2. Methodological framework

2.1 Research sources and criteria

The three main methods used to identify documents covered were: 1) computerized data banks and documentary research engines, 2) the "snowball" method, which involved consulting previously obtained text references, and 3) various field experts’ recommendations.

The data banks and research engines consulted included Cambridge Journals Online, Cambridge Scientific Abstracts, Canadian Research Index, ERIC, FRANCIS, Google Scholar, JSTOR, ProQuest, PsycARTICLES, PsycInfo, Research Library, Sage Journals Online and Science Direct.

Thirteen organizations were contacted for their profiling expertise. By the report deadline, four of these organizations had accepted our request for collaboration, while two had declined. No response was received from the other organizations.

We must also acknowledge the exceptional collaboration of the Ontario Provincial Police (Angela Eke, Kathryn J. Lines and Jim Van Allen), the International Criminal Investigative Analysis Fellowship (Glenn Woods) and the Belgian Federal Police (Françoise Godefroid).

Organizations Contacted and Their Response, October 31, 2008

The following subsections will describe our research criteria and classify the 277 documents obtained.

2.1.1 Document origin

Texts originated from a range of disciplines, primarily criminology (200), psychology (70) and law (42). Documents from the field of education dealt specifically with the phenomenon of school shootings.

Document Origin ¹

2.1.2 Year of document publication

Year of Document Publication

Documents consulted covered a period from 1965 to 2008. The scientific community’s interest in profiling has continued to soar since the early nineties. The volume of publications obtained grew from five between 1980 and 1989 to 188 between 2000 and 2009.

2.1.3 Type of document

Various types of documents were consulted; most were scientific articles in refereed journals (196). This type of publication, as well as reports submitted to various governmental organizations, constituted the most relevant and, generally, most credible data sources.

Types of Documents

2.1.4 Language of publication

The research study team consulted 262 English-language and 15 French-language documents. (Documents available in both languages were accounted for in the language in which they were read.)

2.2 Keywords

Document research involved the use of computerized data banks and document search engines such as Google Scholar, PsycInfo, Science Direct and so on. The key words used in the searches included behavioural profiling, crime mapping, crime pattern, crime scene profiling, criminal assessment, criminal profiling, distance decay, ethnic profiling, geographic profiling, investigative psychology, offender profiling, personality profiling, profiles, profiling, prospective profiling, psychological profiling, racial profiling, recidivism, repeat offending, school shootings and terrorist profiling.

2.3 Document classification

2.3.1 Subject treatment

Documents were initially classified based on their treatment of the subject of profiling. Empirical treatment, which assumes the gathering and analysis of data, was the treatment preferred for evaluating the effectiveness of profiling (144 texts).

Subject Treatment

2.3.2 Approach

The documents were then classified based on their approach to the subject. The most frequent empirical approach was the quantitative approach (93).

Subject Approaches ²

2.3.3 Country

The documents consulted covered profiling experiences in ten countries over four continents: North America (Canada and the United States), Europe (Belgium, Finland, France, the United Kingdom and Sweden), Asia (South Korea and Japan) and Oceania (Australia).

2.3.4 Profiling criteria

Of the texts consulted, 198 dealt with behavioural profiling, 127 were based on sociodemographic criteria (including racial profiling) and 49 dealt with geographic profiling. More than one type of profiling criteria may be covered in the same text.

2.3.5 Type of intervention targeted

Type of Intervention Targeted ³

The type of action targeted by profiling may be divided into six categories: 1) criminal arrest, 2) crime prevention without credible information, 3) crime prevention with credible information on the potential crime, 4) crime prevention with information on the crime and the potential criminals, 5) assessment of an offender’s dangerousness or recidivism risk, and 6) expert testimony during the trial phase.

2.4 Analysis criteria

The analysis of documents gathered, particularly empirical research findings, was based on two criteria: level, and strength of evidence. Studies with major deficiencies in methodology that undermined credibility were systematically eliminated.

2.4.1 Level of evidence

Level of evidence is the first criterion used to judge the methodology proposed by the analyzed study. Documents were classified on the basis of three levels of evidence:

Theoretical: the document does not cover the effectiveness of profiling directly –it may not even cover profiling directly – but it empirically covers the links between crime scenes, criminal behaviour, personality traits and so on. These links contribute to the establishment of theoretical bases suggesting the possible or probable effectiveness of profiling. Salfati’s work is an example that belongs in this category;

Restricted empirical: the document deals directly with the effectiveness of profiling but in an artificial, laboratory context. The study of profiling outside its natural context of use allows at best for support of possible or probable empirical effectiveness. Most of the empirical articles on profiling, including Kocsis’ work, belong in this category (when they are sufficiently solid in methodological terms);

Broad empirical: the document is a direct study of the effectiveness of profiling in an authentic situation; that is, the analyzed data comes from the practice of real profiling by police. Copson’s study (1995) is one of the rare examples of this type of text.

2.4.2 Strength of evidence

Strength of evidence describes the magnitude of support provided by the results of research in favour of the hypothesis that profiling is effective. Here again, documents deemed sufficiently credible were classified on the basis of three hierarchical levels:

1. No support: the results are credible and suggest that profiling does not constitute an effective method of investigation or prevention. This is especially the case for studies on racial profiling;
2. Possible support: the results are credible but do not satisfy all the scientific criteria to establish a causal relationship between the use of profiling and an improvement in the results of police investigations or crime prevention. Nevertheless, the effectiveness of profiling remains one of the hypotheses that could explain the results;
3. Strong support: the results are credible and meet the scientific criteria to establish a reasonable cause-and-effect relationship between profiling and the success of investigations and crime prevention. In this case, there is no doubt as to the effectiveness of profiling.

1. Some of the documents fall under more than one discipline.
2. More than one approach may be used in the same document.
3. More than one type of intervention may be used in the same document.

Previous Page Table of Contents Next Page