You are on page 1of 21


What I Think and Feel


The RCMAS is a 37-item self-report inventory used to measure anxiety in

children, for clinical purposes (diagnosis and treatment evaluation), educational
settings, and for research purposes. The RCMAS consists of 28 Anxiety items
and 9 Lie (social desirability) items. Each item is purported to embody a feeling
or action that reflects an aspect of anxiety, hence the subtitle, What I think and
Feel. It is a relatively brief instrument, which has been subjected to extensive
study to ensure that it is psychometrically sound. However, it is also advisable
that the RCMAS only be used as part of a complete clinical evaluation when
diagnosing and treating a childs anxiety (Gerard and Reynolds, 1999, p.323).


The Revised Childrens Manifest Anxiety Scale was developed by Reynolds and
Richmond (1978) to assess the degree and quality of anxiety experienced by
children and adolescents (Gerald and Reynolds, 1999, p. 323). It is based on
the Childrens Manifest Anxiety Scale (CMAS), which was devised by
Casteneda, McCandless and Palermo (1956). The Revised version of the
CMAS deletes, adds and reorders items from the CMAS to meet psychometric
standards. Reynolds and Richmond (1978) also renamed the instrument, What
I Think and Feel, although subsequent papers primarily refer to it as the
Revised Childrens Manifest Anxiety Scale (RCMAS).

History of Development:

Castaneda, McCandless and Palermo (1956) first reported a scale with

standardised data, which could be used to measure anxiety in children, the
Childrens Manifest Anxiety Scale (CMAS). The CMAS was based on a trait
theory of anxiety. It was an amended version of an instrument used to measure
manifest anxiety in adults, Taylors (1951) Manifest Anxiety Scale. The Manifest
Anxiety Scale was a compilation of items from the Minnesota Multiphasic
Personality Inventory.

While the CMAS was widely used and published, Reynolds and Richmond
(1978) reported a number of issues with the CMAS that prompted the revision.
Reynolds and Richmond (1978) hoped to revise the CMAS to identify items that
meet Flannigan, Peters and Conrys (1969) criteria for a good test item
(p.272), to improve the psychometric properties of the instrument (according to
Guilford, 1954), and to meet the American Psychological Association (1954)
guidelines for psychological tests.

Reynolds and Paget (1981) also noted the need to develop an instrument that
could measure a broader range of anxiety and treatment effects, and that could

reflect research that suggests that anxiety is multidimensional (, p352).
Reynolds and Richmond (1978) also wanted a scale that could be administered
in less time, with individuals or groups of children, aged from 6 to 19 years. The
wording of items had to be adjusted to accommodate the younger children and
poor readers.


Several types of reliability can be demonstrated with the RCMAS, in terms of

the internal consistency of the instrument, stability, and possibly equivalence,
but not in terms of the inter-rater reliability. Reynolds and other researches have
focused on developing an instrument that was psychometrically sound and that
could be used by a variety of practitioners (clinicians, teachers and
researchers), without attention to potential variations with application or
interpretation in its use.


Split-Half/Cronbachs Alpha:

Reynolds and Richmond (1978) argue that the 33 % reduction in the length
of the CMAS scale and reduction in administration time does not detract
from the reliability of the RCMAS. The Kuber-Richardson (KR) analysis of
variance method was used to establish coefficients of internal consistency.
Reynolds and Richmond (1978) report that with the 37 items selected for the
RCMAS, a KR20 reliability estimate of .83 is yielded, confirming internal
consistency of the RCMAS.

When making technical recommendations for psychological tests and

diagnostic techniques, the APA (1954) note the risks associated with
computing reliability and validity estimates on the same sample analysed to
select the test items. Hence, Reynolds and Richmond (1978) conducted a
second, cross-validation assessment of 167 children, from grades two, five,
nine, ten and eleven, in a different school district. This second assessment
group yielded a KR20 reliability estimate of .85, and further support for
internal consistency.

Reynolds, Bradley and Steele (1980) administered the RCMAS to 97

kindergarten children and demonstrated reliability with coefficient alpha (a =
.79 with males, a =.85 with females, and a =.82 for the total sample).
These correlations are high and similarly indicate internal consistency when
the RCMAS is used with younger children.

Gerard and Reynolds (1999) also report that with few exceptions, relatively
high coefficients alpha for the total Anxiety Scale score (a = .80 range), are
indicative of good sampling of the general domain of potential anxiety
items (p.327) and internal consistency according to Cronbach (1951).

Test-Retest Reliability:

Wisniewski, Mulick, Genshaft and Coury (1987) examined the test-retest

reliabilities of the RCMAS with 161 children in Grades 6 to 8. Analyses of
retesting after one and five weeks indicated good reliability (Pearson
correlations from .60 to .88, significant at p .01, p. 67) and an insignificant
difference between test and retest mean raw scores. These results would
support the stability of the scale over brief periods.

With retesting after a substantial longer period, nine months, Reynolds

(1981) found a .68 correlation between RCMAS Anxiety Scale scores and a .
58 correlation with the Lie Scale scores, for 534 children in Grades 4 to 6.
This would be indicative of relatively high temporal stability.

Alternate Form Reliability:

The establishment of temporal , test-retest reliability negates the need to use

alternate forms in many instances.

However, favourable comparisons can be made between the reliability

coefficients with the RCMAS and the CMAS. Reynolds and Richmond
(1978) report KR20 reliability estimate of .83 with the RCMAS and cite
comparative estimates with the CMAS. In particular, Kitano (1960) reported
a reliability coefficient of .86; Finch, Montgomery and Deardoff (1974) of .77;
and by Allison (1970) of .84 for boys and .88 for girls. However, there may
be some question about whether the RCMAS and the CMAS are different
tests or different forms of an instrument (one an abbreviated and arguably
an improved version, and one a longer version, respectively).

Inter-rater Reliability:

With Reynolds and Paget (1981), the 4,972 children were variously tested by
clinical psychologists, school psychologists, classroom teachers and school
administrators. However, this is primarily a self-report measure and no statistical
comparisons were made regarding the different raters.


There is substantive research confirming the validity of the RCMAS as a

measure of chronic manifest anxiety in children, dating back to the
original article reporting the development of the RCMAS (Reynolds and
Richmond, 1978). In addition, the RCMAS is frequently used in
research to validate other instruments and to measure treatment

Content Validity:

In the draft version of the RCMAS, Reynolds and Richmond (1978) added
twenty items to cover areas that teachers and clinicians identified as not being
covered by the previous scale. The wording of the resultant 73 items was
modified by reading specialists, to meet the reading level of Grade three
children and to reflect general changes in word usage since the test was first

Two item statistics were computed for the 73 draft items, the difficulty index, p,
and biserial correlation of the item to the total test score, rbis. With the Anxiety
Scale, items were eliminated if they did not meet both criteria, with the difficulty
index (.3 p .7) and with biserial correlation (rbis .4). Lie Scale items were
eliminated if they correlated .30 or higher with the anxiety scale or if they failed
to correlate significantly with any other lie item.

A total of 28 anxiety items (25 from the CMAS and three new items), and nine
Lie items were retained to form the current 37 items of the RCMAS. The results
would suggest that the 28 anxiety items that were finally selected, adequately
represent all aspects of the anxiety construct, thereby indicating content validity.

Construct Validity:

Reynolds and Richmond (1979 ) conducted a factor analysis with the

Anxiety Scale items, with the test development sample of the RCMAS.
Three anxiety factors were identified and named, based on item content:
physiological, worry/oversensitivity, and concentration. However, there
were a number of anomalies with the results that were attributed to the small
sample size of 329 subjects compared with the 28 variables.

Reynolds and Paget (1981) attempted to replicate Reynolds and

Richmonds (1979) study with a much larger sample size of 4,972 children,
and to extend the analysis to the Lie Scale items. Responses were factor
analysed through the method of principal factors, which identified a large
general factor on which substantive loadings were found for all 28 Anxiety
Scale items but no loadings above .21 for the Lie Scale items. This would
lend support to the RCMAS being a measure of one construct, anxiety (Ag).

Factors extracted through the factor analysis were then rotated orthogonally
through the varimax procedure, to maximise the variance accounted for by
each factor and to increase the distinction between factors. Examination of
Scree plots and eigenvalues identified three to eight factors, but further
analysis and examination of the item content of the factorial groupings
indicated the five factor solution as the most psychologically meaningful and
interpretable solution.

The pattern of factor loadings with the five-factor solution revealed two Lie
Scale factors (accounting for 75% and 25% of the variance) and the three
distinct Anxiety Scale factors (accounting for 34%, 42% & 24% of the
variance). The three Anxiety Scale factors identified by Reynolds and Paget
(1981) also resembled the three found by Reynolds and Richmond (1979)
but without the anomalies found in the earlier research.

Further analysis and additional research has found that the five-factor
solution is consistent across gender, ethnicity (race), and intelligence
(Gerard and Reynolds, 1999).

The five factors confirmed by Reynolds and Paget (1981) are as follows:

Anxiety Scale Factors: Item Numbers

The Physiological Factor 1, 5, 9, 13, 17, 19, 21, 25, 29, 33

The Worry/Oversensitivity Factor 2, 6, 7, 10, 14, 18, 22, 26,

30, 34, 37

The Concentration Anxiety Factor 3, 11, 15, 23, 27, 31, 35

Lie Scale Factors:

Lie 1 4, 8, 12, 16, 20, 24

Lie 2 28, 32, 36

Reynolds and Paget (1981) calculated two indices of factorial similarity, the
coefficient of congruence (rc, Harman, 1976; Mulaik, 1971) and the salient
variable similarity index (s , Cattell, 1978). Reynolds and Paget (1981)
reported rcs ranging from .91 to .99, and highly significant s-values (p < .
01), which would suggest that the five factors were invariant in relation to
gender or race (ethnicity).

The same invariance was found with the general anxiety factor (Ag) which
would lend support to the construct validity of the RCMAS, or the internal
consistency of the RCMAS. Reynolds and Paget (1981) indicated the need
for further analysis of the internal validity of the RCMAS and that they were
currently preparing normative data across gender, race and age (see
Reynolds and Paget, 1983).

Pina, Silverman, Saavedra and Weems (2001) analysed the RCMAS Lie
Scale with 284 anxious children. Results indicated that the Lie Scale scores
were also predictive of the childrens anxiety levels, and that they could
distinguish between children with anxiety disorders and children with
externalising disorders. Hence, the Lie Scale was said to have utility in terms
of predicting a childs anxiety level.

Debate in the literature also tends to lend support to the Lie Scale being a
measure of social desirability in anxiety, especially with younger children
(Reynolds and Richmond, 1978). Dadds, Perrin and Yule (1998) report
indications that social desirability levels partly explain the discrepancies
found between child and adult reports of anxiety in youth. Hagborg (1991)
also reported favourable findings that support the concurrent validity of the
RCMAS Lie Scale as a measure of social desirability.

Using concept mapping and confirmatory factor analysis, White and Farrell
(2001) compared the empirically derived factor structure of the RCMAS with
theory-driven models derived from 8 experts on child anxiety (p.333), with
898 Grade 7 children, primarily black Americans (94%). White and Farrell
(2001) report analyses that identify three-factor models with both the
RCMAS and the expert-derived model (anxious arousal, social evaluation-
oversensitivity and worry). However, empirical support for a higher order
factor was only found with the expert-derived model, which excluded items
of dysphoric mood. White and Farrell (2001) argue that the RCMAS includes
items that are related to a construct or dimension that resembles dysphoric
mood (p.336).

There are some limitations with White and Farrells (2001) study, including
the narrow age range (10.8 to 14.1 years), similar ethnicity of the subjects
(94% African American or black), and subjects who also report lower levels
of anxiety compared to the standardised sample. Such limitations would limit
generalisability of White and Farrells (2001) results. However, White and
Farrell (2001) also claim that the expert-derived model can meaningfully
contribute towards improving our understanding of the assessment of
anxiety using the RCMAS. White and Farrell (2001) suggest that future
research consider refining the domains assessed by the RCMAS, and
possibly including domains reflecting anxious apprehension and behavioural
avoidance, as indicated by the experts.

Convergent or Concurrent Validity:

While investigating the construct validity of the RCMAS, Reynolds (1982) also
found large positive correlations between the RCMAS and the trait measure of
anxiety, the STAIC, but not with the state measures of anxiety. These findings
are consistent with earlier findings by Reynolds (1980), who found high
correlations between the RCMAS and STAIC trait (r = .85, p .05) but not with
STAIC state measures (r = .24, p > .05).

Reynolds (1985) replicated these results with a sample of 465 gifted children
(IQs higher than 130), who tended to score significantly lower on all measures
of anxiety compared with normative samples. The convergent and divergent
validity of the RCMAS was assessed in relation to measures of trait and state
anxiety with the State-Trait Anxiety Inventory for Children (STAIC). Reynolds
(1985) found that the RCMAS scores correlated highly with a Trait Anxiety scale

(r = .78, p < .001) but not with a State Anxiety scale (r = .08). This would lend
support for the convergent and divergent validity of the RCMAS.

Lee, Piersel, Friedlander and Collamer (1988) examined the concurrent validity
of the RCMAS with ninety 13 to 17 year olds, and found that it correlated well
with a similar measure of anxiety using the Minnesota Multiphasic Personality

However, Lee, Piersel and Unruh (1989) evaluated the concurrent validity of the
RCMAS Physiological subscale with parent and teacher behavioural ratings of
anxiety/somatic complaints, depression and aggression with eighty 10 to 17
year old males who had academic or behavioural problems. In contrast to
earlier (and later) findings, Lee et al (1989) found a lack of convergent and
discriminant validity between the RCMAS and behavioural ratings. This
anomaly might reflect problems with the alternative comparative measure used
in this study, the behaviour ratings, and the different perspective of adults and

Muris, Merckelbach, Mayer, van Brakel, Thissen, Moulaert and Gadget (1998)
compared the Screen for Child Anxiety Related Emotional Disorders
(SCARED), the Fear Survey Schedule for Children Revised (FSSC-R) and the
RCMAS. They found that scores on all three tests were positively related, in a
theoretically meaningful manner, and hence evidence of concurrent validity.

Muris, Merckelbach, Ollendick, King and Bogie (2002) examined the

psychometric properties of six anxiety scales to be used with children: the
RCMAS; the trait anxiety version of the State-Trait Anxiety Inventory for
Children (STAIC); the Fear Survey Schedule for Children Revised (FSSC-R);
the Multidimensional Anxiety Scale for Children (MASC); the Screen for Child
Anxiety Related Emotional Disorders (SCARED); and the Spence Childrens
Anxiety Scale (SCAS).

With a sample of 521 normal adolescents (p. 753), Muris et al (2002) found
that the childhood anxiety scales were reliable in terms of internal consistency
(Cronbachs alphas generally well above .60). Convergent validity was indicated
by the substantial correlations between the anxiety scores on the six
questionnaires. Particularly strong correlations were found between the RCMAS
and the STAIC (r = 0.88), and between the RCMAS and the SCARED (r =
0.85), which would suggest that they are likely to be tapping highly similar or the
same construct(s). Correlations between the RCMAS and the FSSC-R were
said to be moderate (r = 0.63), with suggestions that the FSSC-R was tapping
slightly different aspects of anxiety (such as specific fears and phobias).

Muris et al (2002) found that the RCMAS, the STAIC, the SCARED and the
SCAB all correlated highly with an index of depression (CDI), (r s in the .70
range), and slightly less correlations between the CDI and the other two anxiety
measures, the FSSC-R and the MASC. This would be indicative of considerable
overlap between anxiety, as indexed by these measures of anxiety and
depression. However, the correlations between the six anxiety scales in Muris et
als (2002) study were higher than the correlations between the anxiety scales

and the depression measure, which underlines the divergent validity of the
childhood anxiety measures (p.767).

Muris et al (2002) conclude that the six questionnaires all have satisfactory
psychometric properties. The only differentiating factor between the new and old
anxiety scales might be the new scales closer links with the DSM diagnostic
system, which could assist clinical communications about anxiety problems with
children (citing Chorpita, Yim, Moffit, Umemoto and Francis, 2000).

Discriminant Validity:

Reynolds and Richmond (1978) note that evidence of discriminant validity will
be necessary for the revised instrument but that it was not yet available

Mattison, Bagnato and Brubaker (1988) studied the clinical relevance of the
RCMAS as a tool used to discriminate between children with a DSM-III anxiety
disorder and other DSM-III psychiatric diagnoses. They found that the RCMAS
Worry/Oversensitivity subfactor of the Anxiety Scale can significantly
discriminate between those children who have a diagnosable anxiety disorder
and those children who do not have an anxiety disorder. The use of the RCMAS
was recommended by Mattison, Bagnato and Brubaker (1988) as part a multi-
method of assessment for identifying children with anxiety disorders.
Accordingly, the RCMAS could be considered useful as a diagnostic tool and for
screening those children who may be in need of counselling.

Perrin and Last (1992) compared discriminant validity of the RCMAS, the
Modified State-Trait Anxiety Inventory for Children (STAIC-M) and the Fear
Survey Schedule for Children Revised (FSSC-R). With their sample of 213
youth, Perrin and Last (1992) found that the FSSC-R could not discriminate but
the RCMAS and the STAIC-M could distinguish between youth who had never
been given a psychiatric diagnosis and those who had a diagnosis. The
discriminate failure with FSSC-R may reflect anomalies noted with the FSSC-R
in Muris et als (2002) study, in particular, that the FSSC-R may be tapping into
slightly different aspects of anxiety. Furthermore, while the RCMAS and STAIC-
M could distinguish between who had a diagnosable problem and those who did
not, they could not distinguish between who had a diagnosis of an anxiety
disorder or Attention Deficit Hyperactivity Disorder (ADHD). This would indicate
the need for some caution and the importance of the RCMAS being used as
part of a clinical assessment.

Dierker, Albano, Clarke, Heimberg, Kendall, Merikangas, Lewinsohn, Offord,

Kessler and Kupfer (2001) examined the disrciminative accuracy of three rating
scales for detecting anxiety and depression with children (the RCMAS; the
Center for Epidemiological Studies-Depression Scale or CES-D; and the
Multidimensional Anxiety Scale for Children or MASC). Scores with the three
scales were compared with diagnostic interviews for 632 youth. The MASC
scores were said to be most strongly associated with individual anxiety

disorders, and the CES-D composite score was linked with a diagnosis of a
major depression, while the RCMAS was said to be the least successful tool for
discriminating between anxiety and depression.

Stark and Laurents (2001) used a joint factor analysis with the RCMAS and the
Childrens Depression Inventory (CDI) to identify which items uniquely identified
depression and anxiety with 750 children in Grades 4 to 7. Stark and Laurent
(2001) identified an abbreviated version of the scales (with nine unique
depression items and seven unique anxiety items, which was validated with a
separate sample of 131 students. Stark and Laurent (2001) also suggested the
need to explore alternative ways to score the RCMAS and CDI to eliminate
potential problems with overlapping items.

The results of research by Perrin and Last (1992), Dierker et al (2001) and
Stark and Laurent (2001) indicate the need for caution and further research with
respect to the discriminant validity of the RCMAS to distinguish between anxiety
and depression in children. It may be as White and Farrell (2001) suggest, the
RCMAS needs further refinement and possibly the exclusion of items which
relate to dysphoric mood or items which tap into depression.

Criterion (or Predictive) Validity:

Hadwin, Frost, French and Richards (1997) found in a sample of 40 children

aged 7 to 9 years, that levels of anxiety as measured by the RCMAS, could
significantly predict the childrens interpretations of ambiguous stimuli as

Stallard, Velleman, Langsford and Baldwin (2001) conducted a univariate

analysis of variance to determine whether the number of coping strategies used
by children involved in everyday traffic accidents was affected by Post
Traumatic Stress Disorder (PTSD), the childs age or gender, the presence of
depression in the child, or the presence of anxiety in the child (as measured by
the RCMAS). They found that only the childs age and PTSD were significantly
linked to the number of coping strategies used, not anxiety. A logistic regression
analysis also found that anxiety measured by the RCMAS was not predictive of
PTSD at six-weeks post accident. Gender was found to be independently
predictive of PTSD.

Future research might consider the predictive validity of anxiety, as measured

by the RCMAS, on academic achievement. Gaudry and Spielberger (1971)
found a negative relationship between anxiety and academic achievement, but
it seems performance and achievement has a more complex relationship with
anxiety. There is also the interactive effect of intelligence to be considered when
making any predictions about academic or vocational success.

Another consideration for future investigations would be the interactive effect of

anxiety and group membership (such as ethnicity or gender) when predicting
behavioural problems or future adjustment in children.


The RCMAS is suitable for individual or group administration, by clinicians,

researchers or teachers, with 6 to 19 year old children. The scale is best read
out to children in Grades one and two (or to children who have an equivalent
reading age). Grade three and older children need to be monitored carefully as
they read the items themselves, with explanations given for words that they do
not understand.

Reynolds and Richmond (1978) advocate for the use of the RCMAS with
children in Grades three to twelve, and more tentatively (probably satisfactory,
p. 279), for Grades one and two, or with children functioning intellectually in that
lower range.

Caution is recommended for the younger children because of the relatively

higher Lie Scale scores (Reynolds and Richmond, 1978). Reynolds, Bradley
and Steele (1980) found that the younger age group understand and respond
reliably when the items are read to them.


Each item is given a score of one for a yes response, yielding a Total Anxiety
score (Ag). Three empirically derived Anxiety Subscales scores (Physiological
Anxiety, Worry/Oversensitivity, and Social Concerns/Concentration) and Lie
Scale scores can be calculated. The Lie scale is best thought of as a social
desirability scale as it does not directly and conclusively detect lying.

Stallard, Velleman, Langsford and Baldwin (2001) recommend that an overall

cut-off point of 19 out of 28 be used to identify children experiencing clinically
significant levels of anxiety (p.200).

Reynolds and Richmond (1978) suggest that scores within one standard
deviation of the mean, at the appropriate grade level, be used to indicate
scoring within the normal range of variability (see below for norms of means and
standard deviations or sources for norms).

Scores falling at least one standard deviation from the mean (T 60) are
thought to be of clinical interest. However, T-scores above 70 should be
interpreted with caution. The childs response pattern should be examined with
respect to a problematic pattern of endorsement or reading difficulties.

High scores on the sub-scales can represent different aspects of anxiety, which
can be used to develop hypotheses about the origin and nature of a childs

(1) High scores on the Physiological Factor (items 1, 5, 9, 13, 17, 19, 21,
25, 29, 33) can indicate physiological signs of anxiety (eg sweaty hands,
stomach aches).

(2) High scores on the Worry/Oversensitivity Factor (items 2, 6, 7, 10, 14,
18, 22, 26, 30, 34, 37) would suggest that the child internalises their
experiences of anxiety and that he or she may feel overwhelmed and

(3) High scores on the Concentration Anxiety Factor (items 3, 11, 15, 23,
27, 31, 35) would suggest that the child is likely to feel that he or she is
unable to meet the expectations of other important people, inadequate
and unable to concentrate on tasks.



Standardisation sample populations for the RCMAS are thought to be large,

diverse and representative.

Reynolds and Richmond (1978) computed means and standard deviations

(SD) for 329 school age children., who were all tested on the same day. The
researchers were not given permission to collect data about the socio-
economic status of the children, but to ensure representation of the sample,
there was random selection of classes at each grade level from an urban
school district. However, Reynolds and Richmond (1978) also note the need
for further study to determine the generalizability of the instrument to other

Reynolds and Paget (1981) tested 4,972 children, aged six to nineteen
years, from thirteen different states in the USA and eighty school districts.
While socio-economic data was also not available to Reynolds and Paget
(1981), they argued that their sample was representative of cross-section of
the school attending population because rural and urban areas were equally
represented, including inner city and high poverty districts; and specific
neighbourhoods with known SES composition to ensure the representative
nature of the sample (p.353).

Reynolds and Paget (1983) also note that the 4,972 children aged 6 to 19
years, from thirteen states in the USA, are representative of all geographic
regions in the USA. From the data, Reynolds and Paget (1983) established
separate norms for gender, race and age for the three Anxiety subscales,
the two Lie Scales and the total Anxiety Scale. Using the method of rolling
weighted averages standard score distributions were derived from the raw
score distributions, whereby a Total Anxiety score becomes a T-score with a
mean of 50 and a standard deviation of 10. The subscale scores have a
mean of 10 and standard deviations of 3.

Means and Standard Deviations - Sources:

(1) Mertin, Dibnah, Crosbie & Bulkeley (2001)

British Sample (8 to 12 year olds)

Mertin, Dibnah, Crosbie and Bulkeley (2001) questioned the applicability of

North American norms with the RCMAS to a British population. By computing
means by age and gender, for 575 English children aged 8 to 12 years, Mertin
et al (2001) found that eight year old English girls were less anxious than their
North American equivalent; and that most English males were less anxious than
their North American equivalent. Mertin et al (2001) also note language
differences and advise that the RCMAS be used as part of a structured
interview rather than as a self-report questionnaire.

(2) Reynolds and Paget (1981; 1983)

North American Sample (6 to 19 year olds)

(3) Reynolds, Bradley & Steele (1980)

- Preliminary Norms North American Sample (kindergarten


(4) Reynolds and Richmond (1978)

North American sample (6 to 19 year olds)

Reynolds and Richmond (1978) data:

The Anxiety Scale Mean = 13.84 SD = 5.79

The Lie Scale Mean = 3.56 SD = 2.37

Reynolds and Richmond (1978) suggest that scores within one standard
deviation of the mean, at the appropriate grade level, be used to indicate
scoring within the normal range of variability. Reynolds and Richmond (1978,
p.276) also note that the Anxiety scale correlated significantly with the Lie scale,
r(327) = .15; p .01.

Means and SD were also obtained for the RCMAS by Grade, Race and Gender
(Reynolds and Richmond, 1978, Tables II and III, pp. 276-277):

Anxiety Scale Lie Scale

Mean SD N Mean SD


1 13.70 4.85 23 6.00 1.95

2 16.13 6.42 30 4.63 2.55

3 12.78 6.50 32 3.97 2.18

4 16.64 5.70 28 2.25 1.65

5 12.52 5.33 33 2.70 2.47

5 13.82 5.28 28 4.18 2.04

6 11.85 5.27 26 1.93 1.67

8 14.50 5.22 30 2.57 1.87

9 13.25 6.27 40 3.70 1.84

10 13.23 5.85 22 3.68 2.48

11 13.96 5.87 28 3.68 2.75

12 13.67 4.58 9 4.33 2.29


Females 14.97 5.60 173 3.66 2.45

Males 12.58 5.75 156 3.45 2.28


Blacks14.09 5.30 172 4.02 2.09

Whites 13.56 6.29 157 3.06 2.56

With the Anxiety Scale and the Lie Scale, Reynolds and Richmond (1978)
computed a three way ANOVA for grade, race and gender, and submitted
the variances to separate F tests. With the Anxiety Scale, no significant
effect was found for grade or race, but females scored significantly higher
than males (F (1,283) = 10.87; p .001), (p277). This may reflect
speculation that females more readily admit to anxiety than males (Sarason,
Davidson, Lighthall, Waite and Ruebush, 1960, cited in Reynolds and
Richmond , 1978). It was also consistent with previous research using the
CMAS (Bledsoe, 1973, cited in Reynolds and Richmond, 1978).

With the Lie Scale, there was no significant effect with gender in the
Reynolds and Richmond (1978) sample, but blacks reportedly scored
significantly higher than whites on the Lie Scale, for which there was no
explanation. Grade anomalies were also noted. With the exception of
Grades two and twelve, Grade one children scored significantly higher than
all other grade children (Duncans multiple range test procedure, p .05);
and with the exception of Grades five and eight, Grades seven and four
scored significantly lower scores than all other grade children (p .05),

(p278). Reynolds and Richmond (1978) suggest that Grade variations in the
Lie Scale may reflect the unique characteristics of the population, or an
indication of defensiveness or social desirability, especially with younger
children. Hence, while a high Lie score of six or more may invalidate a high
Anxiety score, it might also provide clinical information about the childs
response style, or personality characteristics. This would be true for most lie

Cross Cultural Validity:

Studies of cross-cultural validity of the RCMAS have tended to focus on issues

of validity with respect to gender and ethnicity. Reynolds, Plake and Harding
(1983) found that the RCMAS does contain some potentially biased items in
terms of different gender and race response, but the difference was not clinically
significant. Reynolds and Paget (1981)demonstrated equivalence with the factor
structure for different genders and race.

Wilson, Chibaiwa, Majoni, Masukume and Nkoma (1990) found that the
RCMAS was a modestly reliable measure with 961 Zimbabwe children but a
factor analysis failed to establish the validity of the RCMAS as a diagnostic or
research tool with Zimbabwe children.

Boyd, Kostanski, Gullone, Ollendick and Shek (2000) looked at the prevalence
of anxiety and depression in 1,299 adolescents in Melbourne using the RCMAS
and the Reynolds Adolescent Depression Scale and found striking differences
(p.479) between the prevalence in different countries, which might have
implications for the use of norms from different countries. Self-reported rates of
depression and anxiety in Britain, Canada and the United States were
considered to be similar or comparative, with comparative rates in Asian
countries but the highest rates of anxiety and depression were found in Eastern
European countries. Australian data, which identified rates 14.2% of
adolescents being depressed and 13.2% being anxious, were said to be
comparable with Canada and Hong Kong.

Boyd et als (2000) results would tend to lend support to use of the North
American and British normative data with Australian children, but the ethnic
background of the child might also need to be considered.


American Psychological Association (APA) (1954) Technical Recommendations

for psychological Tests and Diagnostic Techniques (Washington, APA)

Allison, D.E. (1970) Test Anxiety, stress and intelligence-test performance,

Canadian Journal of Behavioural Science, vol. 2(11), pp. 26-37, cited in
Reynolds and Richmond (1978).

Bledsoe, J.C. (1973) Sex and grade differences in childrens manifest anxiety,
Psychological Reports, vol. 32, pp. 285-286, cited in Reynolds and Richmond

Boyd, C.P.; Kostanski, M.; Gullone, E.; Ollendick, T.H.; and Shek, D.T.L. (2000)
Prevalence of anxiety and depression in Australian Adolescents; Comparisons
with worldwide data, Journal of Genetic Psychology, vol. 161(4), pp. 479-492.

Castaneda, A.; McCandless, B.R.; and Palermo, D.S. (1956) The childrens
form of the manifest anxiety scale, Child Development, vol. 27(3), pp.317-326,
cited in Reynolds and Richmond (1978).

Cattell, R.B. (1978) The scientific use of factor analysis in the behavioural and
life sciences, (New York, Plenum Press), cited in Reynolds and Paget (1981).

Chorpita, B.F.; Yim, L.; Moffit, C.; Umemoto, L.A.; and Francis, S.E. (2000)
assessment of symptoms of DSM-IV anxiety and depression in children: A
revised child anxiety and depression scale, Behaviour Research and Therapy,
vol. 38, pp. 835-855, cited in Muris et al, 2002.

Cronbach, L. (1951) Coefficient alpha and the internal structure of tests,

Psychometrika, vol. 16, pp297-334, cited in Gerard and Reynolds, 1999.

Dadds, M.R.; Perrin, S.; and Yule, W. (1998) Social desirability and self-
reported anxiety in children: An analysis of the RCMAS Lie Scale, Journal of
Abnormal Child Psychology, vol. 26(4), pp. 311-317.

Devilly, G.J. and Sanders, M.R. (1993) Hey dad, Watch Me: The Effects of
Training a Child to Teach Pain Management Skills to a Parent with Recurrent
Headaches, Behaviour Change, vol. 10(4), pp. 237-243.

Dierker, L.C.; Albano, A.M.; Clarke, G.N.; Heimberg, R.G.; Kendal, P.C.;
Merikangas, K.R.; Lewinsohn, P.M.; Offord, D.R.; Kessler, R.; and Kupfer, D.J.
(2001), Screening for anxiety and Depression in early Adolescence, Journal
of the American Academy of Child and Adolescent Psychiatry, vol. 40(8), pp.

Finch, A.J.; Montgomery, L.E. and Deardoff, P.A. (1974) Childrens Manifest
Anxiety Scale with emotionally disturbed children, Psychological Reports, Vol.
34, p.68, cited in Reynolds and Richmond (1978).

Flannigan, P.J.; Peters, C.J. and Conry, J.L. (1969) Item analysis of the
Childrens Manifest Anxiety Scale with the retarded, Journal of Educational
Research, vol. 62(10), pp. 472-477, cited in Reynolds and Richmond (1978).

Gaudry, E.; and Spielberger, D. (1971) Anxiety and Educational Achievement,

(Sydney, Wiley), cited in Reynolds and Richmond (1978).

Gerard, A.B.; and Reynolds, C.R. (1999) Characteristics and applications of the
Revised Childrens Manifest Anxiety Scale, in Maruish, M.E. (ed.) The use of

psychological testing for treatment and planning and outcomes assessment (2nd
edition, Mahwah, Lawrence Erlbaum Associates), pp. 323-340.

Guilford, J.P. (1954) Psychometric Methods (2nd edition, New York, McGraw-
Hill), cited in Reynolds and Richmond (1978).

Hadwin, J.; Frost, S.; French, C.C.; and Richards, A. (1997) Cognitive
processing and trait anxiety in typically developing children: Evidence for an
Interpretation Bias, Journal of Abnormal Psychology, vol. 106(3), pp. 486-490.

Hagborg, W.J. (1991) The revised Childrens Manifest Anxiety Scale and Social
Desirability, Educational and Psychological Measurement, vol. 51 (2), pp423-

Harman, H.H. (1976) Modern Factor Analysis, (2nd edition, Chicago, University
of Chicago Press), cited in Reynolds and Paget (1981).

Kitano, H.L. (1960) Validity of the Childrens Manifest Anxiety Scale and the
modified revised California inventory, Child Development, vol. 31, pp. 67-72,
cited in Reynolds and Richmond (1978).

Lee, S.W.; Piersel, W.C.; and Unruh, L. (1989) Concurrent Validity of the
Psychological subscale of the Revised Childrens Manifest Anxiety Scale: A
multitrait-multimethod analysis, Journal of Psychological Assessment, vol.
7(3), pp. 246-254.

Lee, S.W.; Piersel, W.C.; Friedlander, R.; and Collamer, W. (1988) Concurrent
Validity of the Revised Childrens Manifest Anxiety Scale (RCMAS) for
adolescents, Educational and Psychological Measurement, Vol. 48(2), pp.

Mattison, R.E.; Bagnato, S.J.; and Brubaker, B.H. (1988) Diagnostic utility of
the Revised Childrens Manifest Anxiety Scale in children with DSM-III anxiety
disorders, Journal of Anxiety Disorders, vol 2(2), pp. 147-155.

Mertin, P.; Dibnah, C.; Crosbie, V.; and Bulkeley, R. (2001) Using North
American instruments with British samples: Norms for the Revised Childrens
Manifest Anxiety Scale in the UK, Child Psychology and Psychiatry Review,
vol. 6(3), pp. 121-126.

Mulaik, S.A. (1971) The foundations of factor analysis, (New York, McGraw-
Hill), cited in Reynolds and Paget (1981).

Muris, P.; Merckelbach, H.; Ollendick, T.; King, N.; and Bogie, N. (2002) Three
traditional and three new childhood anxiety questionnaires: their reliability and
validity in a normal adolescent sample, Behaviour Research and Therapy, vol.
40(7), pp 753-772.

Muris, P.; Merckelbach, H.; Mayer, B.; van Brakel, A.; Thissen, S.; Moulaert, V.;
and Gadet, B. (1998) The screen for Child Anxiety Related Emotional

Disorders (SCARED) and traditional childhood anxiety measures, Journal of
Behaviour Therapy and Experimental Psychiatry, vol. 29(4), pp.327-339.

Perrin, S.; and Last, C.G. (1992) Do childhood anxiety measures measure
anxiety, Journal of Abnormal Child Psychology, vol. 20(6), pp. 567-578.

Pina, A.A.; Silverman, W.K.; Saavedra, L.M.; and Weems, C.F. (2001) An
analysis of the RCMAS lie scale in a clinic sample of anxious children, Journal
of Anxiety Disorders, vol. 15(5), pp. 443-457.

Reynolds, C.R. (1980) Concurrent validity of What I Think and Feel: the
Revised Childrens Manifest Anxiety Scale, Journal of Consulting and Clinical
Psychology, vol. 48(6), pp. 774-775.

Reynolds, C.R. (1981) Long-term stability of scores on the Revised Childrens

Manifest Anxiety Scale, Perceptual and Motor Skills, vol. 53(3), pp. 702.

Reynolds, C.R. (1985) Multitrait validation of the Revised Childrens Manifest

Anxiety Scale for children of high intelligence, Psychological Reports, vol.
56(2), pp. 402

Reynolds, C.R. (1982) Convergent and divergent validity of the Revised

Childrens Manifest Anxiety Scale, Educational and Psychological
Measurement, vol. 42(4), pp. 1205-1212.

Reynolds, C.R.; Bradley, M.; and Steele, C. (1980) Preliminary norms and
technical data for use of the Revised Childrens Manifest Anxiety Scale with
kindergarten children, Psychology in the Schools, vol. 17, pp. 163-167, cited in
Reynolds and Paget (1981).

Reynolds, C.R. and Paget, K.D. (1981) Factor Analysis of the Revised
Childrens Manifest Anxiety Scale for Blacks, Whites, Males and Females with a
National Normative Sample, Journal of Consulting and Clinical Psychology, vol.
49(3), pp. 352-359.

Reynolds, C.R. and Paget, K.D. (1983) National normative and reliability data
for the Revised Childrens Manifest Anxiety Scale, School Psychology Review,
vol. 12(3), pp324-336.

Reynolds, C.R., Plake, B.S. and Harding, R.E. (1983) Item bias in the
assessment of childrens anxiety: Race and Sex interactions on items on the
Revised Childrens Manifest Anxiety Scale, Journal of Psychoeducational
Assessment, vol. 1, pp 135-142, cited in Gerard and Reynolds, 1999.

Reynolds, C.R. and Richmond, B.O. (1978) What I think and Feel: A Revised
Measure of Childrens Manifest Anxiety, Journal of Abnormal Psychology, vol.
6(2), pp. 271-280.

Reynolds, C.R. and Richmond, B.O. (1979) Factor Structure and Construct
Validity of what I think and Feel: The Revised Childrens Manifest Anxiety
Scale, Journal of Personality Assessment, vol. 43, pp. 281-283.

Sarason, S.B.; Davidson, K.S.; Lighthall, F.F.; Waite, R.R.; and Ruebush, B.K.
(1960), Anxiety in elementary school children, (New York, Wiley), cited in
Reynolds and Richmond , 1978.

Stark, K.D.; and Laurent, J. (2001) Joint factor analysis of the Childrens
Depression Inventory and the Revised Childrens Manifest Anxiety Scale,
Journal of Clinical Child Psychology, Vol. 30(4), pp. 552-567.

Stellard, P.; Velleman, R.; Langsford, J. and Baldwin, S. (2001) Coping and
psychological distress in children involved in road traffic accidents, British
Journal of Clinical Psychology, vol. 40, pp. 197-208.

Taylor, J.A. (1951) The relationship of anxiety to the conditioned eyelid

response, Journal of Experimental Psychology, vol. 42, pp. 183-188, cited in
Reynolds and Richmond (1978).

White, K.S.; and Farrell, A.D. (2001) Structure of anxiety symptoms in urban
children: Competing factor models of the revised childrens manifest anxiety
scale, Journal of Consulting and Clinical Psychology, vol. 69(2), pp. 333-337.

Wilson, D.; Chibaiwa, D.; Majoni, C.; Masukume, C.; Nkoma, E. (2002)
Reliability and factorial validity of the Revised Childrens Manifest Anxiety Scale
in Zimbabwe, Personality and Individual Differences, vol. 11(4), pp. 365-369.

Wisniewski, J.J.; Jack, J.; Mulick, J.A.; Genshaft, J.L.; and Coury, D.L. (1987)
Test-Retest reliability of the Revised Childrens Manifest Anxiety Scale,
Perceptual and Motor Skills, vol. 65(1), pp.67-70.

The Revised Childrens Manifest Anxiety Scale


What I think and Feel

Read each question carefully. Put a circle around the word YES if you think it is
true about you. Put a circle around the word NO if you think it is not true about

1. I have trouble making up my mind. Yes /


2. I get nervous when things do not go the right way for me. Yes /

3. Others seem to do things easier than I can. Yes /


4. I like everyone I know. Yes /


5. Often I have trouble getting my breath. Yes /


6. I worry a lot of the time. Yes / No

7. I am afraid of a lot of things. Yes /


8. I am always kind. Yes / No

9. I get mad easily. Yes / No

10. I worry about what my parents will say to me. Yes /


11. I feel that others do not like the way I do things. Yes / No

12. I always have good manners. Yes /


13. It is hard for me to get to sleep at night. Yes /


14. I worry about what other people think about me. Yes / No

15. I feel alone even when there are people with me. Yes /

16. I am always good. Yes / No

17. Often I feel sick in the stomach. Yes /


18. My feelings get hurt easily. Yes / No

19. My hands feel sweaty. Yes /


20. I am always nice to everyone. Yes /


21. I am tired a lot. Yes /


22. I worry about what is going to happen. Yes /


23. Other children are happier than I am. Yes /


24. I tell the truth every single time. Yes /


25. I have bad dreams. Yes / No

26. My feelings get hurt easily when I am fussed at. Yes /


27. I feel someone will tell me I do things the wrong way. Yes /

28. I never get angry. Yes / No

29. I wake up scared some of the time. Yes /


30. I worry when I go to bed at night. Yes / No

31. It is hard for me to keep my mind on my schoolwork. Yes /


32. I never say things that I shouldnt. Yes / No

33. I wriggle in my seat a lot. Yes / No

34. I am nervous. Yes / No

35. A lot of people are against me. Yes /


36. I never lie. Yes / No

37. I often worry about something bad happening to me. Yes /


Above written by: Ms. Sharon Gilroy

Reviewed and edited by: Dr. Grant J. Devilly