Conners' Rating Scales-Revised (CRS-R)

Frances Mueller, Richard Brozovich, Oakland Schools

Carol Barnes Johnson, Michigan State University

Author
C. Keith Conners, Ph.D., Duke University

Publisher
Multi-Health Systems Inc., 908 Niagara Falls Boulevard, North
Tonawanda, New York 14120-2060. Phone 1-800-456-3003 or in
Canada: 65 Overlea Boulevard, Suite 210, Toronto, Ontario M4H IP1.
Phone 1-800-268-6011

Price
Complete Parent and Teacher Rating Scales Kit $152.00
with technical manual, 25 Quickscore Forms for
Parent Long Form, Parent Short Form, Teacher
Long Form, and Teacher Short Form
Technical Manual 54.00
User's Manual 46.00
25 Parent Long Quickscore Forms 31.00
25 Parent Short Quickscore Forms 31.00
25 Teacher Long Quickscore Forms 31.00
25 Teacher Short Quickscore Forms 31.00
25 Adolescent Self-Report Long Quickscore Forms 31.00
25 Adolescent Self-Report Short Quickscore Forms 31.00
25 Parent Feedback Long Forms 26.00
25 Parent Feedback Short Forms 26.00
25 Teacher Feedback Long Forms 26.00
25 Teacher Feedback Short Forms 26.00
25 Adolescent Feedback Long Forms 26.00
25 Adolescent Feedback Short Forms 26.00
25 Teacher Information Forms 14.00
Conners' Rating Scales: Computer Program 175.00
Copyright Dates and Revisions
The Conners' Rating Scales-Revised (CRS-R) were published in
DIAGNOSTIQUE 24(1-4)/83-97 (1998-1999)

83
84 DIAGNOSTIQUE VOL. 24, NO. 1-4,1998-1999

1997. Commercial publications began in 1989 and 1990. Informal and

qualitative use of the scales date back to 1960.

Groups for Whom the Instrument is Intended

The CRS,R are intended to be used for reporting problem behav-
iors in Canadian and American youth from 3,17 years of age. Self-report
scales target 12,17 year olds. Forms of the Instrument: Long and short
forms of parent, teacher and self-report scales are available in both a
paper and pencil format and a computer format.

Author's Purpose and Recommended Uses

The main use of the CRS,R is for the assessment of Attention
Deficit Hyperactivity Disorder (ADHD). The CRS,R can be used for
screening, program evaluation, treatment monitoring, research and clin-
ical information. As stated in the manual, when used as part of diagnos-
tic decision making, it should be recognized there is a risk of false posi-
tives and false negatives. To minimize error, the manual suggests using
the tool as one source of clinical information that must be combined
with other reliable sources of information. The CRS,R was not meant to
be the sole criteria for decision making, assessment, or diagnosis.

Dimensions Measured
Overall hyperactivity is assessed on the Conners' Global Index.
The author states on page 4 that the Global Index is "one of the best
short measures assessing general psychopathology...." There is an ADHD
Index score. Other subscales allow for a broader use as they contain scales
for the assessment of conduct problems, family problems, emotional
problems, anger control problems, and anxiety problems. The author jus,
tified the use of the additional scales due to frequent comorbidity of those
problems with ADHD. The symptoms are assessed based on the percep-
tions of the youth's behavior over the past month. Each item of the scales
is rated from zero to three based on how true of a statement it is of the
youth (i.e., NOT TRUE AT ALL = 0, VERYMUCH TRUE =3). Below
each rating scale is described by total number of items and by the sub,
scales and their number of items in parenthesis.

Parent Long Form (80 items) Teacher Long Form (59 items)
Oppositional (10 ) Oppositional (6)
Cognitive Problems (12) Cognitive Problems (7)
CONNERS' RATING SCALES-REVISED (CRS-R) 85
Hyperactivity (9) Hyperactivity (7)
Anxious-shy (8) Anxious-shy (6)
Perfectionism (7) Perfectionism (6)
Social Problems (5) Social Problems (5)
Psychosomatic (6)
Conners' Global Index (10) Conners' Global Index (10)
Restless-Impulsive (7) Restless-Impulsive (6)
Emotional Lability (3) Emotional Lability (4)
ADHD Index (12) ADHD Index (12)
DSM-IV Symptoms Subscale (18) DSM-IV Symptoms Subscale (18)
DSM-IV Inattentive (9) DSM-IV Inattentive (9)
DSM-IV Hyperactive-Impulsive (9) DSM-IV Hyperactive-Impulsive (9)

Parent Short Form (27 items) Teacher Short Form (28 items)
Oppositional (6) Oppositional (5)
Cognitive Problems (6) Cognitive Problems (5)
Hyperactivity (6) Hyperactivity (7)
ADHD Index (12) ADHD Index (12 )

Adolescent Self-Report Lonl: Form (87) Adolescent Short Form (27)

Conduct Problems (12) Conduct Problems (6)
Cognitive Problems (12) Cognitive Problems (6)
ADHD Index (12) Hyperactive-Impulsive (6)
Family Problems (12) ADHD Index (12)
Emotional Problems (12)
Anger Control Problems (8)
Hyperactivity (8)
DSM-IV Symptoms Subscale (18)
DSM-IV Inattentive (9)
DSM-IV Hyperactive-Impulsive (9)

Administration
Qualifications of Administrator
According to the author, the CSR,R could be administered and
scored by an "untrained" person. However, the ultimate responsibility for
interpretation should be assumed by a professional that understands the
limitations of such screening and assessment procedures. The author rec-
ommends at least a BA in psychology, counseling or a closely related
field. Also the author recommends that the person interpreting the CRS,
R have relevant training or coursework in the interpretation of psycho,
logical tests and measurements or verification of membership in a pro,
86 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

fessional association recognized to require training and experience in the

ethical and competent use of psychological tests or licensed or certified
by an agency which does the same. Postgraduate training of at least the
Master's level is expected of the practitioner to allow the proper inter-
pretation of the CRS~R.

Administration Time
The completion of the long form of the self-report, parent and
teacher rating scales is estimated at 15 to 20 minutes. The short form of
those scales takes about 5 to 10 minutes to finish.

Forms ofAdministration
Though typically administered individually, the manual does dis-
cuss group administration of the adolescent self-report, Mailing forms to
parents, teachers or adolescents for completion and return by mail was
described as "remote administration." This format was not used in the
standardization and the manual, therefore, advised this practice would
need to be interpreted cautiously. The manual provided the following
phrase to describe the situation: "the data obtained requires additional
validation because a nonstandard administration protocol was used." The
manual recommended follow-up administration using the standard pro-
tocol.

Administrative Cautions
The manual cautions that the administrator and interpreter of the
rating scales must be aware of the quality of the relationship between the
teacher and the child and the parents and the child, and in the case of
self-reports, the adolescent's level of self-awareness. Informant bias is to
be taken into consideration in interpretation. Further, when teacher and
parent results are in conflict, the manual suggests putting more weight on
the teacher score due to the assumption that the teacher has greater
familiarity with age-appropriate norms of behavior. The manual also sug-
gests conducting interviews with current and past teachers and observa-
tion of the classroom situation. Additionally, the manual recognized that
school psychologists may be in the ideal position to complete such a
comprehensive approach but that for most practitioners the ideal proce-
dure is unrealistic. The adolescent self report is offered since students in
middle school and high school change teachers and do not have the sin-
gular consistent exposure to the one teacher as do elementary students.
CONNERS' RATING SCALES-REVISED (CRS-R) 87

The manual also indicated that the self-report may be better than
teacher or parent scales at identifying internalizing psychopathology.

Summation of Data
Procedures
The paper and pencil form can be hand scored in about 10 minutes.
A mail-in scoring system requires 48 hours plus mail delivery time. The
fax-back scoring system is completed in a few minutes, and the comput-
er scoring system takes several seconds.

Types of Scores Available

Raw scores, linear T scores, and percentiles for standard distribu-
tion and for normative distribution.

Interpretation of Scores
The hand scored results can be transferred to a colored grid. Raw
scores may fall within white, pink, or red area backgrounds indicating
that a score is respectively typical, borderline, or clinically significant rel-
ative to the child's age. On the quick score form the child's score is also
compared to same sex. Score levels are grouped as follows: Average (27th
through 73rd percentile) are in the white band, Borderline or Slightly
Atypical (74th to 85th percentile) are on the edge of the pink area;
Mildly Atypical (86-94th percentiles) Moderately Atypical (95th
through 98th percentile) and Markedly Atypical (above the 98th per-
centile) gradually reach a red tone. The descriptor of "atypical" are also
applied to low percentile scores (below the 16th percentile) and are
descnibed as "good'' an d are" not a concern. " Percenn'1 es are repone d lor
c
a normalized standard theoretical distribution as well as from the norma-
tive data. The normative data is markedly skewed. Percentiles are pro-
vided for age and sex.
The manual discusses ethnic factors that may justify using the
African American norms available in the appendix. Additionally, dis-
cussion regarding the need for multiple sources of information for inter-
pretation is repeated in this section of the manual. To assist in interpre-
tation, case studies are also available in the manual. An eight-step inter-
pretation sequence is provided.
88 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

Standardization
Item Selection
Item pools for the teacher and parent scales were selected from the
previously published CRS, modified from other versions of the Conners'
and the OSM-IV criteria. The self-report scale items were developed
based on face validity for AOHO symptoms and aspects reported in the
literature as concomitants, sequelae, or comorbid features of AOHO.
Items were also selected to cover the OSM-IV symptoms. The initial set
was used informally in clinical practice for several years, presented at a
conference, and reviewed by professionals.
An expert panel review, consisting of psychologists, was conducted
for all scales. The task for the panel was to review items for theoretical
basis, clarity and clinical significance. Items were retained in the scale
based on their psychometric properties as identified by analyzing data
obtained in pre-publication research.

Sampling Procedures
A pilot study of ten sites in eight states (SO, MA, N'I, TX, CA,
CO. MI, and WA) was conducted that included assessments of children
aged 5 through 18. The pilot sample group included 170 parents, 170
teachers, and 91 adolescents. The pilot sample group was 80%
Caucasian. Age and sex effects were noted and interpretation of results
now considers those differences. In 1996 a follow up response to the pilot
item pool was achieved from parents of 2,200 students (1,099 males and
1, 101 females) aged 3 to 17 years. Random assignment of 1,100 partici-
pants into a derivation sample and 1,100 participants into a replication
sample allowed for exploratory work to be done on the former set while
confirmatory work was completed on the later.

Sample Characteristics
The author used a total of approximately 8,000 cases of long form
responses for each of the CRS scales to generate short form reliability and
validity results. The reported 8,000 normative cases was also described as
a database of over 11,000 when short form CRS scales were counted. The
data were obtained from over 200 schools in 45 states and 10 provinces
throughout the United States and Canada. In addition to completion of
the ratings, information regarding ethnicity, sex, age, socioeconomic sta-
tus, and geographic location were obtained. Only students in general
CONNERS' RATING SCALES-REVISED (CRS-R) 89

education classes were assessed; students in special education classes were

excluded. While all parent and teacher ratings were completed individ-
ually, some of the adolescent self-reports were completed through group
administration.
The parent rating scale was normed on 2,482 children between the
ages of 3 and 17. The completers of the scale were identified as 83%
Caucasian, 4.8% African American, 3.5% Hispanic, 2.2% Asian
American, 1.1% Native American, and 4.9% as Other or omitted the
ethnic background question. The median annual income was between
$40,001 and $50,000. The majority of the parent short form normative
sample was derived from rescoring the parent long form with similar eth-
nic groupings. The short form sample consisted of 1,973 cases.
The teacher rating scale was normed on 1, 973 students between
the ages of 3 to 17. The students were identified as 78% Caucasian,
10.2% African American, 5.8% Hispanic, 1.6% as Native American and
2.8% as Other or no information about ethnic membership. Again, the
majority of the cases for the normative sample for the teacher short form
was obtained from rescoring the teacher long form. The 1,897 cases were
similar to the long form ethnic distribution.
The adolescent rating scale consisted of 3,394 adolescents (1,558
males and 1,836 females) between 12 and 17 years of age with 62% rat-
ing themselves as Caucasian, 29.9 % African American, 2.3% Hispanic,
1.6% Asian, 1.3% Native American and 3.1% Other or left that ques-
tion blank. Similarly, the majority of the short form norms of the adoles-
cent rating scale were developed from a rescored version of the long
form. Using 3,486 cases, the short form sample group had similar ethnic
composition to the long form sample.
The number of forms that were not completed, incomplete or com-
pleted but rejected due to inconsistency of responses was not reported.
Reliability
Test-Retest
On page four of the manual, 6 to 8 week test/retest reliability coef-
ficients ranging from about .60 to .90 were reported. On page 116 of the
manual, the test-retest reliabilities were described as "In general, ...mod-
erate to high across the various forms." Using only 49 cases for the long
forms of the parent scale and 50 cases for the teacher and adolescent self
report long form scales (and using the same cases for the short form reli-
ability computations), the manual provided tables and minimal interpre-
90 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

tation. On the CPRS,R:L, four scales were below .6 and on the teacher
long form two scales were .47. The manual summarized this section with
the sentence "...practitioners may use the CRS,R with confidence with
3, to 17,year,0Id children."

Internal Consistency Reliability

In one manual section (page 4), a range of around .75 to .90 inter'
nal consistency coefficients was reported. Internal consistency coeffi-
cients using Cronbach's coefficient alpha were computed across all forms
for separate sex and age groups. In another manual section (pages 110,
112), internal consistency coefficients were reported that were both
below and above the range reported on page four. The Parent Rating
Scale-Long Form had total reliability coefficients ranging from. 73 to .94
and the short form yielded from .85 to .94. The Teacher Rating Scale'
Long Form generated coefficients from .77 to .96 and the short form
ranged from .88 to .95. The Adolescent Self Report-Long Form produced
coefficients ranging from .75 to .92 and the short form coefficients
ranged from .75 to .85. Also, the internal consistency of the Adolescent
Self-Report DSM,IV Symptoms subseale was computed on only a small
sample of 75 cases. The number of cases used for the other scales was not
reported.

Standard Errors of Measurement

Standard error of measurement of observed scores and the standard
error of measurement of prediction was computed using Cronbach's
alpha. The manual provides multiple tables with minimal interpretation.
It appears that the Parent Rating Scale may be weak in standard error of
measurement of prediction on the male DSM Total scale (SEM ranging
from 3.338 to 3.554). The standard error of measurement of prediction
also appeared weak throughout the Adolescent Self Report'Long Form.
The total number of cases used for SEM computations was not reported.

Validity
According to the author (1997) "the CRS,R demonstrates suffl-
dent construct validity to warrant publication and recommendation for
clinical and research use."

Construct
One aspect of construct validity addressed in the manual was facto'
CONNERS' RATING SCALES-REVISED (CRS-R) 91

rial validity. The intercorrelations of the factors of each of the CRS-R

was presented in tables. The intercorrelations ranged from -.02 to .97.

Content
The Parent, Teacher and Adolescent Scales were compared to each
other. The correlations between CPRS-R:L and CTRS-R:L were weak to
moderate ranging from .12 to .47 for males and .21 to .55 for females. As
the manual stated on page 125, "Consistent with previous reports that
parents and teachers perceive the same children and adolescents differ-
ently.....there was considerable variability in the correlations among par-
ent and teacher subscales." The short form comparison had a similar pat-
tern.
The comparison of the CPRS-R:L to the CASS:L yielded similar
findings. "Consistent with reports that parents and children can perceive
problem behavior quite differently....low or nonsignificant correlations
were found between most of the self-report and parent ratings" (page
127). A similar result occurred in the short form comparison.
The comparison of the CTRS-R:L to the CASS:L found low to
nonsignificant correlations which the manual stated .is "consistent with
reports that teachers and children have quite different perceptions about
problem behavior..." The pattern of correlations on the short form was
nearly identical to the long form comparison.

Criterion Related
The CPRS-R:L was compared to the Children's Depression
Inventory (COl), a self-report scale for children 7 to 17 years. The find-
ing indicated moderate to high correlation s between the two measures.
The CTRS-R:L and the CASS:L yielded similarly strong correlations to
the COL
The Continuous Performance Test, developed by the same author
as the CRS-R, was compared to the CRS-R. The comparison produced a
modest correlation on the DSM-IV Symptoms Inattentivesubscale and
was not significantly related to any other dimensions of the CRS-R.

Summary & Conclusions

This section groups the reviewers' concerns according to the relat-
ed area in the above review of the CRS-R.
92 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

Dimensions Measured
The psychosomatic scale was not included on the teacher rating
scale. It is included on the parent rating scale only. If the psychosomatic
scale were included on the teacher scale there would be a more direct
comparison of rating scales. It also appears, pragmatically, that teachers
are good sources for confirming psychosomatic behaviors of students.
"Despite maintaining relatively few items, the CRS~R contains
numerous subscales and offer widespread coverage...."(p. 7). Due to the
large amount of over-lapping items across factors and the resulting high
correlations between factors, there does not appear to be evidence to
support the large number of distinct factors in the CRS-R.
Descriptive statements are rated on a continuum of how true they
are thought to be of an individual (i.e., not very true to very true). The
perception is without any qualification by frequency (Anderson, 1996).
The statement from the Self-Report Scale "I get nervous" is true for
many people. A distinguishing characteristic is not whether it is a little
true or very true but how often or frequent is a person feeling or acting
nervous.

Administration
Readability levels for the parent and teacher long forms is ninth
grade, teacher short form is ninth grade and parent short form is 10th
grade. Both adolescent forms are at the sixth grade reading level. The
reviewers question if the readability level for both parents forms take into
consideration the variety of educational experiences represented by the
parent group. The reviewers question if the readability level is low
enough for the broad range of reading abilities of parents.
Respondents are expected to comment on the youth's behavior and
actions over the past month which can be problematic over the summer
and in January.
Cautions were made in the manual about group administration of
the self-report by adolescents and "remote" administration (i.e., admin-
istration by mail). It remains unknown what a different administrative
format has on a response. Reliability and validity may be affected in
unknown ways. Additional data on the influence of this very common
type of administration is needed. The recommended follow up with the
standard, face-eo-face administration, may not be reasonable or neces-
sary.
CONNERS' RATING SCALES·REVISED (CRS-R) 93

Summation of Data
The "atypical" descriptor can be used for low scores however, the
author states that low scores are "good" and "do not present a concern."
Atypical may be better stated as "rare" and interpreted as not indicative
of family problems, emotional problems or whatever factor is presenting
a low score. The rare and atypically low score represents positive behav-
ioral adaptation and does not warrant concern.
Without a research reference, when low commonality is noted
between teacher and parent rating scale results, the author advises
putting more weight on the teacher rating. Support for this is recom-
mendation is needed. Perhaps, behavior functions differently in different
environments or perceptions of the same behavior differ across environ-
ments. Teacher characteristics are a considerable source of error variance
in ADHD rating scales (Danforth & DuPaul, 1990).
At the top of the page of the profile of scores a box is identified as
the "DSM-IV Symptom Scales." If six or more of the Inattentive
Symptoms are indicated or if six or more of the Hyperactive-Impulsive
Symptoms are marked, the scores fall in a red zone. This red zone or "hot
zone" on the Profile Form purports to "suggest possible DSM-IV diagno-
sis." This is very misleading. The use of scores in such a manner without
other confirming evidence is not a practice recommended in the CRS-R
manual, nor by Conners, Sitarenios, Parker, & Epstein (1998). It is not
recommended by the reviewers.

Standardization
Item wording may be an issue. For example, the self-report asks
teenagers to respond to "My parents' discipline is too harsh." Is the stu-
dent's perception of that issue relevant to a scale that purports to assess
ADHD?
Some items have a negative orientation and may appear vague.
Examples of negative orientation include: "fussy about cleanliness," and
"People bug me and get me angry." Examples of vague items include:
"cannot grasp arithmetic" or "has sloppy penmanship."
Exclusion of special education classes (but perhaps not special edu-
cation students) is an issue. Despite careful reading of the manual and
published studies (Epstein, March, Conners, & Jackson, 1998; Conners,
Sitarenios, Parker, & Epstein, 1998a; and Conners, Sitarenios, Parker, &
Epstein, 1998b), it is unclear if all special education students were
excluded from the norm group. This suggests that about 10 to 12% of the
94 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

student population was deliberately excluded. In that 10 to 12% popula-

tion of special education students, it is not unlikely for students to have
the very behavior problems the CRS~R addresses (Danforth & DuPaul,
1998). The pool of subjects may not be representative of the school pop~
ulation.
With a median annual household income for participating parents
between $40,001 and $50,000, this relatively affluent, majority
Caucasian group of parents is not representative of the U.S. population.
Additionally, while 8.3% of parents describe themselves as Hispanic or
African American, 16% of the teacher ratings were completed on stu-
dents described as Hispanic or African American. It appears that a large
portion of African American parents did not choose to participate. Their
absence is not explained. If this reflects a systematic low response rate
among African American parents, then there is a strong bias in the sam-
ple with unknown effects.The percent of race/ethnicity in the normative
sample relative to the percent of USA and Canada populations was not
mentioned in the manual.
If a child is at an extreme of an age range (youngest of the 12 to 14
year olds), the author advises comparing the age-based results of the 9 to
11 year old group and the 12 to 14 year old and predict that the child's
score would fall somewhere in between (p. 44). That recommendation
would suggest a need for additional age group norms.
Table 7.7 is titled Means and Standard Deviations on the CPRS~
R:L, but the standard deviations are not noted. Further, the means sug-
gest a different interpretation than reported elsewhere. For instance,
males were to have scored significantly higher than females on the
Oppositional scale. Of the six ethnic groups, only two groups, Caucasian
and Native American, had males scoring higher on average than females.

Reliability
The short versions of the CRS~R were based on data developed
from administration of CRS~R long forms. There is no data to describe
the effect of the time element in the administration of the longer version
or shorter version of a CRS~R. Due to the relatively minor differences in
time for completing a long vs. a short form of the CRS~R, the reviewers
suggest using the long version only.
It is not clear why fewer cases were used to develop the reliability
data for the short form versions of the CRS~R. The basis to adding or
excluding cases for the short form reliability determination was not clear
CONNERS' RATING SCALES-REVISED (CRS-Rl 95

to the reviewers. The reviewers suggest using every long version case to
determine reliability of a short version or use random selection of the
long version, or just administer a short version rather than extrapolate
from the long version data. There appears to be a bias with the reduced
sample and basing short form reliability data on long form administra-
tion.
Without citing research sources, the author suggests that teacher
ratings are more reliable and valid than other presumably more "objec-
tive" methods (page 16). This scale is measuring perceptions that may be
unstable across environments, time, and people. Low correlations
between parent, teacher and student perceptions is not unexpected. All
behavior rating scales have these inherent problems.

Validity
The manual provides multiple tables with minimal interpretation.
This was particularly confusing in the validity section and the area of
"factorial validity." When assessing factorial validity, many moderate to
high correlations were listed across all the scales. This suggests the fac-
tors were more alike than distinct from each other in what they purport
to measure.
The various forms of the CRS,R were compared to each other. The
criterion-related validity evidence produced disappointing results. A com,
parison to the Behavior Assessment System for Children (Reynolds &
Kamphaus, 1992) or the Child Behavior Checklist (Achenbach, 1991)
would be desirable and preferred to the comparison of the CRS,R to a self,
report children's depression scale and to a test of sustained visual attention.

Distinguishing Characteristics
Differential age and sex effects in the results of pilot and norm
groups lead to age and sex grouping for normative interpretation. For
example, more males than females were found to score significantly high,
er on the oppositional scale, cognitive problems scale, and hyperactivity
scale. Females scored significantly higher than males on the Anxious,
Shy scale. Age group differences were noted in the Perfectionism and
Social Problems scales.
The intention to develop special norms for ethnic differences is
noteable. "Other potential ethnic differences have been alluded to, and
the generation of other sets of special norms is one of the goals of future
research" (p. 110).
96 DIAGNOSTIQUE VOL. 24, NO. 1-4, 1998-1999

Desirable Features
The two-sided, one sheet form is easy to administer and an easy way
for parents, teachers, and adolescents to respond.
Administration time is short, even for the long form.
The parent and teacher feedback forms that support interpretation
with a graph, are written in easily understandable language. They are
available for future reference by the parent or teacher.
The instrument provides an opportunity to assess the same student
through the perceptions of the parent and the teacher and in some cases,
the student. This can provide a more comprehensive view of student
behavior.

Undesirable Features
The difficulty with scoring has been noted. There is difficulty with
putting raw scores across factors on the NCR paper; and difficulty with
marking the T score on the graph due to the one chart representing mul-
tiple age groups. Examiner error is likely when transferring raw scores on
both NCR paper and T scores to the graph.
The manual does not include an index:
The ability of the CRS-R to predict independent criteria indicative
of ADHD has not been demonstrated. The scale scores have a low
reported correlation with objective measures of continuous performance
of sustained visual attention. The relationship of the rating scores to
actual student behavior has not been established.
Low intercorrelations between parent, teacher, and self ratings sug-
gest that the rating scales are not measuring stable, consistent features of
behavior. The highest inter-rater correlations occur between adults who
presumably have discussed and compared information about the student
being rated. Users should always bear in mind that the scales measure
someone's subjective perception of a student's behavior. The relationship
of that perception to independent measures of behavior is either weak or
has not been demonstrated.

Overall Evaluation
The main, general use of the CRS-R as an efficient screening tool
appears to be justified. The variety of CRS forms established over the
years continues to confuse the consumer. It may be that the rich history
of the tool confuses people when attempting to sort out the psychornet-
CONNERS' RATING SCALES-REVISED (CRS-R) 97

ric data in the manual and literature. The CRS-R appears to have simi-
lar predictive powers as other rating scales. It is convenient and easy to
administer. The feedback forms are helpful. The CRS-R has not con-
vinced us that it is more sensitive or has a greater prediction of ADHD
than other similar scales. Practitioners may want to consider the greater
value of a clinical interview in combination with the ADHD Rating
Scale IV (Duf'aul, Power, Anastopulous & Reid, 1998) when the prima-
ry purpose is to differentially diagnose ADHD. If the practitioner is inter-
ested in general child psychopathology the Child Behavior Checklist
(Achenbach, 1991) provides a broad band of subscales and has conve-
nient ease of use with low cost of administration with a better established
reliability and validity for that purpose.

References
Achenbach, T. M. (1991). Child Behavior Checklist andChild Behavior Profile. Burlington,
VT: Author.
Anderson, P. (1996). Comparing the Technical Aspects of Attention Deficit Disorders Rating
Scales. Columbia, Ohio: Hawthorne Educational Services, Inc.
Connors, C. Keith, Sitarenios, G., Parker, J. & Epstein, J. (1998a). The revised Conners'
Parent Rating Scale (CPRS-R): Factor Structure, Reliability, and Criterion Validity.
Journal of Abnormal Child Psychology, 26 (4), 257-268.
Connors, C. Keith, Sitarenios, G., Parker, J. & Epstein, J. (1998b). Revision and
restandardization of the Conner Teacher Rating Scale (CTRS-R): Factor structure,
reliability, and criterion validity. Journal of Abnormal Child Psychology, 26 (4), 279-
291.
Connors, C. K. (1997). Conners' Rating Scales-Revised Technical Manual. North
Tonawanda, New York.
Danforth, J. & DuPaul, G. (1996). lnterrater reliability of Teacher Rating Scales for chil-
dren with attention-deficit hyperactivity disorder. Journal of Psychopathology and
Behavioral Assessment. 18 (3), 227-237.
DuPaul, G., Power, T., Anastopoulous, A. & Reid, R. (1998). ADHD Rating Scale-IV,
Checklists, Norms andClinical Interpretations. New York:Guilford Press.
Epstein, J., March, J., Conners, C. K., & Jackson, D. (1998). Racial differences on the
Conners' Teacher Rating Scale. Journal of Abnormal Child Psychology, 26 (2), 109-118.
Reynolds, C. & Kamphaus, R. (1992). Behavior Assessment System for Children. Circle
Pines, MN: American Guidance Service, Inc.