SOCIAL SCIENCE IN LAW

    You can have a social science OF something or a social science IN something. This lecture is about social science in law because it's the closest thing to forensic social science.  With the exception of psychology and anthropology (where there actually are professions called "forensic psychology" or "forensic anthropology"), none of the other social sciences lend their names in such a manner. Instead, most social sciences prefer to make the study OF law a specialty area in their discipline. This allows them to remain detached and offer critical perspectives on the whole legal system. On the other hand, most social scientists view social science IN law as "selling out" and allowing the legal system to tell them what their subject matter should be. In general, the social sciences prefer to control their own specialty and interdisciplinary areas, regardless if it's OF or IN something.  A problem with the OF approach is the "sociologizing" of everything -- sociology of law, sociology of medicine, sociology of cyberspace, etc., to use sociology as an example -- and the problem is even worse when you follow the IN approach, which results in some rather strange names -- Legal Sociology, Medical Sociology, Cyberspatial Sociology, etc.  Most of the ones with strange-sounding names either don't exist or are limited to a handful of practitioners that are rather distant from their academic cousins.

    Social science is a branch of behavioral science, which is the study of influences upon human behavior, including biological factors, but Biology long ago separated itself to be treated as a natural science. Here's a list of traditional social sciences:

    Social science in law is defined as the use of social science materials to resolve legal problems. It refers to cases in which lawyers think it might be helpful to draw upon the expertise of social scientists, their research or literature, to help the court better understand some complex issue, argument, or question. This interest by lawyers reached its peak from 1930-1950, is called legal realism or sociological jurisprudence, and was sparked by the words and ideas of five great jurists -- Oliver Wendell Holmes, Louis D. Brandeis, Roscoe Pound, Benjamin Cardozo, and Karl Llewellyn. In its extreme form, legal realism holds that there is no question too complex that social science cannot help the law with. Examples of such questions might be:

    These are, of course, some typical research questions of social science, and they can be answered different ways. The standard legal way would be stare decisis (let the decision stand), by going back over old, relevant cases to see what the legal discourse was about, extrapolating principles, rules, tests, or doctrines, and applying them to the new situation at hand. A moralist, rationalist, theological, or philosophical approach would try to answer by appeal to reason or intuition; a normative, often partisan, method that frequently results in dispelling all doubt that "of course" it's harmful or wrong. Social science, on the other hand, uses an empirical approach, empiricism being the method of direct observation, experience, and experimentation. What makes social science a SCIENCE is its method. That's why regardless of whether you study social science at a tech school or liberal arts institution, you're going to have to learn social science research methods.

    It's often said that social science is a "soft" science, but it can be rather "hard" when it comes to social science in law. Most use of social science in law is of the "legislative fact" or "adjudicative fact" variety, which means that research is used to make the law (as in death penalty or jury selection) or determine legal facts (as in pornography or music as a cause of crime). In recent years, however, social science has found a third use in law -- that of providing a "social framework" which means at least a couple of things:

    Empirical social science methodology is all about prediction and control rather than proof. It rarely happens that social science obtains proof in the form of knowing all necessary and sufficient conditions for a causal X-Y relationship. Most influences on human behavior are of a necessary but not sufficient kind. This doesn't mean they are any less causal. It just means that the search for causes is carried out differently. Prediction is the means by which social scientists compare ideas. A social science theory, for example, may have two or three approaches in it, using different concepts or operationalizations of the same concept. They are compared on the basis of their predictive power -- which concept, when measured, sampled, and data collected, produces the best score on some inferential statistic. The predictive statistic in use at the time is a matter of tradition or convention -- during the 1960s, it was correlation; during the 1970s, it was factor analysis; during the 1980s, regression emerged as the convention used. New inferential statistics are being developed every year, and most social scientists at research universities are highly vested in the ones they know and use (it largely constrains their research agenda).

    Control doesn't mean anything ominous or repressive like "social control" or the policy-making aspects of social science.  Control means knowledge of how to change something, regardless of whether it should be changed or not. It's a search for controlling, underlying factors, the ability to influence or bring about change in outcomes. It's the data mining or fishing expeditions that social scientists go on when they obsessively "milk" the data.  For example, there might be a weak predictive statistic between TV violence and crime for most people, but perhaps a certain type of TV violence has a very strong controlling influence on certain types of people. It's possible that this particular influence might unlock some clue about controlling the phenomenon for most people, but for the social scientist, it's enough to have unlocked one small controlling factor that can be "tweaked" or manipulated, if need be.

    It's also important to recognize a thread in social science that emphasizes understanding, appreciation, or the German word verstehen. This thread has emerged for a variety of reasons, from a variety of sources, has some similarities with anthropological method, and is called qualitative social science.  As opposed to quantitative social science, the qualitative school of thought holds that prediction and control without understanding is futile. The analogy is medicine. You might know, statistically, that a certain kind of diet or lifestyle causes disease, and you might have developed a controlling factor in the form of a pill that works to cure or prevent the disease. However, all you have is prediction and control without understanding. You don't really know how the disease or the pill work. People in medicine, pharmaceutics (and business in general) are perfectly comfortable with this. Social scientists, on the whole, are not. They would like to see corroborating qualitative studies (if you will), and this is often called exploratory or descriptive research. It serves a vital yet underrecognized place in social science.

    Social scientists know that no two operational definitions are exactly alike. Different researchers choose different variables, defining and measuring them differently. They may be interested in keeping with tradition, putting a new twist on tradition, or breaking new ground entirely. This produces a plethora of studies, all of which may or may not be compatible on the same subject. Despite the temptation of something called meta-analysis (which reduces incompatible studies on the same subject to some power statistic), the most appropriate way to distinguish quality among studies on the same subject is to focus on validity and reliability. How well a social scientist handles threats to validity and reliability is the trademark of an expert in this field. Both are important, but validity is somewhat more important than reliability in social science.

    VALIDITY is how well something truly or accurately measures what it is supposed to measure. Two general categories of it are internal and external. Internal validity essentially involves whether the social scientist considered plausible rival hypotheses. External validity involves whether inferences can be generalized across person, place, and time. There's also face and criterion validity. Face validity is when something "on its face" appears to measure the relevant behavior. Criterion validity is whether something predicts what it is supposed to predict (i.e., the criterion).

    RELIABILITY is the extent to which measures give consistent results. It's the idea of consistency in getting the same results if you were to conduct your study on the same people a month later, a year later, ten years later, etc. You can probably see the problem already when you're dealing with people's attitudes, which may change over time. Social scientists strive to use and develop four-point (or five-point) scales or indexes that are reliable. So, if a person's attitude changes, a reliable scale or index will pick up that attitude change reliably as long as you didn't have to redo or reword the items. Various item-subitem and test-retest methods exist to estimate reliability, but the most common method involves calculating something called Cronbach's alpha.

THREATS TO VALIDITY

    Social scientists draw conclusions by making inferences from their data on the sample of people they study to a larger group of people called the population. This process can be described as ensuring that findings within the sample would hold true if everyone in the population had been studied. It is accomplished via the use of a null hypothesis, tests of significance, and estimations of Type I and Type II error. The null hypothesis is what is always being testing in a test of significance. The null hypothesis can be stated in terms of the population as a whole or in terms of different subsets or group splits in your sample. In either case, you're hypothesizing that two populations are alike, or they are different. These are fairly easy probability statements to test, and various statistical techniques, called tests of significance, exist which tell us the likelihood of two populations differing by chance even if, in actuality, there were no differences on the whole. The idea is based on the tenable assumption that no two populations are exactly alike. It's the nature of probability that highly improbable events sometimes happen. 

    If our test of significance indicates that we should reject the null hypothesis (that there are no differences), and we always consider the level of significance when doing so, then we are at risk of making the Type I error (rejecting a null hypothesis when it is true). By convention, the level of significance in social science is .05 or 5%. It's a preset value that tells us we should reject the null hypothesis if its chance of being true is five in one hundred or less. Our test of significance statistic compares the probabilities we obtained in our sample against the probabilities of this preset value in the population. If the final calculation is one that says we could obtain by chance a difference greater than 5:100, then the null hypothesis cannot be rejected, and we are at risk of making the Type II error (accepting a null hypothesis when it is false). The only way of protecting yourself against Type II error is to take larger samples, and the more you protect yourself against Type I error, the more likely you are to make a Type II error. Social science conclusions are always expressed in this way: the null hypothesis was rejected at the .05 level of significance, which means that 5 out of 100 times we say there is a chance of obtaining the statistically significant results found in this study, there in fact are no real differences in the population, only chance variation. 

ADMISSIBILITY OF SOCIAL SCIENCE EVIDENCE

    Before we look at some of the different topical areas where social science has been used in law, it's important to note the different ways social science gets into law. There's really no rigid adherence to standards, rules, or tests like Daubert or Frye, and testimony on the stand is somewhat rare for a social science expert. That's because justices, judges and lawyers often think of themselves as experts on social science (witness the Exclusionary Rule or Miranda warnings as judicial experiments) or can make due with what's called ex parte communication. There's no Federal Rules of Evidence on this practice, except regarding discretion on what to give judicial notice to. Ex parte means witnesses heard in secret, and often judges will send their clerks to professors, or talk with them over the telephone or informally. Here's a list of the various ways social science information gets before the judiciary:

TOPICAL AREAS:
SOCIAL DISCRIMINATION

    Probably the most well-known example of social science in law is the case of Brown v. Board of Education (1954). It marked the first judicial use of modern social science. For review, the case overturned the "separate but equal" doctrine of Plessy v. Ferguson, and mandated a 14th Amendment right to desegregated educational opportunities.  Since then, social science experts have been regularly used in a wide variety of social discrimination cases: female applicants to all-male military academies, reverse discrimination, quota systems, affirmative action, racial discrimination, gender discrimination, and so forth.

    The knowledge base in this area consists mainly of research into the psychological effects of prejudice and discrimination. In Brown, the social science finding was that segregation with the sanction of law fostered low black self-esteem which led to low black achievement and perpetuated institutional racism. Social scientists filed both briefs and provided testimony about the "vicious circle" of prejudice and how it was affecting black students' motivation to learn. One of the more interesting elements of the testimony involved the use of dolls as a measure of self-esteem. Children (both black and white) consistently chose white dolls when asked to pick the doll they like best, the smartest doll, and the doll with a nice color. This was interpreted as a statistically significant finding of racial inferiority among black children. Sociologists then offered testimony that it was impossible to talk about discrimination or segregation in a vacuum, that the problem in schools just exemplifies the structural problem of racial injustice in society. 

    With the issue of same-sex v. coeducational schools, social scientists have been used not so much to talk about prejudice and discrimination, but to interpret conflicting studies that men speak up and dominate more in the classroom. Colleges for women and military academies for men also tend to be small in size as well as drawing upon certain social classes for applicants, so much of the research in this area tries to differentiate a single-sex effect from a size or social class effect. Other uses of social science research have involved measuring the "intangible" features and benefits of a college education. Other social science research involves discrimination in employment.

OBSCENITY AND/OR MEDIA VIOLENCE AS A CAUSE OF DELINQUENCY

    Another common use of social science in law is in the area of obscenity, not so much as to whether the state should control it or whether something qualifies as legally obscene, but on the behavioral effects of reading or seeing obscene material. A few well-known cases are U.S. v. Roth (1956), Paris Adult Theatre v. Slaton (1973), New York v. Ferber (1982), and numerous reports by Commissions on Obscenity and Pornography.

    In Roth, criminologists (the Gluecks) considered the country's leading authorities on the causes of juvenile delinquency found that delinquents read very little and that no one factor explains delinquency. There was some evidence that reading "good" books allegedly influenced someone to become a non-delinquent, but there was no evidence that obscene books and pictures influence a child's conduct adversely.  In Paris, the court commissioned special surveys, sending sociologists out into the Atlanta population to test pornography's effect on "normal" people and "bizarre deviant groups". The research showed that there were no statistically significant differences. In Ferber, at issue was the effect of child pornography on child sexual abuse rates. The research used was inconclusive on this "cycle of violence" issue, but did find some case evidence of emotional trauma later in life by those who posed or participated in the making of child pornography. Various national studies have been specially commissioned and many broke new ground in claiming to find dramatic effects in the relationship between exposure to erotic material and criminal behavior. The strongest effect seemed to be for violent pornography, or so-called "snuff flicks", that are not representative of the entire genre.

PAROLE, SENTENCING, RECIDIVISM, REHABILITATION

    Parole as traditionally known is all but a dead institution nowadays (replaced by Sentencing Commissions), but it used to be the one area where courts took a "hands off" approach to social science expertise. In 1928, a sociologist named Ernest Burgess developed a prediction instrument to determine parole eligibility, and before long, almost every correctional system in the country was using some version of the Burgess instrument for classification and release of inmates. Almost every risk assessment instrument today is still a version of the Burgess instrument; e.g., "no previous work record", "changed addresses 3 times in the last 2 years", etc.....

    The sociological expertise that went into parole prediction also found use in civil commitments, or involuntary mental hospitalizations, under various "dangerousness" statutes. A few related cases were Barr v. U.S. (1976) which dealt with antisocials, psychopaths, and sociopaths as incurable, and Barefoot v. Estelle (1983) which dealt with predictions of future dangerousness in the context of the death penalty. These and numerous other cases resulted (with a few dissents) in the Supreme Court not recognizing social science's ability to predict future dangerousness. There have been many studies and debates on this in criminology (the names of Steadman and Monahan come to mind), and the issues are whether statistical prediction is as good as clinical prediction and/or whether mentally disordered (dangerous) offenders are mad, bad, or different.

    By 1975, it was also apparent that what a criminologist named Martinson said a year earlier ("What Works") was true -- there were no effective correctional treatments ("Nothing Works"). This led to a search for some rationale for corrections, now that rehabilitation was dead. A number of studies focused on deterrence, and the issue came up most notably in Gregg v. Georgia (1976) where it was argued that each act of carrying out the death penalty saved eight lives. These numbers are part of what is known as the Ehrlich study, perhaps the most widely criticized piece of social science on methodological grounds. 

PROFILING CRIMINALS AND VICTIMS

    Long before the popularity of serial killer profiling, social scientists were involved in court cases dealing with the profiling of hijackers, illegal aliens, and drug couriers, and the courts have varied widely in their receptiveness.  The typical case involves someone who "fits" or "matches" the behavioral characteristics of a known group of criminal offenders. A few cases were U.S. v. Lopez (1971) where specially trained airport employees screened passengers for potential hijackers, U.S. v. Martinez-Fuerte (1976) which involved illegal alien screening checkpoints, and many, many other cases involving drug courier profiling. In almost all these cases, social scientists were involved in the training of employees to detect such persons and/or involved in defending the practice in court. Some statistical evidence was offered to show the techniques were at least 60% effective, but primarily courts had problems with justification on the basis of impressionistic police officer experience as data.

    Ironically, social science in law has had more success in profiling victims than offenders. At the same time experts were trying to impress the "rapist profile" on the courts in New Jersey v. Cavallo (1982), something called "rape trauma syndrome" was being developed in State v. Saldana (1982). Acceptance of the rape trauma syndrome was a significant step forward in victimology from what previously existed, post-traumatic stress syndrome.

    Also, some social scientists, especially social workers, have taken the stand to testify in court as an expert witness, investigating cases of possible criminal conduct, and assisting the legal system in such issues as child custody disputes, divorce, child support, juvenile delinquency, spouse or child abuse.  Ethics and liability issues abound in this area, and a few practitioners don't separate their role as an expert witness from their role as the fact finder, or fact witness (in the investigation of a child abuse case, for example).  There's a vast difference between an academic social worker who can summarize the settled research on a topic and one who has been too close to an actual case to divorce themselves from the particulars to speak in terms of generalities.  

STATE OF MIND SYNDROMES

    It seems that in recent years, social science in criminal law has been almost completely pro-defense. The number of emerging defenses in the form of various syndromes is amazing (see this lecture for a near-complete list). Most of these cases don't involve a claim to have an exculpatory mental illness, but a lessening of responsibility based on sociocultural factors that led to the crime. Some pertinent examples are:

INTERNET RESOURCES
Law and Humanities Website
Law and Society Association
Law and Society Movement
NSF Law and Social Science Program
Prof. Jonathan Aleck's Online Lecture Notes
Prof. Valerie Hans Online Lecture Notes

PRINTED RESOURCES
Anderson, P. & T. Winfree. (1987). Expert Witnesses: Criminologists in the Courtroom. Albany: SUNY Press.
Barker, R. (2000). Forensic Social Work. Binghamton, NY: Haworth Press.
Erickson, R. & R. Simon. (1998). The Use of Social Science Data in Supreme Court Decisions. Urbana: Univ. of Illinois Press.
Ewick, P., R. Kagan & A. Sarat. (1999). Social Science, Social Policy and the Law. NY: Russell Sage.
Lempert, R. (1989). An Invitation to Law and Social Science. Philadelphia: Univ. of PA Press.
Monahan, J. & L. Walker. (1998). Social Science in Law: Cases and Materials. Mineola: Foundation Press.

Last updated: 11/12/03
Lecture List for JUS 425
MegaLinks in Criminal Justice