1129837

The Study of Stress and Competence in Children: A Building Block for Developmental
Psychopathology
Author(s): Norman Garmezy, Ann S. Masten and Auke Tellegen
Source: Child Development, Vol. 55, No. 1 (Feb., 1984), pp. 97-111
Published by: Wiley on behalf of the Society for Research in Child Development
Stable URL: https://www.jstor.org/stable/1129837
Accessed: 15-05-2019 03:12 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Society for Research in Child Development, Wiley are collaborating with JSTOR to
digitize, preserve and extend access to Child Development
This content downloaded from 111.68.106.154 on Wed, 15 May 2019 03:12:33 UTC
All use subject to https://about.jstor.org/terms
Empirical Articles
The Study of Stress and Competence in

Children: A Building Block for Developmental
Psychopathology
Norman Garmezy, Ann S. Masten, and Auke Tellegen

University of Minnesota
GARMEZY, NORMAN; MASTEN, ANN S.; and TELLEGEN, AUKE. The Study of Stress and Compe-
tence in Children: A Building Block for Developmental Psychopathology. CHILD DEVELOP-
MENT, 1984, 55, 97-111. This article discusses the building blocks for a developmental
psychopathology, focusing on studies of risk, competence, and protective factors. The current
Project Competence studies of stress and competence are described, with particular attention to
the methodology and strategies for data analysis. The authors present a 3-model approach to stress
resistance in a multivariate regression framework: the compensatory, challenge, and protective
factor models. These models are illustrated by selected data. In the concluding section, an evalu-
ation of the project is offered in terms of future directions for research.
Development has been described as a gram (Garmezy, 1973, 1975, 1978, 1981)
"unifying concept" in the study of psy- were directed toward manifestations of
chopathology, the "crucial link" between competence and incompetence in children
biological and psychological causation and considered to be at risk for psychopathology
the genetic and environmental determinants on the basis of maternal diagnosis (schizo-
of mental disorders (Eisenberg, 1977). Un- phrenia, affective disorder, personality dis-
fortunately, this valuable viewpoint is more order) or manifest disturbance in the child
appropriate as a promissory note than as a (externalizing, internalizing, or hyperactive
realistic assessment of the present. behaviors). The dependent variables were
Yet there are indications of advances primarily those of social competence as in-
dexed by peers and of laboratory measures of
that are clearly developmental in both
attentional functioning. The choice of atten-
orientation and content and speak to major
tion as an important variable to evaluate was
research issues in psychopathology. Three
based on the assumption that deficits in this
such areas have been central in a long-term
area of functioning, so evident in adult
research program conducted by a group of
schizophrenic
University of Minnesota investigators: the patients, might not only be a
significant precursor to later psychopathol-
study of children at risk for psychopathol-
ogy; the effects of stressful life events on the and thus indicative of risk status, but
ogy,
also a dispositional attribute fundamental to
functional competence of children; and
the acquisition of competence. In turn it was
protective/competence factors influencing
stress resistance in children. hypothesized that the competence level of a
child, even if at risk, might serve as a protec-
Project Competence tive factor against the expression of behavior
disorder. These studies have been described
A brief review of the early and recent elsewhere (see Driscoll, in press; Garmezy
research programs reveals the continuity in & Devine, in press; Marcus, 1972; Nuech-
these studies over the past decade. For more terlein, 1983; Phipps-Yonas, in press; Rolf,
than 10 years the efforts of the research pro- 1972).
Preparation of this manuscript was facilitated by grants from the National Institute of Mental
Health (MH-33222), the William T. Grant Foundation, a Research Career Award to the first author
(MH-14914), and the University of Minnesota Computer Center. Request for reprints should be
sent to Norman Garmezy, Department of Psychology, University of Minnesota, N419 Elliott Hall,
Minneapolis, Minnesota 55455.
[Child Development, 1984, 55, 97-111. @ 1984 by the Society for Research in Child Development, Inc.
All rights reserved. 0009-3920/84/5501-0024$01.00]
98 Child Development
In general, results of earlier studies of would be defined and redefined as data
these index children of schizophrenic analysis proceeded. Such multimethod ap-
mothers suggested the presence of outlierproaches to various aspects of each major
groups marked by deficiencies in social construct have required a repeated cycle of
competence as rated by peers and in atten- data reduction and data integration. It has
tional dysfunction, twin factors suggestive ofbeen necessary to analyze or to develop the
a greater potential risk for later maladapta- basic psychometric properties of several
tion. But equally important were the largemajor instruments, in a continuing effort to
numbers of vulnerable children who pro- improve the quality of measurement of the
basic procedures, since even the best avail-
vided little, if any, indication of current signs
of pathology or incompetence. The majority able instruments often have required further
of vulnerable children (except for the anti- analysis of their reliability and validity.
social group) revealed adaptive patterns of
social behavior and work achievement- COHORT IDENTIFICATION
behaviors that had been noted by Bleuler
From the inception of the current
(1978) in children of a large group of schizo-
studies, cohorts of children were sought in
phrenic parents who had served as partici-
which the type of stressors that defined each
pants in his 20-year longitudinal study of the
group was identifiably different. Three such
course and outcome of their schizophrenic
disorder. These facts led to a decision tocohorts
turn are included in the present program.
the Minnesota project to the study of stress
Cohort 1 is a community-based group com-
resistance in children. prising approximately 200 children drawn
from two centralized urban samples (N =
An initial report of the methods, vari- 612), whose parents contributed information
ables, and preliminary findings of the first about
3 their specific exposure to a heteroge-
years of the project recently appeared (Gar-neous sampling of stressful life events.
mezy & Tellegen, 1984). What follows is Cohorta 2 was a small group of 32 children
presentation of the research and data whose infancy and early childhood were
analysis strategies and selected results thatmarked by the stress of a profound life-
illustrate the directions now being taken by threatening congenital heart defect
the research group. (O'Dougherty, Wright, Garmezy, Loewen-
In a concluding section an evaluation son,
is & Torres, 1983). Cohort 3 consisted of a
group of 29 severely physically handicapped
offered of the program's potential utility as
one of many building blocks needed in the children who faced the stress of leaving the
collective effort by investigators to createspecial
a school for the handicapped they had
attended to meet the challenge of being
scientific base for a developmental psycho-
pathology that must follow. "mainstreamed" in a regular classroom of a
larger public school (Raison, 1982; Silver-
stein, 1982).
The Project Competence Studies of
Stress Resistance in Children The Community Cohort Samples
One of the major issues in initiating Thea Cohort 1 study began with two
school
program on stress resistance in children was populations of third through sixth
graders. Both school A and school B, located
to define the major constructs in operational
terms; "stress resistance" was defined asinthe
a single architectural complex, serve a
manifestations of competence in children central city community with a substantial
despite exposure to stressful events. proportion
Sub- of both working-class families
sequently procedures were developed to andas-single-parent households headed by
sess competence, stress, and selected mothers,
qual- many of whom receive government
ities hypothesized to be related to coping assistance. Sampling and assessment began
and resiliency, including individual dispo-in school A. School B studies followed a year
sitional attributes, characteristics of the later
fam- and were designed to be a replication
sample, an opportunity to cross-validate re-
ily milieu, individual developmental char-
sults as well as to correct inadvertent errors
acteristics, and parental attributes.
in sampling and design. The basic pool of
Each of two major constructs, compe-approximately 600 children in these popula-
tence and stress, has been operationally tions has been reduced during the sequen-
defined by multiple measures involving a of sampling involved in gathering
tial stages
series of analyses at increasing levelstheofcompetence/stress data base over a span
comprehensiveness. A "bootstrapping"ofpro- 2 years. These procedures have been de-
cess was anticipated in which constructs scribed by Garmezy and Tellegen (1984).
Garmezy, Masten, and Tellegen 99
Approximately 200 children and their hours of interviewing, each interviewer
families have participated in all of the most rated the stress level of the family for a given
intensive aspects of the project, including 6 year using a five-point scale, ranging from 1
hours of parent interviews, 2 hours of inter-("very little stress") to 5 ("a great deal of
views with the index child, a variety of labo-stress"). Interrater reliability, which aver-
ratory studies focusing on attributes of the aged .75, was satisfactory for this global
index child, and the administration of thescale.
various measures of stress and competence
The correlations between the LEQ
developed by project investigators.
stress scores and interviewers' ratings were
AN OVERVIEW OF METHODOLOGY
as follows: for school A, .45 at time 1 and .49
at time 2; for school B, .59 at time 1 and .55 at
Garmezy and Tellegen (1984) havetimede-2. Although significant, these correla-
scribed the history and development oftions
the are fairly modest for two measures of
cohort 1 measures as well as the sampling
the same construct. However, a strong de-
procedures. Much of the data analysisgreecon-
of relationship between these two mea-
sures was not expected primarily because
cerning individual measures of competence
and stress have been documented in the the measures are quite different. While the
technical reports of Project Competence,
LEQ provides only a count of events, the
whose principal investigator is Norman interviewer was making a composite judg-
Garmezy (Notes 1-6). ment based on detailed knowledge of the
family. The interview not only provided
Measures of Stress
greater knowledge of the stressful events not
The life events method.-The cohort 1
tapped by the LEQ but also led to informa-
sampling process was initiated with a rapid
tion on pervasive chronic stressors (rather
survey of life events, which was sent to all
than those that had occurred during the past
the parents of the third through sixth grad-
year) in the family such as illnesses, poverty,
ers. The Life Events Questionnaire (LEQ)
and disability of the parents as reflected in
developed by the project was based onalcoholism
the or diminished emotional or
work of Coddington (1972a, 1972b). Stress
mental capacity, which were not reported on
scores on the LEQ measure have shown
the LEQ, as well as information on the com-
moderate 12-month stability at both schools
petencies of the family that might mediate
(r = .60 and .53 at schools A and B, re-
the impact of stressful experiences.
spectively).
Interview-based methods.-Mindful of Socioeconomic status (SES).-It was
essential to measure SES in order to address
the limits of life events methods (Garmezy &
two6 important issues: (1) the relationship of
Tellegen, 1984), a set of interviews totaling
SES to the competence level of the child,
hours in length was conducted with the
and (2) the role of SES as a mediator of the
mothers of this cohort to obtain more com-
prehensive information about the reported extent and impact of stressors on the family.
Data pertinent to SES were collected during
stressors. The final interview was designed
to probe the context of all negative and am-the first parent interview. The choice of an
SES index raised some issues discussed both
biguous (i.e., positive or negative content to
in Cummings and Mulrooney (Note 5) and
be determined by inquiry) life events that
by Garmezy and Tellegen (1984). The Dun-
had occurred for a family during the 2 years
can Socioeconomic Index was chosen for
of time covered by two successive Life
Event Questionnaires. The contextual inter- data analysis.
view data, now being analyzed by Linder Measures of School-Based Competence
(Note 7), are providing several types of The initial approach to competence was
stress-related variables, including modifier to assess three areas of a child's functioning
scores, a global contextual life events score, in the school context: academic achieve-
and a "corrected" life events score that in-
ment, classroom behavioral competence,
cludes additional events and multiple occur-
and interpersonal (social) competence. The
rences of the same event and corrects errors rationale for these choices has been dis-
in LEQ responding. cussed by Garmezy and Tellegen (1984).
Given the time-consuming nature of Academic achievement.-Two indexes
analyzing interview data, a third interim ap-of academic achievement were obtained.
proach to assessing stress was planned inCumulative record data including grades
which the mothers' interviewers providedand standardized test scores were secured
subjective global ratings of stress. Based on 6for a subsample of children whose parents
gave permission to review school records. On the basis of these analyses two com-
Achievement was also assessed by means posites
of were created: an engaged-disen-
the Peabody Individual Achievement gaged
Test (Engaged) score summing the relevant
(PIAT) (Dunn & Markwardt, 1970) to obtain
two positive and two negative teacher/peer
scores;
a more systematic and standardized index of and a classroom disruptiveness (Dis-
achievement based on an individuallyruptive)
ad- score. The DESB teacher rating
ministered test. "performance anxiety" was excluded from
the composite scores for several reasons, in-
Classroom behavioral competence.--cluding a reduced number of subjects and its
Behavioral competence in the classroom was weaker psychometric properties. Achieve-
measured by means of teacher ratings on thement was excluded so that ability/
Devereux Elementary School Behavior achievement/SES issues could be examined
(DESB) rating scale (Spivack & Swift, 1967).independently of the global "competence"
In view of the DESB factor structure in scores.
combined school A and school B samples (N

= 648), four scores were computed: disrup- The stability of the two composite
tive-oppositional, poor comprehension-dis- scores was computed by correlating time 1
attention, cooperative-initiating, andand per-time 2 scores for the school A sample.
formance anxiety (see Finkelman & Fer- Both DESB and revised class play scores
rarese, Note 3; Ferrarese & Finkelman, Note were obtained twice in school A, with an
8). interval of 17 months between administra-
tions. Correlations were .73 (N = 159) and
Interpersonal competence.--Peer as- .71 (N = 160), respectively, for the En-
sessments of reputation pertaining to social gaged and Disruptive scores.
competence were obtained by means of a
modified Class Play technique (Bower, Measure of general intellectual abil-
1969), the Revised Class Play (Masten, ity.--Intellectual ability was estimated by
Morison, & Pellegrini, Note 9). Factor the administration of two subtests of the
analysis revealed that three reliable dimen-revised Wechsler Intelligence Scale for
sions of peer reputation are measured by thisChildren (WISC-R), vocabulary and block
instrument: sociability-leadership, aggres- design, the two-test short form that has
sive-disruptive, and sensitive-isolated. shown the highest correlation (r = .88) with
the full-scale IQ score (Silverstein, 1975).
Composite competence scores.--One
Interview-based
aim in the process of data analysis wasMeasures
to Child Competence
reduce the number of variables for mul-
Both the parent and child interviews
tivariate analysis. Toward that aim, re-
lationships were examined among the provide extensive information pertinent to
the competence of the cohort 1 children. The
school-based competence variables, in-
cluding three peer, four teacher, and two parent interview, involving hundreds of
achievement scores. Factor analyses of thevariables, is still being processed through
peer and teacher ratings resulted in two the lengthy and complex data reduction
major factors. Loading highly on the firstcycle required by the extensive qualitative
factor were the following measures: from theand quantitative data that interviews gener-
ate. Data provided by the parents about their
peer measure, sociability-leadership (posi-
tive loading) and sensitive-isolated (negativechildren are being empirically examined to
loading); from the teacher measure, co- derive possible competence scales based on
operative-initiating (positive loading) andparent perceptions. A global child compe-
poor comprehension-disattention (negativetence rating was also obtained through the
loading). (If included, the academic parent interview (comparable to the inter-
achievement scores also loaded on this first viewer's global rating of stress). This is a
factor.) This bipolar factor appears to reflect rating on a five-point scale following 6 hours
of interviewing by interviewers who were
the extent to which the child is actively en-
gaged in the work of the classroom; there-blind to the status of the target child in terms
fore, it was called "engaged-disengaged." of their school-based competence measures.
Loading highly on the second factor were The child interview data in school A
the peer score aggressive-disruptive and the have been analyzed by Finkelman (1983),
teacher score disruptive-oppositional. This who developed a set of 173 rating scales that
factor appears to reflect disruptive or nega- were used by independent clinical judges to
tive social behavior; it has been labeled evaluate the extensive contents of the inter-
"classroom disruptiveness." view. Once the school A interview ratings
were completed, Finkelman employed fac-
pulsivity and delay of gratification. And
tor analysis as one method for reducing theMasten (1982) considered two related areas
data. After a series of analyses three primary
of ability not measured by traditional tests of
factors emerged, two of which resemble theintelligence-namely, creative (divergent)
two school-based competence composites, thinking and humor.
Engaged and Disruptive. The first factor de-
These three doctoral studies, each of
scribes children who are academically
which employed diverse measures of the
oriented, psychologically minded, have a
variables under investigation, represent
broad range of interests, are not isolates,
both a variety of methods of sampling a
have a close relationship with both siblings
given child's performance under relatively
and parents, show a high degree of curiosity
challenging conditions, and a variety of
in their world, and have a good sense of
skills/processes from the "intelligence" do-
humor. In the interview itself, they were ob-
main, using that term in a broad sense.
served to be cooperative and candid, and Separately, each study indicated areas of
they had little difficulty understanding or functioning related to competence beyond
being understood by the interviewer. This the role of the IQ score. However, in many
child interview factor has a moderately cases a substantial degree of shared variance
strong correlation with the global compe-
was also found between these "laboratory"
tence score Engaged: for school A, .52 at measures and the IQ score. The integration
time 1 and .60 at time 2.
of results obtained in these studies may shed
In a comparable fashion the second some light on the pervasive relationship of
child interview factor bears a strong re- intelligence to competence. It may also be
semblance to the composite Disruptive possible to identify through these analyses
specific aspects of intelligent behavior that
score. Children who score high on this factor
get into trouble at school and at home, tend are modifiable and might point to interven-
to be irresponsible about doing their chores,tion strategies, such as training and re-
are not very conscientious, frequently get inforcement of a more reflective cognitive
into fights and disagreements with otherstyle.
children, are physically aggressive, selfish,
cynical, and egotistical, tend not to save theirInterview-based Measures of Milieu and
Child Attributes
money, and in general show a low level of
delay of gratification. The correlation of this The child and parent interviews have
interview factor with the school-based Dis- been discussed above with regard to
ruptive score (school A) is .55 at time 1 and broadening the assessment of the target
.52 at time 2. child's competence. However, these data
sets contain much information relevant to
A third factor included items related tothe search for competence modifiers or pro-
self-esteem, level of aspiration, striving, and
tective factors. For example, it is anticipated
achievement motivation. The aspiration- that the competence of the primary caregiver
motivation score based on this factor shows a is a significant factor in the competence and
pattern of correlations with competence stress-resistance qualities of the child. In-
similar to the first child interview factor ex- formation is also available, for example, re-
cept its correlations with IQ/achievement garding family history and demographics,
scores are lower. social support network, the developmental
The emergence of similar competencehistory, and current behavior of the target
factors from school-based teacher/peer datachild. Data analyses are now underway to
reduce the interview data to a workable set
and the child interview suggests that con-
sistent and meaningful constellations of of major variables related to competence
competent behaviors and qualities are being and/or protective factors.
identified in the project.
MODELS OF STRESS RESISTANCE IN A
"Laboratory" Measures of Child Attributes MULTIVARIATE REGRESSION APPROACH
Three major studies were conducted TO DATA ANALYSIS
with cohort 1 to observe the children in more
controlled conditions in an effort to evaluate The process of data analysis has pro-
ceeded in stages, with a sequential pattern of
factors hypothesized to temper stressful ex-
periences or to enhance competence at all preliminary data reduction (e.g., through
levels of stress. Pellegrini (1980) examined factor analyses when sample sizes permit),
the role of social cognition. Ferrarese (1981) followed by multiple regression analyses
explored the dimensions of reflectivity-im- connecting major variables from different
data domains. In doing so, an effort has been is the simplest one. Stress factors and per-
made both to preserve the exploratory sonal attributes are seen as combining addi-
character of the inquiry while at the same tively in the prediction of competence; that
time maintaining the focus on major sources is, with attributes held constant, competence
of variation and covariation, avoiding exces- could covary negatively with stress level.
sive multiplication of variables and re- Conversely, holding stress constant, com-
lationships, and employing cross-validation petence covaries positively with strength of
designs when possible. adaptive attributes. In this model the impact
of severe stress (across individuals) can
Multiple regression/correlation (MRC)
therefore be counteracted or compensated
methods of data analysis, particularly the
for by personal qualities of strength, hence
hierarchical approach described by Cohen
its name. The model is accommodated by
and Cohen (1975), offer a means for clarifying
the familiar simple linear multiple regres-
complex relationships inherent in correla-
tional data such as those collected on the sion model. In the case of a single attribute A
and a single stress variable S, we may write
project.
for the predicted value of competence vari-
Cohen and Cohen have pointed out the able C: 0 = BjA + B2S + D, where B, and B2
strategic advantages of MRC as a general are regression weights and D is a constant.
method of analysis. It allows one to use the
quantitative variables of interest in continu-
The second model may be called the
challenge model. In this model, stress is
ous form, avoiding the loss of information
treated as a potential enhancer of compe-
inherent in the usual ANOVA procedures. It
tence, provided the degree of stress is not
provides results that indicate both the re-
excessive, by allowing for a curvilinear rela-
lationship of a factor to the dependent vari-
tion between stress and competence. This
ables as a whole and its partial relationship
(or net contribution) after other factors are calls for addition of a quadratic term in the
considered. Moreover, MRC allows for the regression equation: C = BjA + B2S + B3S2
+ D.
ready consideration of nonlinear as well as
linear relationships and of general or con- In the third model there is a conditional
ditional relationships. relationship between stress and personal at-
tributes with respect to adaptation, such that
There are several basic approaches for
MRC analysis of which the hierarchical personal' attributes modulate (dampen or
amplify) the impact of stress as a variable.
model was chosen as the most appropriate.
Thus, when certain positive personal attri-
This method is preferable to the simultane-
ous model when there is a logical or tem- butes (or protective factors) are present,
variations in stress will be less strongly re-
poral priority for ordering the variables
flected in variations in quality of adaptation
under consideration. Moreover, this is the
method of choice for partitioning the total
than when these attributes are lacking,
variance in the dependent variable because suggesting that the personal qualities in
it specifies the unique contribution of each question impart a kind of "immunity"
against stress. The reverse is true for attri-
independent variable when included as a
separate step in the analysis (Cohen & butes of personal vulnerability; in their
Cohen, 1975). presence the impact of variations in stress on
competence becomes more, rather than less,
Although necessarily exploratory, these pronounced. This model may be called the
MRC analyses can be guided by certain immunity-versus-vulnerability model or the
generic models that specify relationships of protective factor model. The corresponding
a certain form. The desirability of adopting regression equation incorporates a linear x
such models as a way of providing focus and linear interaction term: C = BA + B2S +
of controlling proliferation of chance B3SA + D.
findings has recently been discussed at some
length by Tellegen, Kamp, and Watson The last model may become clearer if
(1982). The analysis of relations between the last equation is rewritten as 0 = B,A +
stress and competence may be considered (B2 + B3A)S + D, showing that the weight of
briefly as an example. the stress variable S (that is, B2 + B3A) in-
deed becomes larger or smaller in part as a
Three generic models suggest them- function of the weighted magnitude of the
selves for describing the impact of stress and personal attribute, B3A, which therefore can
personal attributes on quality of adaptation. be said to function as a protective or vulner-
The first of these, the compensatory model, ability-increasing factor, depending on the
positive or negative sign of B3. Equivalently, ronment variables in cohort 1 (sex, age, IQ,
rewriting the equation as 0 = B2S + (B1 + SES, and the life events measure of stress) as
B3S)A + D shows that the role or weight of independent variables.
the personal attribute may become more or
Generally, the hierarchical ordering of
less manifest under increased stress, again
variables in the following analyses was
depending on the sign of B3.
based on several major considerations. (Pos-
The compensatory, challenge, and sible rationales for ordering variables in
immunity-versus-vulnerability models are hierarchical analysis have been discussed by
not mutually exclusive; they may be com- Cohen and Cohen [1975].) Priority was given
bined. Each, however, serves to guide to "organismic" characteristics over "en-
analyses of relations between the major situ-vironmental" characteristics of the individ-
ational and personal variables. The chal- ual case, in keeping with the psychological
lenge and immunity models, in particular,focus of this investigation. Thus, cognitive
variables preceded SES or stress ratings. All
could be referred to as two different types of
"coping" models in that they capture special interactions followed the individual vari-
forms of relationships between personal at- ables as required to obtain their unique
tributes and stressful circumstances on the contribution to the R2 after main effects were
one hand and adaptation on the other, which entered. Thus, main effects took precedence
emerge when adaptation is instigated byover interaction effects. This priority order-
stress itself. ing is not strictly for structural purposes but
is also desirable in view of the more explor-
ILLUSTRATIONS OF THE DATA ANALYSIS atory nature of all interactions as compared
PROCESS: SELECTED RESULTS to main effects. Within the organismic
characteristics, priority was given to simple
In the cohort 1 data set, data analysis has
(e.g., sex, age) over complex characteristics
followed the process, described above, of
(e.g., IQ) and to temporal precedence (e.g.,
data reduction, multiple regression/correla-
sex before age).
tion analyses (MRC), and reexamination
resulting in further reduction and integra- Table 1 presents the simple inter-
tion of sets of variables, aimed at the de-
correlations of the school-based competence
velopment of constructs and hypothesis variables, the basic subject variables, the
generation. For example, the assessment LEQ andof interviewer's stress scores, and
competence began in the school andSES. in-Several general points may be noted.
cluded multiple measures, two of which There is a strong relationship among the
were the DESB teacher rating scale and variables
the tapping engaged classroom be-
Revised Class Play peer assessment instru- havior, achievement, and intellectual ability.
ment. The peer and teacher data were Ratings of classroom disruptiveness in this
separately analyzed to determine the mostsample show very little relationship with
reliable and meaningful scores for each mea-these variables. IQ appears to have more re-
sure (Finkelman & Ferrarese, Note 3; Mas-lationship to achievement/competence mea-
ten et al., Note 9). Six of these scores weresures than does SES, whereas SES has
combined into two composite school com- slightly stronger correlations with stress.
petence scores: Engaged and Disruptive Stress appears to have a small inverse re-
(described above). Currently, other major lationship to the Engaged aspect of school
sources of child competence data, the parent competence and a small positive rela-
and child interviews, are being analyzed tionship to the Disruptive dimension.
with a view to creating new combinations
To refine and amplify on the qualitative
within and beyond school domains of func-
impressions one might draw from these sim-
tioning.
ple correlations, a series of hierarchical
Meanwhile, the school-based compe- MRC analyses have been carried out. Tables
tence scores can be related to stress mea- 2 and 3 present the results of the two hierar-
sures and to SES, IQ, and laboratory chical mea- regression analyses with the school-
sures. To illustrate the application of thecompetence composite scores as the
based
generic models of coping and stress dependent
resis- variables. The first five steps in-
tance described above, examples of regres- clude singular personal attributes and milieu
sion analyses will be presented below, with (SES and stress) that might be re-
attributes
engaged, disruptive, and achievement lated to school competence. Significant ef-
(PIAT) scores as the dependent variables fects would be additive. The most striking
and several of the basic subject and envi- result for the Engaged dimension of compe-
00 - -4 - 00 I
CO
UI I I I
0 0 co.. . .
ou * ****
I 0
<: o o ?
* +*
cnq
z
Or) CO),C#13 C
U z~
U - Ncq0o
.c 00 0**
H00 C O)
bO oo C,6
u 'I
Q)w
zz zt
0 0
"0
00
Un U
-W,
' . lw .
tence is the substantial increment in R2 pro- this competence composite than does SES.
vided by the estimate of intellectual ability Socioeconomic status also is significantly
(IQ). After the negligible effects of sex and and inversely related to the Disruptive di-
age are controlled, the IQ score adds 34% mension.
to the "explained" variation in Engaged. In
Life stress (the LEQ score) does not add
contrast, IQ apparently has very little to do to the multiple correlation as step 5 for En-
with variation in the Disruptive dimension.
gaged, but it is a significant factor in the Dis-
Socioeconomic status also adds a small
ruptive analysis. Higher stress scores appear
to be associated with higher disruptiveness
but significant increment to R2 in additionin to
the classroom.
the effects of sex, age, and IQ. If SES pre-
ceded IQ in this analysis, as step 3, it wouldSex also is significantly related to the
account for more variance than it does as Disruptive dimension, although not to En-
step 4, but it would not nearly rival IQ.gaged.
In Boys have higher Disruptive scores.
It should be noted that this effect is due to a
other words, there is shared variance among
the SES, IQ, and Engaged variables, butsex
IQeffect demonstrable only with teacher
ratings since the peer scores included in this
has a stronger independent relationship with
TABLE 2
RESULTS OF THE HIERARCHICAL MULTIPLE REGRESSION ANALYSIS FOR THE

PREDICTION OF THE ENGAGED-DISENGAGED COMPETENCE COMPOSITE
Overall
Step and Signifi- Signifi-
Independent cance Multiple R2 cance
Variable r F Levela R R2 Change Level
1. Sex ............. .15 2.18 .143 .15 .02 .02 .143
2. Age ............. .04 .17 .678 .16 .03 .00 .316
3. IQ .............. .54 45.92 .000 .60 .36 .34 .000
4. SES ............ .30 4.99 .028 .63 .40 .03 .000
5. Stress (LEQ) .... -.15 1.10 .298 .64 .40 .01 .000
6. Stress x stress ..... .13 .714 .64 .40 .00 .000
7. Sex x stress ......... .64 .426 .64 .41 .00 .000
8. IQ x stress ......... .01 .903 .64 .41 .00 .000
9. SES x stress ........ .25 .617 .64 .41 .00 .000
" These values indicate whether the increment in R2 provided
TABLE 3

PREDICTION OF THE DISRUPTIVE COMPETENCE COMPOSITE
Overall
Independent cance Multiple R2 cance
1. Sex ............. -.24 5.67 .019 .24 .06 .06 .019

2. Age ............. -.03 .14 .713 .25 .06 .00 .062
3. IQ .............. -.06 .99 .323 .27 .07 .01 .089
4. SES ............ -.20 4.14 .045 .34 .11 .04 .032
5. Stress (LEQ) .... .29 4.83 .031 .40 .16 .05 .009
6. Stress x stress .... 3.32 .072 .44 .19 .03 .005
7. Sex x stress ......... .06 .806 .44 .19 .00 .010
8. IQ x stress ......... .37 .544 .45 .20 .00 .016
9. SES x stress ....... 1.00 .320 .46 .21 .01 .020
" These values indicate whether the increment in R' provide
composite are standardized within sexwend
and & Dohrenwend, 1969; Kessler, 1979).
In the two analyses reported here, there is
therefore cannot vary significantly by sex.
no evidence that the effect of stress is con-
Within the model, with respect to the
ditional on SES. This hypothesis will be in-
Disruptive analysis, SES might be viewed
vestigated further in the school B sample
as a "compensatory" environmental attri-
and with respect to additional measures of
bute, serving to make up for stressful life
competence and stress. If such an effect is
event effects. Given equal stress levels,
found, it will be important to examine the
children from higher SES backgrounds
global construct of SES further by evaluating
would be expected to have lower Disruptive
scores.
which aspects of demographic or other en-
vironmental data available best explain the
Step 6 in these regressions relationship
represents of SES
a and stress.
test of the challenge model for this data set.
Finally, it is noteworthy that these 10
One curvilinear function was tested by in-
steps combined account for twice as much
serting a "stress squared" term. This term is
variation in the Engaged dimension as in the
not significant for either Engaged or Dis-
Disruptive dimension.
ruptive.
These two components of school func-
Steps 7, 8, and 9 test for possible inter-
tioning are relatively independent of one
action effects of interest. None are signifi-
another and show different patterns of re-
cant. A significant interaction effect might
lationship with other variables. These early
have suggested a protective/vulnerability
results suggest that stress is more related,
factor where the relationship among com-
petence, stress, and the attribute is con- at least in the short run, to the Disruptive
dimension.
ditional. For example, a stress-resistance or
protective factor might not differentiate the Perhaps engaged behaviors are more
competence of children at lower stress cumulative, or perhaps this aspect of reputa-
levels; but at high stress levels, competence tion among peers and teachers is more re-
could vary as a function of the attribute. sistant to change. It would be interesting to
compare the long-term relationships be-
A stress-resistance/vulnerability "attri-
tween each of these two aspects of school
bute" may also be environmental in nature.behavior and stress.
In this example, it is interesting to consider
step 9, the interaction of SES and stress. So- Table 4 presents the summary table of
cial status has received considerable atten- the correlations between the PIAT achieve-
tion as a moderator of the effects of stress ment measure and the comparable set of in-
(Sandler, 1980). Studies have suggested thatdependent variables. In this case the large
stressful events have a greater impact on role of IQ was predictable. In addition to the
lower-class individuals than those of higher IQ variance, SES and stress add small but
SES (Dohrenwend, 1970, 1973; Dohren- significant increments to R . In this analysis
TABLE 4

PREDICTION OF ACHIEVEMENT (PIAT)
Overall
Independent cance Multiple R' cance
1. Sex ............. -.05 .26 .613 .05 .00 .00 .613
2. Age ............. -.06 .38 .540 .09 .01 .00 .730
3. IQ .............. .75 107.48 .000 .75 .57 .56 .000
4. SES ............ .31 5.04 .028 .77 .59 .03 .000
5. Stress (LEQ) .... -.17 4.73 .033 .78 .62 .02 .000
6. Stress x stress ..... .39 .532 .79 .62 .00 .000
7. Sex x stress ......... .91 .342 .79 .62 .00 .000
8. IQ x stress ......... 4.04 .048 .80 .64 .02 .000
9. SES x stress ....... 3.53 .064 .81 .66 .02 .000
a These values indicate whether the increment in R' provided by
there is, interestingly, also a significant This interaction (step 8) is also
interaction effect. The relationships of IQ significant for the other achievement mea-
and stress to PIAT performance appear to be sure based on school records. If this result is
conditional. replicated in the school B sample, one may
have more confidence that IQ indeed func-
To interpret this interaction, the same
tions as a protective factor in these circum-
simultaneous regression equation at step 8 stances. If this were the case, additional,
(where the IQ x stress interaction term was
more fine-grained analyses would be in
added) was solved four times with repre- order. Intervention would be furthered by
sentative high and low values of IQ and
data pointing to specific and remediable as-
stress, setting all other values at the mean.
pects of intelligent functioning.
"High" and "low" values were arbitrarily
selected at 1 SD above or below the mean. These data are presented to illustrate
The resulting illustrative values are pre- the strategies currently employed in this re-
sented in Figure 1. The graph shows that search
the effort. It would be premature at this
interaction provides an example of the time to draw any but tentative conclusions
immunity-versus-vulnerability model. It from these examples since they illustrate
also indicates a compensatory effect. As for work in progress.
the latter, PIAT performance, regardless of
stress level, covaries with IQ. As for the Future Directions for Research
interaction, IQ also appears to function to
some degree as a protective factor. High-IQ An effort has been made here to presen
children maintain good PIAT performance at the central issues, models, and concepts
both low and high levels of stress, whereas the ongoing Project Competence researc
for low-IQ children, PIAT performance program, particularly through a look at th
drops off as a function of higher stress. community cohort studies. The primary
High
IQ
100
> 90
Low
IQ
80
Low High
STRESS
FIG. 1.-Interaction of stress (LEQ scores) and IQ with respect to achievement (PIAT sc
strength of this program lies in the linkage focus.
of An effort will be made to examine
competence and stress by means of multi- changes and patterns over time with respect
method, multidomain assessments. Consid- to competence, attributes of mastery, and
erable attention has been given to psycho- stress. Outcome will be assessed with regard
metrics necessary for building a sound em- to psychopathology as well as competence.
pirical base in a territory that is at bestIt should be possible to evaluate the relative
only partially developed. For example, in- predictive validity of the extensive stress/
struments such as the Life Events Ques- coping/competence data already collected
tionnaire and Revised Class Play have been and to determine the long-term stability of
developed and subjected to extensive basic measures. Better causal analyses will
analyses. Such basic psychometric work ac- be possible in this short-term longitudinal
companied by extensive cross-sectional design (e.g., third graders last contacted in
analyses is needed to identify important dis- grade 5 will be in grades 9 and 10 at
tinctive parameters of competence and re- follow-up). Children in the samples will
silience in current studies and for future in- have confronted several important mile-
vestigations cast in a developmental-process stones, such as transition into and out of
framework. junior and senior high school and, in some
Analyses now underway in the commu-
cases, to independence from family. They
nity cohort include (1) cross-validation ofwill also have encountered a variety of addi-
school A results in the school B sample, (2)
tional stressful experiences. Continuities
reduction of parent and child interview data,
and discontinuities across this time span
and (3) integrations of multiple perspectives
should prove illuminating.
of assessment within constructs. For exam- Nonetheless, follow-up will not provide
ple, competence data will be integrated from data pertaining to many unanswered de-
three major perspectives: school (including velopmental questions in this area nor will it
teacher, peer, and achievement data), child be possible to observe patterns of coping in
interview, and parent interview. The as- progress. These designs and measures can
sessment of stress data is also being ap- only tap the cumulative products of adapta-
proached from multiple perspectives tion, such as those reflected in peer reputa-
through the use of such diverse methods astion or school achievement. However, by
the Life Events Questionnaire, the con- identifying major parameters of stress, cop-
textual life events parent interview, and theing, and competence and examining these
global interviewer ratings of stress. Analyses
over time, it may be possible to speculate in
linking the basic constructs will continue ata heuristic fashion on the developmental
increasing levels of complexity, basically byprocesses implicated in competent func-
means of the models and data analysis strat- tioning and resilience.
egies presented above.
The current program was also not in-
As a first effort, these studies were tended to provide a comparative examina-
necessarily limited. In future community tion of competence parameters and process
cohort studies, more attention will have toin a group with specific and manifest forms
be given in particular to developmental is-of psychopathology. Instead, it undertook to
sues and to psychopathology. In cohort 1 ofstudy competencies and resilience in a nor-
this program, developmental data were col- mative heterogeneous school population and
in two relatively homogeneous special
lected in the form of retrospective question-
naires and interview questions concerning
populations, a specific heart defect popula-
developmental history and health. However,tion (cohort 2) and a mainstreamed physi-
the design of this study was initially cross-cally handicapped population (cohort 3). By
sectional in nature, with a limited age rangecontrast, the earlier program of research em-
for the participating children. Similarly,
phasized the attributes of children who
although a symptom behavior checklist waseither showed manifest signs of maladapta-
obtained from parents, the major measures tion (externalizing, internalizing, and hyper-
employed in cohort 1 were not designed foractive symptomatology) or were considered
an in-depth assessment of psychopathology,to be at risk for later dysfunction by virtue
nor was the sample chosen to represent of parental disorder.
pathology as in the earlier studies of chil-
dren at risk. Studies of vulnerability, risk, protective
factors, competence, and developmental
In follow-up studies of the community
process in psychopathology each represent
cohort, these two issues will receive more essential building blocks in the foundation
of an empirically based developmental vidual and familial attributes whose neglect
psychopathology. Perhaps it is time to may have caused our predictions of disorder
strengthen the foundation by integrating to go awry. This latter function is one in
some of the results of this work into the which developmentalist and psychopathol
underlying theory and research designs of can join together in research that will, i
gist
the next generation of investigations. That
time, have significant implications for inte
new generation of studies will aim at com-vening to prevent or to amelioriate child
parisons implicating risk, protective, hood
and disorders.
competence factors in different cohorts of
children: the already disordered, those at
risk for behavior pathology, and the adaptive
ones who emerge from adverse as wellReference
as Notes
more advantaged backgrounds.
1. Herzog, J., Linder, H., & Samaha, J. The mea
Risk research, defined in terms of surement of stress: Life events and the inter-
anomalous experience, disordered parent- viewers' ratings (Project Competence Tech-
age, disturbed family and rearing milieus, nical Report No. 1). Unpublished manuscript
and disadvantaged environments, has be- University of Minnesota, September 1981.
come part of the scientific enterprise, en-2.
Masten, A. S., & Morison, P. The measure-
joying status as a relatively new investiga-
ment of interpersonal competence: The Class
tive mode in psychopathology research. The Play (Project Competence Technical Report
study of competence has enjoyed a similar No. 2). Unpublished manuscript, University
climate of acceptance, with major contri- of Minnesota, September 1981.
butions stemming from developmental and3. Finkelman, D., & Ferrarese, M. The mea-
personality psychologists. The incorporation surement of classroom behavioral compe-
of such factors in risk studies and in prog- tence: The Devereux teacher ratings (Project
nostic studies of disordered people provides Competence Technical Report No. 3). Un-
an important bridge between studies of published manuscript, University of Min-
normal and psychopathological groups. nesota, September 1981.
4. Cummings, L., & Mulrooney, S. The mea-
But the study of protective factors-- surement of work competence: The achieve-
those dispositional attributes, environmental
ment ratings of the Peabody Individual
conditions, biological predispositions, and
Achievement Test (Project Competence
positive events that can act to contain the
Technical Report No. 4). Unpublished manu-
expression of deviance or pathology-has
script, University of Minnesota, September
been marked by too long a period of neglect. 1981.
A recent report (Garmezy, in press) of ongo-
5. Cummings, L., & Mulrooney, S. The mea-
ing research in this area suggests that activ-
surement of general intellectual ability (Proj-
ity with regard to such factors is negligible.
ect Competence Technical Report No. 5).
However, the importance of the area and a
Unpublished manuscript, University of Min-
scientific agenda for the future has been
nesota, September 1981.
suggested by Rutter (1979, p. 70): "The ex-
6. Linder, H., Silverstein, P., & Samaha, J. The
ploration of protective factors in children's
measurement of socioeconomic status (Proj-
responses to stress and disadvantage has
ect Competence Technical Report No. 6).
only just begun. We are nowhere near the
Unpublished manuscript, University of Min-
stage when any kind of overall conclusions
nesota, September 1981.
can be drawn. What is clear, though, is that
7. Linder, H. Utilizing a contextual life events
there is an important issue to investigate.
interview as a measure of stress. Doctoral
Many children do not succumb to depriva-
dissertation in preparation, University of
tion, and it is important that we determine
Minnesota, 1983.
why this is so and what it is that protects
8. Ferrarese, M., & Finkelman, D. A factorial,
them from the hazards they face."
reliability, and validity study of the Devereux
Rutter's statement suggests a dual task Elementary School Behavior Rating Scale.
for developmental psychopathologists. The Manuscript in preparation, University of
study of the etiology, symptom expression, Minnesota, 1983.
course, and treatment of childhood disorders 9. Masten, A. S., Morison, P., & Pellegrini, D. S.
will remain a central one. But this more The Revised Class Play: Psychometric prop-
traditional orientation must be com- erties of a peer assessment instrument. Man-
plemented by a "new look," one that uscript
em- submitted for publication, University
phasizes those adaptive and resilient of
indi-
Minnesota, 1983.
References schizophrenia and in children at risk. Journal
of Psychiatric Research, 1978, 14, 3-34.
Bleuler, M. The schizophrenic disorders: Long- Garmezy, N. Children under stress: Perspectives
term patient and family studies. New Haven, on antecedents and correlates of vulnerability
Conn.: Yale University Press, 1978. and resistance to psychopathology. In A. I.
Bower, E. M. Early identification of emotionally Rabin, J. Aronoff, A. M. Barclay, & R. A.
handicapped children in school (2d ed.). Zucker (Eds.), Further explorations in per-
Springfield, Ill.: Thomas, 1969. sonality. New York: Wiley, 1981.
Coddington, R. D. The significance of life eventsGarmezy, N. Stress-resistant children: The search
as etiologic factors in the diseases of children, for protective factors. In J. E. Stevenson (Ed.),
I: A survey of professional workers.Journal of Aspects of current child psychiatry research,
Psychosomatic Research, 1972, 16, 7-18. (a) Journal of Child Psychology and Psychiatry
Coddington, R. D. The significance of life events Book Supplement No. 4. Oxford: Pergamon,
as etiologic factors in the diseases of children, in press.
II: A study of a normal population.Journal ofGarmezy, N., & Devine, V. Project Competence:
Psychosomatic Research, 1972, 16, 205-213. The Minnesota studies of children vulnerable
(b) to psychopathology. In N. Watt, E. J. Anthony,
Cohen, J., & Cohen, P. Applied multiple regres- L. C. Wynne, & J. E. Rolf (Eds.), Children at
sionlcorrelation analysis for the behavioral risk for schizophrenia. Cambridge: Cam-
sciences. Hillsdale, N.J.: Erlbaum, 1975. bridge University Press, in press.
Dohrenwend, B. S. Social class and stressful Garmezy, N., & Tellegen, A. Studies of stress-re-
events. In E. H. Hare & J. K. Wing (Eds.), sistant children: Methods, variables and pre-
Psychiatric epidemiology. New York: Oxford liminary findings. In F. Morrison, C. Lord, &
University Press, 1970. D. Keating (Eds.), Advances in applied de-
Dohrenwend, B. S. Social status and stressful life velopmental psychology (Vol. 1). New York:
events. Journal of Personality and Social Academic Press, 1984.
Psychology, 1973, 28, 225-235. Kessler, R. C. Stress, social status, and psychologi-
Dohrenwend, B. P., & Dohrenwend, B. S. Social cal distress. Journal of Health and Social Be-
status and psychological disorder. New York: havior, 1979, 20, 259-272.
Wiley, 1969. Marcus, L. M. Studies of attention in children
Driscoll, R. M. Intentional and incidental learning vulnerable to psychopathology. Unpublished
in children vulnerable to psychopathology. In doctoral dissertation, University of Min-
N. Watt, E. J. Anthony, L. C. Wynne, & J. E. nesota, 1972.
Rolf (Eds.), Children at risk for schizophre- Masten, A. S. Humor and creative thinking in
nia. Cambridge: Cambridge University Press, stress-resistant children. Unpublished doc-
in press. toral dissertation, University of Minnesota,
Dunn, L. M., & Markwardt, F. C. Peabody indi- 1982.
vidual achievement test. Circle Pines, Mont.: Nuechterlein, K. H. Signal detection in vigilance
American Guidance Service, 1970. tasks and behavioral attributes among off-
Eisenberg, L. Development as a unifying concept spring of schizophrenic mothers and among
in psychiatry. British Journal of Psychiatry, hyperactive children. Journal of Abnormal
1977, 131, 225-237. Psychology, 1983, 92, 4-28.
Ferrarese, M. J. Reflectiveness-impulsivity and O'Dougherty, M., Wright, F. S., Garmezy, N.,
competence in children under stress. Un- Loewenson, R. B., & Torres, F. Later compe-
published doctoral dissertation, University of tence and adaptation in infants who survive
Minnesota, 1981. severe heart defects. Child Development,
Finkelman, D. G. The relationships of children's 1983, 54, 1129-1142.
attributes to levels of competence and famil- Pellegrini, D. The social-cognitive qualities of
ial stress, Unpublished doctoral dissertation, stress-resistant children. Unpublished doc-
University of Minnesota, 1983. toral dissertation, University of Minnesota,
Garmezy, N. Competence and adaptation in adult 1980.
schizophrenic patients and children at risk. In Phipps-Yonas, S. Visual and auditory reaction
S. R. Dean (Ed.), Schizophrenia: The first ten time in children vulnerable to psychopathol-
Dean Award lectures. New York: MSS In- ogy. In N. Watt, E. J. Anthony, L. C. Wynne,
formation Corporation, 1973. & J. E. Rolf(Eds.), Children at risk for schizo-
Garmezy, N. The experimental study of childrenphrenia. Cambridge: Cambridge University
vulnerable to psychopathology. In A. DavidsPress, in press.
(Ed.), Child personality and psychopathol-
Raison, S. B. Coping behavior of mainstreamed
ogy: Current topics (Vol. 2). New York: physically handicapped students. Unpub-
Wiley, 1975. lished doctoral dissertation, University of
Garmezy, N. Attentional processes in adult Minnesota, 1982.
Rolf, J. E. The academic and social competence ofSilverstein, A. B. Validity of WISC-R short forms.
children vulnerable to schizophrenia and Journal of Clinical Psychology, 1975, 31,
other behavior pathologies. Journal of Ab- 696-697.
normal Psychology, 1972, 80, 225-243. Silverstein, P. R. Coping and adaptation in
Rutter, M. Protective factors in children's re- families of physically handicapped school
sponses to stress and disadvantage. In M. W. children. Unpublished doctoral dissertation,
Kent & J. E. Rolf (Eds.), Primary prevention University of Minnesota, 1982.
of psychopathology. (Vol. 3): Social compe- Spivack, G., & Swift, M. The Devereux elementary
tence in children. Hanover, N.H.: University school behavior rating scale. Devon, Pa.: Dev-
Press of New England, 1979. ereux Foundation, 1967.
Sandler, I. N. Social support, resources, stress, andTellegen, A., Kamp, J., & Watson, D. Recognizing
maladjustment of poor children. American individual differences in predictive structure.
Journal of Community Psychology, 1980, 8, Psychological Review, 1982, 89, 95-105.
41-52.

1129837

Uploaded by

Copyright:

Available Formats

1129837

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1129837

Uploaded by

Copyright:

Available Formats

The Study of Stress and Competence in Children: A Building Block for Developmental

The Study of Stress and Competence in

Norman Garmezy, Ann S. Masten, and Auke Tellegen

combined school A and school B samples (N

RESULTS OF THE HIERARCHICAL MULTIPLE REGRESSION ANALYSIS FOR THE

" These values indicate whether the increment in R2 provided

RESULTS OF THE HIERARCHICAL MULTIPLE REGRESSION ANALYSIS FOR THE

1. Sex ............. -.24 5.67 .019 .24 .06 .06 .019

" These values indicate whether the increment in R' provide

RESULTS OF THE HIERARCHICAL MULTIPLE REGRESSION ANALYSIS FOR THE

a These values indicate whether the increment in R' provided by

You might also like