Printer Friendly


The Case of Temporally Intensified Shoplifting Behaviour

This article examines the prevalence and correlates of shoplifting phases among Finnish adolescents. The data are based on the Finnish Self-Report Delinquency Study (FSRD-96) whose target population consisted of 9th grade (15-16-year-old) students. The identification of phases was based both on behavioural sequences and subjective interpretations. Consequently a shoplifting phase was defined in a relative fashion as a period during which the respondent shoplifted more often than usually. Of the respondents, 23 per cent had experienced at least one such period. The typical (modal) shoplifting phase began at the age of 13 and ended at the age of 14. It lasted for 26 weeks and amounted to 15 thefts. Among the correlates of phasic shoplifting behaviour were participation in other offences and low level of parental control.

Juvenile delinquency is almost by definition a period in a person's life. In this article, I take a closer look at the `phasic' nature of adolescent crime. Roughly, there are two senses in which we can speak of delinquent phases. First, in the objective sense the time between onset and termination of offending is a phase in the life of the offender. Secondly, there is a subjective sense in which a person can see his or her delinquent career as a past or ongoing `phase', or even to expect such a `phase' in the future. This article aims at drawing attention to the extent to which juveniles experience periods of intensified shoplifting behaviour.

Why shoplifting? According to self-report data, shoplifting is among the most prevalent crimes committed by Finnish adolescents (see Appendix). In a series of qualitative studies at the beginning of the 1990s, I concluded that shoplifting was the `culturally paradigmatic' crime: if asked to write about their crimes, the adolescent respondents typically chose to write about theft. Some respondents even volunteered the interpretation that they had experienced a stealing phase while others commented that it is normal for adolescents to experience episodes of deviant behaviour (Kivivuori 1994). The concept of delinquent phase seemed to emerge, so to speak, from the social world, unimposed by the researcher. The unobtrusive beginning of these early studies is reflected in an attempt to explore, in this article, relatively defined periods of intensified shoplifting. Such phasic episodes are temporary phases against the backdrop of each adolescent's typical or `normal' crime involvement.

Delinquent Phase as a Group-level Phenomenon

Crime typically peaks during adolescence, although there is some cultural and historical variation in the shape of the age-crime curve. Criminologists have debated the question as to which is more important: the stability of the overall pattern, or the instability of the curve's detailed outline. There is, however, no dispute about the fact that crime and adolescence go together (Farrington 1986; Gottfredson and Hirschi 1990). Moreover, the link between the two has grown stronger in the post-war period (Steffensmeier et al. 1989).

Figure 1 shows the age distribution of persons(1) suspected of shoplifting in Finland in 1996. The male curve peaks at ages 16 and 18, while the female curve peaks at the age of 16. Females seem to desist earlier than males. However, the outline of the curve is similar. Shoplifting peaks during adolescence. Persons aged 13 to 19 committed 23 per cent of all cleared shoplifting offences in 1996.


The increase in crime during the teenage years reflects changes in the prevalence of offending (the number of persons committing crimes increases) (Farrington 1986; Sampson and Laub 1992). According to Moffitt (1993), there are two offender types: life-course-persistent offenders and adolescence-limited offenders. While persistent offenders begin their career of antisocial behaviour early in childhood, a much larger group of adolescence-limited offenders engage in crime only in their teens. The concept of adolescence-limited delinquency implies that delinquency is, for the majority of delinquent adolescents, a period or a phase in life with identifiable moments of onset and termination. Similarly, DiLalla and Gottesman (1989) have spoken about transitory delinquents who go through `delinquent phases'.

Adult social bonds, such as marriage and employment, help adolescence-limited offenders to desist from crime (Sampson and Laub 1990), while for some the prospect of these may be enough (Mulvey and LaRosa 1986). Graham and Bowling (1995: 51-66) found that males are less likely than females to desist from offending as a result of acquiring adult status, a British finding that could explain the sex difference in Figure 1, which of course reflects Finnish conditions. It has been suggested that life-course-persistent offenders who continue their criminal careers into adulthood may be genetically predisposed to do so (DiLalla and Gottesman 1989: 347).

Aggregated age-crime curves (such as the ones in Figure 1) peak during the teenage years. It does not follow from this that individual adolescence-limited offenders experience just one peak period of offending. It is likely that there are crime-free periods in the middle of individual criminal careers (Farrington 1986), some of which are unrelated to periods of incarceration. In spite of this possibility, the unimodal (one-phase) model of onset and termination has dominated the study of delinquent careers. A straight line drawn between the median ages of onset and termination symbolizes this paradigm.

There is, however, a sense in which criminal careers research has taken shorter `phases' into account. According to the thesis of heterotypic continuity, various forms of offending and problem behaviour are changing manifestations of a single underlying tendency of antisocial behaviour. This implies that criminal careers consist of stages characterized by different offence types: offenders `switch crimes' by desisting from an offence type in the middle of their criminal career (Loeber and Le Blanc 1990; Farrington 1992: 258-9). There are thus `phases' within delinquent careers.

The traditional emphasis on onset and termination can be supplemented with a perspective that focuses on the lower-level pattern of shorter phases structuring a delinquent career. Such patterns can be reconstructed from official data or observed in follow-up studies. For example, Farrington et al. (1986) found in a longitudinal study that periods of unemployment are also periods of (property) crime for youths who have recently left school. Another and more controversial example is the link between so-called premenstrual syndrome (PMS) and female criminality (Russell 1995: 101-5). The PMS hypothesis predicts very short but frequent phases. According to Mawson (1987), physiological factors may have a role not only in PMS but in all transient crime. He attributes out-of-character criminal episodes to a socio-bio-social chain: stressful environmental changes cause people's cognitive maps to deteriorate, which in turn brings about various physiological changes and intensified sensation seeking.

Phase as a cultural interpretation

The way in which an offender understands his or her career in crime is, to some extent, different from the offender's objective career. The offender may never see or speak of his or her career as a `period' or a `phase' in his or her life. He or she may do so retrospectively. He or she may also understand his or her behaviour as a `phase' before or while actually experiencing it. In other words, periods may be understood as periods prospectively, in advance of their commencing or termination. It is possible that juveniles have a culturally mediated preconception about adolescent `phases' or periods, and that these conceptions are `lived up to'.

There is no shortage of information about the `turmoil phase' of adolescent psychological and physical development, at least in Finland. School textbooks, for example, provide descriptions about adolescent problem phases. In a recent survey of teacher attitudes,(2) 27 per cent of the upper grade comprehensive school teachers (N = 897) held that `during puberty, it is normal for youths to commit illegal acts now and then', while 18 per cent agreed with the statement that `it is understandable that `juvenile creativity sometimes leads to illegal acts'. In an earlier preliminary study, adolescent respondents mentioned `puberty' and `experimentation' as causes of `juvenile shoplifting phases (Kivivuori 1994).

In a longitudinal qualitative study of 60 German adolescents, Matt (1995: 155) described their involvement in crime as episodic. He also noted that the juveniles themselves understood their delinquency as a phase in the `normal' course of adolescent development. Israeli car thieves studied by Hazani (1991:185) `justified their offences by saying that young manhood is a stage when crime is biologically natural. It seems safe to say that, in the western world, young people tend to know the developmental stages of adolescence.

To a degree, developmental psychology, as a cultural phenomenon, `produces what it means to learn' (Walkerdine 1984: 191). This idea, however, goes too far, if it means that developmental processes are fully `constructed'. Biological (hormonal) processes also influence adolescent behaviour (Buchanan et al. 1992). The traditional `storm and stress' conception of adolescence is not, however, fully reducible to biological realities. Buchanan et al. argue that such stereotypical ideas about adolescent phase behaviour can function as self-fulfilling prophesies, because beliefs moderate or fail to moderate reaction to hormonal cycles (p. 69). Phase as a cultural interpretation influences how juveniles perceive the normality of various behavioural options.

To sum up the argument so far, the phase phenomenon has two aspects. On one hand the aggregated `juvenile crime peak reflects the weakening of informal social controls during adolescence. Juveniles are no longer closely supervised by their parents, and they do not yet have wives (or husbands) and employers to supervise them. Deliquency results from a lack of control, not from the presence of social norms to commit illegal acts (Gottfredson and Hirschi 1990).

On the other hand, `phase' is a cultural interpretation and, as such, a `positive norm' which may push individual would-be offenders towards delinquency. Phase interpretations reflect the reflexive character of human activity: we do not merely behave, our acts are meaningful under familiar descriptions.

Can these sense-making interpretations be themselves interpreted as causal agents? At least according to some forms of control theory, `descriptions favourable to the violation of law', for example excuses and justifications, play a causal role in delinquency (Matza 1964). As a cultural rhetoric, `phase' carries excusing connotations by indicating that something `out of character' was taking place during the phase. In this sense, phase interpretation is used as a `neutralization' (Sykes and Matza 1957). This analysis, however, is one-sided because it ignores the role of desistance in the constitution of periodical offending. Phase interpretations are also descriptions favourable to law-abiding behaviour, since phases must end in order to count as phases. Phase interpretations sensitize individuals to the prospect of desistance: offenders who tend to interpret their participation in delinquency as a temporary `phase' find it natural to desist. This effect is difficult to verify, because those who desist for other reasons are more likely to see their delinquency retrospectively as a temporal phase.(3) This of course does not mean that they desisted because they interpreted their delinquency as a phase.

The Data

According to Loeber and Le Blanc (1990: 399), the study of changes in individual offending requires at least two data points if a developmental sequence is to be identified. There is, however, another approach that can be used in the study of shorter phases. We can ask the offenders whether they have experienced delinquent phases, a method used in the present article.


The research is based on the Finnish Self-Report Delinquency Study (FSRD-96) conducted by the National Research Institute of Legal Policy, in which 4,204 ninth-grade pupils from 45 municipal comprehensive schools completed a self-administered questionnaire in spring 1996. The schools were a random cluster sample with geographical area and community habitation density as stratification criteria. The schools were selected from a register in which only Finnish-language municipal schools, representing 87 per cent of the total pupil population, were included. Swedish-language, private and state (teacher education) schools were excluded.


In each school, all ninth grade(4) pupils comprised the target population, including those referred to special education classes because of disciplinary or learning problems. The questionnaire was completed during regular school hours. The classes were supervised by a liaison teacher, who had been thoroughly briefed about the required procedures in the administration of a self-report study. The pupils were advised not to sign the questionnaire, which, when completed, was sealed into an envelope by the pupil. The sealed envelopes were sealed into a larger box by the liaison teacher; pupils were allowed to watch this procedure. According to Bjarnason (1995), using teachers, instead of outside researchers, to administer a self-report survey does not reduce the validity of the results, if a sealed-envelope system is adopted. According to Kreuzer et al. (1992: 92-3), undertaking a survey in a group (classroom) situation does not compromise the validity of the self-report method compared to the more private mail survey.


At any time in normal school work, some pupils are absent from classes. In the present study, 9.7 per cent (439) of the target sample (4,539) were absent. From the point of view of criminological investigation, absenteeism potentially reduces the validity of the results, since it is possible that the most delinquent adolescents truant more often from classes. In order to assess this effect, the liaison teachers were asked to evaluate the absent pupils. Forty-three per cent were described as average; 41 per cent as more disciplined than average, and 15 per cent as less disciplined than average. These figures indicate that `good' (more disciplined than average) pupils are over-represented in the absent group, mainly because their good behaviour seemed to be associated with a family-oriented lifestyle, and consequently with greater likelihood of permitted leave (running errands, vacation with parents, etc.). If sex is controlled for, `good' female pupils are disproportionately absent, while in comparison the distribution of male pupils is more even ([chi square] = 27.9 df = 2, p = .000). It is possible that research in the Finnish school context undersamples `nice' (especially female) students and consequently overestimates the prevalence of offending. At the same time, incidence figures may be underestimated by the absence of some of the `bad' pupils. To some extent, these effects may counterbalance one another.


The 18 offence types included in the FSRD-96 survey are shown in the Appendix. The self-report questions were based, with some modifications, on the questionnaire used in the International Self-Report Delinquency Study (ISRD, see Junger-Tas et al. 1994). The theoretical background of the ISRD questionnaire was control theory with an emphasis on social relations. Questions probing crime-related attitudes and shoplifting phase behaviour were developed and added to the FSRD-96 questionnaire by the present author.


The reliability of the standard self-report questions relies on previous methodological research (for example, Hindelang et al. 1981: 75-86). Long (such as lifetime) recall periods arguably weaken the reliability of the results. Although I cannot prove it, I believe that it is easier for a 15-year-old person to remember correctly whether he or she has ever experienced a shoplifting phase than it is to recall individual offences or the number of non-phase offences. The phrase `more often than usually' in the question emphasizes out-of-the-ordinary episodes.

Although there has been no test-retest study, the phase question has been used in two very similar samples. In 1995, a pilot study (FSRD-95) was conducted in 13 of the FSRD-96 schools. The 1995 phase prevalence was higher (26.5 per cent, N = 1195) than in 1996 (21.1 per cent, N= 1088).(5) However, a change in the same direction, and of the same magnitude, took place in the standard self-report item of lifetime theft participation (from 58.5 per cent to 54.5 per cent). The decrease may reflect real differences between two consecutive grade cohorts or some kind of systematic error, but there is no indication that the phase question behaves differently from the standard self-report item.

The purpose of the following analysis is, first, to describe the essential features of phasic shoplifting among Finnish adolescents and, secondly, to explore if and how phase behaviour is related to sex, age, other offences and social control variables. The analysis is cross-sectional and tentative. As such it cannot show that culturally available phase interpretations cause juveniles to start or terminate shoplifting.

Prevalence of Shoplifting Phases

In the FSRD-96, the relevant filter question on phases was: `Have you ever had a phase during which you stole from stores or kiosks more often than usually?' The question is relative, asking the respondent to recall whether he or she has experienced a period of intensified shoplifting. The resulting data describe the extent and correlates of out-of-the-ordinary shoplifting sequences. The prevalence of temporally intensified shoplifting behaviour call be calculated in two ways. First, we may ask how many of all respondents have experienced at least one shoplifting phase. Second, we may ask how many of those students who have shoplifted at least once have ever experienced a period of intensified shoplifting. In order to assess the quantitative significance of the phase phenomenon, both figures are shown in Table 1. Phase prevalence is moderately higher among females than among males. It is also higher among students living in urban areas than among students who live in semi-urban and rural areas.

TABLE 1 The lifetime prevalence of temporally intensified periods of shoplifting (%)(*)

                                  All        Females      Males

Phase prevalence (%), of all      22.6       23.2         21.9
respondents (N)                   (4,176)    (2,075)      (2,073)


Phase prevalence (%),
of respondents                      40.5     44.4         36.9
who have shoplifted at            (2,309)    (1,082)      (1,218)
least once (N)
                                             p < .001

                                       Community type

                               Urban        Semi-urban   Rural

Phase prevalence (%), of all   25.5         20.3         16.5
respondents (N)                (2,295)      (1,217)      (655)

                               p < .001

Phase prevalence (%),
of respondents                 43.0         38.8         33.0
who have shoplifted at         (1,351)      (634)        (324)
least once (N)
                               p < .01

(*) `Have you ever had a phase during which you stole from stores or kiosks more often than usually?' Of all FSRD-96 respondents, 16 per cent had a single phase experience. Five per cent of the pupils had experienced two periods of intensified shoplifting. Only 2 per cent reported more than two phases. The more phases a pupil has gone through, the more thefts he or she committed during the most recent phase.

Specialization in the choice of objects

As shown in Table 2, 44 per cent of those who had experienced a phase reported that they had stolen sweets (candy) during the most recent phase. Twenty-six per cent had specialized in some other type of goods, the most popular choice being the category of `make-up, jewellery and clothes' (10 per cent).(6) About one-third reported some combination of goods, typically candy and tobacco, or candy and jewellery or clothes (non-specialized shoplifting).

TABLE 2 Specialized and non-specialized stealing during the most recent shoplifting phase, by sex and phase incidence (%)

                                 All     Females        Males

Specialized shoplifting
  candy                          43.7      35.2          53.0
  other                          25.7      29.0          22.0

Non-specialized shoplifting      30.7      35.8          25.1
  All                           100.1      100          100.1
  (N)                           (877)     (458)         (419)

                                           p < .001

                                   Lifetime phase incidence

                                 One          Two        More

Specialized shoplifting
  candy                          50.5         29.4       23.3
  other                          23.5         31.4       30.1

Non-specialized shoplifting      26.0         39.2       46.6
  All                            100           100        100
  (N)                           (608)        (194)       (73)

                                 p < .001

The `only candy' and `non-specialized shoplifting' categories were associated with the lifetime incidence of phases. The diphasic and multiphasic shoplifters not only stole more during the latest phase (see above), but typically stole a wider variety of objects, not specializing in any single product type (Table 2). From another point of view, this means that first-time phasic shoplifters tended to concentrate on just one type of goods, typically candy.

Duration and intensity

The typical (modal) duration of the shoplifting phase was 26 weeks. Sixty-two per cent of the (most recent) phases lasted between 1 and 26 weeks. There were few differences between males and females in this respect. However, males were over-represented in the category of phases lasting more than a year ([chi square] = 6.2, df = 1, p = .012).

During the typical phase, 15 thefts were committed. Since phases differ in their duration, it is more useful to look at the intensity of the phase (the average number of thefts per one phase week). The mean intensity in the FSRD-96 sample was 1.7 thefts (1.57 for females and 1.85 for males). It seems that the female phases were typically slightly shorter and less intense than male phases, although these differences were statistically insignificant. In fact, it may be worth emphasizing that the differences are so small. There seem to be no important sex-based differences in the way the concept `phase' is understood, so that males and females would think of different things when responding to the phase questions.

Onset and termination

The typical (modal) shoplifting phase began at the age of 13 and ended at the age of 14. Figure 3 shows the numbers of FSRD-96 respondents beginning and ending a shoplifting phase at each age level. Only students with one phase experience are included. Follow-up questions were about the onset and termination of the most recent phase, so that the inclusion of students with multiple phases would have been problematic in a retrospective age-phase analysis. (However, the shape of the curve is similar even if they are included).


The earliest phases began around the age of 5. After this, the number of beginners gradually increased, but remained at a low level. Between ages 9 and 10 the number of phase beginners tripled and then rose steeply. Phase onset peaks at the age of 13. At the age of 14, the number of phase terminations exceeded for the first time the number of phase onsets.

The influence of sex on phase onset and desistance is controlled for in Figures 4 and 5. The differences between males and females were marginal and insignificant in both phase onset and desistance. This is in line with Figure 1, which shows that the first (`upward') part of the age-crime curve does not differentiate between males and females. Differences began to show in late adolescence, for which we lack phase data. However, the data suggest that if delinquency is measured by the prevalence of self-reported phases of intensified shoplifting, instead of cleared offences, the adolescent peak still appears.


The duration of the phase was related to the age of onset. The earlier the onset, the longer the phase lasted (for both females and males). This effect was partly caused by technical data truncation, since phases continuing at the time of the survey were excluded. It is, however, possible that early (childhood) and late (puberty) phases require separate explanations.

Self-reported Reasons for Onset and Desistance

The pupils were asked why their latest shoplifting phase began. Reasons that were often cited were lack of money (26 per cent of those with at least one phase) and the excitement of stealing (26 per cent). The wish to experience something new was cited by 13 per cent of the pupils with at least one phase. A compulsive need to steal was reported by 8.9 per cent of females, while 3.6 per cent of the males described their latest phase as a compulsive experience ([chi square] = 10.8, df = 1, p = .001). All in all, the FSRD-96 respondents shoplifted for roughly the same reasons as the respondents of studies on non-phase shoplifting behaviour (Belson 1975; Klemke 1982; Cox et al. 1990).

The pupils were also asked to assess why their latest phase came to an end. The most often cited reasons were getting bored with stealing (32 per cent), getting caught (21 per cent) and growing wiser (11 per cent). In other words, spontaneous desistance was very prevalent, accounting for two phase terminations in five. Of the phases ending at the age of 14, 47 per cent were terminated because the offender `got bored' or `grew wiser'. On the other hand, apprehension (`getting caught') was also an important factor leading to phase termination.

Females were no more likely than males to terminate their latest phase because of an actual apprehension, suggesting that the phasic aspect of their shoplifting did not reflect more stringent external control. However, they were somewhat more likely than males to terminate their latest phase as a result of fear of apprehension. Females desisted from their latest phase for this reason(7) in 11.6 per cent of the cases, while 7.5 per cent of the males did so ([chi square] = 4.5, df = 1, p = .034). Males were unlikely to terminate their phases for this reason at any age. ,Among female respondents, late phases were more likely to be terminated because of fear of apprehension (age-based comparisons were statistically not significant). It is possible that the minimum age of criminal responsibility (15 in Finland) has a greater deterrent effect on females, especially if they perceive ill-repute as being more costly than do males (as suggested by Heidensohn 1985: 183-87).

It needs to be stressed that the above-described reasons (for both phase onset and termination) are self-reported reasons, reflecting how the young people themselves experienced their latest shoplifting phase. They should not be equated, without reflection, with the causal sources of shoplifting phase behaviour. It is likely, however, that reported reasons for phase termination are better estimates of real causes (of termination) than reasons for phase onset are (of onset).

Compulsive shoplifting phases

Six per cent of the students with at least one phase experience reported that they had stolen because of a compulsive need to steal. This type of phase resembles the phenomenon of `kleptomania'. In what follows, I compare these students (N = 58) with students (N = 861) who entered a shoplifting phase for other, more common reasons such as excitement or lack of money.

There were more female (42) than male (16) students among the compulsive shoplifters. Compulsion as a self-reported reason for shoplifting was more common in urban and semi-urban areas than in rural areas, although this difference was not significant. Compulsive; behaviour was not more prevalent among children of unemployed or absent fathers. Compulsive phases did not differ from other kinds of phases with respect to onset, termination, or the number of thefts committed during the phase. The choice of objects did differ from the non-compulsive group: the category `make-up and jewellery' was more popular among the compulsives than among the non-compulsives, an emphasis that is probably related to the female majority of the former group.

Of the compulsive shoplifters, 47 per cent had experienced more than one shoplifting phase, while the same figure for the rest was 30 per cent ([chi square] = 7.34, df = 1, p = .007). It seems that compulsive shoplifters have a higher risk of multiple bouts of intensified shoplifting. However, they do not participate more in other offences.

Compulsion as the self-reported reason for a shoplifting phase does not necessarily mean that the cause of the phase is psychological. Although the tendency to multiphasic shoplifting behaviour is consistent with psychological problems, other interpretations should be considered. For example, if females are expected to behave in a more law-abiding fashion than males, nonconforming females may feel the need to explain their moral guilt away by saying that their competence was reduced during the phase. As a retrospective interpretation, compulsion is an excuse, especially in comparison with other self-reported reasons for phase behaviour (excitement, etc.).

Finally, it should be noted that compulsive shoplifting phase behaviour is a peripheral aspect of both phase behaviour and shoplifting in general. Of those with phase experience, 6 per cent had experienced a compulsive bout of shoplifting. Of those who had shoplifted at least once, 2.5 per cent had had a compulsive phase.(8) Of all FSRD-96 respondents, only 1.4 per cent reported a compulsive shoplifting phase.

Phase Risk Factors: A Tentative Analysis

A logistic regression model was used to assess the effects of variables stemming from control theory on the probability of experiencing relative intensification of shoplifting behaviour (phase). In this tentative analysis, only students who had participated in shoplifting were included.

The respondents who reported at least one phase during their lifetime scored 1 on the dependent variable, while others scored 0. Of the independent variables, lifetime variety of participation(9) in delinquency was measured by a sum variable with a range from 0 to 17 offence types. Shoplifting was not included in this construct. In the regression analysis, this sum variable was recoded into four categories containing (approximately) one-quarter of the respondents. Table 3 disaggregates by sex the prevalence of phases in the resulting categories. It shows among other things that three in four females with a high participation level had also experienced an intensified shoplifting phase.

TABLE 3 The prevalence of temporally intensified shoplifting periods by variety of delinquent participation and sex
                               Lifetime variety of delinquent

                              0-3      4-5       6-8       9-17

Female phase prevalence, %   22.3     39.8      52.4       74.5
(N)                          (323)    (249)     (290)      (184)
Male phase prevalence, %     22.2     28.7      36.8       54.7
(N)                          (248)    (254)     (342)      (331)

p                             n.s.    < .01     < .001    < .001

(a) The number of offence types the respondent has participated in during his or her lifetime, shoplifting excluded. Note that this construct does not tell what the offences are. The higher the variety, the more serious are acts likely to be included (see Appendix).

Parental relations were measured with four questions on how the respondent gets on with his or her mother/father, and whether his or her mother/father knows with whom the respondent spends his or her leisure time (four items, alpha .62). In the analysis of Table 4, the resulting sum variable was recoded into three categories containing (approximately) one-third of the respondents. In the chosen indicator coding scheme, odds ratios reflect relative changes of phase probability in comparison with the omitted reference category, controlling for other variables in the equation.

TABLE 4 The estimated effects of lifetime offence variety, sex, parental relations, family type and urbanization level, on the probability of experiencing a period of temporarily intensified shoplifting(a)
                               B              SE

Lifetime offence variety(b)
  low                          0.56          0.21
  intermediate                 1.01(***)     0.19
  high                         1.77(***)     0.19

Sex (female)                   0.50(***)     0.09

Parental relations(c)
  intermediate                 -0.01         0.10
  close                        -0.30(*)      0.13

Both parents present(d)        -0.25(*)      0.10

Level of urbanization(c)
  urban area                   -0.33(*)       0.14
  semi-urban area               0.18          0.15

                                Wald      Odds ratio

Lifetime offence variety(b)
  low                           14.9         1.7
  intermediate                  81.8         2.7
  high                         223.2         5.9

Sex (female)                    31.0         1.7

Parental relations(c)
  intermediate                   1.1         1.0
  close                         22.3         0.7

Both parents present(d)         13.5         0.8

Level of urbanization(c)
  urban area                     5.2         1.4
  semi-urban area                0.4         1.2

(a) Logistic regression model (N = 2033). (*) p < 0.05; (**) p < 0.01; (***) p < 0.001) (forward stepwise inclusion. -2LL = 2518.1, goodness of fit = 2038.0, model [chi square] = 225.4, df= 9, p = .0000)

(b) Reference category = 0-3 offence types. Low = 4-5 offence types, intermediate = 6-8 offence types, high = 9-17 offence types; shoplifting not included.

(c) Reference category = distant relations with parents.

(d) Reference category = other family type.

(e) Reference category = rural area. Missing data were not replaced by estimates.

The odds of experiencing a period of intensified shoplifting are almost six times higher among students who had wide experience in various offence types, compared with those who had less experience in this respect. Clearly, pupils with shoplifting phase experience were not specialized in shoplifting. Compared with males, females had a higher phase risk, controlling for the other

variables in the equation. Close relationships with parents reduced the phase risk. Students whose natural parents were present in the household were less likely than other youths to go through a shoplifting phase. Finally, the odds of going through a shoplifting phase were 1.4 times higher among city youths compared with youths living in rural areas.

The equation shown in Table 4 correctly predicts the observed values of the dependent variable in 66 per cent of the cases. Since the variables were measured cross-sectionally, the model rests on auxiliary assumptions concerning the temporal order of independent and dependent variables. For example, parental relations are here interpreted as influencing the phase risk. These variables can be related in a reverse order, or recursively.


The main objective of this article has been to draw attention to the fact that juveniles experience periods of intensified shoplifting, and that these phases reflect both behavioural sequences and cultural interpretations. The analysis is based on the retrospective method, with its attendant shortcomings and benefits. The shortcomings include problems of recalling past events. The benefits are related to the comparatively low costs of the method.

Needless to say, both the behavioural and the subjective-interpretive aspect of phase behaviour call for more research. At least two basic research strategies could further illuminate the phase phenomenon. First, a longitudinal panel study consisting of consecutive self-report surveys with short (perhaps one or two month) intervals would produce objective data on the incidence of thefts across time. At the end of the study, the respondents would be asked if they had experienced phases during which they stole more often than usually. This would also illuminate subjective phase interpretation criteria. Another, and equally important, research approach would use qualitative analysis to study the meaning of phasic episodes to adolescents. Are intensified shoplifting periods corollaries of other, perhaps more serious problems of life? The results of the present analysis suggest that this could be the case. The whole personal con text of phasic shoplifting could be further illuminated by using in-depth interviews.(10)

The correlates of shoplifting phase behaviour seem to be the same as in other forms of anti-social and criminal behaviour. To use Klein's (1984) metaphor, a shoplifting phase is one of the many `cafeteria' items selected by juveniles whose ties to sources of social control have been weakened. The results of the phase risk analysis were also in line with control theory predictions. Social control (close family ties, presence of both parents, informal controls in rural areas) reduced the risk. Individual propensity to commit offences (lack of self-control) was associated with higher phase risk even when peer and parent control was controlled (Table 4).

If only those students who had shoplifting experience are analysed, females were more prone to experience shoplifting phases (Tables 1 and 4). Males were over-represented in the long phase group (Figure 2), an ominous finding in relation to Figure 1. This points to the conclusion that female shoplifting is somewhat more prone than its male counterpart to take a phasic form, a difference that could not be attributed to differential external control. As a theoretical and preliminary interpretation, I suggest that females exercise self-control in order to avoid a bad reputation (cf. Heidensohn 1985), something that tends to punctuate their shoplifting to form temporary phases. The mass level phenomenon shown in Figure 1 may be partly related to the female propensity (or capacity) to exercise self-control, for example, by applying phase interpretations.


From the point of view of guarding and surveillance, it would seem to be irrelevant whether the shoplifters are phasic or not. On a closer inspection, however, this may be relevant. Since phasic shoplifters potentially steal more than occasional non-phase shoplifters, prevention efforts are likely to be effective if phasic shoplifters are apprehended, and even more effective if non-specialized multiphasic offenders are caught. As many as one :in five FSRD-96 students with phase experience terminated his or her latest phase because of apprehension. Moreover, students with high phase theft intensity were somewhat more likely to be apprehended (and indeed terminate the phase for that reason) than pupils with low or intermediate theft intensity ([chi square] = 6.3, df = 2, p = .044). This and the overall prevalence of `getting caught' as a self-reported reason for desistance suggest that surveillance is more cost-effective when and if it can be targeted at phasic shoplifters.

Although general surveillance seemed to be rather effective at its present (Finnish) level, the number of spontaneous phase terminations was equally impressive. An actual apprehension was not the only or necessary way out of phasic shoplifting. To some extent, spontaneous termination may reflect the indirect effects of deterrence and the reflexive use of phase interpretations as definitions favourable to law-abiding behaviour.

Intensified periods of shoplifting behaviour in general should not, in my opinion, be interpreted as symptoms of psychological pathology, although some juveniles may enter them because of personal problems. Adolescence-limited phasic behaviour can be seen as `normal' for at least three reasons: because of its relatively high prevalence, because adolescents are expected to loosen their ties to social control, and because our culture is full of descriptions under which periodical `crises' and `phases' make sense. This last aspect means that adolescents themselves reflexively engage in, and terminate, phasic episodes.

(1) Finnish crime statistics are based on offences, not on persons, meaning (for example) that one frequent offender can artificially boost the `number of suspected persons'. This makes it impossible to differentiate between changes in incidence and prevalence.

(2) The 1997 teacher survey was administered in the 45 FSRD-96 schools, the 9th grade pupils which provide the data used in this article (see below). No report of the teacher survey is available in English at the present time.

(3) According to Hirschi (1969: 208), neutralizations both rationalize past behaviour and function as causes for later acts. Analogously, phase interpretations may both describe past desistance and make future desistance easier.

(4) Sixty-one per cent of the respondents were 15 years old, 38 per cent were 16 years old, and 1 per cent 17 years old.

(5) The 1996 reading differs from corresponding figures reported elsewhere in this article, because here only the 13 pilot schools are considered.

(6) Of the 85 pupils who concentrated in this product category, 81 were females (95 per cent).

(7) Fear of apprehension here includes a friend's apprehension as a self-reported reason for desistance.

(8) The prevalence estimates of `kleptomania' in psychiatrically oriented research are not much higher than this (Sarasalo et al. 1997: 2). One FSRD-96 respondent wrote that her shoplifting phase wax related to anorexia.

(9) Lifetime variety of participation was used because it was the best available measure of general propensity to commit delinquent acts. Incidence figures based on a one-year recall period could not be used as predictors of lifetime phase prevalence.

(10) In both of these lines of research, the possibility of phasic participation in other offence types (violence, drugs, truancy, etc.) could be noted.


BELSON, W. (1975),Juvenile Theft: The Causal Factors. London: Harper and Row.

BJARNASON, T. (1995), `Administration mode bias in a school survey on alcohol, tobacco and illicit drug use', Addiction, 90: 555-9.

BUCHANAN, C., ECCLES J. S. and BECKER J. (1992), `Are Adolescents the Victims of Raging Hormones? Evidence for Activational Effects of Hormones on Moods and Behaviour at Adolescence', Psychological Bulletin, 111 / 1: 62-107.

COX, D., COX, A. and MOSCHIS, G. (1990), `When Consumer Behaviour Goes Bad: An Investigation of Adolescent Shoplifting', Journal of Consumer Research, 17:149-59.

DILALLA, L. and GOTTESMAN, I. (1989), `Heterogeneity of causes of delinquency and criminality: Lifespan perspectives', Development and Psychopathology, 1/4: 339-50.

FARRINGTON, D. (1986), `Age and Crime', in M. Tonry and N. Morris, eds., Crime and Justice. An Annual Review of Research, vol. 7. Chicago and London: The University of Chicago Press.

-- (1992), `Explaining the Beginning, Progress, and Ending of Antisocial Behaviour from Birth to Adulthood', in J. McCord, ed., Advances in Criminological Theory Vol. 3: Facts, Frameworks, and Forecasts. New Brunswick: Transactions.

FARRINGTON, O., GALLAGHER, B., MORLEY, L., LEDGER, R., and WEST, D. (1986), `Unemployment, School Leaving, and Crime'. British Journal of Criminology, 26/4: 335-56.

GOTTFREDSON, M. and HIRSCHI, T. (1990), A General Theory of Crime. Stanford: Stanford University Press.

GRAHAM, J. and BOWLING, B. (1995), Young people and crime, Home Office Research Study 145. London: Home Office.

HAZANI, M. (1991), `Aligning Vocabulary, Symbols Banks, and Sociocultural Structure', Journal of Contemporary Ethnography, 20/2:179-202.

HEIDENSOHN, F. (1985), Women and Crime. London: Macmillan.

HINDELANG, M., HIRSCHI, T. and WEIS, J. (1981), Measuring Delinquency. Beverly Hills and London: Sage.

HIRSCHI, T. (1969), Causes of Delinquency. Berkeley and Los Angeles: University of California Press.

JUNGER-TAS, J., TERLOUW, G-J. and KLEIN, M., eds. (1994), Delinquent Behaviour among Young People in the Western World. Amsterdam and New York: RDC, Ministry of Justice/Kugler Publications.

KIVIVUORI, J. (1994), `On the Meaning of Delinquency', ill Scandinavian Research Council for Criminology: Report from 36: research seminar. Norway: NSFK.

KLEIN, M. (1984), `Offence Specialisation and Versatility' among Juveniles', British Journal of Criminology, 24/2, 185-94.

KLEMKE, L. (1982), `Exploring Juvenile Shoplifting', Sociology and Social Research, 67/1: 59-77.

KREUZER, A., GOERGEN, T., ROEMER-KLEES, R. and SCHNEIDER, H. (1992). `Auswirkungen unterschiedlicher methodischer Vorgehensweisen auf die Ergebnisse selbstberichteter Delinquenz', Monatsschrift fur Kriminologie und Strafrechtsreform, 75/2, 3: 91-104.

LOEBER. and LE BLANC, M. (1990), `Toward a Developmental Criminology', in M. Tonry and N. Morris, eds., Crime and Justice. A Review of Research, vol. 12. Chicago and London: The University of Chicago Press.

MATT, E. (1995), `Episode und "Doppel-leben": Zur Delinquenz Jugendlicher', Monatsschrift fur Kriminologie und Strafrechtsreform, 78/3: 153-64.

MATZA, D. (1964), Delinquency and Drift. USA: John Wiley and Sons.

MAWSON, A. (1987), Transient Criminality. A Model of Stress-Induced Crime. New York: Praeger.

MOFFITT, T. (1993), `Adolescence-Limited and Life-Course-Persistent Antisocial Behaviour: A Developmental Taxonomy', Psychological Review, 100/4: 674-701.

MULVEY, E. and LAROSA, J. (1986), `Delinquency Cessation and Adolescent Development: Preliminary Data', American Journal of Orthopsychiatry, 56/2:212-24.

RUSSELL, D. (1995), Women, Madness and Medicine. Cambridge: Polity Press.

SAMPSON, R. and LAUB, J. (1990), `Crime and Deviance over the Life Course: The Salience of Adult Social Bonds', American Sociological Review, 55/5: 609-27.

-- (1992), `Crime and Deviance in the Life Course', Annual Review of Sociology, 18: 63-84.

SARASALO, E., BERGMAN, B. and TOTH, J. (1997), `Kleptomania-like behaviour and psychosocial characteristics among shoplifters', Legal and Criminological Psychology, 2/1: 1-10.

SYKES, G. and MATZA, D. (1957), `Techniques of Neutralization: A Theory of Delinquency', American Sociological Review, 22/6: 664-73.

STEFFENSMEIER, D., ANDERSEN ALLAN, E., HARE, M. and STREIFEL, C. (1989), `Age and the Distribution of Crime', American Journal of Sociology, 94/4: 803-31.

WALKERDINE, V. (1984), `Developmental psychology and child-centred pedagogy: the insertion of Piaget into early education', in J. Henriques, W. Hollway, C. Urwin, C. Venn and V. Walkerdine, eds., Changing the Subject. Psychology, social regulation and subjectivity. London and New York: Methuen.


The 18 offences of the Finnish Self-Report Delinquency Study (FSRD-96).(a) Lifetime and last year prevalence figures (weighted) of ninth grade students(%). Respondent age distribution: see n. 4.
                                         Lifetime    Last year(b)

Shoplifting(c)                              57             19
Truancy                                     56             44
Bullying in school                          54             21
Stealing at school                          50             26
Stealing at home                            36             22
Fare evasion                                35             21
Graffiti writing/drawing                    34             19
Driving without licence                     33             25
Vandalism outside school                    31             13
Taking part in a fight                      25             13
Vandalism at school                         23             13
Drunken driving                             21             11
Buying stolen goods                         18              7
Beating up someone                          12              8
Misuse of legal medicine                    12              8
Running away from home                      11              8
Use of soft drugs                            8              7
Auto theft                                   2              1

Did not participate in any of the
above-cited offences                         9             26

                                         4,204          4,204

(a) The offence repertory was selected from the ISRD questions by analyzing the Finnish (Helsinki-based) ISRD sample of 1992.

(b) 12 preceding months. The survey was administered in all schools between 16 and 26 April 1996.

(c) Stealing from shops or kiosks. The kiosks were not included in the ISRD question.

JANNE KIVIVUORI, National Research Institute of Legal Policy, Helsinki, Finland. This article is based on research financed jointly by the National Research Institute of Legal Policy and the Finnish Board of Education. The author would like to thank Kauko Aromaa and Honkatukia for their helpful comments on this paper.
COPYRIGHT 1998 Oxford University Press
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:British Journal of Criminology
Geographic Code:4EUUK
Date:Sep 22, 1998

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters