I

I
I
I
I
I
I
I
I AN INTERPRETIVE MANUAL
for The Child Abuse Potential Inventory
I Joel S. Milner, Ph.D.

I
I
I
I
I
I
I
CD-24282
--
I
I
I AN INTERPRETIVE MANUAL
for The Child Abuse
I Potential Inventory
I
I
MAY 15
I
I Joel S. Milner, Ph.D.
I
I NATIONAl. CLEA{~(r{GH,)U;;:E ON CHILD
ABUSE AND NEGLEC',j ~NFORMATION

I
I Distributed by:
PSVTEC INC.
P.O. Box 564
DeKalb, IL 60115

I PSYTEC INC.
1-815-758-1415

I
I
 I
I
I
I
I
I
I
I
I
I
© Copyright, 1990; Joel S. Milner, Ph.D.

All rights reserved. No part of this manual may be reproduced by any process, electronic or
I
mechanical, including photocopy, audio and/or visual recording, duplication in an informational
storage and retrieval system, without written permission of the copyright owner. Printed by Blue
Ridge Printing Company, Inc., Asheville, NC. Published by Psytec Corporation, Webster, NC,
DeKalb, IL. Printed in United States of America.
I
Library of Congress Catalog Card Number: 89-062385
I
ISBN Number: 0-940929-09-2

I
I
I
I
I ACKNOWLEDGEMENTS

I The development of this applied manual is the direct result of feedback from
professionals in the field. Input from direct services providers, administrators, and
researchers indicated the need for a user's manual for the Child Abuse Potential

I Inventory that would supplement the existing technical manual. Professionals re-
quested a simple reference manual that would describe test applications and limitations.
In addition, there were requests for specific scale interpretation hypotheses to facilitate
use of the Child Abuse Potential Inventory scales in clinical applications. The author
I expresses his appreciation to these professionals for their suggestions and input.

Malcolm Gordon, Ph.D., and Jim Breiling, Ph.D., National Institute of Mental

I Health, warrant special recognition for their assistance in the initial conceptualization
and planning of this applied manual.

A number of other professionals deserve recognition for their assistance in the

I development of this manual. Specifically, the following individuals deserve special
thanks for their assistance in developing and/or editing different sections ofthis guide:
Ronald Wimberley, Ph.D., Department of Sociology, Anthropology, and Social Work,
I North Carolina State University; Kevin Robertson, Ph.D., University ofNorth Carolina
School of Medicine; Stan Ray, Ph.D., Payne County Guidance Center, Stillwater, OK;
Dian England, M.S.W., MCH Social Worker, Payne County Health Department, Still-

I water, OK; Jane Morgan, M.S.W., Child Welfare Supervisor, Oklahoma Department of
Human Services; Wendy Hawkins, M.A., and Alice Mayall, M.A., Department of
Psychology, Northern Illinois University; Sherry Christensen, B.A., Sycamore, IL; the
Protective Services staff, Onslow County Department of Social Services, Jacksonville,
I NC; Suzanne Jessamyn, B.A., Program Director, Rainbow Family Learning Center,
Tampa, FL; and other reviewers.

I The National Institute of Mental Health also is acknowledged for its support of a
substantial portion of the research on the development of the Child Abuse Potential
Inventory and for additional support for the development ofthis applied manual (MH-

I 34252).

I
I
I
iii
I
 I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
IV

I
I
I TABLE OF CONTENTS

I Page

Acknowledgements III

I o
F 19ures Xl
•

I PREFACE

I CHAPTER ONE: INTRODUCTION

Overview 1

I Applications of the Child Abuse Potential Inventory

Screening suspected physical child abusers
2
2
Selection of at-risk individuals 2
I Assessment of clients prior to treatment
Treatment or program evaluation
3
3
Research applications 4

I Limitations of the Child Abuse Potential Inventory 5

I Issues Related to the Use of the Child Abuse Potential Inventory

Detection of response distortions
Relationship between the abuse score and its intended purpose
6
6
7
False positive and false negative classifications 8
I Probabilistic nature of test scores
Standard error of measurement ofthe abuse scale
10
10
Variation of classification rates 10

I Child abuse screening and population base rates

Multiple stage child abuse screening
Using screening data from multiple sources
11
13
13
Screening versus diagnosis 14
I
I
I
v

I
 I
CHAPTER TWO: SCALE DESCRIPTIONS Page I
Introduction 17

Abuse Scale 20 I
Abuse Factor Scales
Distress
Rigidity
21
21
21
I
Unhappiness 21
Problems with child and selL
Problems with family
22
22
I
Problems from others 22

Validity Scales
Lie scale
23
23
I
Random response scale
Inconsistency scale
23
23 I
Response Distortion Indexes (Validity Indexes) 24
Faking-good index
Faking-bad index
24
24 I
Random response index 24

Special Scales
Ego-strength scale
25
25
I
Loneliness scale 25
I
I
I
I
I
I
vi

I
I
I CHAPTER THREE: TEST ADMINISTRATION Page

Introduction 27

I Administrator Characteristics 27

I Evaluator Characteristics

Examinee Characteristics
27

28

I Readability Level 28

Test Materials 28

I Test Format 28

Time Limit 29
I Testing Procedure 29

I Group Administration 32

I
I
I
I
I
I
I
vii

I
 I
CHAPTER FOUR: SCORING THE SCALES Page I
Introduction 33

Scoring Procedures 33 I
Hand Scoring Templates
Abuse scale and six factor scales
Validity scales
34
35
36
I
Response distortion indexes (validity indexes) 37
Special scale scores 37
I
Computer Scoring 38

I
I
I
I
I
I
I
I
I
I
viii

I
I
I CHAPTER FIVE: SCALE INTERPRETATIONS Page

Introduction 39

I Scale Item Blanks

Interpretation hypotheses for excessive scale blanks
40
40

I Possible courses of action

Validity Scales
41

42
Lie scale (L scale) 42
I Random response scale (RR scale)
Inconsistency scale (Ie scale)
42
42

I Response Distortion Indexes (Validity Indexes)

Faking-good index
Interpretation hypotheses for an elevated faking-good index
43
43
43
Possible courses of action 44
I Intervention and treatment related issues
Other issues
44
45
Faking-bad index 46
I Interpretation hypotheses for an elevated faking-bad index
Possible courses of action
46
47
Intervention and treatment related issues 47

I Other issues
Random response index
Interpretation hypotheses for an elevated random
47
48

response index 48
I Possible courses of action
Intervention and treatment related issues
49
49
Other issues 49

I Abuse Scale Score

Interpretations of elevated abuse scale scores
50
50

I What does the abuse scale measure?

Low base rates
Possible courses of action
51
53
53
Other issues 53
I
I
I
ix

I
 I
Page I
Abuse Scale Factor Scores 54
Distress factor scale
Interpretations of elevated distress scale scores
Rigidity factor scale
54
54
54
I
Interpretations of elevated rigidity scale scores 54
Unhappiness factor scale
Interpretations of elevated unhappiness scale scores
54
54
I
Problems with child and selffactor scale 55
Interpretations of elevated problems with child and self scale scores
Problems with family factor scale
55
55 I
Interpretations of elevated problems with family scale scores 55
Problems from others factor scale
Interpretations of elevated problems from others scale scores
55
55. I
Special Scales 56
Ego-strength scale
Interpretations of elevated ego-strength scale scores
56
56
I
Loneliness scale 56
Interpretations of elevated loneliness scale scores 56
I
REFERENCES 57
I
APPENDIX
Appendix A: Item composition of the CAP Inventory (Form VI) Scales
Appendix B: Selected Readings
59
61
I
I
I
I
I
I
x

I
I
I FIGURES Page

I Figure 1.1

Figure 1.2
Four possible screening outcomes

Assessment stages leading to diagnosis and outcome

9

15

I Figure 2.1 Decision Tree 18

I
I
I
I
I
I
I
I
I
I
I
Xl

I
 I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
xii

I
I
I PREFACE

I The purpose of this applied manual is to provide test administrators and test users
with information that will guide the general use and interpretation of the Child Abuse
Potential Inventory. This manual provides information on the appropriate applications
of the Child Abuse Potential Inventory and suggests interpretation hypotheses for the
I Inventory scales.

It is important for users to note that this manual is a guide to test use and

I interpretation but it is not a substitute for the more comprehensive Child Abuse
Potential Inventory technical manual (Milner, 1986). The technical manual contains
extensive information on the development, reliability, and validity ofthe Child Abuse Po-
tential Inventory. Test users are responsible for evaluating the psychometric data
I provided in the technical manual to determine ifthe Child Abuse Potential Inventory is
appropriate for the intended use. Thus, the applied manual should be used only after test
users are thoroughly familiar with the psychometric data provided in the technical
I manual. The information supplied in this applied manual, therefore, is intended only
as a supplement to the documentation provided in the technical manual.

I
I
I
I
I
I
I
I
xiii

I
 I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
xiv

I
I
I CHAPTER ONE
I INTRODUCTION

I
OVERVIEW
I
The Child Abuse Potential Inventory was designed primarily as a screening scale

I for physical child abuse. The abuse scale, descriptive factor scales, and special scales
contained in the Child Abuse Potential Inventory measure variables that represent ele-
ments of the psychiatric and social-interactional models of physical child abuse. The

I questionnaire development and associated validity and reliability data are described in
a comprehensive technical manual (Milner, 1986). Prior to use ofthis applied manual,
the test user should be familiar with the documentation provided in the technical
manual. This applied manual is not a substitute for the technical manual.
I
I
I
I
I
I
I
I
1

I
 I
APPLICATIONS OF THE CHILD ABUSE POTENTIAL
INVENTORY
I
Screening Suspected Physical Child Abusers
I
The Child Abuse Potential Inventory was designed primarily for use as a screening
tool for the detection of physical child abuse in social services agencies and similar
settings where individuals are reported and investigated for suspected physical child
I
abuse. When the abuse scale is used for screening suspected abusers, the questionnaire
should always be used in conjunction with evaluation data from other sources including
interviews, direct observations, collateral interviews, and other test data.
I
Research indicates that at the time of testing individuals with elevated abuse scale
scores have characteristics, traits, and parenting styles similar to confirmed physical
I
child abusers (see the section in Chapter Five entitled "What does the abuse scale meas-
ure?"). Validity data also suggest that the abuse scale provides a meaningful increase
in the concurrent and future prediction of physical child abuse (see Chapter Five in the
technical manual, Milner, 1986).
I
Selection of At-Risk Individuals
I
The Child Abuse Potential Inventory can be used as a screening tool for the selection
of individuals who are at increased risk for physical child abuse. I
1. Identification of at-risk individuals with the Child Abuse Potential Inventory is
useful in secondary prevention efforts where resources are limited and profes-
sionals need to offer their services to a group thought to be at increased risk for
I
physical child abuse. This application of the Child Abuse Potential Inventory is
especially appropriate in situations where labeling an individual as an abuser is
not an issue and individuals with elevated abuse scores are merely offered
I
services.

2. The Child Abuse Potential Inventory is useful in determining that an individual

designated as at-risk for abuse by some other criteria is actually at-risk. This
I
application is useful in situations where funded programs are required to
demonstrate that individuals receiving services are actually at-risk for abuse. In
these cases, the Child Abuse Potential Inventory provides an independent and
I
objective evaluation of the at-risk status of those receiving services.
I
I
2

I
I
I Assessment of Clients Prior to Treatment

In this application, the Child Abuse Potential Inventory is not used for screening

I but is used to provide clinical information for treatment purposes. The six descriptive
abuse scale factor scores and the two special scale scores are used to provide an indication
ofpossible problem areas for the client. Indicated problem areas can be explored through

I additional testing and/or interviews.

Treatment or Program Evaluation

I Another application of the Child Abuse Potential Inventory is its use in treatment
or program evaluation. Typically, the Child Abuse Potential Inventory is administered

I at the beginning of an intervention/treatment program, at the end of the intervention/

treatment program, and at a follow-up meeting that usually occurs six months to a year
after the intervention/treatment program has been completed.

I Use of the Child Abuse Potential Inventory for evaluation purposes provides
information on the validity ofthe evaluation process through use of the validity indexes
as well as an overall estimation of abuse potential change across the intervention/
I treatment program. In addition, the six descriptive abuse scale factor scores and the two
special scale scores provide information on the client characteristics that show the
greatest and least changes across the intervention/ treatment period. Thus, the factor

I scale scores indicate areas where the intervention/treatment program is most effective
and areas where additional interventions need to be developed.

I
I
I
I
I
I
3

I
 I
Research Applications
I
The Child Abuse Potential Inventory can be used in a variety of research applica-
tions. Studies needed to generate additional test validity data are discussed in the
technical manual (Milner, 1986). Other research applications involve using the Child
Abuse Potential Inventory abuse scale as a criterion for testing theories and models of
I
physical child abuse. More specifically, since groups ofidentified physical child abusers
are usually difficult to obtain, hypotheses can be tested initially on groups with elevated
and normal abuse scores. Promising results can be replicated on groups of recently
I
confirmed, untreated physical child abusers.

Caveat: Analog studies can produce findings that are not meaningful because of I
attenuation in the criterion used (in this case the Child Abuse Potential Inventory).
However, in many cases, use of the abuse scale will allow research to be conducted that
otherwise might not be possible. I
I
I
I
I
I
I
I
I
I
4

I
I
I LIMITATIONS OF THE CHILD ABUSE POTENTIAL INVENTORY

I The test user should be aware that the Child Abuse Potential Inventory has a
number of limitations. The abuse scale provides the most incremental validity when
used in screening for physical child abuse. Although the abuse scale has significantly

I elevated scores for other child maltreatment groups, the individual classification errors
increase when child sexual abusers and child neglectors are screened. Thus, these types
of child maltreating parents are more likely to be missed (i.e. false negatives) when
screened with the physical abuse scale. Further, since a number of variables overlap
I among child maltreatment groups, the overall physical abuse scale score cannot be used
to differentiate between various types of child maltreatment. While the descriptive
factor scales have been shown to load differentially on the different types of child

I maltreatment, these factor scales have yet to be adequately validated for use in the
differential screening of child maltreatment groups.

The Child Abuse Potential Inventory most accurately describes the state
I ofthe parent at the time oftesting. While elevated abuse scores are predictive oflater
physical child abuse, retesting is recommended with the passage of time. Retesting is
especially recommended in situations where the living and/or family conditions have
I changed. For example, ifstress levels increase and/or ifa child is returned to the parent,
these events may affect the potential for child abuse and retesting is recommended.

I A number of psychometric problems can limit the usefulness of any test data
including the data obtained from the Child Abuse Potential Inventory. Some ofthe most
important psychometric considerations are discussed in the next section.

I
I
I
I
I
I
5

I
 I
ISSUES RELATED TO THE USE OF THE CHILD ABUSE POTEN-
TIAL INVENTORY
I
Detection of Response Distortions
I
All self-report scales that are used for screening physical child abuse, including the
Child Abuse Potential Inventory, can be affected by examinees' attempts to distort their
responses to the questionnaire items. The most common response distortions include at-
I
tempts to fake-good, to fake-bad, and to answer in a random pattern.

The detection of response distortions by examinees is important because response

I
distortions can invalidate the screening instrument scales. The Child Abuse Potential
Inventory contains three empirically derived response distortion indexes (i.e., the
faking-good index, the faking-bad index, and the random response index) that should be I
scored and evaluated prior to scoring and interpreting the abuse scale, factor scales, and
special scales.
I
I
I
I
I
I
I
I
I
6
I
I
I Relationship Between the Abuse Score and Its Intended Purpose

In order for the test user to determine the appropriateness of the Child Abuse

I Potential Inventory for a specific application, the test user should evaluate the available
validity and reliability data in the technical manual (Milner, 1986) to determine ifpsy-
chometric data support the intended use.

I Construct validity is defined as the extent to which the characteristics and/or traits
purported to be measured by the instrument are actually measured. In the worst case,
a test will not measure the construct suggested in the title. Even in optimal cases,
I however, the test instrument will only approximately measure the construct. Thus, even
the "best" tests have some measurement error.

I The challenge, then, is for the test user to determine the adequacy of a test to
measure the construct suggested by its title. If a test is used to assess a characteristic
not indicated by the title and not supported by the existing documentation published in
the technical manual, then the test user must take special care to determine that data
I exist to support the special application. In some cases the test user may need to generate
the required data before making the new test application.

I In the case of the Child Abuse Potential Inventory, the title indicates that the
questionnaire measures the examinee's potential for child abuse. Inspection of the
technical manual (Milner, 1986) and the list ofapplications discussed in the prior section

I of this chapter indicate that the Child Abuse Potential Inventory is designed primarily
as a screening tool for the detection ofphysical child abuse in social services agencies and
similar settings where individuals are reported and investigated for suspected physical
child abuse. The target population is adult caretakers who are suspected physical
I child abuse perpetrators.

The construct validity data for the abuse scale are described in detail in the technical

I manual (Milner, 1986). A descriptive summary ofthe construct validity data is provided
in Chapter Five ofthis manual under the heading "What does the abuse scale measure?"
The data indicate that the abuse scale measures traits and parenting styles that are
characteristic ofphysical child abusers. Concurrent and predictive validity data indicate
I that those with elevated abuse scores are usually physical child abusers and are likely
to physically abuse children in the future.

I
I
I
7

I
 I
False Positive and False Negative Classifications
I
During the process of screening an individual, two correct and two incorrect
classifications are possible. The four screening outcomes are presented in Figure 1.1.
The screening procedure may identify correctly an abuser as abusive, or the screening
may misclassify the abuser as not abusive. The probability that a scale will identify
I
correctly an abusive individual is known as the degree of sensitivity of the screening
procedure. The likelihood that the screening scale will designate a nonabuser as abusive
is known as the false positive classification rate. The likelihood that the screening
I
scale will designate an abuser as nonabusive is known as the false negative classifi-
cation rate. The probability that a scale will identify correctly a nonabuser is known as
the degree of specificity of the screening procedure. I
The Child Abuse Potential Inventory user should remember that even the most
psychometrically sound screening instruments misclassify individuals. That is, all
screening scales incorrectly classify nonabusers as abusers (false positive classifica-
I
tions) and incorrectly classify abusers as nonabusers (false negative classifications).
When used with social service populations, the Child Abuse Potential Inventory abuse
scale typically produces more false negative than false positive classifications.
I
To a large extent, the type of classification error that should be avoided is
determined by the consequences that result from a false positive or false negative
classification. For example, iflabeling is not an issue and the intervention that follows
I
screening involves the offer of such things as advice, education, counseling, therapy,
child care, and/or the referral to community resources and services, false positive clas-
sifications may not be considered very damaging when weighed against the possible pre-
I
ventive benefit to children who may otherwise be mistreated. In these cases, the
practical consequence of a false positive classification is that a nonabusive parent may
be offered assistance in parenting that is not needed. On the other hand, iflabeling and
I
possible court action may follow the evaluation, false positive classifications are criti-
cally important and should be avoided. In these cases, the practical consequences of a
false positive classification include the possible removal ofthe child and other legal con-
sequences for the parent.
I
I
I
I
I
8

I
I
I Figure 1.1

Four Possible Screening Outcomes

I
ACTUAL
I Abuse No Abuse

I
I Abuse A B

I SCREENED

I
No Abuse C D
I
I
I A: Correct classification of abuse (Le., sensitivity)
B: Misclassification of abuse (Le., false positive classification)
C: Misclassification of abuse (Le., false negative classification)
I D: Correct classification of no abuse (Le., specificity)

I
I
I
9

I
 I
Probabilistic Nature of Test Scores I
Misclassifications are caused by the probabilistic nature of test scores. Test scores
are known to be affected by errors in test design and by situational variables, such as
those found in the testing conditions. These errors can be estimated using the standard
error of measurement.
I
Standard Error of Measurement of the Abuse Scale
I
Each test score is an estimate ofthe actual or true test score. All of the Child Abuse
Potential Inventory test scores contain measurement error that has been estimated (i.e., I
the standard error ofmeasurement). Thus, each test score can be said to represent a true
score. The obtained test score, at some level of probability, falls within a range of scores.
Overall, the Standard Error ofMeasurement for the abuse scale is about 19 scale points
(see page 58 in the technical manual, Milner, 1986) on an abuse scale containing 486 scale
I
points. This means that the actual score is likely within plus or minus 19 points of the
obtained abuse score. This variation is just under plus or minus four percent ofthe total
number of abuse scale points.
I
Variation of Classification Rates I
The misclassification rates will vary with the type of examinee tested. Thus, the
Child Abuse Potential Inventory abuse scale should be used only for screening purposes
with clients who are similar to those subjects used in the validation research where
I
misclassification rates were determined (see technical manual, Milner, 1986). If the
abuse scale is used with other groups, the test user is responsible for providing additional
cross-validation data and norms for the new target populations.
I
I
I
I
I
I
10
I
I
I Child Abuse Screening and Population Base Rates

The test user should be aware of the population base rate for the examinee being
I evaluated. The base rate refers to how often a behavior (e.g., physical child abuse) occurs
in the population represented by the examinee. The base rate of recent physical child
abuse in a group of general population parents may be less than 5%; whereas the base

I rate of physical child abuse in a group of parents reported for physical child abuse tends
to range from 30 to 50% (i.e., percent of confirmed cases). Any test, including the Child
Abuse Potential Inventory abuse scale, provides the most incremental validity or
additional information when the target population base rate is about 50%. Thus, one of
I the recommended applications of the Child Abuse Potential Inventory, the screening of
suspected physical child abuse cases, involves a situation where base rates are optimal
for increasing predictive validity.

I If the Child Abuse Potential Inventory is used in low base rate situations, the
amount ofincremental validity provided by the Inventory decreases. In situations where

I the target population base rates are very low, the use of the Child Abuse Potential
Inventory may be inappropriate for certain applications.

For example, if the base rate of physical child abuse is 50% in a target population
I and if the screening scale classification rate for abusers is 80% and for nonabusing
parents is 80%, then 80% of the abusers and 80% of the nonabusers will be correctly
classified resulting in an overall correct classification rate of 80%. However, in another

I target population, if the base rate for abusers is only 5% (meaning that 95% are not
abusing) and the screening scale has the same classification rates (i.e., 80% for abusers
and 80% for nonabusers) as in the prior example, the classification outcomes will be much
different. That is, when the base rate for physical child abuse is 5% in a target group and
I the test has an 80% abuser classification rate, then 4 out of the 5 expected abusers in
a group of 100 subjects will be correctly classified resulting in 1 false negative classifi-
cation. In addition, 80% ofthe 95 nonabusing subjects will be detected, which is 76 out
I of 95 subjects, resulting in 19 false positive classifications. In summary, with a child
abuse base rate of 5% in a group of 100 subjects, 4 of those detected will be abusive, but
19 with elevated abuse scores will be nonabusive. Thus, in the 23 (4 plus 19) cases

I screened as abusive, 19 individuals or 82.6% of the 24 cases will be misclassified (false

positives) by the screening scale because of the low population base rates. This result is
found despite the fact that the test has an overall correct classification rate of80% when
target population base rates are 50%. It also is true, however, that in the 76 cases
I screened as nonabusive, only 1 or 1.3% will be misclassified (false negatives) as
nonabusive. In this example, because ofthe low population base rate, the group screened
to be abusive and the group screened as not abusive are both less likely to abuse than the

I test classification rates would indicate.

I
11
I
 I
In the situation where the target population base rate is low, the problem
of false positive classifications can be resolved partially by additional evalu-
I
ation of the group initially screened as abusive (see the next section on multiple
stage screening).
I
I
I
I
I
I
I
I
I
I
I
I
I
12
I
I
I Multiple Stage Child Abuse Screening

Although the Child Abuse Potential Inventory abuse scale and other screening

I scales have limited utility when they are used to detect child abuse in populations with
low base rates, the Child Abuse Potential Inventory can still be used to screen for physical
child abuse in low base rate situations ifa multiple stage screening process is employed.

I For example, the Child Abuse Potential Inventory can be used initially to screen for a
group of at-risk individuals. Then, as a second screening step, the at-risk group can
receive a more intense evaluation using other criteria to identify an even smaller group
that has a greater likelihood of physical child abuse.
I Alternatively, if a professional first uses some existing criterion to identify a group
of high-risk individuals, the Child Abuse Potential Inventory could be used as a second

I screening criterion to select a smaller group ofindividuals who are more likely to engage
in physical child abuse. Either way, a multiple stage screening approach will reduce the
number offalse positive classifications that result from a single stage screening in a low
base rate situation. The problem of excessive false positive classifications is resolved
I partially by the multiple evaluations because each screening stage will raise the base
rate for the remaining group of examinees that will be evaluated at the next stage.

I Using Screening Data from Multiple Sources

Even when base rates are adequate, the Child Abuse Potential Inventory user
I should remember that the abuse scale score alone should never be used to make a
diagnosis or to label an examinee. This limitation applies to any test that is used for child
abuse screening. The abuse score must be used in conjunction with evaluation data
I from other sources such as interviews, case histories, direct observations, collateral
interviews, medical data, and other psychological test data. Multiple types ofevaluation
data serve to increase the probability of making a correct individual classification.

I It should be mentioned that in some cases where assessment data are incomplete
or unclear, too much weight may be inappropriately placed on the Child Abuse Potential
Inventory abuse score. For example, in several instances where case data were ambigu-
I ous and decisions had to be made to remove or return a child, there was pressure from
the court to use the Child Abuse Potential Inventory abuse score as the final basis for the
placement decision. This is always an inappropriate action because even the most psy-

I chometrically sound test contains measurement error. Multiple sources of evalu-

ation data must be used.

I
I
13
I
 I
Screening versus Diagnosis I
The Child Abuse Potential Inventory was developed to be used as a screening instru-
ment rather than as a diagnostic instrument. In assessment, screening is a term used
to describe a rapid, and often rough selection process. Usually screening refers to a
preliminary attempt to determine if a characteristic or behavior is present or absent in
I
a given individual or in a group. In most situations, the screening activity occurs as the
initial stage of a multiple stage assessment process. Diagnosis, a term used to describe
a different part of the assessment process, is an intensive and comprehensive evaluation
I
activity that usually occurs in the last stage of assessment. A paradigm that describes
the typical stages found in the assessment process is outlined in Figure 1.2.
I
The assessment process begins with screening (Stage 1) and, through several steps,
leads to diagnosis (Stage IV). The assessment paradigm provides a step-by-step strategy
for the professional to follow. The strategy defines what the professional will do in any
number of possible situations given different outcomes. The assessment paradigm ap-
r
proaches the decision-making process sequentially, since new information or confirma-
tion data are gained as the professional moves from one stage to the next. Until the end
ofthe multiple stage assessment process, no commitment to a specific diagnosis is made.
I
The addition ofdata at any stage may result in a modification ofthe final diagnosis.
Even at the last stage of assessment, ifthe information gathered remains ambiguous or I
incomplete, the assessment process provides for a return to a previous stage to obtain
needed data. Thus, ifduring the case staffing/diagnosis stage (Stage IV) it is determined
that additional history, personality testing, and/or medical tests are needed, the assess-
ment process should return to the appropriate prior stage. Flexibility and complete-
I
ness are features of this assessment model so that misclassifications are minimized
when the final diagnosis is made. I
This assessment paradigm provides a clear distinction between screening and di-
agnosis. Screening occurs in the initial stage of assessment and usually involves a
technique that places the examinee into one of two categories. That is, either the I
individual is believed, because of the screening criteria, to have a specified characteris-
tic (risk for abuse) or the individual does not have the characteristic. No other outcome
is available. Screening scales are usually designed to be relatively brief in contrast to
more comprehensive and time consuming diagnostic procedures.
I
I
I
I
14
I
I
I Figure 1.2
Assessment Stages Leading to Diagnosis and Outcome

I
I CLIENT

I
I • If ?

I
[ SCREENING (optional) If NO..
STAGE I a. Subjective (e.g., rating scale) STOP
I
I ,
b. Objective (e.g., CAP Inventory)

~ If YES

I STAGE II
[ INTERVIEW
a. Sociological Evaluation
b. Psychological Evaluation
If NO
I
I
STOP

I ,
I If? or YES

I STAGE III
[ TESTS (optional)
a. Psychological Tests
b. Medical Tests
~

I , ~ Continue
If?
STAGE IV [ CASE STAFFING AND DIAGNOSIS If NO I STOP
I
I ~ If YES

INTERVENTION DECISION
I ,
OUTCOME
•
I Continue
,
I EDUCATION
THERAPY, ETC.
-. ADJUDICATION
DECISION f::
CHILD REMOVED

CHILD REMAINS

I
I
15
I
 I
The type ofclassification error allowed varies from stage to stage in the assessment
process. During the initial screening stage (Stage 1) of the assessment process, false
I
negative classifications are usually avoided because at this stage the goal is to avoid
missing any actual cases while reducing the pool of clients who will continue to the next
stage of assessment. Since it is assumed that subsequent stages will eliminate any false
positive classifications, the primary concern is that the screening scale cutting score be
I
set low so that the screening procedure will avoid missing actual abuse perpetrators. In
contrast, during the diagnostic stage (Stage IV) of the assessment process, it is critical
to avoid false positive classifications and the accompanying false accusation of an
I
innocent client because there will not be any subsequent stages to confirm or refute
questionable cases. Thus, during the diagnostic stage, the criteria for inclusion in the
maltreatment group are expected to be more comprehensive and complete, correspond-
I
ingly cutting scores are usually set higher.

I
I
I
I
I
I
I
I
I
I
16
I
I
I CHAPTER TWO
I SCALE DESCRIPTIONS

I
INTRODUCTION
I
The 160-item Child Abuse Potential Inventory (Form VI) contains a total of ten

I standard scales. The primary screening scale is the physical child abuse scale. The
abuse scale can be divided into six descriptive factor scales: distress, rigidity, unhap-
piness, problems with child and self, problems with family, and problems from

I others. The Child Abuse Potential Inventory also contains three validity scales: the lie
scale, the random response scale, and the inconsistency scale.

In addition the Child Abuse Potential Inventory validity scales are used in various
I combinations to produce three response distortion (validity) indices: the faking-good
index, the faking-bad index, and the random response index. Finally, two special
scales have been developed from Child Abuse Potential Inventory items: the ego-

I strength scale and the loneliness scale.

A decision tree that can be used to guide the user in the proper interpretation ofthe
Child Abuse Potential Inventory scales is provided in Table 2.1. The specific items
I composing each of the ten standard Child Abuse Potential Inventory scales and the two
special scales are provided in Appendix A. This chapter provides a brief description of
each of the standard scales, the response distortion (validity) indices, and the special
I scales.

I
I
I
I
17
I
 I
Figure 2.1 I
Decision Tree

if NO
I
STEP 1 Are there more than 10% scale blanks?

,~
if YES
I
See interpretation hypotheses and possible courses
of action. Pages 40 and 41. I
I
STEP 2 Is the faking-good index elevated? -- if NO

, if YES
I
See interpretation hypotheses and possible courses
of action. Pages 43 and 44. I
I
STEP 3 Is the faking-bad index elevated?
-- if NO

,. if YES
I
See interpretation hypotheses and possible courses
of action. Pages 46 and 47. I
I
STEP4 Is the random response index elevated? - if NO

if YES
I
"
See interpretation hypotheses and possible courses
of action. Pages 48 and 49.
I
"
(see next page)
I
18
I
I
I Figure 2.1 continued

(from prior page)

I -- I
STEPS Is the abuse scale score elevated? if NO

I ,Ir
if YES

I See interpretation hypotheses and possible courses

of action. Pages 50-53.

I ,Ir if YES

STEP 6 Are any of the factor scores elevated?

-
~

if NO

I if YES
,Ir

I See interpretation hypotheses and possible courses

of action. Pages 54-55.

I ,Ir
if YES

I
STEP 7 Are any of the special scales elevated? (optional) - if NO

if YES
,Ir

I See interpretation hypotheses and possible courses

of action. Page 56.

I ,Ir
Optional

I Summarize protocol interpretations. -

I
I
19
I
 I
ABUSE SCALE I
The 77-item physical child abuse scale contains clusters of items that are best
described by the six descriptive factor scales discussed below. The first three descriptive
factor scales measure personality characteristics and/or psychological difficulties of the
I
respondent. The remaining three factor scales measure interpersonal and interactional
problems experienced by the respondent. For further information, the reader is referred
to the section in Chapter Five entitled "What does the abuse scale measure?"
I
I
I
I
I
I
I
I
I
I
I
I
20
I
I
I ABUSE FACTOR SCALES

I Distress

This factor represents a general theme of personal distress. The scale is a measure

I of distress perceived by the respondent. This factor scale, however, is not simply a
measure of global distress. The distress is relatively specific to interactional problems
between the parent and the child. Generally, this factor indicates the presence of
personal adjustment problems which result from parenting stress that appear to be
I related to abusive behavior.

I Rigidity

This factor identifies a rigid parenting style. The rigidity measured by this scale is

I specific to the respondent's attitudes toward the appearance and behavior of children.
That is, the examinee has many rigid expectations (of behavior and affect) related to
children. These beliefs may be expressed through the forceful treatment of children to
make them fit a rigid mold defined by the respondent.
I
Unhappiness

I This dimension describes a polarity of happiness/unhappiness. The items describe

a general unhappiness with life and a specific unhappiness related to problems in
interpersonal relationships. The personal characteristics and interpersonal skill defi-
I cits that produce unhappiness in the respondent's life and relationships contribute to the
likelihood of difficulties in the respondent's interactions with children.

I
I
I
I
I
21
I
 I
Problems with Child and Self
I
This dimension measures the degree to which the respondent describes children
and his/her self in a negative manner. The factor focuses on perceptions of having a
problem child and perceptions of having a child oflimited ability and competency. The
factor also indicates a perceived limited physical ability in the respondent. The
I
perception of having a problem child, the belief that children have problems because of
limited ability, and the beliefin the limited physical ability ofone's selfcontribute to the
likelihood that the respondent will maltreat children.
I
I
Problems with Family

This factor measures a variety of family interactional problems. The factor

generally indicates that the respondent's family has many problems, has difficulty
I
getting along, and experiences fighting. A problem laden, disturbed, and possibly violent
family increases the likelihood that the respondent's child will be maltreated. I
Problems from Others

This dimension measures general difficulties in social relationships. Relationships

I
are viewed as a source of personal problems, unhappiness, and pain. Relationships are
viewed as a cause of disappointment rather than a resource because others cannot be
counted on for support. These attitudes and interpersonal problems contribute to the
I
respondent's perceived isolation and contribute to difficulties in the parent-child rela-
tionship.
I
I
I
I
I
I
22
I
I
I VALIDITY SCALES

I Lie Scale

The 18 item lie scale was designed to detect individuals who distort their responses

I in a socially desirable manner. The items include statements that represent the presence
or absence of generally valued attitudes and behaviors that are not fully achieved by
anyone. The statements include items such as "I never act silly" and "I sometimes think
of myself before others." The lie scale is balanced with an equal number of agree and
I disagree scored responses.

I Random Response Scale

The 18 item random response scale was designed to detect individuals who respond

I to the questionnaire items in a random fashion. The random response scale is balanced
with an equal number of agree and disagree scored responses.

I Inconsistency Scale

The inconsistency scale contains 40 items configured in 20 item-pairs. The

I inconsistency scale provides a conceptually different measure from the random response
scale. The scale measures the degree ofinconsistent responding to items that are usually
answered in a predictable (expected), consistent fashion. The inconsistent response
scale is balanced in the following manner. It contains five agree-agree scored response
I pairs, five disagree-disagree scored response pairs, five agree-disagree scored response
pairs, and five disagree-agree scored response pairs.

I
I
I
I
I
23
I
 I
RESPONSE DISTORTION INDEXES (VALIDITY INDEXES) I
Faking-Good Index

The faking-good index is constructed from two validity scales: the lie scale and the
random response scale. The faking-good index is considered elevated when the lie scale
I
score is elevated and the random response scale score is in a normal range.
I
Faking-Bad Index

The faking-bad index is constructed from two validity scales: the random response
I
scale and the inconsistency scale. The faking-bad index is considered elevated when the
random response scale score is elevated and the inconsistency scale score is in a normal
range. I
Random Response Index I
The random response index is constructed from two validity scales: the random
response scale and the inconsistency scale. The random response index is considered
elevated when both the random response scale score and the inconsistency scale score are I
elevated.

I
I
I
I
I
I
I
24
I
I
I SPECIAL SCALES

I Ego-Strength Scale

The ego-strength scale contains 40 items that provide a measure of emotional

I stability. The ego-strength scale measures both personal and interpersonal components
of emotional stability.

The scale assesses the examinee's perception of hislher emotional stability and
I feelings of adequacy. These perceptions ofemotional stability and adequacy are related
to the examinee's evaluation ofhislher ability to maintain emotional stability and to his/
her success in interpersonal relationships. The ego-strength scale, therefore, is not a

I global measure of ego-strength but is a measure of one major aspect of ego-strength or

self-esteem (i.e. emotional stability and interpersonal adequacy) thought to be important
to the etiology of family violence.

I Loneliness Scale

I The loneliness scale contains 15 items that provide a measure of loneliness. The
loneliness scale measures the degree ofthe examinee's perceived loneliness rather than
the actual degree of social isolation. The loneliness scale provides a measure of the

I examinee's view ofthe available social support. The perceived lack of social support and
related feelings ofloneliness have been viewed as contributing to the increased likelihood
of family violence.

I
I
I
I
I
I
25
I
 I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
26

I
I
I CHAPTER THREE
I TEST ADMINISTRATION

I
INTRODUCTION
I Chapter Three is designed to provide the test administrator with instructions on
how the Child Abuse Potential Inventory should be administered. Attention to test

I administration issues is essential in order to obtain the most accurate and representa-
tive examinee test score results.

I ADMINISTRATOR CHARACTERISTICS
The test administrator should have a knowledge ofthe administration procedures
I outlined in this chapter and should be familiar with the Child Abuse Potential Inventory
format, the test instructions, the test content, and other characteristics of the test
booklet. The test administrator should have a knowledge of basic rapport building

I techniques. Since the questionnaire is simple to administer, a nonprofessional (e.g.,

agency secretary, agency volunteer, nurse's aid, etc.) who is trained and supervised by
a qualified professional can administer the Inventory.

I EVALUATOR CHARACTERISTICS

I The evaluator who uses and interprets the Child Abuse Potential Inventory scale
scores should be a professionally trained individual, such as a social worker, psychiat-
ric nurse, counselor, psychologist, or other professional who has received training in

I assessment and test interpretation procedures. The test should be interpreted only
by those with appropriate training and experience because of the potential
negative impact of test misapplication.

I
I
I
27
I
 I
EXAMINEE CHARACTERISTICS I
The Child Abuse Potential Inventory is designed to be used with parents and other
adult primary caretakers suspected of physical child abuse. At the time of testing, the
examinee should not be psychotic, under the influence of alcohol or drugs, or excessively I
fatigued. The questionnaire should not be used with individuals whose characteristics
(e.g., cultural backgrounds) are outside the range ofpopulation characteristics described
in the validity studies reported in the technical manual and other relevant literature. I
READABILITY LEVEL I
The readability level of the Child Abuse Potential Inventory is grade three. At
present, no data are available on the effects of reading the questionnaire to illiterate or
visually impaired examinees. If it is necessary to read the items, first read aloud the
instructions, then read aloud the items without explanation or comment. Ifpossible, the
I
examinee should still circle hislher own responses on the test booklet out of the
examiner's direct view. I
TEST MATERIALS I
In order to administer the Child Abuse Potential Inventory, a test booklet is
necessary. The test booklet has a combined question and answer format, so no separate
answer sheet is required. A pencil or a pen is needed so that the examinee can circle his/
her responses on the test booklet. The examinee should be seated at a table or given a
I
hard writing surface.
I
TEST FORMAT
The Child Abuse Potential Inventory is available as a four-page test booklet. At I
the top of the first page of the test booklet is a personal data section. This demographic
section is followed by the test instructions, that are read by the examinee. The 160 items
that make up the questionnaire begin in the middle of the first page of the test booklet
and continue through to the bottom ofthe fourth, and last, page. For ease ofreading and
I
responding, the test items are printed in large print, and the items are grouped in
clusters offive items. Each item is answered in a forced-choice, agree-disagree format.
Answers are printed as A (agree) and DA (disagree) on the test booklet to the right ofeach
I
test item.

I
I
28

I
I
I TIME LIMIT
There is no time limit for completing the Child Abuse Potential Inventory. The

I examinee should be instructed to respond to the test items in an easy, unhurried manner.
However, it is recommended that the questionnaire be completed in one testing session.
The questionnaire usually takes a high school educated examinee 15 to 20 minutes to

I complete and takes a college educated examinee 12 to 15 minutes to complete.

TESTING PROCEDURE
I The procedures for administering the Child Abuse Potential Inventory are similar
to procedures used to administer other self-report questionnaires, which include estab-

I lishing a positive testing environment. Specifically, the following steps should be

taken.

I 1. The test room should be relatively comfortable and free from distractions such
as children, noise, intrusions, etc.

2. Attempts should be made to establish rapport with the examinee. This may
I require discussing relatively trivial topics and issues unrelated to the test (e.g.,
ability to find the agency, the weather, etc.).

I 3. Occasionally, examinees are concerned about the purposes ofthe testing. In all
circumstances, the examinee has the right to receive a clear and com-
plete statement of the purpose of the testing. Examinees should be told if
the test results will be used as part of a case evaluation, program evaluation,
I research project, or if they will be used for some other purpose. Examinees
should be informed about the precautions taken to maintain confidentiality and
to ensure security of test protocols and test results.
I 4. Prior to reading the test instructions and responding to the test items, the
examinee should be asked to fill in the personal data section at the top ofthe test

I booklet. In cases (e.g., research projects) where it is important for the examinee
to remain anonymous, the test administrator might use a protocol identification
number in place of the examinee's name. If necessary, the examinee can be
given assistance in completing the personal data section. Before the examinee
I begins the test, it is important for the test administrator to inspect the per-
sonal data section to make sure that all personal data have been provided.
Complete demographic data will permit later correct identification of the

I protocol and, where needed, will provide information on the examinee's per-
sonal characteristics for comparison to reference groups and for research
purposes.

I
29
I
 I
5. Mter the personal data section has been completed, the examiner should inform
the examinee that:
I
a. There is no time limit on the questionnaire.

b. There are no "right" or "wrong" answers.

I
c. He/she should circle the answer that best represents hislher attitudes and/
or feelings.
I
6. The examinee should be instructed to read carefully the test instructions
printed on the first page of the test booklet. Since the test instructions are
I
simple, brief, and self-explanatory, it is usually not necessary for the examiner

7.
to read the test instructions to the examinee.

When the examinee has finished reading the test instructions, the examiner
I
should ask the examinee if he/she has any questions. If it appears that the
examinee did not fully understand the test instructions, then the test adminis-
trator should read the instructions aloud to the examinee and explain any area
I
of confusion.

8. Before the examinee is allowed to begin reading the test items, the test admin- I
istrator should emphasize the need for the examinee to be conscientious and
truthful in responding to the questionnaire items. The examinee should be told
that a serious and honest approach to the questionnaire produces the most
accurate and meaningful results. The test administrator should be aware that
I
coercing the examinee in any way increases the likelihood that an examinee will
distort responses and/or will cause the examinee to skip items. I
9. Mter the aforementioned issues have been discussed, the test administrator
should ask once again if there are any questions. If not, the examiner can
instruct the examinee to begin answering the questionnaire items. I
10. Occasionally during the completion ofthe questionnaire, an examinee may com-
plain about the forced-choice format of the test. In these cases, the test admin-
istrator should indicate that, "While it sometimes may be difficult to fully
I
agree or disagree with a particular item, you should circle the answer
that best represents your attitudes, feelings, and/or beliefs."
I
I
I
30

I
I
I 11. Individuals may express surprise, concern, and/or annoyance about the repeti-
tion ofitem content. In these cases, the test administrator should indicate that
helshe is aware that some of the items have similar content (e.g., ''Yes, some
I of the questionnaire items do look the same" or ''You are right, some of
the items ask about the same things"). Following these statements, the test
administrator should provide brief reassuring comments that:

I a. None of the items are exactly the same (e.g., "Even though the items
appear to be the same, they are not exactly the same").

I b. All of the items are important (e.g., "Please remember that each ques-
tion on the questionnaire is important").

I c. Responses to each of the items will provide the examiner with a more
complete understanding of the examinee's feelings, attitudes, and beliefs
(e.g. , ''Your response to every item will help us get a better picture of

I you").

12. During the completion ofthe questionnaire there should be no distractions. The
test administrator should not discuss the meaning ofindividual test items with
I the examinee. Even when an examinee has problems responding to a specific
test item, the test administrator must not make any comments related to the
item content and should simply ask the examinee to make hislher own best

I choice. Likewise, the examinee should not be permitted to discuss test items
with other individuals including agency staff, spouse, friends, etc. In general,
the test administrator is responsible for preventing environmental variables
from distorting test results.
I
I
I
I
I
I
31
I
 I
GROUP ADMINISTRATION
I
While the individual administration of the Child Abuse Potential Inventory is rec-
ommended, the simplicity and self-report nature of the questionnaire lends itself to
group administration. No major changes in the previously mentioned test procedures
are required for group administration. The test administrator, however, should be
I
aware of any special needs that an individual examinee may have. Prior to administra-
tion of the Child Abuse Potential Inventory, the examiner should structure the seating
arrangement in the testing room to maximize the confidentiality ofeach examinee's
I
responses. Further, the examiner should specifically request that the individuals taking
the test not interact with each other or comment on the questionnaire items in
any way. It is recommended that each examinee remain seated until all members ofthe I
group have completed the questionnaire.

I
I
I
I
I
I
I
I
I
I
32
I
I
I CHAPTER FOUR
I SCORING THE SCALES

I
INTRODUCTION
I This chapter provides information on scoring procedures that are available for
scoring the Child Abuse Potential Inventory scales. Careful attention should be given

I to proper scoring of the Child Abuse Potential Inventory scales so that scoring errors do
not add to measurement error.

I SCORING PROCEDURES
There are two procedures available for scoring the Child Abuse Potential Inventory

I scales: hand scoring templates and a computer scoring program (CAPSCORE). Each
scoring procedure yields the same scale scores because both use the same item-weights.

Use of the hand scoring templates to determine scale scores for each of the Child
I Abuse Potential scales is more time consuming and is usually less accurate than
computer scoring because human addition errors can occur when item scores are
summed. Hand scoring, however, can be used in agency and field settings where a
I computer is not available.

I
I
I
I
I
33

I
 I
HAND SCORING TEMPLATES I
The hand scoring approach uses a series of transparent scoring templates to
generate the scale scores. The templates are made of a heavy grade of plastic and are
durable, allowing for repeated use. The name of each Child Abuse Potential Inventory
scale and the associated items to be scored are indicated on each template.
I
As a scoring aid, four small circles are printed horizontally at the top and the bottom
of the right-hand side of each page of the test booklet. These circles are printed above
I
and below the answer columns (see the test booklet). One of the four circles on each page
ofthe test booklet is solid black. The position ofthe solid black circle in the series ofcircles
corresponds to the booklet page number. Thus, on page one of the test booklet, the first
I
of the four circles on the top and bottom ofthe answer column ofthe test booklet is solid
black (e.g., • 0 0 0) indicating page one.

Four corresponding circles are printed at the top and bottom of the scoring sections
I
on each template (see scoring template). However, on each scoring template, the solid
and open circles are reversed from those printed on the test booklet so that on the
template the position ofthe open circle indicates the booklet page number (e.g., 0 • •
I
• indicates page one). Thus, when the template is properly aligned over the test booklet,
not only do the four template circles cover the circles on the test booklet, but the open
circle on the template is filled by the solid circle on the booklet so that all circles will
appear to be solid. This provides correct alignment and indicates that the correct page
I
is being scored.

Once the transparent hand scoring template for a given scale is properly aligned
I
over the test booklet, scoring of the scale can begin. Between the sets offour circles on
each scoring template that are used for template alignment, open squares are printed on
the template that enclose either agree (A) or disagree (DA) item responses on the test
I
booklet. These squares, which are associated with scale items, represent the examinee
responses that are scored for the scale named on the template.
I
I
I
I
I
34

I
I
I Abuse Scale and Six Factor Scales

Weighted item scores appear next to the squares on the abuse scale template and

I the six factor scale templates. When an examinee's response appears in the square
associated with an item, the item score printed next to the square is recorded. The
weighted scored responses are summed for each scale to obtain the scale raw score total.

I NOTE: When weighted scale scores are summed, great care must be taken to avoid
errors in addition. Even experienced, well-trained clinicians make mistakes. Prior
research indicates addition errors are common, occurring in about 25% of all protocols
I checked. The data suggest that, when the weighted abuse scale and six factor scales are
hand scored, the addition of the scored responses should always be checked at
least once.

I
I
I
I
I
I
I
I
I
I
35

I
 I
Validity Scales I
A nonweighted scoring procedure is used to score responses on the lie scale, the
random response scale, and the inconsistency scale. On the lie scale and random
response scale scoring templates, a score of one appears next to all the squares printed
on the templates. The one-point responses (i.e., those item responses appearing within
I
the squares) are summed for each scale to obtain the raw score totals.

While a nonweighted scoring procedure also is used to score the inconsistency scale,
I
the scoring is more complex and requires a scoring template and an associated scoring
sheet. Pairs of squares appear next to each item that is to be scored. No score values
appear next to the squares on the scoring template. An inconsistency scale scoring sheet
I
lists the item-pairs that make up the scale and provides instructions for scoring the scale.
The inconsistency scale is scored by locating 20 item-pairs. The 40 items that make up
the 20 item-pairs are located by using the inconsistency scale scoring template. The
inconsistency scale scoring sheet indicates items (by number) in each item-pair and how
I
the associated item responses are to be scored. Inconsistent item-pair responses receive
a score ofone point and the score is recorded on the inconsistency scale scoring sheet. The
one-point item-pair responses are summed to obtain the inconsistency scale raw score
I
total.
I
I
I
I
I
I
I
I
36

I
I
I Response Distortion Indexes <Validity Indexes)

The faking-good, faking-bad, and random response indexes are determined using

I the raw score totals of different pairs of the individual validity scales. No additional
scoring of Child Abuse Potential Inventory items is required to determine the three
validity indices.

I Special Scale Scores

I A nonweighted scoring procedure is used to score the ego-strength and loneliness

scales. On the ego-strength and loneliness scale scoring templates, an item score of one
appears next to all ofthe squares printed on each template. The one-point responses (i.e.,

I those responses appearing within the squares) are summed to obtain the raw score total
for each special scale.

I
I
I
I
I
I
I
I
I
37

I
 I
COMPUTER SCORING I
A computer scoring program, CAPSCORE, is available for scoring the Child Abuse
Potential Inventory scales. A detailed description of the CAPSCORE program with
examples of computer screens generated by the program is presented in the technical
manual (Milner, 1986). The CAPSCORE program provides the following information:
I
1. The number of total blanks, the number of scale blanks, and the number of
nonscale blanks.
I
2. The raw score totals for the lie, random response, and inconsistency scales.
I
3. A determination of the validity indexes.

4. The raw score totals for the abuse scale and the six factor scores. I
5. The raw score totals for the special scales (optional program).

A research version of the CAPSCORE program is also available. The research

I
version includes all of the information provided on the CAPSCORE scoring program,
and it contains a demographic entry section. The research version records all of the
individual item responses, summed scale scores, demographic data, and other informa- I
tion in a file. The user simply enters the item responses and demographic data when
prompted to do so. All ofthe individual data entered, as well as the computed scale scores,
are stored in a data file for later use. I
I
I
I
I
I
I
38

I
I
I CHAPTER FIVE
I SCALE INTERPRETATIONS
I
INTRODUCTION
I This chapter provides the test user with some suggestions for the interpretation of
elevated Child Abuse Potential Inventory scale scores. Specific interpretation hypothe-

I ses for the validity indices, the abuse scale, the factor scales, and the special scales are
provided. For each index and scale, primary interpretations are given first. Primary
interpretations generally should be considered before other interpretations. Following
the presentation ofinterpretative hypotheses for each elevated index and the abuse scale,
I several courses of action or recommendations for follow-up are given.

For a more complete understanding of the meaning of elevated Child Abuse

I Potential Inventory scale scores, the test user should refer to the technical manual
(Milner, 1986), which contains extensive information on scale reliability, standard errors
ofmeasurement, validity, and classification rates. As noted in the interpretation section

I ofthe technical manual (Milner, 1986), the test user is responsible for determining ifthe
test score is reliable and valid for its intended use.

I
I
I
I
I
I
39

I
 I
SCALE ITEM BLANKS I
For scoring purposes, scale blanks refers to unanswered and double-answered (i.e.,
agree and disagree) test items. With more than ten percent scale blanks, the test
user should consider the scale invalid. Scale blanks will lower the Child Abuse
Potential Inventory scale scores and will substantially affect scale validity. In cases
I
where the scale blanks are less than ten percent, a prorated total score is recommended.
A prorated score can be determined by computing the average item score for each ofthe
answered items. The average item score is then assigned to each unanswered item. The
I
computer scoring program (CAPSCORE) provides a prorated abuse scale scoring as an
option.
I
Interpretation Hypotheses for Excessive Scale Blanks

Scale blanks may indicate one or some combination of the following examinee
I
conditions.

1. (Primary) Inability to read and/or comprehend test items.

I
2. (Primary) Inability or unwillingness to handle the forced-choice item response
format on selected items. I
3. (Primary) A lack of motivation or an intentional lack of cooperation in respond-
ing to the scale items. I
4. A defensive response set associated with an unwillingness and/or inability to
reveal personal attitudes and feelings. I
5. Psychomotor retardation and/or lethargy associated with depression.

6. Confusion and/or paranoid ideation associated with psychotic states. I

7. Confusion and/or inattention due to drug intoxication.
I
I
I
I
40

I
I
I Possible Courses of Action

1. Check the examinee's ability to read English and to comprehend the question-

I naire items.

2. Determine if there are specific motivational issues that need to be addressed.

I 3. Determine if the examinee is depressed.

4. Determine if the examinee is psychotic.

I 5. Determine if the examinee is intoxicated.

I If conditions 1, 2, 3, 4 and/or 5 are not found to exist, the examiner can indicate to
the respondent that it is important for him/her to complete the missing items with the
best possible answer so that the examiner can have the most accurate information about
the examinee. Then, the examinee can be asked to complete the unanswered and/or
I double-answered questionnaire items or, if a large number of items are blank, the
examinee can be asked to retake the complete inventory.

I
I
I
I
I
I
41
I
 I
VALIDITY SCALES I
Lie Scale (L scale)

No interpretations ofthe lie scale scores alone are normally made. The lie scale is
I
used as part ofthe faking-good index (see the interpretation suggestions for the faking-
good index). I
Random Response Scale (RR scale)
I
No interpretations of random response scale scores alone are normally made. The
random response scale is used as part of the faking-bad index and the random response
index (see the interpretation suggestions for the faking-bad index and the random
response index).
I
Inconsistency Scale (Ie scale)
I
No interpretations of inconsistency scale scores alone are normally made. The
Inconsistency scale is used as part of the random response index (see the interpretation I
suggestions for the random response index).

I
I
I
I
I
I
I
42

I
I
I RESPONSE DISTORTION INDEXES (Validity Indexes)

I FAKING·GOOD INDEX
(Elevated Lie and Normal RR scale)

I Interpretation Hypotheses for an Elevated Faking-Good Index

An elevated faking-good index may indicate one or some combination of the

following examinee conditions:
I 1. (Primary) An attempt to give socially desirable responses in order to hide
negative personal characteristics. The goal is to present oneself in a favorable

I light, to appear adequate, and/or to create a positive image.

2. (Primary) A denial ofeven minor faults because ofan excessive concern with the
consequences of revealing negative attitudes and feelings.
I 3. (Primary) Marked social naivete and/or a lack of psychological sophistication.
This may be due to limited intellectual abilities, a lack ofeducation and/or social
I deprivation.

4. Excessive subscription to conventional societal norms and values, e.g., a moral-

I istic attitude.

5. Excessively rigid and controlled personality.

I 6. Unrealistically high evaluation of self-worth.

7. Excessive use of selected defense mechanisms, such as suppression, repression,

I denial, rationalization, and lack of insight into own behavior.

8. Antisocial characteristics in an individual with limited intellectual abilities.

I Caveat: Research indicates that in the general population as sociopathic tenden-

cies increase, Lie scale scores decrease.

I
I
I
43

I
 I
Low base rates (i.e., low frequency of occurrence) may be a problem with this index
giving rise to an excessive number offalse positive classifications offaking good behavior.
I
Under conditions where the examinee has reason to give socially approved responses
(such as in an investigation of a physical child abuse report), base rate problems are less
of a concern. However, under conditions where the examinee has no reason to give
socially desirable responses, low base rates may produce an excessive number of false
I
positive classifications. In either case, but especially the latter situation, the examiner
should have additional data to support a faking-good interpretation based on the
elevation of this index.
I
Possible Courses of Action I
If the faking-good index is elevated, the Child Abuse Potential Inventory scales can
be scored and then either of two courses of action regarding interpretations can be
followed. I
1. If the abuse and factor scores are in the normal range, all results (except for the
faking-good index scale) should be ignored due to the possibility that the low
scale scores were obtained because of a successful attempt to fake good on the
I
Child Abuse Potential Inventory.

2. If the abuse score is elevated, the examiner can interpret the elevation despite
I
the evidence of the examinee's attempt to fake good. While the exact abuse scale
score may not represent the actual magnitude of the examinee's abuse potential
because ofthe effects offaking good, the fact that the abuse scale score still shows
an elevation despite attempts to fake good has clinical meaning. Studies indicate
I
that under these conditions abuse scale score elevations can be interpreted as an
indication of abuse potential for classification purposes. I
Intervention and Treatment Related Issues

1. Clients with a high faking-good index are less likely than most clients to admit
I
and to discuss personal inadequacies.

2. It may be difficult to establish rapport with clients with a high faking-good index. I
It also may be difficult to get these clients into and to keep them in treatment.
If these clients do remain in treatment, they may be resistant to intervention.
I
I
I
44

I
I
I Other Issues

1. The examiner should consider the possibility that all of the interview data and

I other evaluation data may be distorted because of an attempt to give socially

desirable responses.

I 2. Faking-good behavior may be a special concern during program evaluation

attempts. During pre-treatment evaluations, the client may have a special need
to give socially desirable responses or to presentoneselfin a favorable light, to
appear adequate, and/or to create a positive image because they are entering a
I new situation and feel threatened.

I
I
I
I
I
I
I
I
I
I
45
I
 I
FAIUNG-BAD INDEX
(Elevated RR scale and Normal Ie scale) I
Interpretation Hypotheses for an Elevated Faking-Bad Index

An elevated faking-bad index may indicate one or some combination ofthe following
I
conditions:

1. (Primary) A desire to indicate numerous personal and interpersonal problems.

I
The goal is to present oneself in an undesirable light and to create a negative
image.
I
2. (Primary) A desire to subscribe to a wide variety of problems in order to gain
sympathy and/or assistance. A plea for help. Sometimes associated with the ter-
mination of treatment, especially if the client desires to continue. That is, the
client may fake bad in an attempt to gain additional assistance.
I
3. An exaggeration of existing problems in order to gain attention, such as found
in the histrionic personality.
I
4. Oversubscription to problems because of extreme guilt, depression, poor self-
concept, and/or unrealistic expectations and evaluation of self. I
5. The existence of an excessive number of perceived problems due to an adjust-
ment reaction to extreme stress. I
6. Extensive personal dysfunction due to a psychotic state.

Low base rates are frequently a problem for this index and produce an excessive
I
number of false-positive classifications. The test user should have additional data to
support the elevation ofthis index and/or should only use the index in situations where
base rates might be expected to be higher than in the general population, such as in
treatment situations.
I
I
I
I
I
46
I
I
I Possible Courses of Action

1. Determine if the examinee is subscribing to negative attitudes/behaviors and is

I indicating problems in order to gain attention and/or to gain assistance.

2. Determine if the examinee is subscribing to negative attitudes/behaviors and is

indicating problems as a consequence of guilt, depression, low self-esteem, and/
I or unrealistic expectations and evaluation of self.

3. Determine ifthe examinee is subscribing to negative attitudes/behaviors and is

I indicating problems due to unusual and/or extreme life stress.

I Intervention and Treatment Related Issues

1. Clients with an elevated faking-bad index are more likely than most clients to
admit and to exaggerate personal problems and to discuss personal inadequa-
I cies.

2. While a client with an elevated faking-bad index is likely to accept treatment, it

I may be difficult to determine which problems and/or personal inadequacies

should be the focus of treatment.

I Other Issues

1. Ifthe faking-bad index is elevated, the examiner should consider the possibility
I that all of the interview data and other evaluation data may contain nonexist-
ent and/or exaggerated problems due to the examinee's tendency to subscribe to
any problems about which the examinee is asked.

I 2. Faking-bad behavior may be a special concern during post-treatment evaluation,

when clients may have a special need to fake bad in order to achieve additional
assistance.
I
I
I
I 47

I
 I
RANDOM RESPONSE INDEX
(Elevated RR scale and Elevated IC scale) I
Interpretation Hypotheses for an Elevated Random Response Index

An elevated random response index may indicate one or some combination ofthe
I
following conditions:

1. (Primary) A lack of reading ability.

I
2. (Primary) An inability to understand item content and meaning. This inability
to comprehend items may be related to limited reading ability and/or to limited
intellectual abilities.
I
3. (Primary) A distortion of the test results through the use of some type of atypical
response set, such as a random response pattern, a partial random response
I
pattern, or a planned pattern of agree/disagree responses (e.g., five agree, and
five disagree responses, alternate agree/disagree responses, all agree or all
disagree responses). The essential element is that the elevated random response
I
index indicates that the examinee is not responding to item content.

4. Inability or unwillingness to handle the forced-choice item response format ofthe

questionnaire.
I
5. A lack of motivation or intentional lack of cooperation in responding to the item
content.
I
6. A defensive response set associated with an unwillingness and/or inability to
reveal personal attitudes and feelings. I
7. Cognitive distortions and confusion caused by a variety offactors such as fatigue,
distractions including poor testing conditions, depression, psychotic ideation,
and drug intoxication.
I
Low base rates are not a major problem in the case of the random response index
because two separate scales must both be elevated for the index to indicate a random
I
response pattern. Thus, the probability of a chance elevation of the index is the chance
elevation of one scale times the other (e.g., .05 x .05 = .0025).
I
I
I
48

I
I
I Possible Courses of Action

1. Check examinee's ability to read English and to comprehend the items.

I 2. Determine if the examinee is suffering from conditions that could produce

cognitive distortions and confusion such as fatigue, poor testing conditions,

I depression, psychotic ideation, and intoxication.

3. Ask examinee about his/her willingness to respond to personal items, such as

those on the questionnaire.
I 4. If a reasonable explanation (e.g., fatigue or poor testing conditions) for the
random response pattern can be found, the examiner can indicate to the

I respondent that it is important to respond carefully to the questionnaire items

so that the examiner can have the most accurate information about the exam-
inee. Then, if the examinee is willing, another questionnaire can be adminis-
tered.
I
Intervention and Treatment Related Issues
I 1. Clients with an elevated random response index who can read are likely to have
a noncooperative and/or defensive style.

I 2. Clients with an elevated random response index who can read are less likely than
most clients to accept treatment. Ifthese clients do accept treatment, they may
be difficult to keep in treatment.
I
Other Issues

I The examiner should consider the possibility that all of the interview data and
other evaluation data may be distorted because of the examinee's tendency to respond
randomly.
I
I
I
I
49

I
 I
ABUSE SCALE SCORE I
Interpretations of Elevated Abuse Scale Scores

An elevated abuse scale score indicates that the examinee has an array of personal
and interpersonal characteristics that are similar to characteristics of known physical
I
child abusers (see descriptive section below entitled "What does the abuse scale
measure?"). Given these characteristics, an examinee's elevated abuse scale score
indicates an increased risk of physical child abuse.
I
Data indicate that as abuse scores increase so does the likelihood of physical child
abuse. Thus, high abuse scale scores should be viewed as very meaningfuL I
The abuse scale has fewer false positive than false negative classifications. This
means that the abuse scale less often will indicate that a nonabusive caretaker is
abusive. More often it will incorrectly classify an abuser as nonabusive. While this is true
I
for both the 166 and 215 abuse scale cutoffs, the lower abuse scale cutoff score will have
relatively fewer false negative classifications (but more false positive classifications),
while the higher abuse scale cutoff score will produce relatively fewer false positive
I
classifications (but more false negative classifications). As previously discussed, the
cutoff score selected will be determined by the consequences of the test use. While two
abuse scale cutoff scores are provided, extensive norm data as a function of education,
gender, and ethnic background, as well as comparison data on over 5,000 maltreating
I
and comparison subjects, are available in the technical manual (Milner, 1986).

Research shows that elevated abuse scale scores are predictive of present and
I
future physical child abuse. The abuse score, however, best predicts the current status
of the individuaL Thus, the greater the intervening time since an abuse incident, the
greater the likelihood that intervening events (e.g. reduced stress) may reduce abuse I
scale scores. Likewise, an abuse scale score may be in the normal range and may increase
in the future iflife events change (e.g. return of a child to the home). When the examinee's
life situation has changed and/or several months have passed since the last evaluation,
readministration of the Child Abuse Potential Inventory is recommended to obtain the
I
most current and accurate estimate of abuse potentiaL
I
I
I
I
50

I
I
I What Does the Abuse Scale Measure?

Collectively, the data presented in the technical manual provide a descriptive

I picture ofindividuals who obtain elevated Child Abuse Potential Inventory abuse scores.
All of the following descriptive statements, taken from the technical manual, are based
on research findings.

I Those with elevated abuse scores are more likely to report a history of childhood
abuse; and, as abuse scores increase, the reported childhood abuse tends to be more
chronic. These individuals also are less likely to report the existence of a caring adult or
I caring friend in their childhood environment. It is likely that individuals with elevated
abuse scores had parents who lacked warmth, were rejecting and hostile, and/or were
inconsistent and unpredictable in their parenting behaviors.

I Individuals with elevated abuse scores tend to have low self-esteem and poor ego
development. They have a poor self-image and feel little personal worth. They feel

I inferior to others, feel guilty, believe that personal pain and misery are beneficial, and
believe in punishment for wrongdoing. They tend to be withdrawn and perceive
themselves as socially isolated.

I Examinees with elevated abuse scores report less social support, less life satisfac-
tion, more loneliness, and more life stress. These individuals report fewer positive life ex-
periences relative to the number of negative life experiences. They tend to have more

I somatic complaints, report more physical illness, and are more likely to have a history
of emotional problems.

People with elevated abuse scores tend to be immature, moody, restless, self
I centered, evasive of responsibility, lonely, and frustrated. In general, they exhibit more
worry, pessimism, and depression. They are more anxious and report more uncomfort-
able emotional feelings.

I Those with elevated abuse scores tend to be touchy, overreactive, and irritable.
They have low frustration tolerance, poor impulse control, and they become emotional

I when frustrated. They are more likely to be sensitive to anxiety-evoking or threatening

stimuli. Not only do these people tend to be more labile, but they are slow to return to a
normal state after emotional arousal. They are, however, more reactive to parent-child
related stressors than non parent-child related stressors. Those with elevated abuse
I scores show more evidence of confused thinking patterns and unique perceptions of the
world.

I
I
51
I
 I
Caretakers with elevated abuse scores have more difficulties in interpersonal
relationships including family discord, and they manifest higher levels of parent-child
I
interactional stress. Those with elevated abuse scores are more physiologically reactive
to children, perceive more behavior problems in their children, perceive less child
compliance, and report the use of more physical methods of discipline. They are less
available to their children, are less responsive to temporal changes in their child's
I
behavior, and form less adequate child attachments.

Those with elevated abuse scores are unhappy with the extent to which they meet
I
life's demands and obtain what they want. These individuals are less competent and
have less self-control. They show less adaptability and personal resourcefulness, and feel
distress because oftheir inability to handle situations. Those with elevated Child Abuse
I
Potential abuse scores tend to have poorly developed cognitive mastery and coping skills.
While they exhibit few appropriate assertiveness skills, those with high abuse scores are
more aggressive, more assaultive, more prone to temper outbursts, and are more "likely
to fight." Taken together, these personal and interactional characteristics, which are
I
related to elevated Child Abuse Potential abuse scores, appear to set the foundation for
parental perceptions and parenting styles that result in acts of physical child abuse. I
I
I
I
I
I
I
I
I
52

I
I
I Low Base Rates

When the abuse scale is used in low base rate situations, the amount ofincremental

I validity provided by the scale decreases. In situations where the population base rates
are very low, the use of the abuse scale may be inappropriate for certain applications.
Incremental validity is greatest when population base rates approach fifty percent.

I Possible Courses of Action

I The higher the obtained abuse score, the more important it is for the
examiner to follow-up with additional evaluations, including interviewing, behav-
ioral observations, collateral interviews, medical tests, other testing, etc. Elevated abuse

I scores indicate an increased likelihood that the client is in need of secondary or tertiary
prevention efforts.

I Other Issues

The abuse score most accurately represents the characteristics ofthe parent at the

I time of testing. While an elevated abuse score does predict the likelihood of future
physical child abuse, retesting is recommended after the passage of time. Retesting is
especially recommended when the living and/or family conditions have changed. For ex-

I ample, if stress levels increase and/or ifa child is returned to a parent, these events may
affect the potential for child abuse and retesting is necessary.

I
I
I
I
I
I
53

I
 I
ABUSE SCALE FACTOR SCORES I
DISTRESS FACTOR SCALE

Interpretations of Elevated Distress Scale Scores

I
An elevated distress scale score suggests the examinee has a high degree of
personal distress and personal adjustment problems. In aggregate, the distress dimen-
sion reveals a pattern of feeling frustrated, sad, lonely, depressed, worried, afraid, out-
I
of control, confused, mixed-up, upset, worthless, rejected, misunderstood, and angry.
High scores indicate a poor reaction to perceived stress. Research also indicates high
distress scores are related to high levels ofanxiety, depression, and loss ofemotional and
I
behavioral control.
I
RIGIDITY FACTOR SCALE

Interpretations of Elevated Rigidity Scale Scores I

An elevated rigidity scale score suggests the examinee has rigid attitudes towards
the appearance and behavior ofchildren. Specifically, an elevated rigidity score suggests
beliefs that children should be neat, be orderly in their behavior, be obedient, never cause
I
trouble, never disobey, stay clean, be seen and not heard, be quiet and attentive, and not
talk back. An associated beliefis that children need strict rules. A second theme indicated
by elevated rigidity scores involves beliefs that a home should be spotless and everything
should always be in its place. As beliefs relate to behavior, they may be expressed through
I
the forceful effort to make children fit a rigid mold defined by the respondent. Research
has shown that elevated rigidity scores are related to authoritarian styles. I
UNHAPPINESS FACTOR SCALE

Interpretations of Elevated Unhappiness Scale Scores

I
An elevated unhappiness scale score suggests the examinee has a general
unhappiness with life. Unhappiness may be related to difficulties in relationships such
as not being in love, lacking a good sex life, and not having close friends. The content of
I
the unhappiness factor scale also includes not having a good or happy life, not being a
happy person, rarely laughing, feeling inferior to others, and feeling unlucky. Research
indicates that elevated unhappiness factor scores are related to a negative outlook on life
I
and the lack of emotional ties.
I
I
54

I
I
I PROBLEMS WITH CHILD AND SELF FACTOR SCALE

Interpretations of Elevated Problems with Child and Self Scale Scores

I An elevated problems with child and self scale score suggests the examinee tends
to describe children in a negative manner. An elevation in this scale indicates examinee
perceptions of having a child who is slow, has special problems, gets into trouble often,
I and/or is bad. Elevated scores also may indicate that the examinee has physical
problems such as having a physical handicap or not always being strong and healthy.
Overall, elevations in this factor scale suggest the examinee has a perception that his/
I her childCren) have limited ability and/or competency, and that the examinee also has
limited physical ability.

I PROBLEMS WITH FAMILY FACTOR SCALE

Interpretations of Elevated Problems with Family Scale Scores

I An elevated problems with family scale score suggests that the examinee has
numerous difficulties in familial relationships, such as the family having problems

I getting along and the family having a lot of fights. Elevations in this scale suggest a
highly conflictual and possibly yiolent family.

I PROBLEMS FROM OTHERS FACTOR SCALE

Interpretations of Elevated Problems from Others Scale Scores

I An elevated problems from others scale score suggests the examinee has general
difficulties in social relationships both within and external to family members. Overall,

I elevations in this scale indicate that social relationships are perceived as a cause ofthe
examinee's personal difficulties, unhappiness, and pain. Other people are not seen as
dependable. Thus, relationships are viewed negatively rather than as a resource.

I
I
I
I
55

I
 I
SPECIAL SCALES I
EGO-STRENGTH SCALE

Interpretations of Elevated Ego-Strength Scale Scores I

The Child Abuse Potential Inventory ego-strength scale measures dimensions of
the individual that represent emotional stability. An elevated ego-strength scale score
suggests that the examinee lacks the ability to maintain emotional stability and feelings
I
of worth in social relationships. With respect to relationships, examinees with elevated
ego-strength scores are less likely to be talkative, cheerful, expressive, alert, active,
genial, and friendly. They are more likely to feel depressed, to be afraid, and to be upset I
without knowing why.

LONELINESS SCALE
I
Interpretations of Elevated Loneliness Scale Scores I
The Child Abuse Potential Inventory loneliness scale tends to measure the degree
of perceived loneliness more than the objective degree of social isolation. An elevated
loneliness scale score indicates the examinee has perceptions ofbeing isolated and alone.
He/she feels that others are not available; and, even when other individuals are available
I
as resources, they tend not to use these people effectively.
I
I
I
I
I
I
I
56

I
ERRATA

On page 56, the paragraph describing the EGO-STRENGTH SCALE should read as
follows:

Interpretations of Elevated Ego-Strength Scale Scores

The Child Abuse Potential Inventory ego-strength scale measures dimensions of

the individual that represent emotional stability. An elevated ego-strength scale score
suggests that the examinee hM the ability to maintain emotional stability and feelings
of worth in social relationships. With respect to relationships, examinees with elevated
ego-strength scores are more likely to be talkative, cheerful, expressive, alert. active,
genial, and friendly. They are ~ likely to feel depressed, to be afraid, and to be upset, .
without knowing why.
I
I REFERENCES

I Milner, J. S. (1986). The Child Abuse Potential Inventory: Manual (2nd ed.). Webster,
NC: Psytec Corporation.

Milner, J. S. (1986). Assessing child maltreatment: The role of testing. Journal of

I Sociology and Social Welfare, 13,64-76.

Milner, J. S. (1989). Applications and limitations of the Child Abuse Potential

I Inventory. In J. T. Pardeck (Ed.), Child abuse and neglect: Theory. research and
practice. New York: Gordon and Breach Science Publishers.

I
I
I
I
I
I
I
I
I
I
57

I
 I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
58

I
I
I APPENDIX A
I ITEM COMPOSITION OF THE CAP INVENTORY (FORM VI)
SCALES
I Lie scale: 18 items
Items: 12
I 110
34
146
35
149
44
150
46
155
57
157
62
159
66
160
70 106

I RRscale: 18 items
Items: 1 11 16 27 31 33 43 53 58 59
60 61 65 72 89 114 116 119

I Ie scale: 20 item-pairs
Item-pairs: 3-76 4-6 5-9 38-41 44-70
I 52-63
83-94
58-72
85-158
62-65
87-141
75-118
90-152
78-98
95-107
100-151 105-120 122-127 124-133 143-145
I
Abuse scale: 77 items

I Items: 3
23
41
5
24
45
7
25
47
9
26
49
13
28
52
14
29
54
17
32
56
18
36
63
19
38
67
22
39
68
69 73 74 75 76 77 78 80 81 83
I 84
105
90
107
93
108
94
109
95
111
98
112
99
113
100
115
102
118
103
120
122 127 128 129 130 132 134 138 141 143
I 145 147 148 151 152 153 154

Distress scale: 36 items

I Items: 5
41
9
47
17
49
18
52
22
56
23
63
25
73
28
78
29
84
36
93
95 98 99 102 103 105 109 111 112 118
I 120 138 143 145 153 154

I
59
I
 I
Rigidity scale: 14 items
Items: 7
122
9 24 26 32 54 68 80 108 115
I
127 130 132

Unhappiness scale: 11 items

I
Items: 14 38 75 77 81 90 107
152
134 141 147
I
Problems with child and self scale: 6 items
Items: 3 45 69 76 113 128 I
Problems with family scale: 4 items
Items: 39 83 94 148
I
Problems from others scale: 6 items
I
Items: 13 67 74 100 129 151

Ego-strength scale: 40 items

I
Items: 14 17
49
93
52
95
18
56
98
20
63
99
22
65
102
23
67
103
25
73
105
28
74
107
41
78
109
47
84
112
I
118 120 136 138 140 143 145 151 153 154
I
Loneliness scale: 15 items
Items: 6
87
18
98
22
141
23
143
47
145
49 67 73 74 78
I
I
I
I
I
60

I
I
I APPENDIXB
I SELECTED READINGS

I Atten, D. w., & Milner, J. S. (1987). Child abuse potential and work satisfaction in day
care employees. Child Abuse & Neglect, 11, 117-123.

I Ayoub, C., Jacewitz, M. M., Gold, R. G., & Milner, J. S. (1983). Assessment ofa program's
effectiveness in selecting individuals "At Risk" for problems in parenting. Journal of
Clinical Psychology, aft, 334-339.

I Ayoub, C., & Milner, J. S. (1985). Failure-to-thrive: Parental indicators, types, and
outcomes. Child Abuse & Neglect, ft, 491-499.

I Haddock, M. D., & McQueen, W. M. (1983). Assessing employee potentials for abuse.
Journal of Clinical Psychology, aft, 1021-1029.

I Kirkham, M. A., Schinke, S. P., Schilling, R. F., Meltzer, N. J., & Norelius, K L. (1986).
Cognitive-behavioral skills, social supports, and child abuse potential among mothers
of handicapped children. Journal of Family Violence, 1, 235-245.

I Melton, G.B., & Limber, S. (1989). Psychologists' involvement in cases of child

maltreatment: Limits of role and expertise. American Psychologist, 44, 1225-1233.

I Milner, J. S. (1982). Development of a lie scale for the Child Abuse Potential Inventory.
Psychological Reports, 50, 871-874.

I Milner, J. S. (1986). Assessing child maltreatment: The role of testing. Journal of

Sociology and Social Welfare, 13, 64-76.

I Milner, J. S. (1986). The Child Abuse Potential Inventory: Manual (2nd ed.). Webster,
NC: Psytec Corporation.

Milner, J. S. (1988). An ego-strength scale for the Child Abuse Potential Inventory.
I Journal of Family Violence, a, 151-162.

Milner, J. S. (1989). Applications of the Child Abuse Potential Inventory. Journal of

I Clinical Psychology, 45, 450-454.

I
61

I
 I
Milner, J. S. (1989). Applications and limitations ofthe Child Abuse Potential Inven-
tory. In J. T. Pardeck (Ed.), Child abuse and neglect: Theory, research and practice.
I
New York: Gordon and Breach Science Publishers. Also reprinted in Early Childhood
Development and Care (1989), 42, 85-97.

Milner, J. S. (1989). Additional cross-validation ofthe Child Abuse Potential Inventory.

I
Psychological Assessment: A Journal of Consulting and Clinical Psychology, 1, 219-
223. I
Milner, J. S. (in press). Additional issues in child abuse assessment. American
Psychologist. (Research comment.)
I
Milner, J. S., & Ayoub, C. (1980). Evaluation of"At Risk" parents using the Child Abuse
Potential Inventory. Journal of Clinical Psychology, 36, 945-948.

Milner, J. S., Charlesworth, J. R, Gold, R G., Gold, S. R, & Freisen, M. R (1988).

I
Convergent validity of the Child Abuse Potential Inventory. Journal of Clinical
Psychology, 44, 281-285. I
Milner, J. S., & Gold, R G. (1986). Screening spouse abusers for child abuse potential.
Journal of Clinical Psychology, 42, 169-172.
I
Milner, J. S., Gold, R G., Ayoub, C., & Jacewitz, M. M. (1984). Predictive validity ofthe
Child Abuse Potential Inventory. Journal of Consulting and Clinical Psychology, 52,
879-884. I
Milner, J. S., Gold, R G., & Wimberley, R C. (1986). Prediction and explanation ofchild
abuse: Cross-validation ofthe Child Abuse Potential Inventory. Journal ofConsulting
and Clinical Psychology, 54, 865-866.
I
Milner, J. S., & Robertson, K. R (1985). Development ofa random response scale for the
Child Abuse Potential Inventory. Journal of Clinical Psychology, 4.1, 639-643. I
Milner, J. S., & Robertson, K. R (1989). Inconsistent response patterns and the
prediction of child maltreatment. Child Abuse & Neglect, 13, 59-64. I
Milner, J. S., & Robertson, K. R (in press). Comparison of physical child abusers,
intrafamilial sexual child abusers, and child neglectors. Journal of Interpersonal
Violence.
I
I
I
62
I
I
I Milner, J. S., Robertson, K. R, & Rogers, D. L. (in press). Childhood history ofabuse and
adult abuse potential. Journal of Family Violence.

I Milner, J. S., & Wimberly, R C. (1979). An inventory for the identification of child
abusers. Journal of Clinical Psychology, 35, 95-100.

I Milner, J. S., & Wimberley, R C. (1980). Prediction and explanation of child abuse.
Journal of Clinical Psychology, 36, 875-884.

Ochotorena, J., M~dariaga, M., &~J¥.ilner, J.S. (in press). Validacion de ~na version
I Espanola del ChIld Abuse Potentia:! Inventory para su uso en Espana. Child Abuse &
Neglect.

I Pruitt, D. L., & Erickson, M. T. (1985). The Child Abuse Potential Inventory: A study
of concurrent validity. Journal of Clinical Psychology, 41, 104-111.

Robertson, K. R, & Milner, J. S. (1983). Construct validity ofthe Child Abuse Potential
I Inventory. Journal of Clinical Psychology, 39, 426-429.

Robertson, K. R, & Milner, J. S. (1985). Convergent and discriminant validity of the

I Child Abuse Potential Inventory. Journal of Personality Assessment, 49, 86-88.

Robertson, K. R, & Milner, J. S. (1987). An inconsistency scale for the Child Abuse

I Potential Inventory. Psychological Reports, 60, 699-703.

Robertson, W. K. T., & Milner, J. S. (1985). Detection of conscious deception using the
Child Abuse Potential Inventory Lie Scale. Journal of Personality Assessment, 19,
I 541-544.

Robitaille, J., Jones, E., Gold, R G., Robertson, K. R, & Milner, J. S. (1985). Child abuse

I potential and authoritarianism. Journal of Clinical Psychology, 41, 839-843.

Stringer, S. A, & LaGreca, AM. (1985). Correlates of child abuse potential. Journal of

I Abnormal Child Psychology, 13,217-226.

Thommasson, E., Berkovitz, T., Minor, S., CassIe, G., McCord, D., & Milner, J. S. (1981).
Evaluation of a family life education program for rural "high risk" families. Journal
I of Community Psychology,.9., 246-249.

Wolfe, D. A, Edwards, B., Manion, I., Koverola, C. (1988). Early intervention for parents

I at risk of child abuse and neglect: A preliminary investigation. Journal of Consulting

and Clinical Psychology, 56, 40-47.

I
63

I
 I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
64

I
I
I NOTES

I
I
I
I
I
I
I
I
I
I
I
I
I
65

I
 I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
66