrument for compiling and coding the information (such as with a survey questionnaire), and (g) procedures for gathering information (telephone, mail, face to face, reviewing archival data, etc.).
- The data collection phase consists of completing/obtaining completed instruments for all cases in the sample.
- With the data compiled, the information should be checked for accuracy during the recording procedures. Computers are used for the purpose of data cleaning.
- The data are examined in order to test each research hypothesis. This step involves the computation of statistics that help to summarize large quantities of data in order to test the hypotheses of interest and to describe the empirical relationships involved. Like the data collected for a study, statistics are also pieces of information, although they are designed to help make sense out of the data collected. It is up to the investigators, how ever, to apply and to interpret these statistics correctly in order to derive accurate conclusions regarding their data.
- The use of quantitative methods for criminological and criminal justice research has steadily increased since the 1940s, due in part to the growing number of techniques, the avail ability of technology which facilitates data collection and analysis, and the proliferation of graduate programs and methods courses in the field. Ethnography remains a more powerful tool for theory construction, however, and many scholars place a high priority on combining the qualitative and quantitative.
The growing popularity of quantitative research has been met with resistance on the part of some qualitative researchers. Some individuals believe that the inability to numerically measure and evaluate many of the key concepts and processes that are critical to the field will produce misleading information regarding the validity of these ideas. When faced with having to operationalize highly abstract theoretical concepts, researchers can only measure observable proxies for the concepts of interest. These proxies may not capture the full essence of the original idea, as when researchers use structural attributes such as the poverty rate or the unemployment rate to proxy the more complex process of structurally induced strain. This problem is exaggerated when a researcher does not fully understand the theoretical concepts and/or the procedures and limitations of complicated statistical techniques used in order to examine the data. The current state of graduate education in criminology and criminal justice programs contributes further to concerns over knowledge destruction, since these programs are often void of courses in theory construction and offer a very limited number of courses in research methodologies.
Some of the more common problems in extant criminological research include a lack of objectivity in theory construction (e.g., ‘‘convict criminology’’), model misspecification (often due to poor conceptualization of the relevant theories), poor operationalization of concepts (e.g., unidimensional measures of multidimensional concepts such as low self control or social capital), inappropriate units of analysis (e.g., testing neighborhood level theories with county or state level data), inappropriate samples for the target populations (e.g., testing routine activities theory with college freshmen enrolled in a criminology class), and misapplications of statistical techniques (such as meta analyses of quasi and non experimental findings, multi level analyses with insufficient samples, and over corrections for spatially correlated error).
Read more about Criminology.
- Blalock, H., Jr. (1969) Theory Construction: From Verbal to Mathematical Formulations. Prentice- Hall, Englewood Cliffs, NJ.
- Campbell, D. & Stanley, J. (1963) Experimental and Quasi Experimental Designs for Research. Rand McNally, Chicago.
- Farrington, D. P. (1983) Randomized Experiments in Crime and Justice. In: Tonry, M. & Morris, N. (Eds.), Crime and Justice: An Annual Review of Research, Vol. 4. University of Chicago Press, Chicago, pp. 257 308.
- Lofland, J. (1984) Analyzing Social Settings. Wadsworth, Belmont, CA.
- Luckman, T. (Ed.) (1978) Phenomenology and Sociology. Penguin, New York.
- Moser, C. A. & Kalton, G. (1972) Survey Methods in Social Investigation, 2nd edn. Basic Books, New York.
- Popper, K. R. (1959) The Logic of Scientific Discovery. Basic Books, New York.