Title Page
 Table of Contents
 Appendix A: Socioeconomic...
 Appendix B: DSM-III-R structured...
 Appendix C: PCIT questionnaire
 Appendix D: Demographic questi...
 Appendix E: DPICS coding categ...
 Appendix F: Recruitment letter...
 Appendix G: Tables not appropriate...
 Biographical sketch

The Long term effectiveness of parent-child interaction therapy with behavior problem children and their families
Full Citation
Permanent Link: http://ufdc.ufl.edu/AA00009772/00001
 Material Information
Title: The Long term effectiveness of parent-child interaction therapy with behavior problem children and their families a two-year follow-up
Alternate title: Two-year follow-up
Physical Description: ix, 392 leaves : ; 29 cm.
Language: English
Creator: Newcomb, Katharine Payne, 1955-
Publication Date: 1995
Subjects / Keywords: Research   ( mesh )
Child Behavior Disorders -- psychology   ( mesh )
Child Behavior Disorders -- therapy   ( mesh )
Mother-Child Relations   ( mesh )
Family Therapy   ( mesh )
Longitudinal Studies   ( mesh )
Questionnaires   ( mesh )
Treatment Outcome   ( mesh )
Department of Clinical and Health Psychology thesis Ph.D   ( mesh )
Dissertations, Academic -- College of Health Related Professions -- Department of Clinical and Health Psychology -- UF   ( mesh )
Thesis: Thesis (Ph.D.)--University of Florida, 1995.
Bibliography: Bibliography: leaves 375-391.
General Note: Typescript.
General Note: Vita.
Statement of Responsibility: by Katharine Payne Newcomb.
 Record Information
Source Institution: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: aleph - 002302456
oclc - 49849413
notis - ALQ5755
System ID: AA00009772:00001


This item has the following downloads:

Newcomb ( PDF )

Table of Contents
    Title Page
        Page i
        Page ii
        Page iii
    Table of Contents
        Page iv
        Page v
        Page vi
        Page vii
        Page viii
        Page ix
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
        Page 155
        Page 156
        Page 157
        Page 158
        Page 159
        Page 160
        Page 161
        Page 162
        Page 163
        Page 164
        Page 165
        Page 166
        Page 167
        Page 168
        Page 169
        Page 170
        Page 171
        Page 172
        Page 173
        Page 174
        Page 175
        Page 176
        Page 177
        Page 178
        Page 179
        Page 180
        Page 181
        Page 182
        Page 183
        Page 184
        Page 185
        Page 186
        Page 187
        Page 188
        Page 189
        Page 190
        Page 191
        Page 192
        Page 193
        Page 194
        Page 195
        Page 196
        Page 197
        Page 198
        Page 199
        Page 200
        Page 201
        Page 202
        Page 203
        Page 204
        Page 205
        Page 206
        Page 207
        Page 208
        Page 209
        Page 210
        Page 211
        Page 212
        Page 213
        Page 214
        Page 215
        Page 216
        Page 217
        Page 218
        Page 219
        Page 220
        Page 221
        Page 222
        Page 223
        Page 224
        Page 225
        Page 226
        Page 227
        Page 228
        Page 229
        Page 230
        Page 231
        Page 232
        Page 233
        Page 234
        Page 235
        Page 236
        Page 237
        Page 238
        Page 239
        Page 240
        Page 241
        Page 242
        Page 243
        Page 244
        Page 245
        Page 246
        Page 247
        Page 248
        Page 249
        Page 250
        Page 251
        Page 252
        Page 253
        Page 254
        Page 255
        Page 256
        Page 257
        Page 258
        Page 259
        Page 260
        Page 261
        Page 262
        Page 263
        Page 264
        Page 265
        Page 266
        Page 267
        Page 268
        Page 269
        Page 270
        Page 271
        Page 272
        Page 273
        Page 274
        Page 275
        Page 276
        Page 277
        Page 278
        Page 279
        Page 280
        Page 281
        Page 282
        Page 283
        Page 284
        Page 285
        Page 286
        Page 287
        Page 288
        Page 289
        Page 290
        Page 291
        Page 292
        Page 293
        Page 294
        Page 295
        Page 296
        Page 297
        Page 298
        Page 299
        Page 300
        Page 301
        Page 302
        Page 303
        Page 304
        Page 305
        Page 306
        Page 307
        Page 308
        Page 309
        Page 310
        Page 311
        Page 312
        Page 313
        Page 314
        Page 315
        Page 316
        Page 317
        Page 318
        Page 319
        Page 320
        Page 321
    Appendix A: Socioeconomic index
        Page 322
        Page 323
    Appendix B: DSM-III-R structured interview
        Page 324
        Page 325
        Page 326
        Page 327
        Page 328
        Page 329
        Page 330
        Page 331
    Appendix C: PCIT questionnaire
        Page 332
        Page 333
        Page 334
        Page 335
        Page 336
    Appendix D: Demographic questionnaire
        Page 337
        Page 338
    Appendix E: DPICS coding categories
        Page 339
        Page 340
    Appendix F: Recruitment letter to parents
        Page 341
        Page 342
    Appendix G: Tables not appropriate for the body of the text
        Page 343
        Page 344
        Page 345
        Page 346
        Page 347
        Page 348
        Page 349
        Page 350
        Page 351
        Page 352
        Page 353
        Page 354
        Page 355
        Page 356
        Page 357
        Page 358
        Page 359
        Page 360
        Page 361
        Page 362
        Page 363
        Page 364
        Page 365
        Page 366
        Page 367
        Page 368
        Page 369
        Page 370
        Page 371
        Page 372
        Page 373
        Page 374
        Page 375
        Page 376
        Page 377
        Page 378
        Page 379
        Page 380
        Page 381
        Page 382
        Page 383
        Page 384
        Page 385
        Page 386
        Page 387
        Page 388
        Page 389
        Page 390
        Page 391
    Biographical sketch
        Page 392
        Page 393
        Page 394
Full Text







May we open to a deeper understanding

and a genuine love and caring

for the multitude of faces,

who are none other than ourself.

-- Wendy Egyoku Nakao

This dissertation is lovingly dedicated

to the spirit of CDI, to the quiet truths

that become larger from looking for what is

best in other people, especially in children.


I wish to thank the people who mattered the most:

my husband Ward Newcomb for his unfailing kindness and

support in matters large and small, both of my families

for their encouragement, my committee members Dr.

Sheila Eyberg, Dr. Russell Bauer, Dr. Stephen Boggs,

Dr. John Graham-Pole, and Dr. Mary Lou Koran, for the

unique contributions they have each made to this

project, for the specific forms of help when needed,

and for the sharing of their expertise and time in the

critical review of this manuscript; my coders Janet

Bessmer and Peter Faust for many long hours of careful

coding; Dr. Jim Sturges and Dr. John Lipscomb for much

assistance in data analyses and coding equipment; my

colleagues Dr. Beverly Funderburk, Dr. Toni Hembree-

Kigin, and Dr. Cheryl Bodiford McNeil for humor,

perspective, and practical suggestions; and a special,

final thanks to Dr. Eyberg for timely combinations of

all of the above.



ACKNOWLEDGMENTS ................................... iii

ABSTRACT ..................................... ...... vii


1 INTRODUCTION ................................... 1

Correlates of Conduct Disordered Behavior
in Children ............................... 5
Course of Disruptive Behavior Disorders....... 12
Treatment of Conduct Disordered Behavior
In Children ............................... 14
Temporal Generalization of Parent Training
with Pre-Adolescent Children ............. 19
Treatment Failures in Parent Training........ 89
Factors Contributing to Outcome in the
Treatment of Behavior Problem Children .... 90
Conclusions from the Literature.............. 100
Behavioral Framework for Follow-up Assessment 103
Statement of Problem and Study Hypotheses ... 106

2 METHOD ...................................... 114

Participants ................................ 114
Assessment of Child Functioning............. 124
Assessment of Parent Functioning............. 136
Assessment of Dyadic Functioning............. 143
Procedures .................................. 148

3 RESULTS ..................................... 151

Reliability ................................. 151
Preliminary Analyses ........................ 162

Outcome for the Combined Sample of Families.. 163
The Maintenance of Treatment Effects for
CDI-First Families vs. PDI-First Families.. 198
Self-Report of the Use of Parenting Skills in
Successful vs. Nonsuccessful Families ..... 229
The Effects of Socioeconomic Status on
Parenting Skills .......................... 241
Parental Locus of Control and Treatment
Outcome ................................... 244
The Social Validity of Parent-Child
Interaction Therapy ...................... 249

4 DISCUSSION .................................. 255

Follow-Up Attrition ......................... 256
Families Who Did Participate in Follow-Up.... 259
Reliability ................................. 264
Treatment Maintenance for the Combined Sample 265
The Maintenance of Treatment Efects for
CDI-First Families vs. PDI-First Families.. 282
Self-Report of the Use of Parenting Skills
in Successful vs. Nonsuccessful Families .. 300
The Effects of Socioeconomic Status on
Parenting Skills .......................... 305
Parental Locus of Control and Treatment
Outcome ................................... 308
Limitations of this Study .................. 310
Contributions of this Study ................. 319


A SOCIOECONOMIC INDEX ......................... 322


C PCIT QUESTIONNAIRE .......................... 332

D DEMOGRAPHIC QUESTIONNAIRE ................... 337

E DPICS CODING CATEGORIES ..................... 339


BODY OF THE TEXT .......................... 343

REFERENCES .......................................... 375

BIOGRAPHICAL SKETCH .............................. 392

Abstract of Dissertation presented to the Graduate School of
the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Katharine Payne Newcomb

May 1995

Chairperson: Sheila Eyberg, Ph.D.
Major Department: Clinical and Health Psychology

The long-term effects of treatment with Parent-Child

Interaction Therapy were evaluated one and two years

posttreatment for 13 mother/child dyads with children (ages

two to six) referred for severe conduct problems. All

referred children met the DSM-III-R criteria for either

Oppositional Defiant Disorder, Attention Deficit

Hyperactivity Disorder, or Conduct Disorder. Families

received 14 weekly sessions of PCIT. Seven Child-Directed

Interaction (CDI) sessions focused on improving the parent-

child relationship, and seven Parent-Directed Interaction

(PDI) sessions were directed toward modifying noncompliant

and disruptive behaviors. Families were evaluated at

pretreatment, posttreatment, one- and two-year follow-up,


using parent-report, child-report, and observational


On average, the 13 families in this study moved from

outside of normal limits at pretreatment to within normal

limits at posttreatment and one-year follow-up on measures

of compliance, conduct problems, activity level, and

parenting stress. After therapy, mothers interacted more

affectionately and engaged in fewer negative interactions

with their children. Mothers reported a high degree of

satisfaction with treatment at posttreatment and both


At posttreatment, 11 families were fully successful,

achieving clinically significant change on both parent

report measures and observational data. At Follow-Up I,

eight families were fully successful, four were partially

successful, and one was unsuccessful. At Follow-up II, nine

families remained fully successful, two families were

partially successful, and two were unsuccessful.

To evaluate the long-term impact of treatment phase

sequence, seven families received CDI first and six families

participated in PDI first. Groups were matched on

demographic characteristics and degree of child behavioral

disturbance, and were compared after 14 weeks of therapy,

and at one-year and two-year follow-ups.

The two treatment groups were equally successful at

posttreatment, achieving clinically significant change on

both parent report measures and observational data. At


Follow-up I, four (57%) CDI-First families and four (66%)

PDI-First families were classified as fully successful.

However, at Follow-up II, fully successful treatment

outcomes included six of seven (86%) CDI-First families but

only three of six (50%) PDI-First families. Reasons for the

discrepant outcomes and clinical and research implications

of these results were discussed.



The DSM-IV (American Psychiatric Association, 1994)

recognizes a subclass of behaviors known as the Disruptive

Behavior Disorders, which include Oppositional Defiant

Disorder, Conduct Disorder, and Attention-deficit

Hyperactivity Disorder. The behaviors associated with these

disorders are usually manifested in childhood or adolescence

but may persist into adulthood.

The essential features of each of the disorders as

outlined in the DSM-IV are unchanged from the essential

features of the three categories in the DSM-III-R (American

Psychiatric Association, 1987), the manual on which the

diagnostic procedures of this study were based. The

essential feature of Oppositional Defiant Disorder is a

pattern of negativistic, hostile, and defiant behavior that

involves no serious violations of the rights of others and

is usually manifested in the home. In contrast, the

essential feature of Conduct Disorder is a persistent

pattern of behavior in which the rights of others, as well

as major age-appropriate societal norms, are violated.

Conduct disordered behaviors are typically present in the

home, at school, with peers, and in the community. Finally,

Attention-deficit Hyperactivity Disorder is characterized by

developmentally inappropriate degrees of inattention,

impulsiveness, and hyperactivity, manifested in most

situations. In both clinic and community samples, the

symptoms of these disorders covary to a high degree (APA,

1987, pp. 49-50), and a child may meet the criteria for ADHD

simultaneously with either Oppositional Defiant Disorder or

Conduct Disorder.

Many different labels are used to describe the behavior

of these children, including "antisocial," "conduct-

disordered," "conduct problem behavior," "oppositional,"

noncompliantt," or "externalizing" (McMahon & Wells, 1989),

which makes comparison across populations and studies

difficult. In this paper, the terms "disruptive behavior"

and "disruptive behavior disorders" will be the preferred

descriptors for the behavior of these children. The terms

"antisocial" or "antisocial conduct" will appear when their

use is more accurate.

Disruptive behaviors range from relatively minor but

annoying behaviors, such as whining, yelling, and temper

tantrums, to physical destructiveness, aggression, and

stealing (McMahon & Wells, 1989). For any given behavior,

the frequency, intensity, repetitiveness, and chronicity may

vary widely across children. When disruptive behaviors

present as a syndrome, they may involve overt or confrontive

behaviors (e.g. fighting, arguing), covert or concealed

antisocial acts (e.g. stealing, lying, fire setting), or

some combination of both (Loeber & Schmaling, 1985a).

Noncompliance appears to be of central importance in the

clinical picture of both disruptive behavior disorders and

antisocial conduct and children's failure or inability to

comply to school rules is regarded as the leading cause of

children falling behind their age groups in school (as

reviewed in Reid, 1993).

Estimates of prevalence are particularly hard to

obtain, in part because of the inconsistent use of

diagnostic labels given these children. Given this

limitation, the prevalence of disruptive behavior disorder

in the general population has been estimated to be between

3% and 10% (Kazdin, 1987), and boys are two or three times

as likely as girls to manifest and be diagnosed with some

form of disruptive behavior disorder (Quay, 1986).

Gender differences have been observed among even very

young children with conduct problems. In a study of the

development of aggression in toddlers, 89 mother-child dyads

(52 boys, 37 girls) were observed in laboratory assessments

when the child was 18 months old, and again when the child

was 24 months old, and the frequency and pervasiveness of

aggression was coded. Across the two observations, stability

of aggression was moderate, especially for aggression

observed under low-stress situations, and few sex

differences were found in the frequency or stability of

aggression. However, marked differences were found in the

correlates and predictors of aggression for boys and girls,

suggesting that gender-specific interactional models of the

development of aggression may be useful (Keenan & Shaw,

1994). In another study of the preschool-age antecedents of

conduct problems, 79 children 4 to 5 years of age were

examined on measures of social competence, self-control, and

vocabulary (Olson & Hoza, 1993). As they made the transition

from pre-school to kindergarten, boys with high behavior

problem scores performed more impulsively on a delay-of-

gratification task, responded more aggressively to

hypothetical peer-conflict situations, and received more

negative peer nominations than others. Conduct problems in

the girls were associated with motor impulsivity, low

vocabulary scores, and with measures of both positive and

negative peer status (Olson & Hoza, 1993). Behavior problems

in girls have been associated with an increased risk for

both internalizing and externalizing problems later in life

(Robins, 1986), but less is known about conduct problems in

girls than in boys (Reid, 1993).

Correlates of Conduct Disordered Behavior in Children

The prevailing theoretical formulation of the

development and maintenance of disruptive behavior disorders

has emphasized the primacy of the family socialization

process (Patterson, 1986). A variety of behavioral,

cognitive, and personality characteristics of parents, as

well as family and other contextual variables, are

correlated with disruptive behavior disorders and are

thought to have important implications for identifying and

implementing effective treatments.

Observation of oppositional preschool children in their

home settings reveals that these children are highly

aggressive toward their parents, demonstrating over twice

the rate of aversive behaviors in interactions with family

members than control children do (Reid, 1978). In contrast

to parents of normal children, parents of behavior problem

children report greater numbers of problem behaviors from

their children and have been found to engage in higher

numbers of commanding and critical behaviors toward their

children (McMahon & Wells, 1989). Mothers and siblings of

socially aggressive children engage in more coercive

interactions than do members of nonreferred families, and

parents of children who steal are reported to be more

distant and less involved in interactions with their

children (Patterson, 1982). Poor monitoring skills and

maternal rejection were demonstrated by families with

children who fought and stole (Loeber & Schmaling, 1985b),

and the fathers of children who steal appear to be less

involved in their children's discipline than fathers of

socially aggressive or normal children (Loeber, Weissman, &

Reid, 1983).

Mothers of behavior problem children report higher

levels of parenting stress, more external locus of control,

and less satisfaction in their parenting role than parents

of nonreferred children (Mouton & Tuma, 1988). The parents

of children who demonstrate high rates of disruptive or

deviant behavior report more negative perceptions of their

children's behavior and experience more depression, anxiety,

and marital distress than parents of nonreferred children

(McMahon & Wells, 1989). Depression has been shown to

affect mothers' and fathers' perceptions of their children

(Webster-Stratton & Hammond, 1990) and parental monitoring

skills (Patterson, 1982). Behavior problem children are

generally more maladjusted when their mothers are depressed

(Dumas, Gibson, & Albin, 1989). Maternal depressive symptoms

have been found to interact with children's behavior toward

parents and siblings (Dumas & Gibson, 1990) and to affect

not only the quality of a mother's care but the quality and

organization of the home environment as well (Egeland,

Kalkoske, Gottesman, & Erickson, 1990). In a recent study of

96 families with behavior problem children between the ages

of 4 and 9 years old, home observational data generally

supported previous findings that depressed mothers report

higher levels of child maladjustment and behave in a more

aversive and controlling manner toward their children than

nondepressed mothers, although the children in this sample

were not more aversive or less compliant than the children

of nondepressed mothers (Dumas & Serketich, 1994). While

children's behavior during observations was not related to

the degree of their mothers' depressive symptoms,

symptomatic mothers did display more aversiveness and more

aversive affect toward their children, and maternal

irritability in turn predicted child aversiveness and

aversive affect. Data also suggested that socioeconomic

disadvantage played a crucial role in the association

between maternal depressive symptoms and child

maladjustment, and that in fact maternal depressive symptoms

appeared to reflect socioeconomic adversity more than actual

psychiatric dysfunction or children's actual behavior toward

their mothers. Tannenbaum and Forehand (1994) found that for

a sample of 282 nonreferred 11- to 15-year old adolescents,

maternal depression was associated with poorer functioning

in terms of internalizing problems, externalizing problems,

and grade point average, but for internalizing and

externalizing problems, a good father-adolescent

relationship served to buffer the child from the adverse

effects of the mother's depression.

Marital distress has been positively related to

observed levels of both child deviant behavior and negative

parental behavior (Rutter, 1970; Webster-Stratton & Hammond,

1990), although the marital-child adjustment relationship

weakens when social desirability is controlled for (Robinson

& Anderson, 1983). At least one group of investigators

failed to find a relationship between marital discord and

overall levels of observed child deviance, parent-reported

child deviance, or maternal aversiveness (Dadds, Schwartz, &

Sanders, 1987); however, in several other studies,

children's behavior problems have been associated with

marital distress, negative parental affect, disagreements

over child rearing, and ineffective marital communication

(Grych & Fincham, 1990; Jouriles et al., 1991). Researchers

have moved beyond the view that family structural variables

alone necessitate the development of psychopathology in

children, broadening the focus of attention to include the

events that accompany marital discord and/or dissolution,

now identified as highly salient correlates of children's

adjustment (Shaw, 1991). The role of triadic parent-child

relationships in marital discord-behavior problem

associations was examined by observing how effectively

mother-father pairs worked together to support their child's

efforts during problem-solving tasks, and results provide

preliminary support for a model in which triadic mother-

father-child processes mediate the associations between

marital adjustment and children's functioning (Westerman &

Schonholtz, 1993). In a sample of 112 preschool boys, family

adversity, as well as lower IQ and severity of symptoms at

intake, discriminated boys with continuing behavior problems

at a two year follow-up from those with less serious

difficulties, and control group boys with emerging behavior

problems were also characterized by more family problems and

discord when compared to nondistressed controls (Campbell,

1994). Finally, in studies of over 400 families with young,

oppositional, behavior problem children, Webster-Stratton

(1994) notes that 75% of the parents reported having been

divorced at least once or described their marriages as

highly distressed, and half of the couples reported

experiences of spouse abuse.

In contrast to marital distress, the experience of

extreme levels of social isolation in the parenting role may

also be related to child deviant behaviors. In a one-year

follow-up study of a 10-week parent training program,

single-parent status was found to be associated with more

negative behaviors and higher child deviance (Webster-

Stratton & Hammond, 1990). There is also evidence that

maternal social isolation, apart from single-parent status,

may function as a setting event for dysfunctional parent-

child interactions in some families. Maternal "insularity"

is defined as a specific pattern of social contacts within

the community that is characterized by a high level of

negatively perceived coercive interchanges with relatives

and/or helping agency representatives and by a low level of

positively perceived supportive interchanges with friends

(Wahler & Dumas, 1984). Maternal insularity is positively

related to negative parent behaviors directed toward the

child, in that when a mother has a large proportion of

coercive or aversive interactions outside the home, her

interactions with her child within the home are likely to be

aversive as well (Wahler, 1980).

More recently, the role of stress-induced deficiencies

in maternal attention in dysfunctional mother-child

interactions has been examined and it is proposed that some

mothers, living in the midst of extreme environmental

stressors, are unable to recognize and/or respond to the

cues offered through their children's behaviors (Wahler &

Dumas, 1989). Stressors may disrupt parenting practices by

causing some parents to be more irritable, critical, and

punitive, which in turn increases the likelihood of conduct

problems and further aversive parent-child interactions

(Webster-Stratton, 1990a). High levels of maternal stress

for parents of 6 month-old children have been associated

with less cooperative, more intrusive, more disjointed

parent-child interactions at 6 months and with

noncompliance, negativity, and avoidance of the mother in

teaching tasks (particularly by daughters) for children at

42 months of age (Pianta & Egeland, 1990). In a study of 66

single nonreferred mothers with at least one 5- to 12-year-

old child living in the home, mothers reportedly perceived

child behavior as more problematic, upsetting, and

warranting of a more intense behavioral response when it

occurred in stressful contexts compared to nonstressful

contexts, and daily stressors were associated with

significantly more intense and negative perceptions and

reactions for these parents than were major life events

(Krech & Johnston, 1992). In another sample of 41

nonreferred mother-child dyads, maternal perceptions of high

daily stress were related to mothers' reports of child

behavior problems and symptoms and to children's reports of

their own adjustment (Thompson, Merritt, Keith, Murphy, &

Johndrow, 1993).

In addition to the effects of early social environment,

evidence from twin and adoption studies and investigations

of temperamentally difficult children (Webster-Stratton &

Eyberg, 1982; Olweus, 1980) raise the question of a genetic

predisposition to the development of disruptive behavior

disorders. Cadoret, Cain, and Crowe (1983) found a greater

likelihood of adolescent conduct disordered behavior when

both genetic and environmental influences were present,

suggesting that both child and parent/family characteristics

contribute to the clinical picture, although it is not

possible at this time to answer the question of


Course of Disruptive Behavior Disorders

Disruptive behavior disorders, in their advanced forms,

are characterized by antisocial behavior which tends to be

relatively stable over time. In a sample of 29 hard-to-

manage preschoolers first assessed at age 3, 67% of those

showing clinically significant problems at age 6 also showed

problems of clinical significance at age 9, and earlier

child behavior was predictive of diagnostic status at age 9

(Campbell & Ewing, 1990). Loeber (1982) noted that

antisocial behavior which is more extreme on at least one of

the following dimensions is likely to persist into

adolescence and adulthood: density of the behavior,

occurrence of the behavior in more than one setting, degree

of variety of antisocial acts, and age of onset of

antisocial behavior. In particular, early onset of behavior

problems is a powerful predictor of the frequency and

severity of behavior problems in adolescence. Adolescent

boys typified by their mothers as difficult children at ages

1 to 5 were found to display offender rates twice that of

boys described as easy children to care for at those ages,

with aggression and hyperactivity playing prominent roles in

the later development of antisocial behavior (Loeber,

Stouthamer-Loeber, & Green, 1991).

That antisocial behavior represents a tremendous cost

to society is evident both in specific ways (e.g. vandalism

and firesetting) and more generally in the fact that

antisocial youth often remain in continued contact with

criminal justice and mental health systems well into

adulthood (Alexander & Parsons, 1982). Antisocial behavior

in childhood has been found to predict similar behavior in

one's offspring, traceable into the third generation (Glueck

& Glueck, 1968).

In addition to being a very costly societal problem,

antisocial conduct is also a personal problem characterized

by chronic maladjustment and unhappiness. There is

increasing evidence that childhood behavior disorders, if

left untreated, are associated with impaired functioning in

later life (Quay, 1986). As adults, not only are children

who exhibit disruptive behaviors more likely to engage in

more serious antisocial behavior, but they are at increased

risk for psychiatric impairment, poor occupational

adjustment, low educational attainment, marital

distress/disruption, less social participation, and poor

physical health (McMahon & Wells, 1989). Conduct disordered

children are also at increased risk for substance abuse,

chronic unemployment, dependence on welfare services, and

generalized patterns of reduced attainment and competence,

and the behaviors that characterize children with a conduct

disorder are powerful predictors of subsequent delinquency

and criminal offense (Reid, 1993). Kazdin (1987) has noted

that the diversity of behaviors and range of dysfunction

associated with the disruptive behavior disorders present "a

remarkable challenge for treatment."

Treatment of Conduct Disordered Behavior in Children

An array of treatments of disruptive behavior disorders

are available and are often classified according to who is

actually targeted for behavior change (child, child and

family, community, etc.), by the level of cognitive

functioning or age of the child, and/or by behavior problem

type (overt, covert, mixed). Available forms of treatment

include individual, group, family and behavior therapies,

early educational interventions, problem-solving skills

training, cognitive and social skills training, parent

training programs, residential treatment, and

pharmacotherapy (Kazdin, 1987; Dumas, 1989).

Parent Training Programs

Studies of family-based interventions with pre-

adolescent behavior problem children (i.e. parent training)

comprise the largest and most sophisticated body of

treatment research in this area and have presented the most

promising results (O'Dell, 1985; Kazdin, 1987; McMahon &

Wells, 1989; Dumas, 1989; Reid, 1993). Parent training

programs vary on dimensions such as treatment

characteristics, length, format (individual, group),

location (clinic, school, home), and extent of therapist

training (Dumas, 1989), but the basic parent training model

is designed to alter the patterns of interchanges between

parent and child so that prosocial rather than coercive

behavior is directly reinforced within the family.

Positive Outcomes in Parent Training

Investigations of the basic format of parent training

programs and various types of generalization have supported

the effectiveness of behavioral parent training programs

with noncompliant young children (Kazdin, 1987; Dishion &

Patterson, 1992). Controlled studies of the short-term

efficacy of parent training for the overt problem behaviors

of pre-adolescents have demonstrated improvements in child

and parent behaviors over the course of treatment. Parents

have been shown to decrease their overall use of commands

and questions, make their commands clearer and more direct,

and increase their use of rewards and contingent praise

after treatment (Forehand & McMahon, 1981). Following

treatment, parental attitudes toward their children have

improved (Eyberg & Matarazzo, 1980) and parents have been

found to be less depressed (Griest & Wells, 1983).

Additionally, increases in child compliance and decreases in

child deviant behaviors have been shown to accompany parent

changes (Eyberg & Robinson, 1982; Webster-Stratton,

Hollinsworth, & Kolpacoff, 1989; Eisenstadt, Eyberg, McNeil,

Newcomb, & Funderburk, 1993).

Eyberg and Ross (1978) demonstrated that problem

behaviors of these children can be brought to within normal

limits following treatment. Some evidence of the

generalization of treatment effects may also be seen for

nontargeted behaviors and for siblings of treated children

(Eyberg & Robinson, 1982; Webster-Stratton, 1990b). McNeil,

Eyberg, Eisenstadt, Newcomb, and Funderburk (1991)

demonstrated that treatment effects generalized to lower

rates of behavior problems in treated children's preschool

classrooms, and treatment gains have been noted in hospital

settings (Kazdin Esveldt-Dawson, French, & Unis, 1987) as

well as multiple settings (home, school, and peer settings)

for the same child (Chamberlain, 1990).

A recent review of treatment outcomes at the Oregon

Social Learning Center (OSLC) further supports the focus on

family management to remediate antisocial behavior

throughout early and middle childhood (Dishion & Patterson,

1992). The sample consisted of all families of preadolescent

children referred to the OSLC outpatient clinic between 1965

and 1978 for treatment of antisocial or disruptive behavior

problems, who completed treatment (n = 73 of 75 referrals).

Specifically, the research examined the extent to which

clinically significant improvement in parent training varied

as a function of children's ages (early childhood, 2.5- to

6.5-years-old, vs. middle childhood, 6.6- to 12.5-years-

old). Analyses revealed a statistically reliable effect for

pre- to posttreatment improvement of observed aversive

behavior and a general effect for age of the child. In this

sample, 63% of the distressed children 6-years-old and

younger showed clinically significant change (defined as a

reduction in a targeted behavior of at least 30% of the

baseline level), compared to 54.3% of the distressed

children 6-years-old and older, a surprising finding in

light of facts suggesting that the severity and

pervasiveness of both child and family problems increase

over time if left untreated. These data suggest that parent

training is effective with both younger and older behavior

problem children, although preliminary analyses revealed

that dropping out of treatment occurred more frequently

among families with older children.

Positive treatment effects have been obtained for

videotape self-training, and enhanced effects have been

demonstrated for videotape self-training combined with group

discussion with other parents (Webster-Stratton,

Hollinsworth, & Kolpacoff, 1989) and for videotape self-

training combined with therapist consultation (Webster-

Stratton, 1990b). Recently, positive results have extended

findings with the basic 12- or 13-week videotaped parenting

skills training program to include an additional 14-week

treatment component ("ADVANCE") which trains parents to cope

with interpersonal distress through improved communication,

problem solving, and self-control skills (Webster-Stratton,


Temporal Generalization of Parent Training
with Preadolescent Children

Investigations of the temporal generalization of parent

training efforts also support the effectiveness of

behavioral parent training programs with noncompliant young

children within many families (Kazdin, 1987), and long-term

treatment outcomes have been reported for follow-up periods

ranging from 6 weeks posttreatment (Eisenstadt et al., 1993)

to 14 years posttreatment (Long, Forehand, Wierson, &

Morgan, 1994). Long-term follow-up studies from three major

outpatient parent training programs will be reviewed,

including the work of Patterson, Webster-Stratton, and


Patterson's training program for families of children

3-12 years of age was summarized in Reid, 1978, and

behaviors were coded using the Family Interactional Coding

System (FICS, Patterson, Ray, Shaw, & Cobb, 1969).

Patterson's approach was used to treat 13 children who were

consecutively referred for conduct problems, and it was

found that 9 of the 13 families demonstrated improvements

greater than or equal to a 30% reduction from baseline

levels of observed problematic behaviors using behavioral

observation data from the FICS (Patterson, Cobb, & Ray,

1973). Additional evidence for the reduction of problematic

behaviors has been reported for periods of up to 1 year

posttreatment (Patterson, 1974; Patterson & Fleischman,

1979). Using a more standardized set of treatment procedures

and more standardized termination criteria that evolved over

time, results have reportedly been replicated (Fleischman,

1981; Fleischman & Szykula, 1981).

Patterson (1974) reported immediate and 1-year follow-

up treatment results for a second sample of 5- to 12-year-

old boys (M = 8.4, SD = 2.0) exhibiting conduct problems in

the home (n = 27) and, for a subset of the boys (n = 14), at

school. Adults in each setting were trained to alter the

child's deviant behaviors. Parents' descriptions and reports

from other agencies were used to verify that all of these

children had extensive, long-term histories of behavior

problems, and 5 children in the sample had been diagnosed as

brain damaged, or hyperkinetic. Standardized screening

instruments were not used for inclusion in or exclusion from

this study.

Changes in child deviant behaviors in the home were

assessed in two ways. First, target child behaviors in the

home were assessed by means of direct observations yielding

rates-per-minute of 14 noxious responses ("total deviant

behavior"). Second, beginning with Case 24, outcome was

assessed by parents' daily report of symptom occurrences in

the form of a list of problem behaviors, specific to each

individual child, that was updated at each assessment by

parents' reports to therapists of current problems.

Beginning with case 32, this parent report criterion was

administered on each occasion that observers went to the

home during baseline, treatment, and follow-up. Modest

convergence was found between these two measures of outcome.

Bi-weekly reliability checks were performed in the home, and

reliability was calculated using an event-by-event analysis

which required agreement for category and sequence of

behaviors and yielded an average agreement of 74.2%.

The school observational code contained 19 behaviors of

the target child and 19 reactions of peers and/or the

teacher, and yielded an "appropriate behavior" summary

score. Observations of school behavior sampled group and

individual academic work of the target child as well as 6

randomly selected peers. For school observations, data were

collected in 21 sessions for 11 subjects and peers, and

produced 84.5% agreement for coding subject behavior.

Baseline observations were extensive: 6 to 10

observations were conducted in the home prior to treatment,

and a minimum of 5 observations were conducted in the

classroom prior to treatment. Observations were made

following the parents' reading of the programmed text, at

weeks 4 and 8, at termination, and at monthly intervals for

6 months follow by semi-monthly intervals for 6 additional

months, yielding a total of one year of follow-up data.

Treatment procedures were reported to have been

moderately successful in producing reductions of troublesome

behaviors; children's deviant behavior means fell within the

normal range at termination, and improvement rates in

children's behaviors persisted at follow-up. Comparing

baseline observational data to the last data point available

showed in both settings that approximately 2 out of 3 cases

were statistically and clinically improved ("reductions of

30% or more from the baseline level"), and behavioral

changes were also accompanied by changes in parents'

perceptions among unspecified subgroups of this sample,

using bi-polar adjective checklists.

Although these results seem favorable, several

problematic points were reported. During the month following

termination, half of the families required "booster

sessions" (an average of 1.9 hours of additional therapy)

due to increases in noxious behaviors. Data sets for all of

the analyses appear to have been incomplete (observational

data through follow-up were available for only 20 families,

and parent daily report was available for only 14 families).

Thirty-seven percent (10 of 27 families) did not participate

in follow-up; finally, due to a lack of standardized

screening procedures, the sample was found to include some

boys who had been mislabeled and in fact showed no problem

behaviors (p. 479). The parent report criterion was

apparently useful for tracking progress within individual

families; however, its usefulness as a group treatment

outcome criterion is limited in that it did not represent a

standardized measure of problem behaviors, making

comparisons across children problematic. Further, the

dataset from this outcome criterion was incomplete because

the method was developed during the study.

Several factors therefore affect conclusions regarding

the reported outcome (2/3 of sample improved), including

which cases were included in statistical analyses. The

conclusions reported above were based on all families

completing at least 4 weeks of treatment, but when data were

re-analyzed (Fleischman, 1981) with only the families

completing treatment and cooperating with follow-up, the

"overall effectiveness and durability of treatment was

questionable" in that statistically significant F tests were

not obtained (ps < .15), although planned t-tests of

baseline to posttreatment and baseline to follow-up data

were significant at the .05 level. The rate of follow-up

attrition (37%) in this study seems relatively high for a

12-month follow-up period, but differences in families who

did and did not participate in this follow-up apparently

were not evaluated and the possible effects on the results

are therefore unknown. Other sources of variability that may

have obscured treatment effects include the lack of

standardization of treatment length (M = 26.9, SD = 22

sessions, for an average of 31.2 hours of therapy) or

setting (some sessions were conducted in the clinic, while

others were conducted at home), "subject variability" (e.g.,

some families had more than one treated child, some children

in the sample had apparently been previously treated by

other community agencies, sample included 5 children who had

been diagnosed as "brain damaged or hyperkinetic," and

observational data revealed that some of the boys "seemed to

be mislabeled in that the data showed no behavior

problems"), and variability introduced through assessment

procedures. The first of the two outcome criteria,

behavioral observations, was conducted with all other family

members present and thus potentially affecting interactions

under observation, but parent observational data were not

reported in this study. A second outcome criterion, parents'

report of symptom, were available only on the later cases

(50% of sample assessed on behavioral observations only),

and each parent reported on an individualized list developed

separately for each treatment child, thus limiting

comparisons across children. Finally, classroom observations

were unstandardized across children, occurring in non-

classroom settings for some but not all of the children. The

classroom outcome measure, proportion of appropriate

behavior, was described as "a minimally reliable measure of

performance skills relevant for adjustment in the classroom

setting," and it allowed only inferences about changes in

the targeted, noxious behaviors that brought the families

into treatment. Thus, the conclusion that 2\3 of the

sample showed improvements relative to baseline levels were

based on only one outcome criterion (child deviant scores),

and the claim of successful treatment is also weakened by

the fact that roughly 50% of the sample requested booster

sessions within one month of termination, although repeated

measures analyses on follow-up data (rates per minute of

child deviant behaviors) for 20 families who had relatively

complete follow-up data did reveal significant improvement

over baseline levels. Unstandardized parent symptom report

data showed a nonsignificant trend toward improvement over

follow-up. Fleischman's 1981 re-analysis reported an average

reduction in total child aversive behavior from baseline to

termination of 43% for this sample.

An additional index of treatment outcome was defined as

the mean baseline total deviant score, minus the mean at

termination, divided by the baseline mean. For families on

whom data were available in this sample, negative treatment

outcome using this index was associated with lower social

class, with mothers' single-parent status, and with mothers'

F, K, and Hypomania scores on the MMPI. The "most difficult

case" to treat was reportedly the welfare mother, living

alone with her children, who perceived herself unable to

cope with the myriad of crises impinging upon her.

A later sample of 35 children (girls and boys) ages 3

to 12 (X = 7.6, SD = 2.9) who were displaying aggressive

and/or antisocial behavior were treated with the Patterson

assessment and treatment format through an autonomous

clinical facility in an attempt to fully replicate

Patterson's 1974 results (Fleischman, 1981). Assessment

again consisted of naturalistic observations in the home

(mean interrater reliability 75.8%), individualized lists of

behavior problems that parents reported on during semi-

weekly phone calls, parental attitudinal descriptions of the

child using the Becker Bi-Polar Adjective Checklist, and

parents' own observational data on two positive and two

problem behaviors (Fleischman, 1981). On the observational

measure of total aversive behavior and on parents' symptom

report, using outcome data for families completing at least

6 weeks of treatment as well as for families completing all

of treatment, statistically significant results were

reported for the replication sample, thus leading the author

to conclude that it was possible to replicate Patterson's

work in applying a social learning-based treatment to

families with conduct problem children.

For families in the replication study completing at

least 6 weeks of therapy, Fleischman reported an average

reduction in total child aversive behavior from baseline to

termination of 44% (p < .01), but given the methodology, a

basic question remains as to the significance of these

changes: who were these families and how difficult were the

problem behaviors to begin with? Unfortunately, it appears

that pretreatment screening and quantification of presenting

problems and exclusionary criteria were still not rigorous:

children were accepted into treatment if presenting problems

centered on the child's aggressive or antisocial behavior

and if neither the child nor his or her parents appeared

retarded, physically handicapped, or autistic. The

pretreatment severity and pervasiveness of the problems were

not assessed with standardized measures, and the sample was

again inadequately described. Thus, although the 44%

reduction in total child aversive behavior from pre- to

posttreatment was obtained on observations in the home, it

is not clear to whom these results might be generalized.

The Patterson approach to treating conduct problem

behaviors, as reported in the two studies reviewed thus far,

by definition was tailored to fit the clinical needs of

individual families, but unstandardized treatment procedures

resulted in variability in both the content and process of

therapy and problems for clear evaluation of outcome. In

the Fleischman study, there was less variation in the length

of therapy across subjects as compared to the 1974 study,

with families receiving a mean of 14.7 weeks (SD = 7.6) or

an average 13.7 hours of therapy. However, the treatment

format in the replication study varied across families, with

7 of the 35 families receiving group treatment, and the rest

receiving individualized treatment. Therapy was again open-

ended, parents were reported to follow roughly equivalent

programs and assignments, occasionally being helped with

problems other than childhood aggression, such as

bedwetting, excessive crying, and helplessness. Even

termination status was somewhat unclear: some families

terminated with their therapist's approval, some initiated

termination against therapist's advice, and for some

families, the on-going level of cooperation with the demands

of treatment represented a de facto termination. Data from a

particular child were included in statistical analyses if

the family completed at least 6 weeks of therapy (compared

to 4 weeks in the 1974 study), with 5 weeks being the

minimum amount of time necessary for parents to be taught

the basic program components. Therefore, even if unambiguous

results were obtained with this design, the variability in

what each family actually experienced as treatment makes it

difficult to identify what a successful (or unsuccessful)

outcome could be attributed to.

Additional confusion in interpreting the meaning of the

results (again) stems from the different numbers of families

completing treatment, participating in follow-ups, and

completing measures (e.g., 31 out of 52 parents completed

the adjective checklist at baseline and termination, and 23

at 12-month follow-up; who did not complete this measure,

and why?). Twenty of 35 families thus completed treatment,

and 16 subjects participated in at least two of the three

observations conducted during the year posttreatment, and

some families completed some measures. First, it is not

clear which families were thus included in the pre- to

follow-up outcome analyses, and second, systematic

investigation of possible differences in families who (1)

did and did not finish treatment and (2) did and did not

cooperate with follow-up were not reported. Thus questions

remain as to who completed an instrument or an assessment,

why or why not, and how might this have affected the


Fleischman and Szykula (1981) further tested the OSLC

treatment program in a community-based treatment facility in

Helena, Montana, using therapists not directly affiliated

with OSLC. Target children were 62 children (64% males) from

3 to 12 years old exhibiting conduct problems and/or

aggressive behaviors at home and/or at school (50% of

sample). Pretreatment quantification of presenting problems

was not reported. Outcome measures included naturalistic

home observations using the FICS (total aversive child

behaviors), with a mean interrater agreement of 76.4%,

parents' standardized reports by means of a 24-hour

telephone recall of five positive and ten negative child

behaviors, and parents' global ratings of change on 2 5-

point Lickert scale questions. Observational data were

available for 23 of 62 children within 18 of 50 families;

observations were not obtained for families who lived too

far away (6), who refused or were too difficult to schedule

(7), or whose target child's behavior mean observed problem

behavior was not outside of normal limits at baseline

observations (X = .306 deviant behaviors per minute for

nonreferred children, SD = .438; .45 being the maximally

discriminating score for referred and nonreferred children;

Patterson, 1974). Given that a number of families refused

observations, rates per minute of total child aversive

behaviors did drop from pretreatment (roughly .70) to

termination (approximately .24), increased at 4-month

follow-up (.37), and declined from pretreatment levels

throughout the rest of the follow-up period (.20 at 12 month

evaluation) for those children on whom data were available

(ps < .001). Statistically significant changes (pre- to

termination, and pre- to 4rth month, 8th month, and 1 year)

in the desired directions were also obtained for both

positive and negative behaviors on parent report for those

families who completed follow-up assessments (ps < .001). At

termination, global ratings were obtained from 31 mothers

and 15 fathers, out of a total of 50 families completing

treatment, whose average rating indicated their child's

behavior was improved and family functioning as somewhat

improved. At one year follow-up, 9 mothers and 5 fathers

rated the program on the two 5-point Lickert scale

questions, and while there was a modest decline in mean

ratings, 14 of these 15 parents continued to report a

beneficial effect of treatment; 1 father reported no change.

Again, these generally favorable results must be

interpreted cautiously because of problems with attrition

and incomplete data sets; although 9 of the 16 subjects who

left follow-up did so only because they moved away from

Helena, it follows that 7 other families refused. These

authors did not report analyses investigating differences in

families who did and did not participate in follow-up.

Additionally, although behavioral observations of

children's deviant behaviors were measured at several points

from baseline through follow-up (event-by event reliability

checks required agreement on both category and sequence;

reliabilities generally above .80), parents' actual

behaviors were not reported, and the clinical significance

of observed changes in outcome measures was not reported in

the Fleischman (1981) or Fleischman & Szykula (1981)

replication studies.

Home and Van Dyke (1983) examined 56 families with an

aggressive target child (mean age = 8.46, standard deviation

= 2.30) who had been referred to the Oregon Social Learning

Center for treatment to determine whether treatment would

reduce aggressive child behaviors and whether reductions

would be maintained for a year. An intake interview and

observations using the FICS were used to gather descriptive

baseline measures and a trained observer coded alternately

and sequentially the behavior of the targeted subject and

then each family member with whom the subject interacted

(specific reliabilities not reported for this study). Child,

mother, and sibling aggressive behavior rates were measured

at baseline, termination (mean length of treatment = 22

sessions, range = 3 to 85), and 1-year follow-up.

Of 119 families who came for treatment, 56 provided

complete baseline data, remained in treatment three or more

sessions, and provided follow-up data after 1 year. Analyses

of the total aversive behavior scores from these 56 families

(persisters), who were of lower occupational status than

nonpersisters (defined as terminating treatment within three

weeks or less), indicated statistically significant

differences (ps < .01) between baseline and termination

means, and between baseline and follow-up means, for target

child, mother, and sibling aggressive behavior rates. No

significant differences were found for termination to 12

month follow-up rates of aggression; target child and mother

aggression rates did increase slightly over this period, but

sibling aggressive behavior rates continued to decrease.

While these results are promising, in view of the lower

overall occupational level of the sample of persisters which

would be expected to result in less favorable outcomes, it

is again difficult to draw a conclusion about treatment,

given that treatment referred to a process that spanned 3 to

85 sessions. Sources of error due to variability introduced

in the assessment process used in this set of Patterson

studies have already been noted, but in addition, the

results for this study were based upon only one outcome


In summary, the social learning based treatment

approach to child behavior problems, as described by

Patterson (1974), has been shown to produce statistically

significant changes in child deviant behaviors and parents'

reports of problem behaviors immediately after treatment and

up to one year posttreatment for up to two-thirds of treated

families on some but not all outcome measures, with patterns

of improvement maintained up to one year for some families

on some measures (inferred, because outcomes for specific

families across measures were not described). Strengths of

one or more of these outcome studies included multiple

method outcome criteria (parent report and observational

data) for combined groups of boys and girls, for mothers and

siblings, and results of some studies (e.g., Patterson,

1974; Patterson & Fleischman, 1981;) have been reported in

terms of both statistical and clinical significance.

Problems have included the lack of a no-treatment control

group in these follow-up studies, unstandardized treatment

formats which may result in widely differing lengths of

contact (e.g., Home & Van Dyke, 1983), and follow-up drop-

out rates of up to one-third or more of families which have

not been systematically investigated as to possible effects

on conclusions about outcome. In addition, the wide age

ranges of the children in these samples (typically 3 to 12

years old), combined with unquantified and somewhat variable

types and levels of presenting problems, may have masked the

effects of this treatment approach with conduct problem

children due to subject variability. In fact, higher success

rates have been reported for the treatment of younger

behavior problem children (63% for 3 1/2 to 6-year old

children) as compared to older behavior problem children

(27% for 6 1/2 to 12-year old children) (Dishion &

Patterson, 1992). Finally, there have been no long-term

follow-up studies of this particular treatment program

beyond 1 year.

A second line of programmatic research on treatment of

conduct problem children and their families has directly

addressed the cost-effectiveness of parent training

treatment programs by systematically investigating the

effectiveness of videotaped modeling of parenting skills,

group treatment formats, and adjunctive procedures designed

to address parental and marital distress. Webster-Stratton

has evaluated the immediate and longer-term effects of

parent training approaches which used videotaped training

sessions and group discussion. The first study (1981a,

1981b) of this performance-based training program based on

videotaped modeling and group discussion led to significant

changes in mother-child interactions immediately

posttreatment for a treatment group (early treatment) as

compared to a control group (received delayed treatment 6

weeks later than the first group). At the end of treatment,

early treatment group mothers rated their children's problem

behaviors as significantly less intense than delayed

treatment mothers did. Thirty-two of the 35 mothers and

their preschool children were re-evaluated one year after

treatment with videotaped observations of mother-child

observations (Interpersonal Behavior Construct Scale; IBCS)

and with parent-completed measures including the Eyberg

Child Behavior Inventory (ECBI), Parent Attitude Survey

(PAT), and a Consumer Satisfaction Survey (Webster-Stratton,

1982). Original participants were 35 lower to upper middle

class mothers with an average of four years of college

education who had been recruited by flyer for a parent

training program designed to help manage child misbehavior.

Children in this sample were 23 boys and 12 girls, 3 to 5

years old; 62% of the group were within normal limits on the

ECBI, and 38% of the group had ECBI scores more than one

standard deviation above the mean. Subjects were randomly

assigned to an early treatment group (n = 16) or a delayed

treatment group (n = 19), which began treatment

approximately 6 to 7 weeks later, after the first group

finished treatment.

Significant pre- to posttreatment reductions in ECBI

intensity scores (p < .02) were obtained from early

treatment mothers after 4 weeks of treatment, but not from

delayed treatment mothers who had not yet received treatment

and served as a control group. Both treatment groups

achieved clinically significant mean decreases (reductions

of 30% or more of baseline) in numbers of behaviors parents

considered problematic (ECBI problem scores) at their

respective posttreatment assessments. Immediately after

receiving treatment, early treatment mothers showed

significantly fewer leadtaking behaviors, dominance

behaviors, and nonacceptance behaviors, and significantly

increased positive affect behaviors as compared to delayed

treatment mothers who had not yet been through the program.

At their posttreatment assessment, early treatment children

showed significantly fewer negative affect and submissive

behaviors, and significantly increased positive affect

behaviors, compared to delayed treatment (wait-list control)


For the 1 year follow-up, 32 of the 35 mother child

dyads were retested on all measures, and since no

significant differences were found between the two groups on

any measures immediately following their respective

treatments, data from the two groups were combined for

follow-up analyses. Mother attitudes at 1 year follow-up

showed a trend toward increased scores over the year

(lowered confidence in ability to manage child's behavior)

for the combined group. The number and intensity of behavior

problems had decreased significantly (p < .01) over the

course of treatment, but the total number of behavior

problems showed a nonsignificant increase from immediate

posttreatment to 1-year follow-up. All the significant

behavioral changes reported for both mothers and children

immediately posttreatment were maintained at 1-year follow-

up except for positive affect behaviors, which dropped

significantly for both mothers and children. Child

nonacceptance and dominance behaviors, which had not changed

by postreatment, showed a significant drop at follow-up

compared to the early assessment, and mothers' nonacceptance

behaviors continued to decrease significantly from immediate

posttreatment to follow-up. Although child behavior problems

as defined by ECBI problem and intensity scores were within

normal limits for 2/3 of this sample at baseline,

statistically significant improvements did occur over the

course of treatment in mother/child interactions and in

maternal perceptions of child behaviors in this sample of

nonreferred mothers. At 1 year follow-up, most behavioral

changes were maintained or continued to improve, although a

slight decrease was found in mothers' confidence in

parenting skills and ability to manage their children, and

96% of the sample stated they would like more classes and

would attend more classes if they were offered again.

Although these behavioral data are promising, the results

may not be representative of outcomes with a clinical

sample, or a sample of parents with less education and fewer


In a second study of this program for which follow-up

data were gathered (Webster-Stratton, 1984), referred

mothers of 35 conduct disordered children were randomly

assigned to a 9 week video modeling therapist-led group

discussion program (VMGD), an individual treatment parent

training program (IPT), or a wait-list control group (CON).

This study included 25 boys and 10 girls between the ages of

3 and 8, mean age = 4 years 8 months, with the primary

referral problem of oppositional behaviors as described by

parents in a preliminary screening phone call and subsequent

interview Nineteen of the 35 children (54%) were from

father-absent families of predominantly lower middle to

lower class standing. Fifteen of the families reported that

they had had recent contact with child protective services

because of child abuse reports. Treatment content and

sequencing were kept as similar as possible across the two

groups, although the VMGD received slightly fewer sessions

(M = 8.5, 1.3) than the IPT group (M = 9.1, .60). VMGD

families had a cotherapist team consisting of a male and

female therapist; IPT families were assisted by either the

male or the female therapist. Assessments were conducted

using (1) parent report measures including the Child

Behavior Checklist (CBCL), the ECBI, and individually

tailored, shortened (unstandardized) Parent Daily Reports

(PDR's) administered in biweekly phone calls from intake to

posttreatment assessment, (2) observational measures of

parent-child interactions using DPICS (home observations),

(3) a teacher report measure of preschool children's

behaviors (Behar Preschool Questionnaire, PBQ), and (4) a

consumer satisfaction questionnaire. Although only mother-

problem child interactions were analyzed, observations were

conducted with all family members present.

Mothers were assessed at baseline, three months after

baseline, and 31 of 35 subjects were again tested on most

measures 1 year later. For each dependent variable, the

Dunn-Bonferroni tables were used to determine critical

values in order to correct for the number of individual

comparisons. It is reported that positive changes were found

in parent attitude measures as well as behavioral data for

the combined treatment groups compared to the control

condition at posttreatment. The combined treatment groups

obtained significant change on six out of six attitudinal

variables; IT mothers obtained significant changes on four

out of six attitudinal variables, VMGD mothers had obtained

significant changes on five out of six attitudinal


Comparisons of pre- to posttreatment observations

indicates that the mothers in the combined treatment groups

decreased the total number of commands, no-opportunity

commands, and critical statements, and increased their use

of praise significantly compared to the wait-list control

mothers. Compared to the control group, IT mothers showed

improvements on two out of five behavioral variables (total

commands and total praise) at posttreatment, and VMGD

mothers showed significant improvements on four out of five

variables (total commands, no-opportunity commands, critical

statements, and praise) at posttreatment. Neither group

obtained significant postreatment increases in the

proportion of direct commands given compared to the control

group, although a direct comparison of the two treatment

groups showed that mothers in the IPT group demonstrated a

significantly higher direct command ratio than the VMGD

mothers, and this was the only significant posttreatment

behavior difference for the mothers in the two treatment


For child behaviors at posttreatment, total child

noncompliance (child's response to all types of commands)

was significantly decreased (E < .05) for the combined

treatment groups children as compared to the control group

children, however, posttreatment improvements in the

proportion of noncompliance to parents' commands (the

noncompliance ratio) and in child deviant behaviors were of

borderline significance (ps < .10) for the combined

treatment group children compared to the control group

children. Posttreatment patterns of change within each group

(VMGD, IPT) of treated children mirrored results for the

combined treatment groups sample: each group showed

significant reductions in total noncompliance, but only

nonsignificant reductions in the noncompliance ratio

(proportion of child's responses to parents' commands that

are noncompliant) and in the composite measure of child

deviant behaviors.

At 1-year follow-up, 97% (31 of 32) families were re-

evaluated using the same measures, and mothers in the two

treatment programs did not differ significantly at 1-year

follow-up on any of the parent attitudinal measures (ECBI,

CBCL, PDR), the teacher report measure (PBQ), or any mother

or child observational variables. Significant improvements

(ps < .001) had occurred on all three parent report measures

(ECBI, CBCL, PDR) for both treatment groups compared to

baseline levels, and no significant changes had occurred in

these measures at follow-up compared to posttreatment


On DPICS home observational measures, there were no

significant changes in mothers' behaviors from posttreatment

to 1-year follow-up for either treatment group, indicating

that the significant improvements in mother's behaviors had

been maintained. One year after treatment, both sets of

mothers had significantly decreased critical statements and

increased use of praise compared to their own pretreatment


For the combined treatment groups, comparing

pretreatment to 1-year follow-up, mothers in both groups

reduced their total number of commands, but not at a

statistically significant level. Although mothers in the IPT

group had demonstrated a significantly higher direct command

ratio than the VMGD mothers at posttreatment, the only

significant posttreatment behavior difference for the

mothers in the two treatment groups, by 1-year follow-up,

the specific type of command given was not significantly

reduced from pretreatment levels for either group.

Interestingly, although mothers in both treatment

groups apparently demonstrated increases in their use of no-

opportunity and indirect commands by 1-year follow-up, both

sets of children demonstrated significantly lower

frequencies of noncompliance, lower proportions of

noncompliant responses to mothers' commands, and fewer

deviant responses compared to their own pretreatment levels

at the 1-year follow-up. In fact, children's noncompliance

ratios and deviant behaviors, which had not shown

significant decreases at posttreatment, did continue to

decrease over the year of follow-up, an important finding in

itself. These data suggest that in this study, children's

compliance at 1-year follow-up, was not simply a function of

parents' use of commands.

Parents in both treatment groups had acquired new

parenting skills and both sets of parents continued to

report being very satisfied with treatment at 1-year

follow-up. No significant differences in treatment

satisfaction were found between groups, and no significant

changes in satisfaction occurred during the follow-up

interval. Three IPT mothers and one VMGD mother still

perceived problems and felt the need for further therapy.

The author noted that the group studied represented a "high

risk" sample due to the high number of single parents, the

low socioeconomic status, the low mean educational level,

the high prevalence of reported or alleged child abuse, and

the deviant nature of the child (pretreatment CBCL behavior

problem scores exceeded the 90th percentile, and ECBI

Problem and Intensity scores also exceeded clinical

cutoffs). According to the author, the two major findings

were (a) that successful behavior changes occurred following

treatment for both therapy formats, which were maintained 1

year later and (b), that few differences between groups were

found on parent report or observational measures at

posttreatment or 1-year follow-up. This finding is more

striking in view of the fact that parents in the videotape-

modeling group-discussion group did not receive any direct

feedback on their interactions with their children, nor did

they overtly rehearse the modeled skills during treatment

sessions. An additional finding, however, is that neither

group of mothers obtained statistically significant change

at posttreatment in the use of direct commands, and neither

group maintained statistically significant changes in the

reduction of total commands or in the type of commands used

at 1- year follow-up.

The next study in this series compared self-

administered videotape modeling (IVM), self-administered

videotape modeling and group discussion (VMGD), group

discussion alone (GD), and a waiting-list control condition

(Webster-Stratton, Kolpacoff, & Hollinsworth, 1988) in

parents of 114 conduct problem children ages 3 to 8 (mean

age = 4.5 years), referred primarily for misconduct of at

least 6 month's duration, with mean ECBI problem scores in

the clinical range (M = 21.3, SD = 6.2). Thirty percent of

the parents were single, yearly income for this sample

ranged from welfare level to above $29,000, and the sample

was comprised of a wide range of social classes. For the 114

families entered into the trial, interview data indicated

that 31% of the mothers had reportedly experienced spouse

abuse and 39% of the families had reported alcoholism or

drug abuse in the immediate family. Twenty percent of the

mothers and 24% of the fathers reported that they were

abused as children, and 13% of the mothers reported prior

involvement with Child Protective Services. Thirty-one

percent of the mothers reported some symptoms of depression

(scores > 10) on the Beck Depression Inventory. Thus, the

group that entered this study also constitutes a sample at

high risk for problems in parenting.

Assessments at pretreatment and 1 month after treatment

included parent's and teachers' perceptions of child

adjustment (CBCL, ECBI, individually tailored PDR's, PBQ

Total Behavior Problem Scale), parents' personal adjustment

on the Parenting Stress Index (PSI) Parent Domain score,

home observations using the DPICS coding system behaviors

that have been shown to discriminate clinic from non-clinic

families (p. 560)(total no-opportunity commands, total

praise, total criticisms, parent affect, and total child

deviant behaviors, which included noncompliance), and a

social validity measure (Consumer Satisfaction

Questionnaire). Home observations, imposing as little

structure as possible, were made in the evenings with family

members requested to do what they would normally do.

Reliability was assessed on 30% of observations, and mean

interrater reliability was 79% (range = 71%-89%).

Training content, sequencing, and duration were held

constant across the groups, as were therapist experience and

training, but VMGD and GD weekly sessions lasted two hours

and IVM weekly sessions lasted approximately 1 hour. Of the

85 mothers and 59 fathers assigned to one of the groups, 78

mothers (91.7%) and 52 fathers (88.1%) completed greater

than 50% of treatment sessions and posttreatment

assessments. The group discussion treatment group had the

highest rate of missed sessions and drop out. After 10 to 12

weeks of treatment focusing on interactional skills,

nonpunitive discipline approaches, and specific operant

techniques, all three treatment groups reported fewer child

behavior problems and increases in child prosocial behaviors

as compared to the control group. Compared with control

parents, mothers and fathers from all three treatment groups

exhibited significant interactional changes with their

children on at least two of the four behavior variables (no-

opportunity commands, praise, criticisms, affect). Children

showed significant reductions in deviant behaviors (a

summary variable that included noncompliance) when

interacting with fathers in all three treatment groups and

with their mothers in the VMGD and GD groups. The VMGD

treatment was the only treatment that resulted in

statistically significant reductions of mothers' parenting-

related stress and father reports of the intensity of child

behavior problems, and this treatment increased both

mothers' and fathers' praise statements. On the consumer

satisfaction measure, the mothers in the VMGD group reported

the highest satisfaction with child behavior improvements as

well as the highest usefulness ratings for the format and

content of treatment.

One year posttreatment, 93% of families (94 mothers and

60 fathers) in the original study were reassessed using the

same measures (Webster-Stratton, Hollinsworth, & Kolpacoff,

1989). Follow-up attrition was equal or nearly equal across

treatment groups (6 VMGD, 5 GD, and 5 IVM families) and

there were no significant differences in demographic data,

pretreatment variables, or posttreatment outcome variables

between families who dropped out of follow-up and the rest

of the sample. As in the 1988 study, DPICS home observation

summary variables were total praise, total critical

statements, total no-opportunity commands, and a combined

total of child deviant and noncompliant behaviors. Parent

nonverbal affect was again rated by coders every 5 minutes

on a 5-point scale ranging from unrestrained negative affect

(5) to neutral affect (3) to exuberant affect (1).

Reliability checks were again performed on at least 30% of

all home observations, and mean overall interrater

reliability was 80% (range 70%-92%).

One year follow-up data, compared to pretreatment data,

indicated that mothers and fathers in all three treatment

groups changed in the same way compared to pretreatment,

continued to report significantly fewer behavior problems

(ECBI, CBCL, and PSI Child Domain score, PDR telephone

reports), and reported significantly increased child

prosocial behaviors. There were no significant changes from

immediate posttreatment results to 1-year follow-up results

for the VMGD or IVM parents' or children's behaviors (p <

.01), and the only statistically significant finding

reported was that GD fathers showed a reduction in no-

opportunity commands from immediate posttreatment to 1-year


While the VMGD group tended to be favored slightly in

the overall results in this study, an interesting finding is

that on average, mothers in all three treatment groups did

not maintain significant reductions in average levels of no-

opportunity commands at 1-year follow-up compared to their

pretreatment levels, although fathers in each of the three

treatment groups did. In addition, VMGD mothers' mean use of

no-opportunity commands increased (nonsignificantly) from

posttreatment to 1-year follow-up, but the average use of

no-opportunity commands by mothers in the other two groups

remained the same (IVM) or decreased nonsignificantly (GD).

Comparing posttreatment means with 1-year follow-up means

(Table 3), group means for total child deviant behaviors

exhibited with fathers decreased in all three treatment

groups, however group means for total child deviant

behaviors exhibited with mothers increased in the VMGD

children, remained the same in IVM children, and decreased

slightly in GD children. The reasons for differences in use

of no-opportunity commands 1 year after treatment for

mothers and father and for mothers in different treatment

formats are unclear, but patterns of change in children's

total deviant behaviors with their fathers and mothers

parallel patterns of change in fathers' and mothers' use of

no-opportunity commands and seem worth noting even if they

failed to reach statistical significance. Unfortunately,

child noncompliant behaviors were not reported separately

from other child deviant behaviors in the 1988 and 1989

comparative treatment studies so it is not possible to

determine whether parents' differential use of commands in

this sample may have been associated with changes in

children's deviant behavior in general, or with subsets of

behaviors, particularly those relating to compliance.

The lapse in mothers' use of no-opportunity commands

one year after treatment in this study are similar to

results of the 1984 Webster-Stratton study comparing

videotaped modeling-group discussion treatment (VMGD) to

individual parent training (IPT), in that mothers in both

1984 treatment groups also failed to maintained

statistically significant changes in the type of commands

used at 1 year follow-up, which included both direct and no-

opportunity commands. In contrast, children's noncompliance

ratios and deviant behaviors (whine, cry, yell, smart talk,

physical negative, and destructive), studied separately in

the 1984 study, that had not shown significant decreases at

posttreatment and actually decreased over the year of

follow-up for both groups in spite of mothers' decreased use

of direct commands and increased use of no-opportunity


Results of the 1989 study 1-year follow-up consumer

satisfaction scores showed that VMGD mothers perceived their

children as significantly more improved and perceived their

treatment program as significantly easier to implement than

IVM mothers did. GD mother consumer satisfaction scores fell

intermediate between the other two treatment group scores.

Similarly, VMGD fathers perceived their treatment program as

significantly easier to implement than IVM fathers did.

The authors of the 1989 study used four criteria to

define clinically significant change: (1) parents had to

report a CBCL T score of 63 (90th %) or lower, this score

being the cutoff point between normalcy and deviancy

(Achenbach & Edelbrock, 1983), (2) mother PDR reports had to

be reduced 50% of baseline levels and reports of spanking

stopped, (3) for the behavior outcome criterion (DPICS

variables), parent criticisms and physical negatives had to

be reduced 30% from baseline and total child deviance had to

be reduced 30% of baseline at posttreatment and 1-year

follow-up. Using these conservative criteria, the authors

noted that approximately two thirds of the sample showed

"clinically significant" improvements on either the parent

report or behavioral criteria at the immediate posttreatment

and 1-year follow-up assessments, although percentages

showing clinically significant change were not reported

separately for each of the three treatment groups.

At posttreatment, percentages of the overall sample,

combined across treatment groups, demonstrating clinically

significant changes were (1) 71% of fathers and 74% of

mothers on the CBCL Total Behavior Problem score, (2) 61% of

mothers reported a 50% reduction in specific target

behaviors on individually tailored PDR's, (3) 51% of mothers

and 61% of fathers showed a 30% reduction in critical and

physical negative behaviors, while 53% of children showed a

30% reduction in child deviant behaviors when interacting

with their mothers, and 57% of the sample showed a similar

improvement in interactions with fathers. Generally

speaking, behavior change was greater in parents than in

children at posttreatment. Although not presented,

percentages of the overall sample showing clinically

significant change on 1-year follow-up data were reported to

be similar to posttreatment percentages. Chi-square analyses

indicated that the percentages of the treatment groups

showing clinically significant improvements did not differ

significantly at either the immediate posttreatment

assessment or the 1-year follow-up (Webster-Stratton,

Hollinsworth, & Kolpacoff, 1989). It is not known how many

families at either assessment demonstrated clinically

significant changes on both parent report measures and

observational variables, or on both parent and child

observational variables.

In a subsequent follow-up assessment, a subset of the

original 1988 sample of families 83 (82%) mothers, and 51

(72.8%) fathers were also assessed three years later

(Webster-Stratton, 1990b). Separate attrition rates for each

of the three treatment groups were not given, but of the 37

parents lost to 3-year follow-up, 12 refused to participate

or simply did not return questionnaires. Families who

dropped out at the 3-year follow-up did not differ from

those who did participate with regard to demographic

variables, age of child, parent or teacher reports of child

adjustment, parenting stress, or life stressors at either

pre- or posttreatment. Comparisons of participants and non-

participants on pretreatment behavioral data or

posttreatment outcome variables were not presented.

The mean age of the children at 3-year follow-up was

7.5 years, with a range of 6 to 11 years. Fifty six percent

of the follow-up sample children were enrolled in first

grade, and the rest of the children were enrolled in grades

two through four. Specific measures used in this 3-year

follow-up study included the CBCL, the TRF, PBQ, the Beck

Depression Inventory, the Dyadic Adjustment Scale, a measure

of the marital adjustment of couples, and the Life Events

Survey. Unfortunately, observational indices of behavior

were not collected.

Results indicated that although parents from the

combined treatment group continued to report fewer child

behavior problems and increased prosocial behaviors at the

3-year follow-up compared to pretreatment levels, only

families in the group discussion-videotape modeling

treatment (VMGD) showed stable improvements according to

mothers' reports. Regarding the clinical significance of

changes, again for the combined sample, approximately 46%

of mothers, 25% of fathers, and 26% of teachers reported

that the children displayed behavior problems in the deviant

range three years later, and more than one third of the

families expressed ongoing concern about aggression,

noncompliance, peer problems, and hyperactivity.

Approximately 10% had received further therapy for

themselves or for their children during the 3 years and 15%-

19% of families requested further therapy at the time of

this follow-up assessment. About half of the children rated

by their preschool teachers as abnormal in the 1989 study

obtained similar ratings from grade school teachers in this


Using only the parent (CBCL) and teacher (TRF) Total

Behavior Problem scores, and whether or not families

continued to request additional therapy for their children's

behavior problems at 3-year follow-up, each treated child

was classified as a responder or non-responder.

Responder/nonresponder percentages were not given separately

for the three treatment groups, but 46% of children were

classified as nonresponders by mothers' CBCL scores, 25% by

fathers' CBCL scores, 26.3% by teacher TRF scores, and 20%

of families (17) requested more therapy for their children's

behavior problems three years after treatment. Children not

doing well at the 3-year follow-up as assessed by mothers'

CBCL scores (46% of the sample) were distinguished in this

sample by single-parent status, and/or increased alcoholism,

drug abuse, and depression among immediate family members.

The author noted that data also suggest that the parents of

nonresponder children had themselves experienced deprived

and nonurturing childhoods with painful memories of

alcoholism and drug abuse in their fathers and depression in

their mothers (p. 148). Children classified as nonresponders

by fathers' CBCL scores were from families with lower social

position scores, as indexed by the education and occupation

of the adult designated as the head of the household.

Children classified as nonresponder by teacher report were

more likely to come from families in which the mother was

single or divorced. Further analyses by Webster-Stratton of

the records of 218 treated families have since revealed that

for the families receiving the videotape modeling-group

discussion parent training (the most successful group), the

most powerful predictors of child deviance at long-term

follow-up were marital distress and lack of a supportive

partner (Webster-Stratton, 1994).

In general, Webster-Stratton's 1990 3-year follow-up

study lacks the rigor of the 1988 report of the immediate

outcome and the 1989 report of 1-year follow-up results.

This study failed to include behavioral observations,

relying instead on a much briefer and consequently less

thorough assessment of children's and families' functioning.

As a result, it is not as clear that 3-year maintenance (the

first purpose of the study) was demonstrated as was 1-year

maintenance for some families, nor is it immediately clear

from the data presented which families in which groups

demonstrated positive outcomes. This in itself is

unfortunate, since some differential 1-year maintenance was

noted for mothers' and fathers' use of no-opportunity

commands, nonsignificant group and group-by-parent gender

differential maintenance occurred for mean totals of child

deviant behaviors, and maintenance of parent behavioral

changes in the GD mothers was lower than maintenance of

behavioral changes for mothers in the VMGD and IVM treatment

groups. Potentially useful information on the possible

interactions of parent gender and treatment format may have

been lost by failing to use the same measures at 3-year


For the data that were presented in the 1990 3-year

follow-up study, 59% of mothers were still concerned about

behavior problems, 19% of mothers requested additional

therapy, and approximately 46% of the children were

classified using mothers' reports and 26% of the children

were classified on teachers'/fathers' reports as displaying

clinically significant problems at 3-year follow-up. These

results were felt by the author to be comparable to other

parent training studies, in which 30% to 50% of treated

families failed to maintain clinically significant

improvements over varying lengths of time.

As for differential group effects 3 years later, the

author concluded that the video modeling-group discussion

treatment (VMGD) was somewhat superior to the other

treatments according to mother reports (CBCL Total Behavior

Problem scores) in terms of producing stable long-term

results. This conclusion is perhaps best viewed tentatively,

given the author's previous observation that "one of the

most common methodological limitations in early parent

training outcome studies was the use of only one outcome

measure, such as parents' verbal or written opinion,

telephone contact, or questionnaire (Webster-Stratton, 1982,

p. 703), as well as the fact that a nonsignificant decrease

in children's deviant behaviors can be associated with a

positive change in parents' perceptions up to 8 months after

treatment (Johnson & Christensen, 1975).

The second purpose of the Webster-Stratton 1990 follow-

up study, which appeared to be given greater emphasis than

definitively documenting 3-year maintenance, was to

determine the psychological, demographic, and familial

characteristic of those children classified as responders

and nonresponders. Given that there was no untreated control

group, the author noted that it was "impossible to determine

whether classifying a family as responder or nonresponder is

related to treatment or to some other intervening variables"

(Webster-Stratton, 1990b, p. 148). It seems that one such

intervening variable may well be the effect of mothers'

depression at follow-up on their reports of child behavior

problems. Since it was noted that children classified as

nonresponders on mothers' CBCL scores had mothers who were

more depressed (p < .02), the question arises of just how

much the responder/nonresponder classification on the

mothers' CBCL criterion reflects actual child behavior

change vs. how much it may reflect a response bias related

to the possible interactions between maternal depression and

mothers' reports of child behavior problems (e.g., Dumas &

Serketich, 1994). In short, it appears that one of only

three classification criteria used may have been

contaminated, or at the very least, may not be an objective

and stable measure of behavioral change outcomes in

children, much less in an entire family. Consequently, in

the conclusions in this study regarding predictors of

treatment success, the apparent relationship of maternal

depression to poor long-term maintenance may actually

reflect the relationship of maternal depression to the

classification criteria as opposed to the actual effect of

maternal depression on maintenance of behavioral changes

over three years in this particular sample.

This problem was less apparent in an earlier effort to

determine predictors of treatment outcome for families of

conduct disordered children (Webster-Stratton, 1985a), in

which separate classifications (responder, nonresponder)

were examined for each of four outcome criteria. Families of

34 referred conduct problem children 3 to 8 years old (25

boys, 9 girls) from lower middle to lower class standing

were treated with a series of 9 weekly 2-hour sessions (as

reported in Webster-Stratton, 1984). The first half of 9

treatment sessions covered positive interactional skills in

a play context, and the second part of treatment focused on

teaching parents a specific set of operant techniques

including praise, ignoring, giving commands, and the use of

Time Out for noncompliance or destructive behaviors. Thirty

two families completed at least 70% of the sessions.

Families were classified as responders or nonresponders

using parent perceptions of child behaviors, observations of

children's behaviors, observations of parents' behaviors,

and a combination of parent and child behaviors. These

variables were assessed using the CBCL and the DPICS for

home observations. Summary variables were used for mothers'

behaviors (Total Critical Statements and Physical Negatives)

and child behaviors (Total Child Deviancy and Total

Noncompliance Behaviors), and assessments occurred at

baseline, 1 month, and 1 year after treatment. For the first

child behavior criterion, both child deviancy and

noncompliance had to be reduced 50% from baseline to be

classified as a responder; for the parent behavior

criterion, both mothers' criticism and physical negative

behaviors had to be reduced 50% of baseline, and mother and

child had to meet these two criteria jointly in order to be

classified as treatment responders on the interactional

criterion. For the parent perception outcome criteria, CBCL

Total Behavior Problem scores had to fall at or below the

90th percentile (at or below a score of 43).

Using these criteria, the author concluded that over

half of the families in this study were helped

substantially, and changes in attitude and behavior were

maintained or even improved at 1-year follow-up (p. 233).

Rates of posttreatment and follow-up responders were as

follows: child noncompliance and deviancy (55% and 79%

respectively for the two assessments), mother critical and

physical negatives (65% for both assessments), CBCL total

Behavior problem scores (56% and 64% respectively), and

combined mother and child behavior criteria (41% and 64%

respectively). One problem with this type of reporting is

that it does not reveal how a given family fared on all

outcome criteria.

The predictor variables (mother's depression, index of

socioeconomic disadvantage, and number of negative life

experiences during the year preceding treatment) accounted

for less variance when measured immediately after treatment

than when measured at 1-year follow-up. Socioeconomic

disadvantage, lack of social support, and negative life

experiences all impacted on the outcome of treatment, but

the author felt the results clearly illustrated that no

single predictor that is responsible for treatment success

or failure with these families. In this sample, single

parent families with economic difficulties and lack of

support emerged as significantly more likely to be

nonresponsive to this form of parent training therapy, data

which mirror Patterson's observation that such families are

also the most difficult to treat.

In order to further understand the relative

contribution of the psychological, interparental, and

extrafamilial or environmental variables that are related to

short-term and long-term treatment outcomes in parent

training, a second study of predictors of outcome for

families of conduct disordered children was conducted using

direct home observations and both parent and teacher report

of children's adjustment from 1-year follow-up data

(Webster-Stratton & Hammond, 1990). Families had been in 10-

12 week parent training programs, and 1-year and 3-year

follow-up were reported (Webster-Stratton, Kolpacoff, &

Hollinsworth, 1988; Webster-Stratton, Hollinsworth, &

Kolpacoff, 1989; Webster-Stratton, 1990b). After

statistically examining outcomes for 101 mothers and 70

fathers with conduct problem children between the ages of 3

and 8, several conclusions were reached. First, it appears

that different predictor variables will emerge depending

upon which criteria are selected to evaluate clinical

outcomes. In this study, for the immediate posttreatment

assessment outcome criteria, higher parental depression

predicted more negative mother and father perceptions of

child adjustment, while single-parent status or marital

conflict predicted more negative behaviors and higher child

deviance on home observations. The next finding addressed

whether the same predictors emerged for the 1-year follow-up

assessment. In this sample, results showed that for both

mothers and fathers, the amount of negative life stress

which occurred during the year following treatment was as

powerful a predictor of parent reports of child adjustment

as parental depression. For parental behavioral outcomes,

marital and socioeconomic status were significant predictors

of maternal criticisms and physical negatives, and

socioeconomic status was a predictor of fathers' negative

behavioral interactions with their children. In terms of

children's behaviors, parents' marital status made the most

significant contribution to predictions of children's

behaviors with their mothers and the amount of negatively

perceived life stress in the year following treatment made a

more significant contribution to predictions of children's

behaviors toward their fathers. Finally, according to

teachers' reports at 1-year follow-up, parents' marital

status made the greatest contribution to the variations in

children's adjustment. Even when these predictor variables

were used in combination, the variance accounted for in

significant sets of predictors ranged from 12-28%,

suggesting the need to look elsewhere for factors

contributing to successful outcomes and maintenance of

effects over time in parent training interventions.

Taken together, the results of Webster-Stratton's

outcome studies have provided evidence of the immediate

posttreatment usefulness of the information offered to

parents in Webster-Stratton treatment program, as well as

some evidence of the differential effectiveness of the

videotaped modeling-group discussion format over group

discussion alone or individually administered videotaped

modeling alone. For about 50-70% of families in the

videotape modeling comparative studies, positive changes

were maintained at 1-year follow-up assessments depending on

the outcome criteria used, and for a sample of college

educated, self-referred, middle class parents whose children

exhibited elevated but nonclinical levels of oppositional

behavior, positive changes were actually enhanced at 1-year

follow-up. At 3-year follow-up, all three groups of parents

continued to report fewer child behavior problems and

increased prosocial behaviors as compared to their own

pretreatment levels, and approximately 54% of mothers, 75%

of fathers, and 74% of teachers described the children's

functioning (number of behavior problems) as falling within

the normal range 3 years later. However, approximately one

third of parents in this combined sample still had ongoing

concerns about their children's aggression, noncompliance,

peer problems, or hyperactivity, and about one third of the

families asked for further therapy either for themselves or

their children. Since Webster-Stratton's 1-year and 3-year

follow-up results did not include comparisons to untreated

wait-list control groups, improvements due to maturation or

regression of extreme levels of behavior to the mean cannot

be ruled out.

To summarize the strengths of the Webster-Stratton

studies, they have included multiple-method, multiple-

observer assessments of parents, children (boys and girls),

and parent-child interactions, conducted over uniform and

extended intervals of time. Adequate descriptions of

standardized treatments and of the samples permit ready

comparisons to other samples and estimates of the

generalizability of the results. Results have been reported

in terms of both statistical and clinical significance, with

clinical significance defined either as a score/percentile

representing the cutoff score between deviant and

nonreferred populations, or a 30-50% reduction from baseline

level for a particular behavior or combination of behaviors.

Wait-list control groups have been incorporated into two of

the studies in the form of "delayed treatment" groups,

allowing posttreatment control group comparisons after 4

weeks or 9 weeks of treatment (the 1982 and 1984 studies).

Finally, the systematic inclusion of fathers in treatment

and follow-up studies has yielded important information

about the influences, both positive and negative, of marital

status and the quality of the marital relationship on the

behavior of conduct problem children over time.

A third treatment program to be reviewed is the work of

Forehand (e.g., McMahon & Forehand, 1984), who has conducted

empirical studies of the Hanf two-stage operant conditioning

model for treating dysfunctional interactional patterns of

multiply handicapped children and their mothers (Hanf, 1969)

as applied to the treatment of noncompliant children.

The program Forehand has tested is typically

administered in two phases of treatment. During Phase I (the

reinforcement phase), parents are taught to increase the

frequency and range of their social rewards, attend to the

child, and eliminate all commands, criticisms, and questions

in the parent-child interaction. By using praise

contingently upon desired child behaviors, Phase I

intervention teaches parents how to increase children's low-

rate prosocial behaviors. In Phase II, the parent is taught

how to use specific, direct, and concise commands combined

with the Time Out procedure, to decrease child noncompliant


The short term effectiveness, generalization, and

social validity of the McMahon and Forehand description of

treatment for families of children with noncompliant

behaviors has been reported for both parent and child

behaviors, as well as for parents' perceptions of their

children (McMahon & Forehand, 1984). The authors reported

that these improvements were found to occur regardless of

SES or age of child (range = 3 to 8 years). The

effectiveness of adjunctive procedures (Parent Enhancement

Therapy) for enhancing generalization and/or the maintenance

of treatment effects has also been reported (Griest,

Forehand, Rogers, Breiner, Furey, & Williams, 1982).

Temporal generalization has been reported in studies

with follow-up intervals ranging from 6 months to 14 years

after termination (Forehand, Sturgis, McMahon, Aguar, Green,

Wells, & Breiner, 1979; Baum & Forehand, 1981; Forehand,

Steffe, Furey, & Walley, 1983; Forehand & Long, 1988; and

Long, Forehand, Wierson, & Morgan, 1994).

The effects of treatment over time were tested in 11

mother-child pairs referred for the treatment of multiple

noncompliant child behavior problems in the home (Forehand,

Sturgis, McMahon, Aguar, Green, Wells, & Breiner, 1979). The

children were 7 males and 4 females, ages 3 to 8 years (M =

5.3 years). The sample parents were described as 91%

married, with 73% of the heads of households employed in

positions ranging from skilled labor to managerial

personnel, with education ranging from some college training

to having obtained college degrees, and 27% of the sample

was high school educated and employed in semi-skilled labor.

No attempt was made to quantify or report the actual levels

of pretreatment noncompliance in this sample.

Treatment consisted of two phases, as described, and

the mean number of sessions was 9.5 and the range was 7 to

12. Treatment outcome was assessed using the Parent's

Attitude Test (PAT) and home observations, both of which

were collected prior to treatment, immediately after

treatment, at a 6-month follow-up, and at a 12-month follow-

up. The Parent's Attitude Test assessed parent's attitudes

toward and perceptions of child behavior, and consisted of a

Home Attitude Scale, a Behavior Rating Scale, and an

Adjective Checklist Scale. Home observations were conducted

with a coding system that permitted the recording of parent

behaviors (rewards, commands, and time-out) and mother-child

interactions (child compliance, child noncompliance, and

contingent parental attention from the parent in response to

the child's compliance) (Forehand, Peed, & Roberts, 1975).

Three to five pre- and posttreatment home observations were

collected, and one home observation was collected at each

follow-up. Psychometric data were not given on either of the

two outcome measures.

For the home observations, reliability checks were

obtained on 21% of the pre- and posttreatment observations

and on 20% of the follow-up observations. Interrater

agreement was reportedly assessed using an event-by-event

analysis that required agreement for both the category of

behavior and the sequence, and average agreements among 8

observers were reported to be 80% for pre- and posttreatment

data and 82% for the 4 follow-up observations that were

checked for reliability. Unfortunately, the time of day for

home observations varied across mother-child pairs, but data

typically were collected at the same time of day for any one

mother-child pair.

One parent refused to participate in the follow-up,

therefore follow-up analyses were based on data from 10

subjects. The authors reported that significant improvements

(gs < .05) were obtained on each of the three PAT Scales

when pretreatment scores were compared to scores from

posttreatment and the 6-month and 12-month follow-ups.

The authors obtained significant F-values for repeated

measures analyses (across four assessments) of rewards, no-

opportunity commands, contingent attention, and children's

compliance (ps < .05). Significant overall change was not

obtained on the measure of alpha commands (commands the

child could obey), which the authors reportedly expected

since they did not train parents to increase or decrease the

use of these commands, but simply trained parents to

decrease beta or no-opportunity commands (commands to which

the child has no opportunity to demonstrate compliance). In

spite of a nonsignificant F ratio for overall changes in

alpha commands, this variable was apparently included in

pairwise comparisons.

Pretreatment to posttreatment and follow-up comparisons

apparently revealed significant changes in rewards, alpha

commands, and children's compliance. Pre- to post changes in

no-opportunity commands and contingent attention are

unknown, however parents' use of no-opportunity commands at

6-month follow-up did not differ significantly from

baseline, and parents' use of contingent attention at 6-

month and 12-month follow-up did not differ significantly

from baseline.

The way that the authors reported analyses of the

behavioral data from home observations made the actual

results very difficult to apprehend, but it appears that

significant change (pretreatment through 12-month follow-up)

was obtained in mothers' use of rewards (praise, approval,

positive physical attention, and descriptions) in

interactions with their children, and that significant

improvements apparently occurred in children's compliance to

alpha commands and to total commands. Apparently no

significant changes occurred in mothers' use of alpha

commands (commands with which the child could comply), and

posttreatment changes in no-opportunity commands and

contingent attention were not maintained 6 months after

treatment. As such, these results provide little evidence

that changes in children's compliance were due to changes in

parents' commanding behaviors. One might logically assume

that improvements in children's compliance were therefore

due to parents' use of the time-out procedure; however,

time-outs occurred too infrequently for analysis, suggesting

that the children were not exhibiting serious behavior

problems at baseline.

The problem of unclear reporting of behavioral data

highlights a more basic problem with the design of this

study, which is the adequacy of the behavioral indices to

assess and document treatment outcome. Parents in this

treatment were reportedly taught skills in two phases

corresponding to two very distinct mother-child

interactional contexts: parent-following and parent-leading

situations. In Phase I, parents were specifically taught to

"eliminate commands and directive behaviors", and in Phase

II, they were taught to "use alpha commands and a time-out

procedure to decrease noncompliant behavior demonstrated by

the child" (p. 10). Thus parents were not only taught new

skills, but were taught when and how to use them, yet the

observational data were apparently not obtained in a

situationally specific context and do not reflect "when and

how" parents employed the skills. This makes it difficult to

evaluate the importance of changes in commands or contingent

rewards (3 of the 4 parent behaviors measured), because any

use of commands or contingent rewards in the Phase I

situation or context is incorrect, therefore increases or

decreases (in that context) are irrelevant. Similarly,

increases or decreases in children's compliance in the

parent-following context are at best confusing.

A second problem in assessment is that the sample's

presenting complaint, noncompliance, was inadequately

quantified at baseline. Coupled with the lack of a no-

treatment control group, the changes in children's

compliance from pretreatment through the 1-year follow-up

may very well have been due to maturation, regression to the

mean, or due to undocumented historical events.

Long-term follow-up was also the purpose of a second

study (Baum & Forehand, 1981) of 34 mother-child pairs

referred for the treatment of child noncompliance. The

children also presented with secondary problems such as

aggression, destruction of property, and negative verbal

behavior (e.g., cursing, back talk, and whining), and some

children displayed problem behaviors in more than one

setting. The actual level of baseline noncompliance was not

reported. The children were 22 males and 12 females, and

children's ages at follow-up ranged from 4.42 years to 12.92

years (M = 7.79 years). The sample parents included 1 mother

who was a welfare recipient.

Treatment, which had occurred between 1975 and 1978,

consisted of two phases as previously described but the mean

number of sessions was not given for this sample. Treatment

outcome was assessed as in the 1979 study using the Parent's

Attitude Test (PAT) home observations, and a consumer

satisfaction questionnaire. The PAT and home observations

were collected prior to treatment, immediately after

treatment, and at the time parents were contacted for


Of the 36 of 40 treated mothers who were located for

follow-up assessments, 34 (94%) agreed to participate by

completing questionnaires, and of these 34 mothers, 20 (59%)

also agreed to home observations. The two groups were

compared on pre- and posttreatment behavioral measures and

the three parent attitude scales, and group comparisons were

also made at each follow-up on three attitude and six

consumer satisfaction measures. No significant differences

were found.

In this study, 6-month to 18-month blocks of time since

treatment were collapsed to form 3 follow-up periods or

intervals (i.e., 1-1.5 years posttreatment, 2-2.5 years

posttreatment, and 3-4.5 years posttreatment), and families

were then grouped according to the follow-up interval into

which they fit based on the length of time since the family

had been treated, presumably determined by when the family

was contacted. Comparisons for each family were therefore

for pretreatment, posttreatement, and one of three follow-up

periods; no family was measured at more than one follow-up

interval. Due to unavailable data that resulted from some

families being willing to complete questionnaires but not

observations, analyses of questionnaire and observational

data were based on different groups and numbers of subjects

even at the same assessment, as well as across follow-up


Home observations were conducted with a coding system

that permitted the recording of parent attends and rewards,

beta or no-opportunity commands, contingent attention,

time-out, child compliance, and child deviant behaviors

(whining, crying, yelling, tantrumming, aggression toward

objects or people, and negative verbal remarks) (Forehand,

Peed, & Roberts, 1975). Three to eight pre- and

posttreatment, and two follow-up home observations were

collected. For the home observations, reliability checks

were obtained on 22% and 25% of the pre- and posttreatment

observations and on 31% of the follow-up observations.

Interrater agreements for categories of behavior were

reported to be .67-.90% for pretreatment data, .83-.99% for

posttreatment data, and .98-100% the follow-up data. For

home observations of mother-child interactions, other

family members could be present but were not coded.

The authors reported that for all questionnaire

measures, children were perceived as better adjusted at post

and follow-up than at pretreatment, and the changes achieved

at post were reportedly maintained at follow-up. Similarly,

for the behavioral data, parents used more attends, rewards,

and contingent attention at posttreatment and follow-up than

at pretreatment. There was a significant decrease from

posttreatment to follow-up in attends and rewards,

contingent attention, and child deviant behaviors. Thus, the

authors concluded that posttreatment gains in children's

compliance and deviant behaviors and parents' use of beta

(no-opportunity commands) were maintained up to 4.5 years

posttreatment. The data and research design of this study do

not actually warrant such a conclusion, because no family

was actually tracked and assessed throughout the follow-up


It was also noted by the authors that the significant

posttreatment increases in parent attending, praise, and

contingent use of praise were not maintained up to 4.5 years

after treatment, but far from viewing this as a treatment

failure, the authors noted that the use of such intermittent

reinforcement is designed to maximize the maintenance of

behavior, and "decreases in positive reinforcement by the

parent are programmed into the training program as parents

are told that the initial frequent use of reinforcement can

be gradually reduced as the child's negative behaviors

decrease" (p. 650). This 1981 observation is particularly

interesting in view of the fact that a similar finding in

the Forehand et al. 1979 study, that during follow-up

parents did not maintain their use of contingent attentions

significantly above the baseline level" indicated at that

time "the need for greater emphasis on this particular

parent behavior during treatment" (p. 15).

A similar methodology was again used to extend the Baum

and Forehand 1981 data to families who had completed the

previously described treatment program between 11 months and

7 years earlier (Forehand, Steffe, Furey, & Walley, 1983).

Two outcome measures were used in this study: the PAT, and 4

questions concerning the mothers' satisfaction with

treatment. Although the authors state their belief in the

introductory remarks that "behavioral observations are the

most important aspect of assessment of parent training" (pp.

339), behavioral observations were not collected for this

study. Sixty-eight mother-child pairs were contacted who had

been treated for noncompliance and other child behavior

problems, and families were described as being in the upper

lower class to low middle class range based on either

pretreatment or follow-up indices. Unfortunately, it was not

specified which indices were used for this classification,

and social status may have changed in the 7.2 years of


Of the 68 mothers who were contacted, 34 mothers

returned the questionnaires, comprising the sample for this

1983 study. A comparison of the 34 responders with the 34

nonresponders showed that the two groups did not differ

significantly on pretreatment or posttreatment PAT scores,

however, group comparisons were not reported on pre- or

posttreatment demographic variables or behavioral

variables. Following initial analyses of the relationship

between PAT scores and time since treatment, the PAT scores

for the 34 mothers were collapsed across the entire follow-

up period and results of a univariate analysis of variance

led to the conclusion that mothers perceived their children

as significantly better adjusted after treatment and this

effect maintained at follow-up. Means were 66.03 at

pretreatment, 40.59 at posttreatment, and 37.24 at follow-

up. Standard deviations were not reported. Given the fact

that differences in the responders and nonresponders were

inadequately assessed, and that treatment outcome was

assessed using only a parent report measure with no

provision made to control for or measure variables that may

have affected mothers' reports of child behavior problems,

the authors' assertions that this study provides evidence of

long-term maintenance appear to be unsupported by the data

they have presented.

A similar methodology was used in the Forehand and Long

(1988) study, which purported to demonstrate adequate

adjustment, compared to a matched comparison sample, of

former clients during their early adolescent years. Follow-

up sample children ranged in age from 11 to 14 years old and

had been treated by Forehand and his associates between 4

1/2 and 10 1/2 years earlier for noncompliance and secondary

behavior problems such as aggression, destruction of

property, and negative verbal behavior. In most cases, only

the mother and child had participated in treatment.

A comparison group consisting of 21 families who had

not participated in parent training and whose early

adolescent aged child had never been in any type of therapy

was selected from a pool of 155 volunteer participants in a

project examining parent-adolescent interactions. The 21

comparison families were matched with the parent training

families in terms of age, gender of adolescent, and family

social status.

For the 21 of 43 families who were contacted and agreed

to participate in this follow-up study, families were

assessed on adolescent functioning (4 areas), parent

functioning (3 areas), and parent satisfaction with the

treatment they had received. Specific measures included (1)

the Revised Behavior Problem Checklist (RBPC) completed by

mothers, fathers, and teachers, (2) the Issues Checklist

(IC), a 2-week recall about 44 issues discussed between

parent and child that can yield an index of the number of

issues and the degree of negative affect of the discussion

(on a 5-point Lickert scale), completed by mothers, fathers,

and adolescent, (3) the 20-item short form of the Conflict

Behavior Questionnaire (CBQ) assessing family communication-

conflict behavior as reported by mother, father, and

adolescent, and (4) the Children's Depression Inventory

(CDI), completed by the adolescent, (5) the Perceived

Competence Scale for Children (PCSC), completed by the

adolescent, (6) the Rating Scale of Children's Actual

Competence (RSCAC) Cognitive and Social Competence

Subscales, completed by mothers, fathers, and teachers, (7)

academic grades, (8) the Parent Competency Inventory (PCI),

completed by parents as to their own competencies as

parents, (9) the Beck Depression Inventory (BDI), completed

by mothers and fathers, (10) the Dyadic Adjustment Scale

(DAS) completed by both parents, (11) the O'Leary Porter

Scale (OPS) which was designed to assess the frequency of

overt parental conflict that occurs in front of the child,

(12) subjective ratings by undergraduate students of

videotaped observations of the mother and adolescent

discussing and attempting to resolve, during a 3 minute

observation, an issue regarding keeping one's bedroom clean

(methodology adapted from Robin & Carter, 1984;

reliabilities across 7 ratings was reportedly above 85%),

and (13) a consumer satisfaction questionnaire. Psychometric

data were described for all of the measures except the IC,

the PCI, the observational system, and the consumer

satisfaction questionnaire. Parents agreeing to participate

were paid $50 per mother-adolescent dyad, and fathers were

paid $15 for participation in the project. Questionnaires

were presented in a randomized order.

Analyses were performed with a preset a level of .01

to reduce chance findings, and in this study, the entire

period of follow-up (4.5 to 10.5 years) was collapsed to

permit two-group comparisons of families who had (n = 21)

vs. had not (n = 21) participated in parent training.

In spite of significant differences in participants and

non-participants (who reported less marital satisfaction

before treatment and lower child compliance to opportunity

commands than participants), and in spite of using a control

group not adequately matched on marital and/or family

compositional variables of interest to the findings of this

study, the authors concluded that few significant

differences emerged for adolescent externalizing or

internalizing problems, that no significant differences

occurred between groups for (self-reported) measures of

parenting skills and personal adjustment, and further, the

authors concluded that mothers in the Parent Training group

reported better marital adjustment scores and fathers in the

parent training group reported less interparental conflict,

than parents in the comparison group.

Acceptance of these conclusions is severely limited by

the participant-nonparticipant difference, suggesting that

the families with less favorable results were not included

in the analyses. Additionally, any conclusions regarding the

relative adequacy of parental functioning and adjustment at

follow-up are precluded by the fact that the "matched

comparison group" controlled for none of the variables which

could be expected to influence marital adjustment and

interparental conflict (e.g. size of family, parents' ages,

number of marriages, length of present marriage, etc.)

Further, the authors concluded that with regard to parenting

skills, marital adjustment, and personal adjustment

(depressive symptoms), "the results in each of these three

areas indicate that parents who had previously participated

in parent training were functioning as well or better than

the comparison families" (p.164). To attribute parents'

positive (or negative) functioning to previous participation

in a parent training program is simply inappropriate, given

the methodology of this study.

Most recently, the authors have used a methodology

similar to the 1988 study in an attempt to document that

noncompliant children who participated in parent training

during their early years are functioning as well as

nonclinic individuals as they move into adulthood (Long,

Forehand, Wierson, & Morgan, 1994). Long et al. (1994)

described a 14-year follow-up study of the 26 of 30 young

adults who could be located. Subjects were between the ages

of 17 and 22 and had participated in parent training with

their mothers at least 14 years earlier. Using a comparison

sample matched on age, gender, and family socioeconomic

status, the authors concluded that the previously treated

sample did not differ from the matched community sample on

measures of delinquency, emotional adjustment, academic

progress, or relationship with parents, suggesting that

noncompliant children who have participated in parent

training during their early years are capable of functioning

as well as nonclinic individuals as they move into adulthood

(Long, Forehand, Wierson, & Morgan, 1994).

In addition to the problems cited in the original

treatment outcome studies from which this treated sample

presumably came (most notably, inadequate documentation of

the actual pretreatment severity and pervasiveness of

specific child behavior problems; e.g. pretreatment alpha

compliance of this sample = 83% and pretreatment deviant

behavior = 10%, pp. 105), problems in this study again

include an inadequately described control group (e.g., were

any control group subjects ever in therapy, ever arrested,

were they actually "well adjusted"?), dubious dependent

variables (i.e. self-report alone used for estimates of

delinquency, consumption of alcohol, and highest grade

achieved), and participant-nonparticipant differences

(participants were of significantly higher pretreatment SES

than nonparticipants) suggesting that data from clients

likely to have less favorable outcomes were not included in

these analyses.

Unfortunately, the results of the Forehand follow-up

studies do no support conclusions of positive long-term

outcomes. Temporal generalization in the Forehand studies is

weakened by a number of factors. Inadequate pretreatment

documentation in each of these studies of child behavior

problems limits the generalizability of results due to a

lack of clarity about the severity and pervasiveness of the

behavior problems that were addressed. Although most studies

included observational data with adequate reliabilities,

inadequate control of observational circumstances could have

led to systematic error in behavioral data (e.g., time of

day varied across mother-child pairs, but was kept constant

for any one mother-child pair; Forehand et al., 1979). The

actual number of observations within two of the studies vary

widely across subjects, also leading to problems with

sampling error (Forehand et al., 1979; Baum & Forehand,

1981). In the Baum and Forehand study (1981), mother-child

interactions were coded between 3 and 8 times at

pretreatment, between 3 to 5 times at posttreatment, and

once each for the follow-up intervals.

In this same study, 6-month to 18-month blocks of time

were collapsed to form 3 follow-up "periods" or intervals:

1-1.5 years posttreatment, 2-2.5 years posttreatment, and 3-

4.5 years posttreatment. Mother-child pairs in this study

were grouped according to length of time since treatment,

and comparisons across time of questionnaire as well as

observational data were therefore based on different groups

and different numbers of subjects (see also the 1983, 1988

and 1994 studies, which used the same methodology).

Although it was generally concluded in the Forehand

follow-up studies that many families appeared to be

functioning better at follow-up than at pretreatment, the

method of combining data from many families by collapsing

data across months or years of time introduces many new

sources of variability and precludes clear statements about

long-term outcomes. This method, chosen by Forehand to

evaluate long-term maintenance, is actually a cross-

sectional survey of many families rather than a true

longitudinal study of specific families, and it does not

therefore give a picture of sequential changes over time for

specific subjects. Using follow-up "intervals" that collapse

across blocks of time introduces considerable subject

variance and variability in the range of ages of the

children across periods of measurement, thus confounding the

contributions of individual differences, developmental and

maturational effects, and treatment effects. Even if no-

treatment control groups had been included, the type of

design used in the Forehand outcome and long-term follow-up

studies simply does not permit one to attribute changes in

behavior over time to treatment alone.

The question of sampling bias due to follow-up

attrition is a potential problem for all researchers

collecting longitudinal data, and in the 1988 study, this

question is unfortunately raised by the fact that the

follow-up non-participants had reported less marital

satisfaction at pretreatment and their children were

observed to be less compliant at posttreatment than the

subjects who did participate in follow-up. Similarly, in the

1994 study, follow-up participants were of significantly

higher SES than the follow-up non-participants. Finally, the

generalizability of findings to disadvantaged and/or single-

parent families is limited, in that these groups were not

often represented in this line of research (percentages

ranged from 3% to 27% in studies where percentages could be


In more positive terms, conducting follow-up studies of

children over such lengthy periods of time means that

investigators will also encounter the considerable problem

of choosing age-appropriate measures, and a strength of the

1988 study is the authors' intent to apply the multi-method

multi-outcome criteria to the conceptualization and

assessment of parent-child relationships with children just

entering adolescence.

Treatment Failures in Parent Training

The parent-training literature with young behavior

problem children has empirically demonstrated successful

interventions with some families for periods of up to three

years after treatment, however, a limitation of this form of

treatment is the fact that there are many families who do

not respond to treatment. Nonresponding may mean failing to

initiate treatment after being referred, dropping out of

treatment after it has begun, or failure to show adequate

improvement over the course of treatment (e.g., Kazdin,

Mazurek, & Bass, 1993; Webster-Stratton, 1985a; Webster-

Stratton, 1994). Approximately 50 to 75% of all children

referred for psychological treatment either do not initiate

treatment, or terminate treatment early if they do begin

(Kazdin, 1990). Child and family therapy dropout rates have

been consistently estimated at 50% or higher (Pekarik &

Stephenson, 1988), and agency-referred families tend to be

more resistant to therapy than self-referred families, with

dropout rates of 45% and less than 0%, respectively, in one

study (Chamberlain, Patterson, Reid, Kavanaugh, & Forgatch,

1984). Studies of parent training have suggested that of

families completing treatment, 30-50% of parents fail to

show clinically significant improvements, defined as changes

of at least 30% of baseline rates of behavior (Wahler &

Dumas, 1984).

Factors That Contribute to Outcome in Treatment of Behavior
Problem Children

Duration of treatment, specific training components,

and therapist training and skill appear to be associated

with the magnitude and durability of treatment effects

(Kazdin, 1987). The effects of parent, child, and family

characteristics have also been examined as predictors of

poor outcomes in the treatment of behavior problem children.

Parental behaviors related to negative treatment

outcomes (drop-outs or inadequate improvement) have included

high levels of maternal aversive and indiscriminate behavior

at baseline (Dumas, 1984) and high rates of maternal

commands (McMahon, Forehand, Griest, & Wells, 1981).

Maternal perception of children's adjustment at pretreatment

was not associated with outcome (Dumas, 1986), however,

reductions in maternal blaming attributions and less global

summary descriptions of children at the end of treatment

have been associated with maintenance of treatment effects

measured 4 months after treatment (Wahler & Afton, 1980).

For families completing treatment, the roles of marital and

personal distress (depression) in treatment outcome and

maintenance are interwoven and appear to be mediated by

other variables such as social support from extended family

contacts and others in the community (Dadds & McHugh, 1992).

The inclusion of a brief, maritally focused treatment with

parent training may be useful in overcoming parent training

treatment failures for maritally discordant families (Dadds,

Schwartz, & Sanders, 1987).

Dumas and Wahler (1983) investigated the relative

predictive power of maternal insularity and disadvantaged

socioeconomic status for treatment outcome, and found that

each variable contributed unique variance to predicting

outcome and that both together accounted for 49% of the

variance in the data. Disadvantaged but noninsular families

(or vice versa) had a 50% chance of successful outcome,

whereas families who were both socioeconomically

disadvantaged and insular were highly likely to be

classified as treatment failures at 1-year follow-up.

Further investigation of the two-variable model of

treatment outcome (maternal insularity and the index of

disadvantaged socioeconomic status) (Dumas and Wahler,

1983), led to the conclusion that "treatment outcome and

involvement appear to be a function of socioeconomic

adversity only" (Dumas, 1986). One explanation is that

parent setting events (e.g., personality) set the stage for

the development and/or maintenance of a variety of adverse