Conduct disorder: a biopsychosocial review.
ABSTRACT To review published works on the epidemiology, risk factors, protective factors, typologies, and genetic aspects of conduct disorder (CD).
Findings from refereed journal articles and current texts in the field are briefly summarized.
CD is commonly encountered in clinical practice. Factors strongly predictive of future delinquency include past offenses, antisocial peers, impoverished social ties, early substance use, male sex, and antisocial parents. Factors that moderately predict recidivism include early aggression, low socioeconomic status (SES), psychological variables such as risk taking and impulsivity, poor parent-child relationships, poor academic performance, early medical insult, and neuropsychological variables such as poor verbal IQ. Mildly predictive variables include other family characteristics such as large family size, family stress, discord, broken home, and abusive parenting, particularly neglect. Protective factors include individual factors such as skill competence (in social and other arenas), adult relationships, prosocial and proeducational values, and strong social programs and supports.
We know a great deal about psychosocial risk factors for CD. Some research into protective factors and genetic contributions exists but is in its early stages. Future work will increase our knowledge about subtypes, developmental pathways, and CD treatment.
- Di ag nos tic evalua tion of the vio lent child and ado les cent. Child and Ado les cent Psy chi at ric Clin ics of North Amer ica. 2000. Oc to ber 815-39..
- Se vere con duct dis or der – some key is sues. 1997. Can J Psy chia try 42 577-83..
- On tario child health study: sum mary of se lected re sults. 1989. Can J Psy chia try 34 483-91..
IN RE VIEW
Conduct Disorder: A Biopsychosocial Review
Lin dley Bas sarath MD, FRCPC1
Ob jec tive: To re view pub lished works on the epi de mi ol ogy, risk fac tors, pro tec tive fac tors, ty polo gies, and ge netic as pects of con -
duct dis or der (CD).
Method: Find ings from refe reed jour nal ar ti cles and cur rent texts in the field are briefly sum ma rized.
Re sults: CD is com monly en coun tered in clini cal prac tice. Fac tors strongly pre dic tive of fu ture de lin quency in clude past of -
fenses, an ti so cial peers, im pov er ished so cial ties, early sub stance use, male sex, and an ti so cial par ents. Fac tors that mod er ately
pre dict re cidi vism in clude early ag gres sion, low so cio eco nomic status (SES), psy cho logi cal vari ables such as risk tak ing and im -
pul siv ity, poor par ent–child re la tion ships, poor aca demic per form ance, early medi cal in sult, and neu rop sy cho logi cal vari ables
such as poor ver bal IQ. Mildly pre dic tive vari ables in clude other fam ily char ac ter is tics such as large fam ily size, fam ily stress,
dis cord, bro ken home, and abu sive par ent ing, par ticu larly ne glect. Pro tec tive fac tors in clude in di vid ual fac tors such as skill
com pe tence (in so cial and other are nas), adult re la tion ships, pro so cial and pro edu ca tional val ues, and strong so cial pro grams
and sup ports.
Con clu sions: We know a great deal about psy cho so cial risk fac tors for CD. Some re search into pro tec tive fac tors and ge netic
con tri bu tions ex ists but is in its early stages. Fu ture work will in crease our knowl edge about sub types, de vel op men tal path ways,
and CD treat ment.
(Can J Psy chia try 2001;46:609–616)
Key words: conduct disorder, delinquency, risk factors, protective factors, genetics, subtypes
For our dis cus sion is on no tri fling mat ter, but on the
right way to con duct our lives.
Plato, The Re pub lic
as pects of CD. Other im por tant as pects, such as as sess ment,
comor bidi ties, neu rop sy chi at ric vul ner abili ties, and treat -
ment have been re cently well- reviewed (1–3). This pa per up -
dates the reader on cur rent think ing about psy cho so cial risk
and pro tec tive fac tors and hints at emerg ing bio logi cal risk
fac tors, such as ge net ics. We con sulted origi nal ar ti cles and
re views in both Med line and Psycinfo, par ticu larly those pub -
lished over the last 15 years. For fur ther in for ma tion, please
con sult the origi nal ref er enced works. Clini cal and pub lic
health im pli ca tions are in di cated, where ap pro pri ate.
his ar ti cle sum ma rizes rele vant find ings on epi de mi ol -
ogy, en vi ron mental con tri bu tions, sub types, and ge netic
CD is 1 of 3 dis rup tive be hav iour dis or ders (4), the oth ers be -
ing op po si tional de fi ant dis or der (ODD) and attention- deficit
hy per ac tiv ity dis or der (ADHD). As in di cated in the DSM- IV,
CD con notes a se vere ex ter nal iz ing dis or der com pris ing se ri -
ous ag gres sive and an ti so cial be hav iours such as fight ing,
bul ly ing, cru elty, rob bery, forc ing sex ual ac tiv ity, fire set ting,
theft, con ning, tru ancy, and other rule vio la tions. An ti so cial
be hav iour de scribes ac tions con trary to the rights of oth ers
and rules of so ci ety. Ado les cent an ti so cial be hav iour that
breaks the law (and gets caught) may re sult in con tact with
po lice and the courts; the terms “de lin quent” and “young of -
fender48 would then ap ply. Thus, CD rep re sents a con stel la -
tion of an ti so cial be hav iours; a sub group of youths with
se vere CD will be de lin quents (5). We will also use the terms
“ag gres sion” and “vio lence” in this over view. Ag gres sion is
de fined as out ward de struc tive be hav iour that re sults from the
con flu ence of longer- term fac tors (for ex am ple, bio logi cal,
psy cho logi cal and per son al ity, fam ily, peer, school, and com -
mu nity), short- term in flu ences (for ex am ple, in ter nal states of
an ger, bore dom, or in toxi ca tion) and situa tional op por tu nity.
Vio lence—a par ticu lar form of overt and in ten tional ag gres -
sion—uses or credi bly threat ens to use physi cal force such as
beat ing, kick ing, chok ing, us ing a weapon, forc ing sex, and
throw ing ob jects.
Can J Psy chia try, Vol 46, September 2001609
Manu script re ceived and ac cepted June 2001.
1Staff Psy chia trist, Cen tre for Ad dic tion and Men tal Health, To ronto; Lec -
turer, Uni ver sity of To ronto, To ronto, On tario.
Ad dress for cor re spon dence: Dr L Bas sarath, Staff Psy chia trist, Cen tre for
Ad dic tion and Men tal Health, 250 Col lege St, Suite 114, To ronto ON M5T
e- mail: lin dley_bas firstname.lastname@example.org
Epi de mi ol ogy
CD is the most com mon rea son for psy chi at ric evalua tion of
chil dren or ado les cents (6). Be tween 30% and 50% of all
child psy chia try re fer rals tend to in volve CD (7). The On tario
Child Health Study in di cated that for ages 4 to 16 years, 5.5%
suf fered from this con di tion (8).
Not all youths with CD have a crimi nal rec ord, and not all
youths with a crimi nal rec ord have CD. Crime sta tis tics there -
fore can be in for ma tive but ex tremely con tro ver sial. In flated
sta tis tics could re sult from popu la tion growth, po lice re port -
ing prac tices, or growth of po lice forces. Equally pos si ble,
un der re port ing may oc cur. This could be due to the ex is tence
of spe cial youth crime units and more community- based po -
lic ing, along with con flict reso lu tion strate gies rather than
for mal prose cu tion. In any event, Sta tis tics Can ada data in di -
cate that youth of fences in creased stead ily un til 1991 but
have de clined since. In 1986, there were 136 787 charges laid
against youth in Can ada; in 1991, there were 171 673; by
1994, these had de clined to 143 337. By 1998–1999, courts
heard 106 665 ado les cent cases in Can ada. While this de cline
in over all num bers is com fort ing, an other dis turb ing trend is
oc cur ring. The pro por tion of youth charged with vio lent of -
fences seems to be in creas ing. In 1986, 408 youths per 100
000 were charged with a vio lent crime (that is, homi cide, at -
tempted mur der, as sault, sex ual of fences, ab duc tion, and rob -
bery). In 1995, the rate was up to 938, rep re sent ing a growth
rate of 13% per an num. We ex pect a 31% in crease in the ju ve -
nile popu la tion over the next 2 dec ades (9). Ac cord ingly, CD
is and will con tinue to be a fairly sig nifi cant pub lic health
prob lem, war rant ing much at ten tion from both the men tal
health com mu nity and the ju ve nile jus tice sys tem.
Sev eral re search ers have ex plored CD sub types and de vel op -
men tal path ways to an ti so cial or vio lent be hav iour. One such
ex am ple is life- course per sis tent (LCP) of fend ing ver sus
adolescence- limited an ti so cial be hav iour (10). The LCP
path way in volves early on set, ac tive and di verse de lin -
quency, in creas ingly se ri ous crime, and con tinua tion into
adult hood. It ac counts for about 5% to 8% of the of fender
popu la tion, but most of fenses. The be hav iours to watch for in
chil dren and youth are as fol lows: bit ing and hit ting at age 4
years; shop lift ing and tru ancy at age 10 years; sell ing drugs
and steal ing cars at age 16 years; rob bery and rape by age 22
years; fraud and child abuse by age 30 years; ad dic tions, un -
der em ploy ment, drunk driv ing, vio lent as sault, do mes tic vio -
lence, child ne glect and aban don ment; and psy chi at ric ill ness
past age 40 years. Adolescence- limited of fend ers tend to have
less pre ado les cence his tory of con duct prob lems. They can
of fend with peers but may be have well in school and at home.
Only 25% of these late start ers con tinue their de lin quent ca -
reers in adult hood. Simi larly, a re cent Ca na dian study found
that 4% of kin der gar ten boys con tin ued through ado les cence
to have a high rate of ag gres sion and that the 2 most pre dic tive
vari ables were teen age preg nancy and low edu ca tional
achieve ment in the mother (11). Thus, more and ear lier psy -
cho so cial in ter ven tion—es pe cially with the early- onset ag -
gres sive and an ti so cial chil dren and their high- risk
par ents—would hope fully save the child, fam ily, and so ci ety
much emo tional, physi cal, and fi nan cial dam age.
Psychosocial Risk Factors
Risk fac tors are those in di vid ual, fa mil ial, or en vi ron mental
fac tors that have been shown in re search to in crease a child’s
risk of fu ture con duct prob lems. While 20% to 30% of ado les -
cents com mit a vio lent act, only 5% to 8% of males and 3% to
6% of fe males are re spon si ble for most such acts (12).
Knowl edge of these youths’ char ac ter is tics—es pe cially if
used to in form ef fec tive in ter ven tions—could be quite thera -
peu tic and pre ven tive in re duc ing their suf fer ing and as so ci -
ated so cial dam age.
A huge body of lon gi tu di nal re search that dis cusses risk fac -
tors ex ists (for ex am ple, 13–16). There are, how ever, a great
va ri ety of meth od olo gies, sam ples, meas ure ments, and out -
come vari ables. Ac cord ingly, metaana ly ses of these pro spec -
tive stud ies also ex ist (17,18). Clini cal im pli ca tions will be
dis cussed where rele vant.
Strongly Pre dic tive Risk Fac tors
Prior An ti so cial Be hav iour. Prior an ti so cial be hav iour (for
ex am ple, steal ing, prop erty de struc tion, early sex ual in ter -
course, and drug sell ing) is proba bly the strong est pre dic tor of
fu ture of fend ing for males, but not nec es sar ily for fe males
(19). Early on set of de lin quent be hav iour has con sis tently
pre dicted more chronic and se ri ous vio lence and of fend ing in
many stud ies (20).
At any age of child hood, past charges, ag gres sion, and vio -
lence raise one’s clini cal sus pi cion of fu ture de lin quency risk.
In one study, 50% of boys con victed for a vio lent crime be -
tween age 10 and 16 years were con victed again of vio lence
by age 24 years, com pared with 8% of those not con victed as
youths (21). Also, dis ci pline prob lems at ages 8 to 10 years,
self- reported con duct prob lems by 14 years, regu lar ciga rette
smok ing by 14 years, and sex ual in ter course by 14 years pre -
dicted later vio lence in males (22). An early his tory of prob -
lem atic be hav iours and de lin quency should there fore be
taken fairly se ri ously, be cause these high- risk chil dren and
youths tend not to grow out of these be hav iours and may re -
quire more in ten sive and sus tained in ter ven tion.
An ti so cial Peers. By ado les cence, de lin quent peers con trib -
ute greatly to the over all spread of an ti so cial be hav iour and
de lin quency. Youths with CD tend to have prob lem atic fami -
lies in the first place, and they are also re jected fre quently by
pro so cial peers. As they sepa rate, they can be come at tached
to de lin quents’ mu tual friends or to other youths with longer
crimi nal his to ries (23). Fur ther, gangs are also re spon si ble for
a sig nifi cant pro por tion of se ri ous and vio lent of fenses (24);
gang in volve ment ap pears to add an ad di tional risk level
610The Ca na dian Jour nal of Psy chia tryVol 46, No 7
above and be yond de lin quent peers and groups. Clini cally, it
may be eas ier to steer chil dren to ward pro so cial peers and ac -
tivi ties be fore ado les cence.
So cial Ties. This cate gory in cludes low popu lar ity and few
so cial ac tivi ties, es pe cially in ado les cence. Chil dren who suf -
fer from early peer re jec tion are more likely to be bul lied, dis -
rup tive in class, and so cially ag gres sive in in ter ac tions (25).
These chil dren of ten al ready have higher lev els of ag gres sive
and ex ter nal iz ing be hav iours; thus, peer re jec tion likely in -
creases their risk for fu ture con duct dif fi cul ties (26). The
clini cal im pli ca tions are sev eral. Get ting a sense of the
youth’s ex tra cur ricu lar hob bies, in ter ests, and ac tivi ties is im -
por tant and will help build a thera peu tic al li ance. In ad di tion,
ex plor ing per son al ity traits, such as neu roti cism or ex trav er -
sion, could be rele vant, as in the Five Fac tor Model of per son -
al ity (27). The pres ence of clus ter A per son al ity
traits—par ticu larly para noid or schiz oid—should be as cer -
tained. Cau tion, how ever, should be used in di ag nos ing per -
son al ity dis or der in chil dren and youth.
Sub stance Use. Early ciga rette smok ing and al co hol or other
sub stance use, par ticu larly be fore age 12 years, con fers in de -
pend ently a greater risk of fu ture an ti so cial ity. This may be
due to any of 3 in ter re lated types of as so cia tions be tween
drugs and crime: psy cho phar ma cologic, eco nomic com pul -
sive, and sys temic (28). Psy cho phar ma cologic mecha nisms
in crease vio lent out comes more di rectly, ei ther through in -
toxi ca tion or with drawal. Economic- compulsive as so cia -
tions re fer to situa tions in which chil dren (or adults) com mit a
crime to ob tain money to pur chase drugs. Sys temic is sues re -
late to in volve ment in the drug trade (for ex am ple, deal ers’
dis putes, prob lems with in form ers, and pun ish ment for adul -
ter ated drugs or un paid debts). The 3 mecha nisms out lined
above sup port the no tion that sub stance use leads to an ti so cial
be hav iour. There are, how ever, many other pos tu lated ex pla -
na tions: youths with CD may self- medicate with mari juana if
they have ADHD; situa tional fac tors such as bars or raves
may in crease the pro pen sity for ag gres sive be hav iours in vul -
ner able youths. No sin gle model has yet been found to neatly
ex plain the re la tion be tween sub stances and con duct prob -
lems, and we likely need to de velop more com plex in ter ac tive
mod els, be cause the rate of comor bid ity be tween sub stance
abuse and CD is as high as 91% (29).
Male Sex. In stud ies and sta tis tics, be ing male has con sis -
tently been shown to con fer much higher risk of CD (the ra tio
of boys to girls is 5:1). Why the sex dif fer ence? Psy cho logi cal
ex pla na tions in clude dif fer ences in so ciali za tion and ad -
vanced adap tive de vel op ment in girls. There are, how ever,
likely hor mo nal, an thro po logi cal, and evo lu tion ary as pects,
as well. Even so, so- called in di rect ag gres sion or re la tional
ag gres sion likely ap plies more to girls (30). This in volves
such acts as spread ing ru mours, shun ning or os tra ciz ing
peers, and con script ing oth ers to “get even” physi cally.
Crimi nol ogy data in di cate that males are more likely to be ar -
rested for vio lent crimes, and fe males for “status of fenses”
(for ex am ple, tru ancy, pros ti tu tion, run ning away, un der age
drink ing, or pur chas ing to bacco). When ag gres sive, girls are
more likely to fight with a par ent or sib ling than with strang -
ers—34% vs 9% for boys (31). Con se quently, a clini cal
search for in ter per sonal or cov ert an ti so cial be hav iours in fe -
males may yield ad di tional in for ma tion.
An ti so cial Par ents. Hav ing a con victed mother, fa ther, or sib -
ling sig nifi cantly pre dicts boys’ con vic tions. Fur ther, this
find ing holds both for self- report and of fi cial crimi nal rec ords
(32). One study that used the Dan ish Na tional Reg is ter found
that men aged 18 to 23 years whose fa thers had at least 2
crimi nal con vic tions were more likely to com mit vio lent
crimi nal acts (33). Whether this re flects ge netic or so cial
learn ing mecha nisms, or both, is be ing de bated (34).
Mod er ately Pre dic tive Risk Fac tors
Early Ag gres sion. Many have noted sta bil ity and con ti nu ity
in early- onset ag gres sive be hav iours in males (35). Hav ing an
ag gres sive style of be hav iour and in ter ac tion may, how ever,
have ear lier pre cur sors. Per sis tent attention- seeking at age 12
months has been found to re late to non com pli ance at age 18
months, which then cor re lates with ag gres sion at age 24
months. This has been re lated to ma ter nal re ports of ex ter nal -
iz ing prob lems at age 36 months (36).
In ter est ingly, in one study, two- thirds of 10- and 13- year- old
boys with high teacher- rated ag gres sion scores had, by age 26
years, a crimi nal rec ord of vio lence. This cor re la tion did not
hold up for girls (37). Some re search ers have opined that by
age 8 years, chil dren have an idio syn cratic ag gres sion level
that is fairly sta ble over time (38). The op po site end of the de -
bate is, how ever, that ag gres sion is not a sta ble trait and is en -
tirely in flu enced by situa tional and con tex tual fac tors (39).
Clini cally, a care ful his tory of the de vel op men tal course of
vari ous op po si tional, ag gres sive, and con duct be hav iours
may sug gest an LCP tra jec tory and its im plied need for in ten -
sive treat ment and re sources.
Low Fam ily So cio eco nomic Status (SES). So many stud ies
have found cor re la tions be tween pov erty and crime that it is
con sid ered by some to be a given that poor fami lies and
neigh bour hoods are as so ci ated with de lin quency. This is
highly con tex tual and not nec es sar ily causal, how ever. In the
US, we found a dif fer ence be tween poor fami lies in inner- city
neigh bour hoods and “other ur ban poor” neigh bour hoods,
with inner- city com mu ni ties hav ing de lin quency rates over
2.5 times the na tional av er age (40). Fur ther, fami lies char ac -
ter ized by so cial iso la tion, bro ken homes, sparse net works,
weak so cial sup ports, and pov erty are much more likely to
physi cally abuse chil dren, in creas ing their risk of ag gres sion
September 2001Conduct Disorder: A Biopsychosocial Review 611
From a pub lic health per spec tive, it has been found that good
par ent ing, sta ble fam ily en vi ron ment, and good early health
and de vel op ment can miti gate against the ef fects of im pov er -
ish ment and un em ploy ment; thus, the per ceived as so cia tion
of pov erty with crime may be jus ti fied only if there are bio -
logi cal and psy cho so cial vul ner abili ties op er at ing to gether
Psy cho logi cal Char ac ter is tics. These in clude a high ac tiv ity
level, risk tak ing, im pul sive ness, and a short at ten tion span.
To gether, these fea tures were as so ci ated in a Swed ish study
of boys at age 13 years with a much higher rate of ar rest for
vio lence by age 26 years than was found for boys with out
these char ac ter is tics (43). A rele vant find ing in the adult lit -
era ture is that an ti so cial per son al ity dis or der (APD) is as so ci -
ated with traits such as high nov elty seek ing (with im pul sive
and tem pes tu ous be hav iour), low harm avoid ance (for ex am -
ple, fear less and dar ing), and low re ward de pend ence (that is,
aloof and in de pend ent) (44).
Pro spec tive and ret ro spec tive stud ies have linked hy per ac tiv -
ity and im pul siv ity to later vio lent be hav iour (45,46). In fact,
sev eral stud ies sug gest that youths with comor bid ADHD and
CD in child hood are at high risk for chronic and per sis tent
con duct and at ten tion prob lems (47,48). This comor bid ity
there fore should be de tected and ag gres sively treated in an at -
tempt to re duce fu ture risk.
Par ent–Child Re la tion ships. These in clude dis ci pline prac -
tices (for ex am ple, mixed, in con sis tent, or pu ni tive), mini mal
in volve ment, poor su per vi sion, low emo tional warmth, and a
nega tive at ti tude to ward the child. Nu mer ous stud ies have
shown that these par ent ing prac tices are con sis tently linked
with later de lin quency and sub stance abuse (49). Simi larly,
low in volve ment and in ter ac tion—par ticu larly be tween sons
and fa thers at age 8 years—was also found to pre dict later vio -
lence (22). In an other study, a high de gree of “nega tive fam ily
la bel ling” of the youth (a meas ure ment of fam ily bond ing)
was found to be as so ci ated with sex ual ag gres sion in later
teen age years from ages 13 to 19 (50). A posi tive par ent ing
style has been found to im prove the out come when en vi ron -
mental risk fac tors such as pov erty ex ist (51). Clini cally,
family- focused treat ments, such as Func tional Fam ily Ther -
apy, Par ent Man age ment Train ing, or Multi sys temic Ther apy
are among the most prom is ing to re duce re cidi vism (1).
School At ti tude and Per form ance. This cate gory in cludes a
low in ter est in edu ca tion, drop ping out, low school achieve -
ment, and tru ancy. Poor aca demic achieve ment in ele men tary
school, and par ticu larly in high school, has been re lated con -
sis tently to later con duct dif fi cul ties (52).
There has been a spe cific link age dem on strated be tween at -
ten tion defi cits and un dera chieve ment (53). Thus, rul ing out
comor bid sub types of ADHD at any age is im pera tive. Re -
view ing school rec ords is time- consuming but in valu able.
We rec om mend in tel li gence and achieve ment test ing (54). If
pos si ble, neu rop sy cho logi cal test ing for ver bal and ex ecu tive
func tion defi cits, among other de vel op men tal de lays, is also
in di cated.
Medi cal and Physi cal Con di tions. These in clude per ti nent
find ings in de vel op men tal his tory, medi cal con di tions, and
physi cal de vel op ment. Pre na tal and de liv ery com pli ca tions
some what pre dict, in par ticu lar, later vio lent of fend ing (55).
In a study to test bio so cial in ter ac tions, birth com pli ca tions
(for ex am ple, for ceps ex trac tion, breech de liv ery, um bili cal
cord pro lapse, preeclamp sia, and long birth du ra tion) greatly
in creased the risk of vio lent of fend ing in adult hood, but only
when as so ci ated with early child re jec tion (for ex am ple, un -
wanted preg nancy, at tempted abor tion, and in sti tu tional care
for at least the first 4 months of life) (56). This has pro found
medi cal and pub lic health im pli ca tions. In simi lar in ter ac -
tional style, large body height and weight at age 3 years, cou -
pled with fear less ness and novelty- seeking traits, were found
to pre dict ag gres sion at age 11 years (57). Other medi cal and
psy chi at ric con di tions worth rul ing out in clude mi nor physi -
cal anoma lies, head in ju ries, neu ro logi cal con di tions, bi po lar
dis or der, dis so cia tive dis or ders, and para noid dis or ders
IQ. Most stud ies have found that IQ scores of chil dren with
CD are on av er age 8 points lower than those of non de lin quent
chil dren, even when other vari ables such as SES, eth nic ity,
aca demic achieve ment, and mo ti va tion are con trolled. An -
other con sis tent find ing is that per form ance IQ is greater than
ver bal IQ (58). This strongly sug gests spe cific lan guage dif fi -
cul ties and pos si ble neu rop sy chi at ric dys func tion in chil dren
and youths with CD. Ac tu ally, when the sub set of early- onset
de lin quents was ex am ined, they were found to have IQ scores
17 points lower on av er age than those of non de lin quent sub -
jects, a find ing also ob served in chil dren with early brain
dam age (59,60). Thus test ing, where avail able, could aid as -
sess ment and in form us about needed edu ca tional re sources.
Mildly Pre dic tive Risk Fac tors
Other Fam ily Char ac ter is tics. These re fer to fac tors such as
high fam ily stress, large fam ily size, and mari tal dis cord. An
early find ing was that boys raised in high- conflict fami lies
were more likely to be con victed later of a vio lent crime (61).
More spe cifi cally, be ing ex posed to vio lence be tween par ents
in creases the risk of later vio lence in the chil dren when adults
(62). This, how ever, has not al ways been found: in the Se at tle
So cial De vel op ment Study, fam ily con flict at age 10 years
was not cor re lated with vio lence, ac cord ing to self- reports, at
age 18 years (63).
Bro ken Home. This cate gory in cludes both fam ily break down
(for ex am ple, di vorce) and sepa ra tion from par ents for other
rea sons. The re la tion be tween bro ken homes and child or
youth ag gres sion is com plex, be cause fami lies sepa rate for
many rea sons. Not with stand ing this, par ent–child sepa ra tion
612The Ca na dian Jour nal of Psy chia tryVol 46, No 7
be fore age 10 years has been found in more than 1 study to
pre dict con vic tions for vio lence in later ado les cence and up to
age 21 years (64).
Abu sive Par ents. This in cludes all cate go ries such as emo -
tional, physi cal, and sex ual abuse, as well as mal treat ment
and ne glect. It has been con sis tently found that, com pared
with those with out an abuse his tory, adults who were sexu ally
abused as chil dren were less likely to com mit vio lent crimes
than were con trol sub jects. Adults who were physi cally
abused were slightly more likely, while those ne glected as
chil dren were the most likely, to be ar rested for vio lence (65).
Thus, ne glect—the most com mon and highest- risk form of
abuse—should be seen as a se ri ous pub lic health is sue with
im pli ca tions for vio lence pre ven tion.
Risk Factors: Synthesis
No sin gle fac tor ac cu rately pre dicts later con duct prob lems.
Fac tors ag gre gate to pro duce in creased vul ner abil ity in these
chil dren. An ex cel lent ex am ple is the Cam bridge Study in
De lin quent De vel op ment (66), in which 411 boys in
working- class Lon don, Eng land, were fol lowed from age 8 to
40 years. Vul ner abil ity scores were cal cu lated based on 5 risk
fac tors: low fam ily in come at age 8 years, large fam ily size by
age 10 years (4 or more bio logi cal sib lings), low non ver bal IQ
at age 8 to 10 years, and poor par ent ing (harsh or in con sis tent
dis ci pline and par ent con flict) at age 8years. For those with no
risk fac tors, con vic tions for vio lence were in creased by only
3%, com pared with con trol sub jects. Hav ing 4 to 5 fac tors
pres ent, how ever, in creased one’s risk by 31%. Thus, the in -
ter ac tion be tween in di vid ual and en vi ron mental risk fac tors
really de ter mines the over all vari ance in the out come of these
Pro tec tive fac tors are not sim ply the ab sence or op po site of
risk fac tors, al though this is some times the case. In fact, de -
spite ex po sure to mul ti ple known risk fac tors, many chil dren
avoid se ri ous an ti so cial be hav iour (67). Pro tec tive fac tors are
best de fined as those vari ables that off set the ef fects of risk
fac tors. Re search, how ever, has largely ig nored these fac tors
in fa vour of elu ci dat ing risk. It is quite likely that more em -
pha sis on these vari ables could sig nifi cantly in flu ence prac -
tice and pol icy.
In di vid ual Pro tec tive Fac tors (13,68)
Fe male Sex. Be ing fe male may be pro tec tive via dif fer ent
par ent ing or so ciali za tion pat terns; also, girls gen er ally ma -
ture and ac quire skills more quickly.
High In tel li gence. This is as meas ured by stan dard IQ test ing.
Such test ing, how ever, has been criti cized as too lim ited with
re spect to types of in tel li gence and cul tural di ver sity.
Posi tive So cial Ori en ta tion. This would likely in clude the ab -
sence of an ti so cial at ti tudes and cog ni tive bi ases, such as in -
ter pret ing so cial cues as nec es sar ily hos tile or threat en ing.
Re sil ient Tem pera ment. This usu ally means pos sess ing good
cop ing skills and the abil ity to en dure stress, hard ship, or
trauma with out men tal de com pen sa tion.
Com pe tence at a Skill. At tain ing at least 1 area of good abil ity
at a skill, hobby, or in ter est has been noted to be pro tec tive,
per haps lead ing to other pro so cial ac tivi ties and in ter ac tions.
Anxi ety. Ac cord ing to some re search, hav ing anxi ety, worry,
and guilt tends to pro tect against the de vel op ment of an ti so -
cial be hav iour.
So cial Fac tors (67,69,70)
Warm, Sup por tive Re la tion ships with Adults. Posi tive in ter -
ac tions, warmth, or men tor ship from adults has been found to
be greatly bene fi cial in in flu enc ing away from CD.
In di vid ual and Fam ily Com mit ment to So cial Val ues Such As
Pro so cial Norms or Aca demic Achieve ment. As pir ing to uni -
ver sal val ues such as posi tive so cial in ter ac tions with oth ers
and school suc cess helps chil dren de crease their risk of con -
duct prob lems and in creases their achieve ments and
op por tu ni ties.
Rec og ni tion for In volve ment in Posi tive Ex tra cur ricu lar Ac -
tivi ties. Not just be ing in volved in pro so cial ac tivi ties but ac -
tu ally be ing ac knowl edged, and even re warded, helps
per petu ate such posi tive be hav iour.
So cie tal Fac tors (71,72)
In creased Eco nomic Equal ity and Na tional So cial Pro gram
Sup port. De vel oped coun tries with larger social- program
spend ing tend to have lower homi cide rates.
So cial Or gani za tion. Strong and sta ble com mu nity in sti tu -
tions (for ex am ple, church, neigh bor hood or gani za tions, and
ex tended fami lies), as op posed to dis or gan ized, cha otic, and
crum bling com mu ni ties, tend to be pro tec tive.
It is well known that psy chi at ric dis or ders of ten run in fami -
lies. Sim ple men de lian in heri tance, how ever, is not found
with such com plex con di tions. Given the di ag nos tic het ero -
ge ne ity of CD, a more ap pro pri ate phe no type, such as ag gres -
siv ity or vio lence, is used in re search. Ge netic stud ies can be
broadly cate go rized into 2 types: those us ing twin, adop tion,
or fa mil ial meth ods, and those us ing mo lecu lar bi ol ogy meth -
ods. Dis cuss ing the sci ence be hind these is be yond the scope
of this ar ti cle. One ex cel lent rep re sen ta tive twin study looked
at the re la tion be tween child be hav iour check list (CBCL)
syn dromes such as mood and anxi ety dis or ders, at ten tion
defi cits, an ti so cial be hav iour, and ag gres sive be hav iour (73)
and tem pera ment rat ings us ing the Emo tion al ity, Ac tiv ity,
September 2001Conduct Disorder: A Biopsychosocial Review 613