Although criminologists often explore the occurrence of offences, the patterns of these offences are rarely examined. However, it is precisely these patterns of offences which could help to predict and gain a greater understanding of offending behaviour and its causes as well as contribute to the improvement of prevention. In this project, offence patterns were determined by looking for offence combinations within the criminal career of individuals. To this end, the research analysed criminal specialisation in several ways. Firstly, specialisation was explored within the offence categories and secondly, the relationship between different offence categories was analysed to define the occurrence of typical offence patterns within that specialisation.

The aim of the re­search pro­ject was to de­term­ine sim­il­ar­it­ies between of­fence types on the basis of data from the Freiburg Co­hort Study. The sim­il­ar­ity of of­fences is de­term­ined em­pir­ic­ally by the fre­quency of the com­mon ap­pear­ance of com­bin­a­tions of of­fences with­in the crim­in­al ca­reers of in­di­vidu­als.

It is as­sumed that of­fences com­mit­ted by one and the same per­son show sim­il­ar­ity. First, a cri­terion for the sim­il­ar­ity of of­fences was cre­ated and fixed. Sim­il­ar­ity can be defined by means of the pen­alty for the of­fence or by means of oth­er cri­ter­ia. In this pro­ject, none of these a pri­ori cat­egor­isa­tions were used. In­stead, the sim­il­ar­ity of of­fences was iden­ti­fied em­pir­ic­ally. The fre­quency of of­fence pairs with­in the crim­in­al ca­reers of in­di­vidu­als were used as a sim­il­ar­ity meas­ure­ment. The fre­quency of of­fence pairs were com­pared with the fre­quency of ran­dom of­fence pairs. If one of­fence pair oc­curs more of­ten than it would in ac­cord­ance with a ran­dom dis­tri­bu­tion of of­fences with­in crim­in­al ca­reers, then these two of­fences are sim­il­ar. Al­tern­at­ively, if one of­fence pair oc­curs more sel­dom than an­ti­cip­ated, then these two of­fences are dis­sim­il­ar. There­fore the Ad­jus­ted Stand­ard­ized Residue (ASR) is suit­able for meas­ur­ing and ana­lys­ing of­fence sim­il­ar­it­ies. In con­trast with oth­er stud­ies on the top­ic of spe­cial­isa­tion, in the present pro­ject the trans­ition from one of­fence to an­oth­er is not im­port­ant. All com­bin­a­tions of of­fences with­in the crim­in­al ca­reer of a per­son are equally im­port­ant, not just the next of­fence.

For this study po­lice and court data dif­fer­en­ti­ated by gender and by na­tion­al­ity was used.

One con­clu­sion of the study is that all of­fences are sim­il­ar with them­selves. So spe­cial­isa­tion is vis­ible. The tend­ency to­wards spe­cial­isa­tion is greatest for fraud, drug and sexu­al of­fences. The study also re­veals a tend­ency for ‘spe­cial­isa­tion’ re­gard­ing court re­gistered traffic of­fences.

The sim­il­ar­ity of of­fences is rep­res­en­ted with the help of the Mul­ti­di­men­sion­al Scal­ing (MDS) meth­od. MDS is a meth­od that rep­res­ents meas­ure­ments of sim­il­ar­ity among pairs of ob­jects (here of­fences) as dis­tances between points of a low-di­men­sion­al mul­ti­di­men­sion­al space, es­pe­cially a two-di­men­sion­al space. MDS at­tempts to mod­el data as dis­tances among points in a geo­met­ric space. The main reas­on for us­ing this ap­proach is to achieve a graph­ic­al dis­play of the struc­ture of the data, one that is much easi­er to un­der­stand than an ar­ray of num­bers. With MDS, the pat­tern of of­fences can be clearly de­pic­ted.

The fol­low­ing fig­ure shows the of­fence pat­tern of court re­gistered Ger­man males pro­duced with MDS. What are im­port­ant in this dia­gram are only the dis­tances between the points in re­la­tion to each oth­er: the factors of "above and be­low" or "left and right" are ir­rel­ev­ant. The closer the points are (short dis­tance) the more sim­il­ar the of­fences are. A dis­tance of 0.7 units relates to an ASR of zero. That means of­fences with a dis­tance less then 0.7 units are sim­il­ar. If the dis­tance is great­er then they are dis­sim­il­ar.

A two-dimensional MDS representation of the offences from court registrations (German men)

A two-di­men­sion­al MDS rep­res­ent­a­tion of the of­fences from court re­gis­tra­tions (Ger­man men)

Sim­il­ar of­fences are (the points have a small dis­tance between each oth­er) as­sault, ag­grav­ated as­sault, li­bel, vi­ol­a­tion of pri­vacy, sexu­al crimes and hom­icide. The sim­il­ar­ity of vi­ol­ent of­fences per­mits the con­clu­sion that there are com­mon causes and motives for these of­fences. Bod­ily in­jury caused by neg­li­gence and vi­ol­ent crime are not sim­il­ar to one an­oth­er. Drug of­fences are sim­il­ar to theft and to fare dodging, which ap­pears to sug­gest a re­la­tion­ship between drug of­fences and drug-re­lated crime. The point for traffic of­fences is far away from the points for the oth­er of­fences. There­fore the dis­sim­il­ar­ity of traffic of­fences from ‘nor­mal’, clas­sic­al and con­ven­tion­al crimes is ap­par­ent. The qual­ity of the two-di­men­sion­al present­a­tion of the of­fences is very good (R² = 0.76).

In dif­fer­ence to court re­gistered cases, po­lice re­gistered ag­grav­ated theft among Ger­man males, is loc­ated re­l­at­ively far away from the oth­er of­fences and is thus dis­sim­il­ar to all oth­er of­fences by po­lice re­gistered Ger­man males. Ag­grav­ated theft (burg­lary) of­ten oc­curs in a series and is there­fore of­ten re­gistered mul­tiple times in the po­lice data. Thereby the de­gree of spe­cial­isa­tion is very high amongst crimes of burg­lary when re­gistered with the po­lice and ac­cord­ingly the fre­quency of com­bin­a­tions from oth­er of­fences with burg­lary is sel­dom. In court data a seri­al set of burg­lar­ies may come to a single de­cision in a single court case. A one-time re­gistered burg­lary in court data could thus be re­gistered mul­tiple times in po­lice data. There­fore the dis­sim­il­ar­ity from ag­grav­ated theft with oth­er of­fences which is ap­par­ent from the data of po­lice re­gistered males is not in fact ap­par­ent amongst court re­gistered males.

Of­fences against the For­eign­ers Act by for­eign na­tion­als sel­dom oc­cur in com­bin­a­tion with com­mon crime. Only for­gery and fare dodging re­veal a re­la­tion­ship to ‘for­eign’ crime.

Spe­cial res­ults were aroused by the ana­lys­is of sep­ar­ate age-groups. Many res­ults were con­sist­ent in dif­fer­ent age-groups, but a few res­ults show that the re­la­tion­ship between dif­fer­ent of­fences var­ies in ac­cord­ance with dif­fer­ent age phases. Es­pe­cially amongst sexu­al of­fences, a tend­ency ex­ists to sim­il­ar­ity between sexu­al crime and vi­ol­ence by young people. This is not ap­par­ent with adult of­fend­ers.

The study also con­cen­trated on de­term­in­ing types of activ­ity over the life course. In a first step, of­fence pat­terns of five-year peri­ods were cre­ated with the help of prob­ab­il­ist­ic cluster ana­lys­is. In a second step, vari­ations were ana­lysed by age.

The prob­ab­il­ist­ic cluster ana­lys­is poin­ted to sim­il­ar­it­ies amongst vi­ol­ent of­fences. The ana­lys­is offered one cluster of vi­ol­ent of­fences. The es­tim­ated prob­ab­il­ity of an age strip in this cluster hav­ing a vi­ol­ent of­fence (as­sault, vi­ol­a­tion of pri­vacy, rob­bery, hom­icide, sexu­al crime) and also li­bel and crim­in­al dam­age is above av­er­age. An­oth­er res­ult is a small cluster of ver­sat­ile of­fences. All of­fences in this cluster are above av­er­age and the av­er­age num­ber of of­fences in an age strip is very high. In this cluster every age strip con­tained on av­er­age six of­fences which is three times more then all age strips on av­er­age con­tained. There­fore this cluster con­tained ‘chron­ic’ of­fend­ers.

The prob­ab­il­ist­ic cluster ana­lys­is pro­duced two clusters for court re­gis­tra­tions of traffic of­fend­ers. One cluster con­tained traffic of­fences without per­son­al in­jury, the oth­er with per­son­al in­jury.

An ad­di­tion­al res­ult of the prob­ab­il­ist­ic cluster ana­lys­is offered an in­sight in­to the routes taken by of­fend­ers in their crim­in­al ca­reers. Two is­sues of the crim­in­al ca­reer – spe­cial­isa­tion and of­fend­ing path­ways – were ana­lysed. A spe­cial­ist of­fend­er is defined as someone who stays with­in the same of­fend­ing cluster. Spe­cial­isa­tion in­crease as of­fend­ers grow older, without tak­ing in­to ac­count de­sisters. For any of­fend­ing cluster a tend­ency of spe­cial­isa­tion was vis­ible, in the sense that the of­fend­ers stay in the same cluster in the five-year peri­od be­fore and in the five-year peri­od after. Nev­er­the­less the biggest group with­in the of­fend­ing path­ways is the group of of­fend­ers who de­sist from crime. Only in the cluster of the ‘chron­ic’ of­fend­ers is the be­fore and after ‘de­sister’ group not the biggest.