by Dr. Maurice Godwin
As a result of the vexing opinions concerning exactly what deductive and inductive profiling entails, I felt that it was necessary to address some important issues surrounding the use of these terms.
Deductive reasoning was first developed by Thales, Pyuthagoras, Aristotle, and other Greeks of the Classical Period (600 to 300 BC). However, anyone who has read the Sherlock Holmes stories by A.C. Doyle is aware that Holmes used a form of deductive reasoning. Holmes, for example, would observe that Watson’s clothes were dry in spite of the fact that it had been raining all day. From this Holmes would guess or deduce that Watson had spent the day at his club – of course Watson could have spent the day at some other location.
Profiles constructed by the FBI profilers, clinical psychologists, criminologists, and the police routinely draw inferences about, for example, serial murderers and their behaviors based solely on work experience, gut feelings, and the motivation of the offender. This form of deductive profiling is where the profiler assumes one or more facts as self-evident about a crime or offender and then, following work experience and hunches, arrives at other facts commonly called conclusions. Hence, the FBI profiles are deductive rather than inductive. However, some argue that the FBI profiling method is inductive. Broadly, the argument put forth for the FBI method being inductive is, since the FBI relies on data collected from interviews with serial murderers, as a foundation for developing their profiles, then their reasoning must be inductive. The basis for this argument is flawed, because the data collected by the FBI has never been empirically analyzed, or has it been properly organized in a systematic manner so that profilers could refer to it in future. Rather, the information has been passed down over the years based on memories of past experiences and inferences gleaned from the interviews in order to arrive at conclusions about a particular case. To be sure, deductive reasoning is from the general (passed down information) to the specific (case).
The ‘truth’ of the conclusions reached by deductive profiling is a contingent truth; that is, it depends upon the truth, or the basis for the truth, for theories formed when the investigator first arrives on scene or during interviews. These statements or beliefs are taken for starting points for the argument. Deductive profiling processes available information by application of personal experiences as opposed to theoretically driven inductive profiling based on all available instances of a of crime. The guarantee of the deductive profiling method is that IF the premises are true and IF the hypotheses are valid, then the conclusions are also true. However, this could hardly be the case, since rarely is an profiler’s deductive opinions about what may have occurred at a crime is empirical or theoretically driven by research and hypothesis testing. Drawing deductive inferences about crime scene behavior produces truths and conclusions out of thin air. Consequently, for deductive profiling to be robust there must be truths that are known a priori, which can only be achieved through empirical research.
It is further argued that there are a number of factors that interfere with adequate empirical evaluation into the variables that contribute to the success or failure of a profile, and what is needed is more of a systematic and empirical approach to offender profiling than currently exists. Contrary to deductive profiling where generalizations guide the profiles, inductive profiling is an empirically based approach where conclusions are derived from scientific analysis. Inductive reasoning is from the specific to the general. For example, inductive profiling derives general principles about the behavior of serial murderers by empirically examining and testing particular facts or instances of a large number of solved cases. Briefly, not including information on unsolved cases in the inductive analysis would not be helpful, because no background information on the offender is available. In contrast to the deductive process, which starts with assumptions about behavior, inductive profiling relies on data gathered from the crime scenes, police reports, psychological evaluations, method examiners’ reports and victimology reports in order to be empirically analyzed and subsequently to support a theory. In all instances the internal validity of the data is confirmed.
An important step in the inductive profiling method is to formalize operational definitions (hypotheses) for testing. Once this step has been completed, the next step involves coding the data for statistical analysis in which the results are supported with theories. The important item that makes inductive profiling more robust than deductive profiling is, through research observable patterns in the data can be found that lead to new theories. Rather than being stagnated with held views of behavior, inductive profiling is like a revolving door – researchers are always striving to look for emerging patterns in crime data, which leads to new ways to assist investigators.
If, for example, an a profiler or investigator in a serial murder investigation deduced from their current ongoing investigation specific knowledge that serial murderers were preferential about selecting particular victims, they may decide only to include cases for consideration where the victim targeting is similar. However, profilers and investigators should not make such an assumption and cases should only be excluded where there is inductive reasoning for exclusion or their is alternative information which can exclude the cases. As previously mentioned, current profiling decisions are made deductively on the basis of experience of the individual decision makers. However, this book argues that such expertise must be validated through the development of theories and statistical models of criminal behaviors, which are based on empirical data relationships. As such, decision-making in criminal investigations should move from deductive to inductive. Adopting this approach would help to minimize the biasing effect of individual experience upon decisions which are based on all the circumstances of any particular case available to the individual making the decision.