One facet of family structure which has been the subject of many studies is the broken home. Some researchers have found a very high incidence of broken homes among delinquents and have attributed much significance to broken homes as a cause of delinquency. Others have given less direct emphasis to the importance of broken homes and have suggested that the broken home may have a differential effect by variables such as sex, area, or family cohesiveness. The controversy over the effect of broken homes on delinquency began with a paper published by Shaw and McKay (1932)? In which they concluded from a study of Chicago school boys and juvenile court cases that only slightly more broken homes appeared in the delinquent group than in the control group (42 percent:36 percent) and that the correlation between high delinquency rate areas and high broken-home rate areas was small. The data they presented contradicted several earlier studies, especially one by Burt (l925) in London which had found delinquents coming from broken homes twice as often as nondelinquents.
Shaw and McKay’s study was criticized as unrepresentative, since it made no attempt to discover delinquents in the control group, and refutations of it soon appeared. A study by Weeks and Smith (l939) in Spokane, Washington, found that 41.4 percent of the delinquents and only 26.7 percent of the controls came from broken homes; their correlation between delinquency and broken homes by area was considerably higher than Shaw and McKay Eleanor and Sheldon Glueck (l950), in their monumental study of 500 matched pairs of delinquents and nondelinquents, found 60.4 percent of the delinquents and 34.2 percent of the nondelinquents with broken homes in their back grounds.
More recently, Monahan (1957)6 reported that delinquents coming from broken homes were more likely to be recidivists than delinquents from unbroken homes. Browning (1960)’ found significantly greater numbers of Los Angeles delinquents coming from ‘disorganized” homes. Slocum and Stone (l963), using the Nye-Short self-report delinquency technique, found a significant correlation between broken homes and delinquent-type behavior. Peterson and Becker (1965) have referred to other studies that have also found a relationship between broken homes and delinquency.
Many researchers, however, have indicated that broken homes have a differential effect upon children — that the delinquency-producing effect is higher for preadolescents than adolescents and for property offenders than authority offenders. Early studies by Barker (1940) and Weeks and Smith (1939)” found significant variations in the correlation between delinquency rates and broken home rates among different areas of a community, and there is similar evidence on rural-urban differences in a study by Ferdinand. But the most common observation about the differential effect of the broken home has been that delinquent girls come from broken homes more often than delinquent boys. An early study by Hodgkiss (l933),’ which repeated Shaw and McKay’s Chicago study on delinquency and broken homes using girls instead of boys, found that 66.8 percent of delinquent girls came from broken homes compared to 44.8 percent of the school girls. Hodgkiss’ ratio of 1.49 to I, delinquent girls to school girls, is considerably higher than the ratio computed by Shaw and McKay ( for boys,l.18 to I. Wattenberg and Saunders (1954)15 studying Detroit juvenile delinquents, also found a greater percentage of delinquent girls as compared to delinquent boys coming from broken homes. Monahan (1957)10 reported much the same results for Philadelphia delinquents — 55.4 percent of white girls and 74.3 percent of Negro girls coming from broken homes, compared to 32.2 and 57.9 percent of white and Negro boys, respectively.
Jackson Toby (1957b)’ has reviewed some of the literature on the differential impact of broken homes, and his paper adds a good deal of clarity to the findings. For example, he introduces age as a differential variable in order to explain the apparent discrepancies between the data of Shaw and McKay and others. He points out that while Shaw and McKay found little overall difference between their delinquent group and control group in the percentage of broken homes, some differences do show up when controlling for age. The delinquents were considerably older than the control group, and it turns out that at the older age groups there is little difference between delinquents and controls in the rate of broken homes, while in the younger age groups there is a good deal of difference. Toby reasons that well-integrated American families generally have less control over their older, adolescent Sons. As a result, family disorganization (broken homes) would have its greatest impact upon younger, preadolescent sons where the well-integrated family could generally exert greater control. Toby’s data lend support to the hypothesis of differential impact with age, and a similar differential effect was observed by Lees and Newson (1954) in their study of British delinquents.
Toby also applies the same reasoning to account for the differential impact of broken homes on boys and girls — in general the family exercises more control over girls, hence they are more affected by a broken home. Toby presents data showing that, as predicted, the impact of broken homes is greater for girls than for adolescent boys, and further, that urban areas and Negroes, assumed to be more characterized by family disorganization, have a disproportionate number of female (and preadolescent) delinquents. But Toby does not have data on types of offenses, and it appears that the relationship between sex and rate of broken homes is eliminated when one controls for type of offense. Weeks (l940), for example, suggested that the differential effect of broken homes on boys and girls was due to their differential distribution according to type of offense — most girls are arrested for ungovernability, running away, and sex offenses, while most boys are arrested for vandalism, theft, and assault. Testing his theory on Spokane, Wash., delinquents, Weeks found that when type of offense is held constant, delinquent boys and girls come from broken homes in nearly the same proportions. Other studies, by Nye (1958) using his self-report delinquency scale, and Ferdinand (1964), using Michigan court data, have also shown that broken home rates vary according to the type of delinquency, being higher for authority” offenses such as ungovernability and truancy.
Ordinal Position, Family Size, and Delinquency
Another aspect of family structure which has often been related to delinquency is the ordinal position of the child in the family. Lees and Newson (1954)22 made an extensive study of the differences among delinquents which could be attributed to sibling position. Their study showed that intermediates children having both older and younger siblings — were scientifically overrepresented in a group of Nottingham, England, delinquents. The explanation they gave for their finding is that the attention parents often give to oldest and youngest children “squeezes” the intermediates out of the family into the gang. Their study has received some support from the findings of both Nye (I958) and the Gluecks (1950) Nye found both youngest and intermediate children overrepresented in his “most delinquent” group. The Gluecks found that 60 percent of their delinquents were intermediate children, compared to only 47.8 percent of their control group. Thus, there would seem to be greater involvement of middle children in delinquent activities.
Family size has also been cited as a factor which can differentiate between delinquents and nondelinquents, with delinquents seen as more likely to come From larger families, The Gluecks’ study gave definite evidence of this, finding that there was a significant difference between a mean of 6.85 children in the delinquents’ families and 5.90 in the nondelinquents’ families. Nye’s ‘most delinquent” boys were also more often from large families, but he found no such relationship for girls.
This differential in the relationship between family size and delinquency is not restricted to sex differences. Several writers have suggested that large families may have a varying relationship to different types of delinquent offenses. Barker and Adams (1963), for example, report a greater incidence of large families in the background of a group of juvenile glue sniffers than in control group taken from the general population of a Colorado correctional institution. Over 50 percent of the glue sniffers came from families with more than eight children, compared to (9 percent in the control group. Furthermore a study by Reiss (1952)20 of delinquents examined by psychiatrists attached to the Cook County Juvenile Court shows that the psychiatric classifications — integrated, weak ego, weak superego — also differentiated the delinquent according to family size. A greater proportion of weak superego delinquents came from large families.
Maternal Deprivation and Delinquency
One aspect of family structure which has received a great deal of attention especially from psychiatrists, is that of mother separation — the absence of the mother for a prolonged period of time when the child is young. Although mother separation is generally regarded as a cause of mental illness, one British psychiatrist, John Bowlby, has also attempted to demonstrate its connection with juvenile delinquency, Studying a group of thieves and controls chosen from patients at a child guidance center, Bow (1952)27 found significantly more mother separation among all the thieves, and the statistical significance is even more pronounced among those thieves he classified as affectionless” — children who were unable to enter into a deep affectionate relationship with anyone. Other studies, of course, have also shown the association between maternal deprivation and an inability to enter an affectionate relationship (Goldfarb, 1943)28
Bowlby, however, has been criticized by several investigators for putting too much emphasis on a factor which is often outweighed by others in the process of delinquency causation. Naess (1959),’ for instance, in her detailed commentary on Bow efforts, accepts the relationship between mother separation and mental illness but criticizes Bowlby for generalizing from his i data on thieves in a psychiatric clinic to his thesis that mother separation is an important cause of juvenile delinquency generally. Furthermore, she presents evidence from a study of Oslo delinquents and their nondelinquent brothers that no such relationship exists. Robert O. Andry (1957, 1962), the most insistent critic of Bowlby’s theory, is unhappy with the emphasis Bowlby gives to the part that mother separation plays in the etiology of delinquency, explaining that the role of the father is of at least equal importance. Michael Hakeem (1958) makes a devastating e’ of Bow by’s study, criticizing it for a lack of scientific sophistication, for generalizing too much from limited data, and for not attempting to eliminate possible diagnostic biases by proper scientific methods.
In addition, Bowlby also theorized that the younger a child was when he was separated from his mother, the more unfortunate were the effects of that separation; but both Nye (1958) and the Gluecks (1950) found that the child’s age at the time of the first break in his home bore no relationship to delinquency, Bowlby’s theory of maternal deprivation and delinquency therefore lacks strongly supportive evidence, Maternal deprivation may very well be causally related to delinquency but it is not “foremost among the causes of delinquent character development.” (1952, p3 Bowlby himself (Row et al., 1956, p. 242) now speaks more cautiously about the consequences of maternal deprivation.
Family Interrelationships, Family Adjustment, and Delinquency
Studies which have focused upon the internal structure of the family have generally shown greater associations with delinquency than studies focusing upon outward structure. This has led Nye to comment, on the basis of his extensive study of juvenile delinquency, that ‘‘the structure of the family ‘itself’ does not cause delinquency.” (1958, p. 34) Similarly, Browning has concluded, after his study of legally and psychologically broken homes in Los Angeles, that the “broken home, as generally defined, is ineffective and probably meaningless as an indicator of family disorganization and other characteristics of family life known to be associated with deviant behavior. It does not include all homes which are sociologically and psychologically broken nor exclude homes which are well integrated.” (1960. p. 43)3 MeCord and McCord ( have indeed shown that quarrelsome and negligent homes lead to more delinquency than broken homes. Nevertheless, we must be careful not to discount the possible influence of a factor like broken homes. As Hirschi and Selvin (l966)7 have ably documented, it may be true that a broken home is not a sufficient cause of delinquency; it may also be true that other variables are more strongly related to delinquency than broken homes; but such reasons do not permit us to conclude that broken homes are not causally related to delinquency. As a result, we must use statements such as those by Nye and Browning to alert us to the fact that variables measuring the quality of family relationships may be more important in the etiology of delinquency than variables measuring the outward structure of the family, and not to conclude that the latter variables are of no causal importance.
One of the most important aspects of family relations, closely related to the concept of broken or unbroken homes, is the quality of parental marital adjustment. Browning (1960) dealt with marital adjustment and family solidarity and found that both of these bore a significant relationship to truancy and auto theft, the two types of delinquency he studied. Nye (1938) reported a very strong association between his self-report delinquency scale and the marital happiness of the child’s parents. Among boys and girls from completely happy” homes — 23 and 22 percent, respectively, were “most delinquent’; from “unhappy homes” — 46 and 49 percent. The Gluecks (l9S0). in their research, found that significantly more delinquents than nondelinquents had parents with poor conjugal relations (31.2 percent: 14.9 percent). They also found that marked family cohesiveness was present in 61.8 percent of the nondelinquents’ homes but in only 16 percent of the delinquents’. Slocum and Stone (l963), in a self-report study of Washington State schoolchildren, report that 52 percent of the ‘most delinquent” boys called their families uncooperative, compared to 16 percent of the boys in the ‘conformist” category. The same relationship, although less pronounced, existed for girls, Dentler and Monroe (l96l) report that adolescent theft is related to the quality of interpersonal relations” in the family. Furthermore, Jaffe (1963)” found that family “anomie,” measured by the amount of disagreement within the family on selected value questions, was correlated with a high score on a “delinquency proneness” scale. In all of these studies, therefore, we find that reported variables such as marital adjustment family agreement, and family solidarity are significantly related to measures of juvenile delinquency.
Parental Discipline and Delinquent Behavior
Probably even more important a factor than parental marital relations in delinquency causation is the quality of parent-child relations. The consistency, “fairness,” and strictness of parental discipline are among the most important family variables related to delinquent behavior. As Peterson and Becker Q965) say, “If one endorses the common assumption that capacities for internal control are complexly but closely related to previously imposed external restraints, then parental discipline assumes focal significance as a factor in delinquency.”
In the Slocum and Stone study (I963) reported fairness of discipline was significantly associated with conforming behavior for boys and for girls. Nye (1958) found that of the children in his study who considered their father’s discipline “always fair,” only 30 percent of the boys and 20 percent of the girls fell into the “most delinquent category while of those who felt their father’s discipline was unfair, 55 percent of the boys and 44 percent of the girls fell into the “most delinquent’ category. Nye also showed that a relationship existed for girls with regard to the strictness of the mother’s discipline, with reported Strictness being related to less delinquency.
The Gluecks (l950) found that lax and erratic disciplinary techniques identified a higher percentage of delinquents than did overstrict techniques, and that firm but kindly techniques were practiced much more frequently by the parents of the nondelinquents. McCord and McCord (1959),40 reviewing the information collected by the Cambridge-Somerville project, concluded that consistent discipline by both parents, whether punitive or love oriented, significant reduced delinquency. Nye (1958) found that 49 percent of both the boys and girls who reported that their mothers “very often” failed to follow through on threatened punishment were in the “most delinquent” category, compared to 30 percent of the boys and 22 percent of the girls who reported that their mother “never” failed to follow through. The overall relationship is significant for girls but not significant for boys.
Exploring the tendencies for parents to use physical punishment rather than reasoning as a disciplinary method, the Gluecks ( found that significantly more parents of delinquents than nondelinquents (mothers, 55.6:34.6 percent; fathers, 67.8:34.7 percent) resorted to physical punishment for discipline, while significantly more parents in the control group than in the delinquent group used reasoning (mothers, 2S.2:16.4 percent; fathers, 244:11.3 percent). Nye (1958), on the other hand, found no relationship between physical punishment and delinquent behavior. He did, however, find a positive relationship between delinquency and love withdrawal as a disciplinary technique although this relationship seemed to disappear when the adolescents feelings of acceptance- rejection were take into consideration. For example Nye showed that love withdrawal by mothers made no difference among adolescents when one controlled for their feelings of acceptance or rejection. Since many studies have a relationship between types of discipline and measures of delinquency, Nye’s findings are worth noting. It must be remembered, however, that Nye’s study also failed to show the usual relationship between delinquency and low socioeconomic status. Furthermore, Kohn (1963) has questioned the association of class with types of punishment.
In short, the data suggest that the consistency of discipline and its fairness are importantly related to nondelinquency. It must be noted, however, that most of these studies (the report by McCord and Mc 1959, is a significant exception) are based upon the perception and recollection of delinquent and controls It is therefore possible that delinquents and controls perceive their parental discipline differently despite similarities in disciplinary techniques; it is also possible that the actual differences in disciplinary techniques are due to the parent’s differential responses to delinquent and nondelinquent behavior, such behavior having been triggered by other variables.
A.H. Maslow and R. Diaz-Guerrero (l960), have pointed to another way in which parental roles and the cultural support for discipline may play a part in the etiology of delinquency. Comparing Mexico and the United States, they suggest that delinquency is much more common in the United States (definitive evidence is lacking, however) despite the greater incidence of poverty and family disorganization in Mexico. They attribute this difference to the fact that parental roles within the family, especially the father’s role, are clear cut in Mexico but muddled in the United States. Mexican children are brought up understanding quite clearly the differences between “right” and “wrong,” and the father’s authoritative and disciplinary role, which the mother supports, fosters the development of strong internal controls in the child which reduce his delinquency. The American father, on the other hand, is faced with a larger number of available roles — friendly, authoritarian, democratic, lenient — and he receives less support from his wife in carrying out his role. As a result, there is more role confusion by the father in the United States, and this creates uncertainty in his children about approved and disapproved behavior. This uncertainty about values, Maslow and Diaz-Guerrero assert, plays a major part in the delinquency of United States adolescents.
Affection, Rejection, and Delinquency
A second aspect of parent-child relations which is often associated with juvenile delinquency is a lack of parental affection. The Gluecks (1950) reported that all the affectional patterns of a home — mother-child, father- child, child-parent, and child-child — bore a highly significant relationship to juvenile delinquency. The most important factor, however, seemed to be the father’s affection for the boy — 40.2 percent of the delinquents but 80.7 percent of the controls had affectionate fathers. Andry (l957) reported a similar finding in his study of London delinquents and schoolboys, with 54 percent of his delinquents but only 7 percent of his controls expressing the opinion that their fathers ought to love them more. Slocum and Stone (1963) found that 52 percent of the boys in their “delinquent” group came from unaffectionate families, compared to only 18 percent in the conformist” group. Nye (1958), choosing acceptance-rejection rather than affection as a variable, found a significant positive correlation between the parents’ rejection of their child and the child’s delinquency. When mutual rejection between mother and child occurred, 48 percent of the group fell into the “most delinquent” category; with mutual acceptance, 14 percent. McCord and McCord (I964) found that the presence of at least one loving parent, coupled with consistent parental discipline, was enough to mitigate the delinquency- producing effect of a criminal father. More generally, they report that both paternal rejection and the absence of maternal warmth were significantly related to delinquency.
The interplay of affection and discipline in parent-child relationships has been cited by Weinberg ( as most important in affecting the personality of the child and predisposing him to select delinquent associates and participate in delinquency. He quotes the Gluecks’ study for evidence and then says that the parents of delinquents by “indifference or hostility hindered their children from acquiring positive attitudes towards authority.” (1958, pp. 126- 127.) Clark and Wenninger’s self-report study (1964)52 of Illinois school children’s attitudes toward the law provides indirect evidence for such a supposition. They found that favorable attitudes toward legal institutions were most closely related to adjustment to maternal and paternal discipline than to social class factors.
Furthermore, Nye (1958) found a strong relationship between the child’s acceptance of parental values and the affection of the parent for the child. Finally, the longitudinal data reported by McCord and McCord (1964) show an interesting interaction effect between discipline and affection. With two loving parents, the type of discipline used has no effect upon the delinquency of Sons; but with one loving parent, erratic or lax discipline produces significantly more delinquency than consistent discipline.
One way in which child-rearing practices are seen to have an influence upon delinquency is through fostering the development of an aggressive personality. Studying a group of aggressive delinquents and their families, and comparing them to a group of nondelinquent controls and their families, Albert Bandura and Richard H. Walters ( 1959 found that a higher proportion of the parents of the aggressive boys have denied the boys an opportunity to express dependency feelings. Parental punishment of the boys’ strivings for dependency gratification, a form of parental rejection, was significantly higher for the group of delinquent boys. Bandura and Walters went on to show that adolescents whose dependency needs had not been met were also less likely to internalize their parents’ standards and values. Becker (I 954)54 reviewing the various studies dealing with the personality consequences of different types of child-rearing practices, put Bandura and Walters’ finding in a wider context and showed them to be consistent with several other studies.
The importance of affection within the family this emerges as an important factor in relation to delinquency The findings of a number of studies point to the significant differences between delinquent and nondelinquent groups in terms of the patterns of affection within the family. In addition, several studies point out that affection is related to the child’s internalization of parental values. Assuming that parental values support conventional rather than deviant behavior, it appears that an affectionate parent-child relationship promotes the internalization of conventional values and thus insulates a child against delinquent behavior, Of course, affectionate family patterns may also o in other ways to reduce delinquent behavior.
Some Psychiatric Theories of Delinquency
The problem of relating the wealth of available data on the association between family variables and juvenile delinquency has been approached theoretically in many ways. Psychiatrists, psychologists, and sociologists have all tried their hand at explaining the connection, with differing degrees of success. We will start our discussion by reviewing several psychiatric theories of delinquency.
One theory, the “superego lacunae’ theory put forward by Adelaide M. Johnson (l949), considers only the delinquencies arising in ‘apparently normal families of good reputation’ unassociated with the influence of sociologic gangs.” She suggest that inconsistent discipline in many middle class families deprives children of an important part of their security (Johnson and Burke, 1955)50 Furthermore, this inconsistency is often the manner in which parents unconsciously foster delinquency in their children. These children grow up with gaps or lacunae in their superego, and play a scapegoat role in their families — the parents project their own problems onto their child and derive vicarious pleasure from the child’s delinquency.
Although the theory may have some validity when it is limited to middle class nongang delinquents, other psychiatrists have extended it almost beyond reason. Ruth S. Eissler (l949), for instance, claimed that not only do the parents of delinquents need and foster delinquency in their children, but also that the whole society needs and fosters such delinquency, as a societal scapegoat. Matza and Sykes (l961) have suggested that many of the values held by delinquents are actually subterranean values held by certain segments of society; but Eissler’s extreme position claims that society needs criminal or delinquent scapegoats, that it seduces individuals into delinquent behavior, and that it interferes with programs which promise to prevent delinquency.
Hakeem (l958 has pointed to several weaknesses of the superego lacuna’ theory. He points out that no experimental or predictive studies have tested it and that no scientific evidence has been put forth to support it. It is merely based upon case studies that have been presented by the theorists. Moreover, these case histories usually deal with emotionally disturbed children whose delinquency is a secondary problem.
David Abrahamsen (1949, 1960) has presented another theoretical view of the connection between the family and crime of juvenile delinquency. According to him all delinquents are emotionally disturbed, and their disturbance results from tensions in the family. Although his position has changed somewhat between 1949 and 1960 so that he acknowledges multiple factors in the causation of delinquency and crime, he still stresses deformed character structure as a basic causal factor, whether dealing with individual delinquents or gang delinquents, lower class delinquents or middle class delinquents. There is little evidence to support his position, however. The Gluecks (1950), for instance, found only 36 psychopaths in their group of 500 delinquents. This finding hardly supports the statement, as Abrahamsen (p. 82) suggests it does, that “basically the persistent juvenile delinquent has a deformed character structure” Furthermore, Abrahamsen’s own evidence in support of his theory about family tension and emotional disturbance is sketchy and unconvincing. He refers to a comparative study he made of 100 criminals and 100 noncriminals who needed treatment. Using data from psychiatric interviews with the criminals and the controls, along with Rorschach tests administered to 31 criminals and 29 of their family members, he concluded that there was much more family tension in the criminal group than in the control group; moreover, he found that criminals always manifested emotional disturbance. Unfortunately, the author worked with knowledge of who was in the delinquent group and control group; he did not attempt to collect similar information from the two groups; and the representativeness of these clinical groups leaves something to be desired. Finally, the author’s acknowledgment that some differences between the groups were not easy to detect except through skilled interviewing and interpretation necessarily puts us on guard about the possible operation of subjective bias.
In general it can be said of the psychiatric theories of delinquency that they are based almost entirely on the subjective clinical experience of the theoretician rather than on objective evidence. Furthermore, they probably apply mainly to that small proportion of the delinquent population which can be termed “psychopathic.” Estimates differ on the size of this ‘psychopathic” group, usually depending upon the diagnostic methods and categories used. As examples, the Gluecks (1950) report 7 percent of their delinquents have serious personality disorders while Reiss (1952) reports over 20 percent with some degree of pathology.