The Cambridge Study in Delinquent Development (CSDD): Why Was the Cohort Set Up?

The CSDD is a prospective longitudinal survey of 411 London males who were first studied in 1961–1962 at age 8–9. Their parents, teachers, peers, female partners, and children have also been interviewed. In the interests of clarity, the original 411 males are now termed generation 2 (G2), their parents are termed generation 1 (G1), and their children are termed generation 3 (G3). However, this cohort profile focuses on the original cohort of G2 males, and so they will merely be referred to as the males.

The CSDD was begun by Professor Donald West (who is now, sadly, deceased) in 1961. Professor David Farrington joined Professor West to work on the CSDD in 1969, and took over the direction of the CSDD in 1982. The latest interviews were carried out in collaboration with Professor Jeremy Coid. Professor Darrick Jolliffe joined the CSDD as co-Director in 2020 and will take charge of future data collections.

On January 1, 1960, Donald West joined the Cambridge University Institute of Criminology as one of the three founding Assistant Directors of Research. The Institute had been founded in 1959, with Professor Leon Radzinowicz as its first Director. Donald West wanted to do research on homosexuality, but Professor Radzinowicz would not allow this. Consequently, Donald West began the Cambridge Study in Delinquent Development (CSDD) in 1961. The project was suggested by a Schools Medical Officer from Bedford, who had heard about the Gluecks’ research, and thought that a longitudinal study on delinquency should be initiated in England. Professor Radzinowicz was initially against it, as he wanted quick results from the new Institute, complaining that everyone would be grey-haired by the time that any conclusions were drawn! However, the Home Office agreed to fund the CSDD, so Professor Radzinowicz agreed that Donald West should launch it.

When he was planning the CSDD, Donald West wrote to the Gluecks and to the McCords to ask about how they measured family and other variables. However, the reply from the Gluecks was unsatisfactory because their measurement of variables seemed to be very subjective and impressionistic. In contrast, the McCords, who had just published their influential book (McCord et al., 1959) on Origins of Crime, gave Donald West very detailed and explicit instructions about the definition and measurement of family factors such as parental attitudes, including clear examples of the coding (e.g. what was “loving normal”, “loving anxious”, “overprotective”, “cruel”, “passive”, or “neglecting”). Not surprisingly, therefore, Donald West was greatly influenced by the McCords’ research and tried to use their measures of family variables as much as possible in the early days of the CSDD.

Since 1961, the CSDD has been mainly funded by the Home Office and the Department of Health, with contributions from the Department for Education, the Rayne Foundation, the Barrow Cadbury Trust, and the Smith-Richardson Foundation.

The original aim of the CSDD was to describe the development of delinquent and criminal behaviour in inner-city males, to investigate to what extent this could be predicted in advance, and to explain why juvenile delinquency began, why it did or did not continue into adult crime, and why adult crime often ended as men reached their twenties. The main focus was on continuity or discontinuity in behavioural development, on the effects of life events on development, and on predicting future behaviour. The CSDD was not designed to test any one particular theory about delinquency but aimed to test many different hypotheses about the causes and correlates of offending. However, a theory was later proposed to explain its findings (Farrington, 2005, 2020c). The results of the CSDD have been described in six books (Farrington et al., 2013; Piquero et al., 2007; West, 1969, 1982; West & Farrington, 1973, 1977), and in seven summary articles (Farrington & West, 1981, 1990; Farrington, 1995b, 2003, 2019a, 2021a; Farrington et al., 2009).

Who Is in the Cohort?

At the time they were first contacted in 1961–1962, the males were all living in a working-class area of South London (which is never identified). The vast majority of the sample was chosen by taking all the males who were then aged 8–9 and on the registers of six state primary schools within a 1-mile radius of a research office which had been established. In addition to 399 males from these six schools, 12 males from a local school for “educationally subnormal” (special needs) children were included in the sample, in an attempt to make it more representative of the population of males living in the area. Therefore, the males were not a probability sample drawn from a population, but rather a complete population of males of that age in that area at that time.

Most of the males (357 or 87%) were White in appearance and of British origin, in the sense that they were being brought up by parents who had themselves been brought up in England, Scotland, or Wales. Of the remaining 54 males, 12 were African-Caribbean, having at least one parent of West Indian (usually) or African origin. Of the remaining 42 males of non-British origin, 14 had at least one parent from the North or South of Ireland, 12 had parents from Cyprus, and the other 16 males were White and had at least one parent from another Western industrialized country.

On the basis of their fathers’ occupations when they were aged 8, 94% of the males could be described as working-class (categories III, IV, or V on the Registrar General’s scale, describing skilled, semi-skilled, or unskilled manual workers), in comparison with the national figure of 78% at that time. A majority of the males were living in conventional two-parent families with both a father and a mother figure; at age 8, only 6% of the males had no operative father and only 1% had no operative mother. This was, therefore, overwhelmingly a traditional White, urban, working-class sample of British origin.

How Often Have They Been Followed Up?

The males have been interviewed nine times, at ages 8, 10, 14, 16, 18, 21, 25, 32, and 48. At all ages except 21 and 25, the aim was to interview all the males who were still alive. Because of inadequate funding, only about half of the males were interviewed at age 21, and about a quarter at age 25. At age 21, the aim was to interview all the convicted males and an equal number of randomly chosen unconvicted males, and 218 of the 241 target males (90%) were interviewed. At age 25, only 85 males were interviewed.

In the other interviews, it was always possible to interview a high proportion of males who were still alive: 405 (99%) at age 14, 399 (97%) at age 16, 389 (95%) at age 18, 378 (94%) at age 32, and 365 (93%) at age 48. At age 48, 17 males had died (of whom 13 were convicted), five could not be traced, and 24 refused, which meant that 365 out of 394 who were alive were interviewed. The males also completed a medical interview at age 48. Out of 343 who completed the age 48 social interview face-to-face in person, 304 (89%) completed the medical interview.

It is important to have a high response rate in criminological studies because the most elusive and uncooperative people tend to be the most antisocial, and therefore the most interesting to criminologists. For example, West and Farrington (1973, p. 77) reported that parents who were rated as uncooperative (5%) or reluctant (5%) in the initial interviews when the male was age 8 were significantly more likely to have sons who were later convicted as juveniles (40% compared to 18%). Similarly, West and Farrington (1977, p. 165) showed that 36% of the males who were the most difficult to interview at age 18 were convicted, in comparison with 22% of the remainder.

What Has Been Measured?

Interviews with the parents were carried out by female psychiatric social workers who visited their homes. These took place about once a year (from 1961 to 1969) from when the male was aged 8–9 until when he was aged 14–15 and was in his last year of compulsory education. The primary informant was the mother, although many fathers were also seen. The parents provided details about such matters as family income, family size, their employment histories, their child-rearing practices (including attitudes, discipline, and parental conflict), their degree of supervision of the male, and his temporary or permanent separations from them. Huge efforts were made to ensure that the psychiatric social workers were as objective and consistent as possible in their ratings.

At ages 8, 10, and 14, the males were assessed in their schools, by CSDD psychologists. The tests in schools measured individual characteristics such as intelligence, attainment, personality, and psychomotor impulsivity. In addition, the boys’ teachers completed questionnaires when the boys were aged about 8, 10, 12, and 14. These furnished data about their troublesome and aggressive school behaviour, their lack of concentration or restlessness, their school attainments, and their truancy. Ratings were also obtained from the boys’ peers when they were in the primary schools at ages 8 and 10, about such topics as their daring, dishonesty, troublesomeness, and popularity.

The males were interviewed in our research office at about ages 16, 18, and 21, and in their homes at about ages 25, 32, and 48, by young social science graduates. The topics covered in the later social interviews were as follows:

  1. a.

    Date, time, and place of the interview, the interviewer, the length of the interview, the age of the male;

  2. b.

    Type of accommodation, number of rooms, whether the home was owned or rented, problems with the accommodation or area, reasons for moving out from or back to London, other persons living in the household, details about the wife or female partner, dates of marriages, places where the male has lived in the last 5 years;

  3. c.

    The male’s employment, average weekly take-home earnings, money from other sources, job of female partner, periods of unemployment in the last 5 years, illnesses, injuries and hospitalization in the last 5 years, money owed;

  4. d.

    Details about the male’s (and his female partner’s) children, including dates of birth, and questions referring to children at ages 3–15: age children are let out in the street on their own, extent to which the male knows where they are when they go out, last resort punishment if they are naughty, agreement between the male and his female partner on controlling children, child problems (lying, stealing, running away, truancy, disobedience, temper tantrums, bullying, destructiveness, restlessness, bed-wetting, sleep disturbance, fears, nervous habits), and contact of children with social agencies;

  5. e.

    Whether the male’s parents and siblings are still alive, dates of deaths, relationships with parents and siblings;

  6. f.

    Relationship with female partner, rows, violence between male and female partner, sexual relationship, intercourse with other females;

  7. g.

    Frequency of going out in the evening, companions and places, money spent on smoking and drinking, amount drunk, four questions designed to detect alcoholism (the CAGE questionnaire), driving after drinking at least 5 pints of beer (or equivalent) in the last 5 years, drug taking in the last 5 years, why drug taking has stopped, physical fights in the last 5 years;

  8. h.

    The same attitude questionnaire was completed by the male at ages 18, 32, and 48;

  9. i.

    Court appearances in the last 5 years, characteristics of offences, opinion of police behaviour and court decision, effects of sentence;

  10. j.

    The same self-reported offending questions were asked from age 18 onwards, plus questions about companions, age each act was first and last committed, estimated chances of being caught;

  11. k.

    The General Health Questionnaire (GHQ), given at ages 32 and 48, was designed to detect non-psychotic psychiatric illness (anxiety/depression);

  12. l.

    The Eysenck Personality Inventory and the Big Five Personality Inventory were completed.

The criminal records of the males (and their relatives) were repeatedly searched from 1964 to 2017, providing information about convictions of the males from age 10 to age 61.

What Has It Found? Key Findings and Publications

It is only possible, of course, to mention a few of the key findings in this cohort profile. About 300 publications from the CSDD are listed on the website of the Cambridge University Institute of Criminology (in David Farrington’s entry). The CSDD has advanced knowledge particularly about the development of offending and antisocial behaviour from childhood to adulthood, about the life success and health of offenders and non-offenders, about childhood risk factors for later offending and antisocial behaviour, and about the effects of life events on the course of development of offending. Later publications have compared risk factors in the two generations of males (Farrington et al., 2015) and have investigated the intergenerational transmission of convictions (Farrington et al., 2017), psychopathy (Auty et al., 2015), and self-reported offending (Farrington et al., 2018), but results obtained with the children of the males (who were interviewed at the average age of 25) will not be reviewed here.

Beginning with conviction careers up to age 61, 44% of the males were convicted (Farrington, 2019b). The probability of a first conviction increased to a peak at ages 14–16 (15%) and then decreased. The most common types of crimes were burglaries, thefts of vehicles, other thefts, and assaults; minor motoring offences were not counted. On average, conviction careers began at age 19.5, ended at age 32.1, lasted 12.6 years, and included 5.3 convictions. The males who started at the earliest ages tended to have the longest criminal careers. The probability of being reconvicted decreased with increasing time since the previous conviction, but was still substantial (19%) even after 15 crime-free years.

During the interviews, the males were asked to self-report offences that they had committed that had not necessarily come to the notice of the police. The most important interviews were at ages 14, 18, 32, and 48. During these four age ranges, almost all of the males (93%) said that they had committed at least one of eight types of offences (burglary, theft of motor vehicles, theft from motor vehicles, shoplifting, theft from machines, assault, drug use, and vandalism), which account for the majority of all conviction offences. However, only 29% had been convicted for at least one of these offences during the same age ranges.

Based on the commission of eight offences, the average age of onset was much earlier according to self-reports (10.3, compared with 19.1 according to convictions up to age 56). Similarly, the average age of desistance was much later in self-reports (35.2, compared with 25.1 according to convictions for these eight offences). The probability of a self-reported offence leading to a conviction was highest for burglary and theft of vehicles (both 28%) and lowest for fraud and theft from work (both 1%; see Farrington et al., 2014).

There have been many studies of childhood risk factors in the CSDD. One of the most recent publications (Farrington, 2020a) showed that the strongest predictors of life-course-persistent offenders up to age 61 (those with a criminal career lasting at least 20 years) were harsh parental discipline, poor parental supervision, a convicted father, and parental conflict. Interestingly, the strongest childhood predictors of criminal career duration were uncorrelated with the strongest childhood predictors of offending (convicted versus unconvicted males). Other analyses have investigated independent predictors. For example, Farrington and Malvaso (2019) found that the strongest independent childhood predictors of violence convictions were harsh parental discipline, a convicted father, and low family income (and incidentally that the independent childhood predictors of violence convictions by the next generation of males were very similar). Still other analyses have studied interactions between risk factors. For example, Farrington (2020b) found that a high resting heart rate acted as a protective factor in nullifying the effect of several childhood risk factors on later convictions and antisocial personality scores.

Farrington et al. (2006) classified the convicted males into desisters who were convicted only before age 21, late-comers to crime who were convicted only at age 21 or older, and persisters who were convicted both before and after age 21. A key question was the extent to which these categories of offenders were successful in different aspects of their lives at age 48. A measure of life success was developed based on nine criteria, including employment and accommodation histories, cohabitation, alcohol and drug use, convictions and self-reported offending, anxiety/depression, and violence. The good news is that a majority of all groups were considered to be leading successful lives at age 48: 95% of unconvicted males, 96% of desisters, 84% of late-onset offenders, and 65% of persisters. The most important finding was that desisters were not significantly different from unconvicted males in eight out of nine areas or in their total success score. Therefore, it seems that those who give up offending eventually become similar to non-offenders in their life success.

Late-onset offenders were significantly worse than unconvicted males in their alcohol and drug use. Interestingly, among the best predictors of late-onset offenders were nervousness and having few friends at age 8–10, teacher-rated anxiety at age 12–14, high neuroticism at age 16, and not having had sexual intercourse by age 18 (Zara & Farrington, 2009). It seems that the protective effects of social inhibition in adolescence may wear off after age 21 when the males leave home. Other research on late-onset offenders was carried out by McGee and Farrington (2010), showing that about one-third of them were self-reported delinquents before age 21, so were fortunate not to have been convicted earlier.

There have been several other studies of adult life success. For example, Ullrich et al. (2007) related personality disorder symptoms to different types of life success, and found that narcissistic and obsessive–compulsive scores were positively related to success in status and wealth (e.g. a high income), whereas high schizoid and borderline scores were negatively related to success in intimate relationships. Ullrich et al. (2008) also found that high psychopathy scores were negatively related to all types of life success, and Theobald et al. (2016b) reported that males who were violent both inside and outside the home were particularly unsuccessful in their lives. Other research on the males who committed intimate partner violence focussed on their childhood and adolescent risk factors (Theobald & Farrington, 2012), their personality disorder symptoms (Theobald et al., 2016a), and their offending trajectories (Piquero et al., 2014b).

Several studies have investigated the relation between offending and physical health. Farrington (1995a) found that males who were convicted up to age 40 tended to have had more accidents and serious illnesses at age 14–16, more injuries and hospitalizations at age 18, and more road accidents at age 32 than other males. However, Shepherd et al. (2002, 2004) discovered that these convicted males were less likely than others to suffer respiratory tract diseases (e.g. colds and flu) at ages 18 and 32, possibly because of the protective effects of their alcohol consumption and their tendency to have physically active manual jobs. Piquero et al. (2011, 2014a) studied the health of different trajectory groups of offenders, and found that high-rate chronics were the most likely to be registered as disabled at age 48 and dead by age 57; 22 out of 31 males who died up to that age were convicted. A later search of death records in 2020 revealed that 54 males had died up to age 65, of whom 30 were convicted. Reising et al. (2019) reported that anxiety and depression at ages 32 and 48 were most prevalent among late-onset offenders (up to age 56), and Skinner et al. (2020) showed that organic illnesses and hospitalizations at age 48 were most prevalent among late-onset and life-course-persistent offenders (up to age 61).

Attempts were made to go beyond the identification of risk factors to test alternative theories about mechanisms and processes relating risk factors to offending. For example, different explanations of the relationship between permanently disrupted families (broken homes) and delinquency were tested, and it was found that, while males from broken homes were more delinquent than males from intact homes, they were not more delinquent than males from intact high-conflict families. Therefore, it seemed that the loss of a parent was not the key factor. Overall, the most important factor was what happened after the family break. Males who remained with their mother after the separation had the same delinquency rate as males from intact low-conflict families. However, males who remained with their father, with relatives or with others (e.g. foster parents) had high delinquency rates. Therefore, it was not true that broken homes always had undesirable consequences (Juby & Farrington, 2001). Later analyses on this topic were carried out by Theobald et al. (2013).

In a longitudinal survey, it is possible to investigate the effects of specific life events on the development of delinquency, by comparing before and after measures of offending in within-individual analyses. For example, the effects on delinquent behaviour of being found guilty in court were studied. If convictions have a deterrent or reformative effect, a male’s delinquent behaviour should decline after he is convicted. On the other hand, if convictions have stigmatizing or contaminating effects, a male’s delinquent behaviour should increase after he is convicted. These hypotheses were tested by studying self-reports of delinquency before and after a male was first convicted. The results suggested that first convictions had undesirable delinquency-amplifying effects on the males (Farrington, 1977; Farrington et al., 1978). Later analyses on this topic were carried out by Murray et al. (2014).

Another investigation of the effect of a specific event on offending focussed on unemployment. The key question was whether the males committed more offences (according to official records) during their periods of unemployment than during their periods of employment. The results showed that the males did indeed commit more crimes while unemployed than while employed. Furthermore, the difference was restricted to offences involving financial gain, such as theft, burglary, robbery, and fraud. There was no effect of unemployment on other offences, such as violence, vandalism, and drug use. These results suggest that the males committed more offences while they were unemployed primarily because they lacked money at these times and were trying to obtain money (Farrington et al., 1986).

It is often believed that a good marriage is one of the most effective treatments for male offending. When the males in their twenties were asked why they had stopped offending, they often mentioned marriage and the influence of females, as well as the fact that they did not hang around so much with delinquent friends. Before-and-after analyses showed that getting married led to a decrease in conviction rates compared with remaining single, whereas separation from a wife led to an increase in conviction rates compared with staying married (Farrington & West, 1995). Later analyses on this topic were carried out by Theobald and Farrington (2009, 2013).

The CSDD generally shows that a combination of childhood adversities, including low family income, poor child-rearing, antisocial parents, delinquent peers, and a bad neighbourhood, tends to produce antisocial and impulsive children who tend to fail in school. Later on in life, these types of children tend to have a combination of adult life problems, including offending and employment, accommodation, substance use, and relationship problems, and they tend to produce another generation of antisocial children. The main policy implication is that early intervention programmes are needed for families and children at risk. There are many effective programmes, including home visiting, parent training, child skills training, preschool programmes, and anti-bullying programmes (Farrington, 2021b). The fact that offenders tend to have multiple life problems means that any effective programme would have a variety of benefits in different areas of life; and therefore, its financial benefits would almost certainly exceed its financial costs (Farrington & Welsh, 2014).

Numerous tests of validity were carried out in the CSDD, in most cases based on comparisons between interview data and external information from records. For example, self-reports of convictions were compared with criminal records of convictions, and the mother’s report of the male’s birth weight was compared with hospital records. It was shown that self-reported delinquency had predictive validity: among unconvicted males, those who reported a particular type of offence had an increased probability of being convicted for it later (Farrington, 1973, 1989). As another example, more than twice as many males who said that they had sexual intercourse without using contraceptives at age 18 subsequently conceived a child outside marriage compared with the remainder (Farrington & West, 1995).

Reliability checks were also made. For example, information about the same topic (e.g. school leaving age) from different interviews was compared, as was information about the same topic from different parts of the same interview. Generally, the males were randomly allocated between our two or three interviewers in each data collection wave in order to investigate interviewer effects, but fortunately these were rarely found (West & Farrington, 1977, pp. 172–175). All these validity and reliability checks suggested that, in the vast majority of cases, the males were genuinely trying to tell the truth.

What Are the Main Strengths and Weaknesses?

The CSDD has a unique combination of features:

  1. a.

    Nine personal interviews with the males have been completed over a period of 40 years, from age 8 to age 48;

  2. b.

    The main focus of interest is on offending, which has been studied from age 10 to age 61;

  3. c.

    The sample size of about 400 is large enough for many statistical analyses but small enough to permit detailed case histories of the males and their families (Zara & Farrington, 2016);

  4. d.

    There has been a very low attrition rate, so that the information is very complete;

  5. e.

    Information has been obtained from multiple sources: the males, their parents, teachers, peers, female partners, and various types of official records;

  6. f.

    Information has been obtained about a wide variety of theoretical constructs, including intelligence, impulsiveness, popularity, socio-economic status, parental child-rearing methods, peer delinquency, school behaviour, employment success, marital stability, substance use, and so on.

The CSDD is unparalleled in its combination of a 50-year follow-up period, a relatively large number of face-to-face interviews, and a very low attrition rate.

Turning to limitations, the CSDD provides information about the development of offending and antisocial behaviour in an inner-city, working class British White male sample born about 1953. To what extent similar results would be obtained with females, Black or Asian children, suburban or rural children, middle or upper class children, children born more recently, or children brought up in other countries is an interesting empirical question. However, results obtained in the CSDD are similar to those obtained with comparable male samples from Sweden (Farrington & Wikström, 1994), the USA (Farrington & Loeber, 1999), and other Western industrialized countries (Farrington, 2015). Also, risk factors for the males and their sons were very similar (Farrington et al., 2015), and risk factors for the brothers and sisters of the males were quite similar (Farrington & Painter, 2004). While parameters such as the prevalence and frequency of offending may be more specific to this sample, it seems likely that relationships (e.g. between risk factors and offending) are more generalizable.

The CSDD has the usual methodological problems of prospective longitudinal surveys. While the problem of attrition was largely overcome, testing effects (the effects on the males of repeated interviews) are not clear. However, the percentage of brothers (who were never contacted) who were convicted was very similar to the percentage of males who were convicted, suggesting that the repeated interviews had little effect on convictions at least. The single cohort design made it difficult to distinguish between ageing and period effects; for example, between ages 14 and 18, the percentage of males who had taken drugs increased from less than 1% to 31%, but this was probably influenced by broader social changes during the time period (from 1967 to 1971).

The sample size was too small to study rare events adequately, such as low birth weight, or sex offenders; for example, only 10 males were convicted of a sex offence up to age 50 (Piquero et al., 2012). Because of intermittent funding, the interviews were too infrequent to establish the exact or relative timing of many life events, and hence to establish developmental sequences between presumed causes and observed effects. Inevitably, some of the early measures, based on interviews by psychiatric social workers, now appear rather old-fashioned. Also, asking the males to recall over a 5-year period was not ideal, but was necessary because of the infrequency of the interviews.

Can I Get Hold of Data? Where Can I Find Out More?

The data collected from age 8 to age 25 has been deposited in the UK Data Archive (University of Essex) and the National Archive of Criminal Justice Data (University of Michigan). The variables measured between these ages are described in detail in these datasets. Other data may be supplied for collaborative research; requests should be addressed to Professor David Farrington (dpf1@cam.ac.uk).

Profile in a Nutshell

  • The CSDD was set up to investigate the development of delinquency and antisocial behaviour from childhood to adulthood.

  • A total of 411 London males (mostly born in 1953) have been followed up from age 8 to age 61, in face-to-face interviews and records.

  • The CSDD is unique in its long follow-up period, its relatively large number of face-to-face interviews from age 8 to age 48, and its very low attrition rate.

  • A great variety of information has been collected, in childhood, adolescence, and adulthood, from the males, their parents, peers, teachers, and various types of official records.

  • Numerous collaborative research projects have been carried out, and future collaborations are possible.