Title: Parental demographic and psychosocial factors, neonatal behaviors, and infant temperament as correlates of infantile colic /
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00099591/00001
 Material Information
Title: Parental demographic and psychosocial factors, neonatal behaviors, and infant temperament as correlates of infantile colic /
Physical Description: xv, 277 leaves : ; 28 cm.
Language: English
Creator: McAuliffe, Marilyn Wiegand
Publication Date: 1984
Copyright Date: 1984
Subject: Infants   ( lcsh )
Colic   ( lcsh )
Parent and child   ( lcsh )
Anxiety   ( lcsh )
Foundations of Education thesis Ph. D
Dissertations, Academic -- Foundations of Education -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Thesis: Thesis (Ph. D.)--University of Florida, 1984.
Bibliography: Bibliography: leaves 266-276.
Statement of Responsibility: by Marilyn Wiegand McAuliffe.
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00099591
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000493704
oclc - 11988662
notis - ACR2564


This item has the following downloads:

parentaldemograp00mcau ( PDF )

Full Text






This work is dedicated to

my parents, Ligia and Joseph,

who provided me with the

opportunity and support;

my husband, John,

who provided me with the

impetus and desire;

my son, Christopher,

who provided me with the

inspiration and motivation;


my soon-to-be-born child,

who provided me with the


with my love and gratitude.


The last 5 years at the University of Florida have been rewarding

in terms of the knowledge and friendships gained. Certain individuals

stand out in my memory for the time spent and interest shown in me as

a student, colleague, and friend. Special thanks are due to my chair-

person, Patricia Ashton, and cochairperson, Stephen Olejnik, who con-

tributed countless hours toward the completion of this work. Their

high levels of professionalism and expertise, as well as their personal

qualities of good humor and wit, made this learning experience an

especially enjoyable one. I would also like to thank Gene Anderson for

the insights and enthusiasm imparted in the study of a phenomenon of

infancy of special interest to us both. Don Avila will always be

remembered for his part in helping me to maintain my perspective

throughout the course of my graduate education. I consider all of

these people to be both mentors and friends.

Childbirth instructors from Shands Hospital, Alachua General

Hospital, the Birthplace, and several private obstetrical practices

in Gainesville, Florida, were most cooperative in permitting me

access to their classes for the purpose of recruiting subjects for

this study. I would like to thank them and the affiliated obstetri-

cians who indirectly supported this research endeavor. Of course, I

am particularly indebted to the mothers and fathers who consented to

participate in this study. It goes without saying that without their

cooperation this study could never have taken place. The photographs

and long letters concerning their children, as well as the many hours

spent completing questionnaires and imparting information over the

telephone,attest to the special interest and caring these parents

have for their children.

I would also like to thank the many friends and family members

who supported me throughout the course of this study. The tedium of

coding thousands of questionnaires and stuffing envelopes would have

been greatly increased without the assistance of my parents, Lee and

Joe, and husband, John. Joni Charm6 and Safieh Javid were most helpful

in their scoring of countless questionnaires. The encouragement and

insights provided by Tish Denny, Kathy and Charlie Wilkinson, and my

uncle, Dr. Paul Taylor, will always be remembered with fondness and


Finally, my heartfelt thanks to go my sister, Virginia Wiegand,

and friends, Lisa Conover, Joni Charm6, and Kim Givens,who provided

my son, Christopher, with tremendous amounts of love and learning while

I was away from him. The special bonds developed between Christopher

and each of his special friends gave me the peace of mind necessary

for relinquishing even a few shared moments of his first two years of



ACKNOWLEDGMENTS . . . . . . . . ... . ... iv

LIST OF TABLES .... . . . . . . . . x

ABSTRACT . . . . . . . . ... . . . . . xiv


I INTRODUCTION. . . . . . . . . . 1

Background of the Problem . . . . . . . 1
Statement of the Problem. . . . . . . . 2
Methodological Problems . . . . . . 3
Theoretical Problems. . . . . . . . 5
Purpose of the Study. . . . . . . . . 8
Definition of Terms . . . . . . . . 9
Significance of the Study . . . . . . .. 10
Summary . . . . . . . . ... .. .. . 14

II REVIEW OF THE LITERATURE. . . . . . . ... 17

Introduction. . . . . . . . . ... . 17
Part I: Infantile Colic. . . . . . . ... 17
Background. . . . . . . . .... 17
Definition of Colic . . . . . . . 19
Characteristics of Colic. . . . . . ... 20
Suggested Etiologies of Colic . . . . .. 25
Management of Colic . . . . . . 26
Colic Research. . . . . . . . . . 28
Part II: Prenatal Anxiety. . . . . . . ... 41
Introduction ................. 41
Prenatal Anxiety and Maternal Factors ..... . 41
Prenatal Anxiety, Obstetrical Complications, and
Childbirth Abnormalities. . . . . . 42
Maternal Blood Pressure and Pregnancy Outcomes. .. 44
Maternal Prenatal Anxiety and Neonatal Behavior . 46
Animal Studies. . . . . . . . . ... 48
Part III: Temperament. . . . . . . .. 50
Introduction. .. . . . . . . . . 50
Historical Perspective. . . . . . . ... 50
Definition of Temperament . . . . . ... 54
Origins of Temperament. . . . . . . ... 55
Continuity and Stability. . . . . . ... 56

Predictive Value of Temperament . . . ... 58
Temperament and Infantile Colic . . . ... 62
Assessment of Temperament . . . . .... 64
Problems in the Assessment of Temperament .... . 65
Validity of Assessment Measures . . . ... 67
Summary . . . . . . . . ... ... .. 75

III METHODOLOGY . . . . . . . . .. . 79

Introduction. . . . . . . . . .. . 79
Sample. .... . . . . . ...... 80
Variables . . . . . . . . .. .. . . 83
Demographic Variables . . . . .... ... 84
Parent Variables. . . . . . . . 84
Infant Variables. . . . . . . . ... 85
Dependent Variable. . . . . . . . ... 86
Procedure . . . . . . . . .. .. . . 86
Instrumentation . . . . . . . .... . 88
Demographic Questionnaires. . . . . . ... 88
Parent Measures . . . . . . . . . 90
Infant Measures ... . . . . . . 97
Measurement of Colic. . . . . . . .. 104
Hypotheses . . . . . . . . . . 111
Demographic Variables . . . . . . ... 112
Parent Variables. . . . . . . . ... 112
Infant Variables. . . . . . . . .. 113

IV RESULTS . . . . . . . . ... . . . 115

Introduction. . . . . . . . . . 115
Characteristics of Colic. . . . . . . . 116
Introduction. . . . . . . . . . 116
Incidence . . . . . . . . . 117
Onset and Disappearance . . . . . ... 120
Feeding Method. . . . . . . . . . 120
Birth Order . . . . . . . . . 122
Periodicity . . . . . . . .... . 123
Allergy . . . . . . . . . 124
Gastrointestinal Distress . . . . . . 125
Relationship Among Variables. . . . . . ... 127
Demographic and Descriptive Characteristics . 128
Parent Characteristics. . . . . . . ... 129
Infant Characteristics. . . . . . . ... 132
Summary of Correlation Analyses . . . ... .133
Hypotheses. . . . . . . . .. . . .. 135
Demographic and Descriptive Variables ...... 136
Parent Variables. . . . . . . . ... 142
Infant Variables. . . . . . . . ... 153
Infant and Parent Variables . . . . ... 166
Summary of Results ............. . .. . 166
Characteristics of Colic. . . . . . .. 168
Hypotheses...... . . . . . . 169
Summary . . . . . . 174

V DISCUSSION. . . . . . . . . . 181

Introduction . . . . . . . . . . 181
Characteristics of Colic. . . . . . . .. 181
Incidence . . . . . . . . . 181
Onset and Disappearance . . . . ... 183
Periodicity of Crying . . . . . . 185
Variables Thought to be Related to Colic. . . ... 185
Feeding Method. . . . . . . . .. 186
Allergy and Cow's Milk Intolerance. . . .. 187
Gastrointestinal Distress ............ 188
Demographic and Descriptive Variables . . ... 189
Father's Education. . . . . ... ... . 190
Birth Order . . . . . . . . ... 192
Race . . . . . . . . . . 193
Parent Variables. ................. . 193
Prebirth Parent Variables . . . . .... 194
Postbirth Parent Variables. . . . . .. 195
Infant Variables. . . . . . . . . .. 200
First Postpartum Week . . . . . .... 200
Three- and Eight-month Temperament. . . .. 202
Summary and Conclusions . . . . . .... 205
Limitations of the Study. . . . . . . ... 211
Recommendations for Future Research . . . ... 213


A INFORMED CONSENT. . . . . . . . . .. 215

B DEMOGRAPHIC SHEET I . . . . . . .... .217

C DEMOGRAPHIC SHEET II . . . . . . ... 219



F CONSENT FORM (PILOT STUDY). . . . . . ... 231

G INTERVIEW SCHEDULE. . . . . . . . ... 232

INFANT (MABI) . . . . . . . .... . 234



K CRY SCALE . . . . . . . .... .. .. 246



REFERENCES . . . . . . . . ... ... .. . 266

BIOGRAPHICAL SKETCH. . . . . . . . . ... . 277



3.1 Mean Age, Education, and Parity for Parents. ... . 81

3.2 Frequency Distribution of Sample by Race. . . .. 81

3.3 Frequency Distribution of Sample by Marital Status. .. 82

3.4 Frequency Distribution of Sample by Mother's and
Father's Occupation . . . . . . . ... 82

3.5 Frequency Distribution of Sample by Mother's Parity . 83

3.6 Frequency Distribution of Sample by Infant Gender . 83

3.7 Variables, Instrumentation, and Timing of Data
Collection. . . . . . . . . ... ..... 89

3.8 Correlation Matrix for Mother's State and Trait
Anxiety Measured at Four Different Points in Time . 93

3.9 Correlation Matrix for Father's State and Trait
Anxiety Measured at Four Different Points in Time . 94

3.10 Stability Coefficients for the Nine Temperament Scales
of the RITQ for the Present Study Sample. . . ... 101

3.11 Convergent Validity Coefficients for RITQ Scales of
Activity, Mood, Rhythmicity, and Adaptability .... 101

4.1 Intercorrelation Matrix of Eight Measures of
Infant Colic. . . . . . . ... . .. 118

4.2 Incidence of Colic for Various Colic Measures (in %). 119

4.3 Means and Standard Deviations for Onset and
Disappearance of Increased Infant Crying. . . .. 120

4.4 Descriptive Statistics for Onset and Disappearance of
Increased Crying for Infants in the Four MYRATE
Colic Groups . . . . . . . . . . 121

4.5 Periodicity of Crying for Infants in Four MYRATE
Colic Groups (in %) . . . . . . . . 123


4.6 Periodicity of Crying Measured by PECOL (in %) ... 124

4.7 Experience of Gas Pains in 6-Week Old Infants in
Four MYRATE Colic Groups (in %) . . . . .. .126

4.8 Frequency of Cramps During Crying Periods in Infants
in Four MYRATE Colic Groups (in %). . . . . ... 127

4.9 Frequency of Gas Passed During Crying Periods of
Infants in Four MYRATE Colic Groups . . . ... 128

4.10 Correlation Coefficients for Demographic and
Descriptive Variables and Total CASS Scores . . .. 129

4.11 Correlation Coefficients for Mother's and Father's
Prenatal and First Postpartum Week State and Trait
Anxiety and Total CASS Scores . . . . ... 130

4.12 Correlation Coefficients for the Seven Subscales of
the Pregnancy Research Inventory and Total CASS Scores. 131

4.13 Correlation Coefficients for the Three Subscales of
the Maternal Support Questionnaire and Total CASS
Scores. . . . . . . . . . . . 132

4.14 Correlation Coefficients for the Four Dimensions of
the MABI and Total CASS Scores. . . . . . ... 133

4.15 Regression Model Estimating the Relationship Between
Prebirth Demographic and Descriptive Variables and
Total CASS Scores .... . . . . . . 137

4.16 Stepwise Regression Analysis Estimating the Relation-
ship Between Prebirth Demographic Variables and
Total CASS Scores . . . . . . . . 137

4.17 Analysis of Covariance Performed on Father's Education,
Race, Mother's Occupation, Father's Occupation, and
Total CASS Scores. .... . . . . . . 138

4.18 Regression Model Estimating the Relationship Between
Postbirth Demographic and Descriptive Variables and
Total CASS Scores . . . . . . . ... .141

4.19 Stepwise Regression Analysis Estimating the Relation-
ship Between Father's Education and Postbirth
Demographic and Descriptive Variables and Total CASS
Scores. .... . . . . . . . .. 141

4.20 Regression Model Estimating the Relationship
Between Mother's and Father's Prenatal State and Trait
Anxiety and Total CASS Scores . . . . . .. 143


4.21 Stepwise Regression Analysis Estimating the
Relationship Between Mother's and Father's Prenatal
State and Trait Anxiety and Total CASS Scores .... . 143

4.22 Regression Model Estimating the Relationship
Between the Subscales of the Pregnancy Research
Inventory and Total CASS Scores . . . . .... 145

4.23 Stepwise Regression Analysis Estimating the Relationship
Between the Seven Scales of the Pregnancy Research
Inventory and Total CASS Scores . . . . ... 146

4.24 Regression Model Estimating the Relationship Between
Mother's and Father's First Postpartum Week State and
Trait Anxiety and Total CASS Scores . . . ... 148

4.25 Stepwise Regression Analysis Estimating the Relation-
ship Between Mother's and Father's First Postpartum
Week State and Trait Anxiety and Total CASS Scores. . 148

4.26 Regression Model Estimating the Relationship
Between Mother's First Postpartum Week Fatigue and
Total CASS Scores. .. . . . . . . . 150

4.27 Regression Model Estimating the Relationship
Between Mother's First Postpartum Week Fatigue and
Total CASS Scores Including Interaction Terms .... . 150

4.28 Stepwise Regression Analysis Estimating the
Relationship Between Mother's First Postpartum Week
Fatigue and Total CASS Scores . . . . .. . 151

4.29 Regression Model Estimating the Relationship
Between Mother's Support and Total CASS Scores. ... 153

4.30 Stepwise Regression Analysis Estimating the
Relationship Between Mother's Support and Total
CASS Scores . . . . . . . .. . .154

4.31 Regression Model Estimating the Relationship
Between the Four Dimensions of the MABI and Total
CASS Scores . . . . . . . .. . .155

4.32 Stepwise Regression Analysis Estimating the
Relationship Between Four Dimensions of the
MABI and Total CASS Scores . . . . . .... 156

4.33 Regression Model Estimating the Relationship
Between Neonatal Irritability and Total CASS Scores. . 159



4.34 Stepwise Regression Analysis Estimating the
Relationship Between Neonatal Irritability and
Total CASS Scores . .. . . . . . 160

4.35 Correlation Coefficients for RITQ Subscales at 3
and 8 Months of Age and Total CASS Scores ...... 162

4.36 Correlation Coefficients for 3- and 8-Month ICQ
Temperament Factors and Total CASS Scores ...... 164

4.37 Correlation Coefficients for 3- and 8-Month IBQ
Subscales and Total CASS Scores . . . . .... 165

4.38 Regression Model Estimating the Relationship
Between the Interactions Between Mother's First
Postpartum Week State Anxiety and Fatigue and
3-Month Infant Threshold to Stimulation and Total
CASS Scores . . . . . . . .... .. .. 167

4.39 Summary Table for Testing of Hypotheses . . ... 175

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



Marilyn Wiegand McAuliffe

August 1984

Chairperson: Patricia T. Ashton
Cochairperson: Stephen F. Olejnik
Major Department: Foundations of Education

Infantile colic is a behavioral phenomenon of unknown origin, oc-

curr'ng during Lhe first 3 months of life, characterized by fussing,

crying, or screaming for which there is no uniformly successful method

of therapy. This study investigated the relationship between colic and

several demographic variables, parents' prenatal and postpartum state

and trait anxiety, mother's fatigue and support, neonatal irritability,

and later temperament. A sample of 163 couples, recruited from prepared

childbirth classes, completed a battery of questionnaires during the 3rd

trimester of pregnancy and the 1st, 6th, 12th, and 32nd postpartum weeks.

Colic was not found to be significantly related to birth order,

feeding method, race, parents' ages, infant's gender, mother's drug use

in labor and delivery, father's work hours, birthweight, gestational

age, infant allergy, or family history of allergy. Although most in-

fants demonstrated increased crying, mostly during the evening hours,

from 2 to 12 weeks of age, the variability of these age parameters was

less for more colicky infants, whose crying invaded any hour of the day.

A differentiation, therefore, should be made between colic and the de-

velopmental phase, characterized by increased crying, that most infants

are thought to experience.

Correlations between temperamental characteristics and colic

indicated that more colicky infants were less rhythmic, less dis-

tractible, more variable in mood, had shorter durations of orienting,

and greater frustration to limitations at 3 months of age than less

colicky infants. The relationship between colic and mood and frustra-

tion to limitations remained significant at 8 months of age.

The best regression model for predicting colic demonstrated a

weak but significant positive relationship between colic and father's

education, mother's first postpartum week fatigue, and neonatal ir-

ritability. Although several factors related to colic were identified,

the small amount of variance explained in colic scores suggests that

the most crucial variables have yet to be recognized. Or, perhaps,

more valid measures of these suggested associated factors, or of

colic itself, could result in a better understanding of this pediatric

problem. The variables identified by this research should, therefore,

not be dismissed as insignificant, but considered in terms of why

they bear any relationship to colic at all.


Background of the Problem

Parental complaints of unsoothable crying in their young infants

are not uncommon to the pediatrician. After ruling out medical prob-

lems as the cause of crying, the physician frequently affixes the

label "colicky" to the wailing infant and reassures the parents that

the condition is self-limiting and will be outgrown by 3 or 4

months of age. Such reassurance may be of some value in calming the

fears of the parents, but it does little to reduce the tension aroused

by persistent and unsoothable infant crying.

There is little consensus regarding the definition, treatment, and

etiology of infantile colic. Colic is more of a symptom than it is a

disease in that one can easily describe the classic profile of the

colicky baby, but the cause of the symptoms has not been determined.

This lack of knowledge has resulted in a conglomeration of descriptions

of characteristics of the colicky baby rather than a single, agreed-

upon, definition. Taylor's (1957) description of a colicky baby is

similar to most others found in the literature.

Infantile colic is a symptom complex characterized by
recurrent attacks of crying or screaming, occurring
in infants under the age of three months, in whom there
is no obvious cause for these attacks, such as improper
feeding or gross physical defects. The crying is
characteristically worse in the evening, but in some
cases may occur at any time of the day or night.

The crying is usually rhythmic in nature, recurring
every 5 to 10 minutes, but it may be continuous over
a period of a few hours. During severe attacks the
infant draws his knees up, tenses his whole body, sweats
profusely, and appears to be in pain. There are grada-
tions in the severity of the condition from the "fussy"
baby to the infant who drives his parents to distraction.
The symptoms usually disappear by the age of three
months, and for this reason the condition has sometimes
been referred to as "three-month colic." (p. 458)

In attempts to adhere to stringent criteria of colic, many re-

searchers (Carey, 1968; Cobb, 1956; Licamele, Palumbo, Quinn, &

Zuckerman, n.d.) have defined the colicky baby as "one who, otherwise

healthy and well fed, [had] paroxysms of irritability, fussing or

crying lasting for a total of more than three hours a day and occurring

on more than three days in any one week" (Wessel, Cobb, Jackson,

Harris, & Detwiler, 1954, pp. 425-426). Other researchers (Paradise,

1966; Stewart, Weiland, Leider, Mangham, Holmes, & Ripley, 1954) have

preferred to view colic on a continuum from none to severe with infants

displaying varying degrees of soothability. Regardless of the exact

definition employed, the one aspect of colic that is uniformly agreed

upon is that it involves infant crying for which no immediate explana-

tion or means of relief can be found.

Statement of the Problem

Numerous theories exist that attempt to identify the etiology of

infantile colic. Allergy, gastro-intestinal disturbances, infant

temperament, and maternal anxiety are only a few of the suggested

causes. Although time has seen many theories fall in and out of favor,

one theory that continues to appear in the literature implicates

maternal anxiety and family tension as causes or contributing factors

in the etiology of colic. Several authors (Brazelton, 1962; Carey,

1968; Cobb, 1956; Ulessel et al., 1954) have suggested that the salience

of environmental tension may be influenced by constitutional or tem-

peramental characteristics of the infant.

One explanation for the confusion surrounding the causes of colic

has been the inadequacy of the measures that have been used to identify

potential etiological factors. The problems to be addressed by the

present study are, therefore, both methodological and theoretical.

Methodological Problems

Much of the research investigating the role of maternal anxiety

in infantile colic has been plagued with methodological problems. In

addition to the procedural shortcomings of retrospective evaluation of

anxiety and colic, there has been a lack of valid, reliable, and objec-

tive instrumentation for measuring both variables. Many authors

(Bonar, 1935; Breslow, 1956; Brody, 1979; Meyer, 1958; Stewart et al.,

1954; Taylor, 1957) have concluded that postnatal maternal emotions

are important, but the assessment of these emotions has frequently

been based on subjective clinical impression or crudely constructed

instruments lacking evidence of reliability or validity.

A major procedural limitation of several studies (Breslow, 1956;

Stewart et al., 1954; Taylor, 1957; Wessel et al., 1954) relating post-

natal maternal anxiety to infantile colic has been the timing of data

collection. These researchers obtained measures of anxiety after

the onset of colic. It is not surprising to find elevated anxiety

scores for mothers living with colicky babies. It would be more in-

formative to correlate colic with anxiety scores obtained prior to the

onset of colic. Paradise (1966) suggested that future research should

investigate maternal emotional factors early in pregnancy "or, prefer-

ably, before conception" (p. 130). A few researchers (Lakin, 1957;

Paradise, 1966) have collected prenatal anxiety data immediately after

birth; however, these studies have been retrospective in that mothers

were asked to report on the anxiety they experienced during pregnancy.

It is questionable whether one can adequately recall the level of

anxiety experienced at an earlier point in time, especially if the

anxiety was not severe.

Researchers focusing on prenatal maternal anxiety have usually of-

fered a more organized assessment of anxiety. However, many of these

studies have failed to deal with the suggestion that anxiety can be

conceptualized as both a trait and a state (Spielberger, Gorsuch, &

Lushene, 1970) and that the less stable state anxiety might be the more

appropriate variable in a study focusing on anxiety in pregnancy (Beck,

Siegel, Davidson, Kormeier, Breitenstein, & Hall, 1980). State anxiety

would also be a more sensitive indicator of maternal emotional fluctua-

tions through the early postpartum months when colic has been found to


Many studies of maternal psychological factors during pregnancy

and the early postpartum period have been univariate in nature focusing

on the single variable of anxiety. Some researchers (Carey, 1968;

Pleshette, Asch, & Chase, 1956) have suggested that the support re-

ceived by the pregnant woman and new mother is among the best predic-

tors of maternal anxiety during pregnancy and the postpartum period.

Because anxiety is a global construct, it provides little information

as to its source. It should be profitable to examine anxiety in con-

junction with more specific and informative measures of maternal

psychosocial experiences.

to pregnancy, labor, and delivery complications in humans (see

McDonald, 1968). Several studies (Chisholm, Woodson, & Da Costa

Woodson, 1978; Farber, Vaughn, & Egeland, 1981; Korner, Gabby, &

Kraemer, 1980; Ottinger & Simmons, 1964; Sontag, 1944; Woodson,

Blurton Jones, Da Costa Woodson, Pollock, & Evans, 1979) have carried

the investigation beyond the pregnancy and birth experiences to the

physical status of the neonate. A few studies have examined the ef-

fects of anxiety on infant behavior.

Although Lakin and Paradise were the only researchers to associate

prenatal anxiety with the specific infant behaviors involved in in-

fantile colic, other researchers (Ottinger & Simmons, 1964) have found

a relationship between prenatal maternal anxiety and the crying behavior

of the neonate. More recently, Farber et al. (1981) found prenatal

maternal anxiety to be related to neonatal behaviors, inchUding infant

irritability, and mother-infant interaction. Other researchers

(Chisholm et al., 1978; Korner et al., 1980; Woodson et al., 1979)

have found a relationship between physiological measures of maternal

stress and neonatal irritability. Chisholm et al. (1978) reported a

significant correlation of .71 between blood pressure during the second

trimester of pregnancy and neonatal irritability.

The fact that studies have found relationships between prenatal

maternal anxiety or blood pressure and neonatal irritability lends

credibility to Sontag, Steele, and Lewis's (1969) statement:

Although emotional stages in postnatal life are rec-
ognized as affecting physiological-endocrinological
states, as for example the effects of anxiety on the
thyroid function and blood sugar levels, there has been
in the past little general recognition of the fact that
such psychosomatic relationships and others might,

crying is attended to, the baby stops crying" within minutes (Anderson,

1980, adapted from Paradise, 1966).

Temperament was defined as behavioral style or the "how" of behavior

rather than the content or motivations of behavior. Temperament was

operationally defined by Carey and McDevitt's (1978) Infant Temperament

Questionnaire (ITQ), the Infant Characteristics Questionnaire (ICQ) of

Bates et al. (1979), and Rothbart's (1981) Infant Behavior Questionnaire

(IBQ). Temperament for the ITQ was operationalized by Thomas and

Chess's (1977) nine temperamental dimensions of activity, rhythmicity,

approach or withdrawal, adaptability, threshold of responsiveness,

energy level of response, mood, distractibility, and persistence. Also

based on Thomas and Chess's categories, the ICQ of Bates et al. (1979)

defines temperament in terms of the four factors of fussy-difficult,

unadaptable, dull, and unpredictable. Rothbart's IBQ measures Thomas

and Chess's nine categories as well as the child's reactivity and

self-regulation through the scales of activity level, smiling and

laughter, fear, distress to limitations, soothability, and duration of


Trimester was defined as one-third of the pregnancy period.

Significance of the Study

The significance of the present study can be viewed from a

methodological, theoretical, and practical perspective. Several

authors have related maternal anxiety to infantile colic; however,

their results have been of questionable value because of the meth-

odological limitations inherent in their studies. The present study

attempted to circumvent these difficulties by using valid and reliable


The signals and responses of infants and mothers are finely

synchronized. The infant cry has been found to cause physiological

changes in the mother. For example, the nursing mother's breasts

fill with milk upon hearing an infant's cry (cited in Eiger & Olds,

1972). Through this primitive vocalization infants are usually able

to ensure the proximity to caregivers that is so necessary for sur-

vival. "Normal" crying, therefore, serves a useful function.

Aldrich, Sung, and Knop conducted the first systematic study of

crying in 1945. Their results, indicating that newborn babies cried

more in the hospital than at home, led them to conclude that the indi-

vidualized care from the mother was a definite comfort to the babies.

Signals, such as crying, can obviously be responded to more rapidly and

consistently when there is a one-to-one relationship between child and

caregiver. Such a relationship rarely exists in a hospital nursery. The

infant's cry is a signal of need and is a normal function of babies.

However, when the cry is prolonged unduly or becomes
the usual pattern of behavior, we feel the need to
intervene and to ameliorate the situation. It should
not be necessary for a baby to cry more than three
minutes to indicate an ordinary need. (Aldrich et al.,
1945, p. 428)

In his study of 80 mothers and their infants, Brazelton (1962)

found that babies cried between 1-1/2 and 2-3/4 hours daily between the

first and sixth weeks of life. He hypothesized that although this

crying may be normal for infants in our culture, parents view any

crying as representing failure. Their attempt to compensate for this

supposed failure

may create unnecessary stimuli for the newborn, and
the resulting tension is added to his innate reasons
for crying. An ever-increasing cycle in the duration
of intensity of the crying may result in the picture
of 24-hour "colic." (p. 579)

Definition of Colic

The literature presents definitions of colic that are numerous

and varied. Several authors do not even define it. Definitions are

usually similar, yet some revision, deletion, or addition has invari-

ably been made.

In 1954, Illingworth reserved the term "three-month's colic"


a clinical entity in which the baby, in the first three
months of life, has rhythmical screaming attacks in the
evenings, which are not stopped when he is picked up,
and for which there is no obvious explanation, such as
hunger. (p. 174)

In the same year, Wessel et al. defined the colicky baby as one who

otherwise healthy and well-fed, had paroxysms of
irritability, fussing or crying lasting for a total
of more than three hours a day and occurring on more
than three days in any one week. (pp. 425-426)

This definition has been employed by many researchers (Carey, 1968;

Cobb, 1956; Licamele et al., n.d.) who have desired a qualitative and

quantitative definition of colic. Other researchers (Paradise, 1966;

Stewart et al., 1954) have preferred to view colic on a continuum from

none to severe with infants displaying varying degrees of soothability.

Some authors' views of the etiology of colic have influenced

their definitions. Glaser (1956) retained much of Wessel et al.'s

and Illingworth's definitions but added that the crying was due to

abdominal pain. He also omitted Wessel et al.'s quantitative criteria

for crying, as well as their emphasis on the evening as the time of day

when colic occurs.

Licamele et al. (n.d.) recommended that a differentiation be

made between a symptom of colic and the colic syndrome. An infant

who, infrequently, cries unsoothably or for unexplained reasons may be

demonstrating symptoms of colic, but the term "colic syndrome" is

reserved for infants whose behavior patterns constantly demonstrate

colic symptoms. Extending Wessel et al.'s definition, Licamele et al.

(n.d.) wrote

This colic syndrome consists of paroxysms of irritability,
fussing and crying lasting more than three hours a day
and occurring on more than three days a week, frequently
occurring in the evening, beginning two weeks past the
EDC [expected date of delivery], lasting 3 months, in
a well-fed, burped and changed infant without vomiting,
diarrhea, constipation, rashes, rhinitis, wheezing or
fever. (p. 11)

This comprehensive definition includes the main points of the most

frequently cited definitions. In addition, it precludes the confusion

between crying for unknown reasons and crying due to allergy and other

reasons of organic origin. Regardless of the exact definition em-

ployed, the one aspect of colic that is uniformly agreed upon is that

it involves infant crying for which there is no consistently workable


Characteristics of Colic

Incidence. The reported incidence of colic is variable depending

upon the particular sample of infants studied and the definition em-

ployed. Cobb (1956) found the incidence in his study to be 50%,

whereas Levin (1950) found only .95% of infants in a New York Foundling

Hospital to be colicky. Paradise (1956) found the incidence of colic

to be 23% in his sample,which included all normal infants delivered at

a large urban hospital over a 5-month period in 1951. In private

practice, colic has been reported to occur in anywhere from 15%

(Breslow, 1957) to 40% (Taylor, 1957) of the cases. Rambar's (1956)

suggestion that colic occurs more frequently in private groups than in

clinic patients seems to be supported by findings that colic occurs

more frequently, or is more often reported, among the children of

highly educated and intelligent parents (Hide & Guyer, 1982; Paradise,

1966). Wessel et al. (1954), however, found no differences in educa-

tional backgrounds of mothers of fussy and contented babies. Meyer

and Thaler (1971) found 11.6% of their sample of low birthweight in-

fants became colicky between the 39th and 44th weeks of gestation.

Licamele et al. (n.d.) have estimated the average incidence in the

literature to be between 20% and 30%.

Onset and disappearance. Infantile colic has been most frequently

described as beginning during the second or third week after birth and

ceasing spontaneously around the third or fourth month of life, herce,

the name "three-month colic." Although most authors have agreed with

these dates for the onset and disappearance of colic, the onset in

Illingworth's (1954) sample of colicky babies was between birth and 5

days of age in 15% of the cases. Brazelton (1962) found that colic

began around the second week after birth. The intensity and duration

of crying increased in severity until the sixth week when it reached

a plateau before disappearing between the eighth and tenth weeks.

Meyer and Thaler (1971) stated that colic usually begins between the

19th and 44th gestation week, whereas Pierce (1948), in his study of

colic in premature babies, found that colic began 2 weeks after the

expected date of birth regardless of the gestational age at birth.

There is much disagreement on the mechanism for the sudden onset

and disappearance of colic at specific points in time. Those authors

who have viewed colic as an allergic manifestation have reasoned that

it takes a few weeks for sensitization to take place, while the disap-

pearance at 3 months is probably due to desensitization to foods or to

loss of allergic reactivity of the gastrointestinal tract (Rambar, 1956).

Those who have attributed colic to immaturity in the infant, whether it

be immaturity of the digestive system or immaturity of the nervous sys-

tem, have suggested that colic disappears by the third or fourth month

by virtue of maturation (Illingworth, 1954; Paradise, 1966). Other

authors' explanations have also been based on the premise that colic

emerges from some form of immaturity in the infant. Benjamin (1961)

stated that infants with low sensory thresholds are prone to colic "dur-

ing the third or fourth weeks when babies go through a maturational

spurt in sensory capacities but have not yet developed an adequate

stimulus barrier" (p. 616). Craven (1979) suggested that coltc stops at

3 months because at this age babies have developed forms of communica-

tion other than just crying. Brazelton (1962) found that as crying

decreased other activities such as sucking, cooing, and turning over

increased. Meyer (1958) felt that the advance in motor development

reduced the frustration by providing infants with other outlets for

their motor drives. Wolff (1969) found that the mother's face began

to exert a calming influence over the infant by the fifth or sixth

month which could explain the cessation of excessive crying. Explana-

tions for the onset and disappearance of colic at a specific time in

an infant's life have been absent from many of the other theories and

studies pertaining to colic.

Feeding method. Method of feeding has been investigated in rela-

tion to the colic syndrome. Meyer (1958) reported that colic was

seldom found among breast-fed infants, and Paradise (1966) found that

there was a lower, but nonsignificant, incidence of colic among infants

maintained on breast feedings. In contrast, Illingworth (1954) found

slightly more infants in the colic group than in the noncolic group to

be breast-fed but considered this difference to be trivial. On the

basis of these empirical findings, it would seem that method of feeding

is not a factor in the emergence of colic.

Birth order. It has often been assumed that colic occurs more

frequently among first-born children due to the inexperience of the

new parents. Some authors (Bakwin, 1956; Bruce, 1961; Rambar, 1956;

Spock, 1944) have written that colic is more common among first chil-

dren; however, they produced no evidence for this conclusion. The

results of more systematic investigations have demonstrated no signifi-

cant differences in the occurrence of colic due to birth order

(Illingworth, 1954; Meyer & Thaler, 1971; Paradise, 1966; Taylor, 1957).

Periodicity. The periodicity of colic has baffled many research-

ers. Although colic may occur at any time of the day or night, it tends

to occur during the late afternoon and evening hours. Explanations

for this phenomenon have ranged from suggestions that the mother's

breast milk is low at the end of the day to the theory that family

tensions are highest at the end of the day resulting in increased

crying. Neither suggestion, however, explains all of the symptoms and

characteristics of the colic syndrome. Those researchers suggesting

etiologies of colic other than hunger or family tension have been

unable to explain the predominant occurrence of colic in the evenings.

Other characteristics. On the basis of clinical observations of

patients, Holmes (1969) viewed infantile colic as being characterized by

excess motor activity, frequent hard crying, much flatus, an apparent desire

to eat more often and to take more at a feeding, above-average weight

gain and growth in length, and continued overactivity and precocity in

motor and mental development. Other pediatricians have concurred with

these observations, but they have offered no empirical support for

their beliefs. Colicky babies have been found to be above average in

weight gain (Holmes, 1969; Illingworth, 1954; Taylor, 1957), although

Paradise (1966) did not find this difference to be significant.

Brazelton (1962) and Levine (1956) found that colicky infants demon-

strated excessive motor activity, and other authors (Meyer, 1958;

Holmes, 1969) have claimed that the colicky infant demonstrates excep-

tional activity in utero and precocity in motor and mental development

after colic ends.

Future development. Although few valid data have been col-

lected, via longitudinal studies, on the emotional and behavioral

development of colicky babies, some authors have offered prognoses.

McGee (1950) described the child who was a colicky infant as being a

problem in school, "as he has difficulty keeping quiet and paying

attention. He often molests other children" (p. 337). No evidence has

been offered in support of this prognosis.

Other authors, offering brighter futures for colicky infants,

have also been unable to provide evidence to support their claims.

Meyer (1958) stated that "the child's future is secure and that the

very symptoms which presently are disturbing will be virtues when

[the child] has grown more mature" (p. 630). Holmes (1969) wrote that

the once-colicky baby is active, restless, and talkative throughout

childhood. He even ventured to make predictions for his later life:

As an adult in sports, business and civic affairs, he
is likely to be an energetic, enthusiastic participant
rather than a spectator and, in turn, he will have
colicky babies of his own when he becomes a father.
(p. 568)

Research is needed to shed light on the future of the colicky baby.

Suggested Etiologies of Colic

The absence of a single agreed-upon definition for colic is easily

understood when one examines the numerous and varied factors considered

to be of etiological importance. Most authors have attributed colic

to multiple causes with one or two being of major importance.

Factors considered to be etiologically important have included

feeding problems (Craven, 1979, McGee, 1950), allergy (Breslow, 1956;

Fries, 1956; Martin, 1956; McGee, 1950; Rambar, 1956; Speer, 1958;

White, 1929), cow's milk intolerance (Harris, Petts, & Penny, 1977;

Jakobsson & Lindberg, 1978, 1983; Lothe, Lindberg, & Jakobsson, 1982),

immaturity of the intestinal tract causing excessive flatus (Stewart

et al., 1954), spasms or kinks of the colon producing localized ob-

struction (Illingworth, 1954), immaturity of the central nervous

system (Lipton, Steinschneider, & Richmond, 1960; Paradise, 1966;

Spock, 1944), discharge of pent-up energy (Brazelton, 1962), constitu-

tional hypertonicity (Meyer, 1958), unsatisfied need for oral gratifi-

cation (Anderson, 1983; Levine & Bell, 1950), progesterone deficiency

(Clark, Gains, & Bradford, 1963), soft tissue swelling around growing

bones (Ditkowsky, 1970), infant fatigue (Stewart et al., 1954), frus-

tration due to immaturity of motor skills (Meyer, 1958), low threshold

for sensory stimuli (Benjamin, 1961; Carey, 1968; Korner, 1971;

Meyer, 1958), inheritance of an overactive nervous system from one or

both parents (Holmes, 1969), and emotional factors, such as family

tension and maternal anxiety transmitted either prenatally or post-

natally (Bakwin, 1956; Bonar, 1935; Brazelton, 1962; Breslow, 1956;

Brody, 1979; Carey, 1968; Cobb, 1956; Harley, 1969; Levine, 1956; Meyer,

1958; Rambar, 1956; Taylor, 1957; Weiland, Leider, & Mangham, 1957;

Wessel et al., 1954). In summary, the most popular suggested etiologies

of colic have related colic to allergy or cow's milk intolerance, im-

maturity of the gastrointestinal tract, central nervous system im-

maturity, or environmental tensions such as maternal anxiety.

In his extensive review of the literature, Illingworth (1954)

cited a list of less popular causes for colic. Included in this list


congenital malformations of the alimentary tract,
inguinal hernia, urethral colic, appendicitis, foreign
bodies in thIe alimentary tract, lead poisoning, anal
fissure, imperforate anus, peptic ulcer, disease of
the gall bladder, respiratory tract or osseous system,
congenital syphilis, volvulus, intussusception, renal
colic, nasopharyngitis, otitis, pyelitis, tension
developed IN UTERO from a hypothetical uterine handicap
or transmitted from a high strung mother's nervous
system, hyperacidity, exposure to cold, chilling of
the extremeties, abdominal binders, toxins from the
mother, tension due to lack of oral satisfaction,
acidosis, introversion and accumulation of uric acid
in the kidneys. (p. 167)

Such a list suggests that one must be extremely careful to differenti-

ate between excessive unsoothable crying due to medical problems and

truly unexplained colic.

Management of Colic

Specific treatments suggested for the management of colic re-'lect

the author's views concerning the etiology of the condition. The

literature presents a variety of treatments, most of which can be

grouped into three major categories: dietary therapy, pharmacological

therapy, and psychological therapy. Dietary treatments have included

elimination of cow's milk products from the diets of nursing mothers

(Jakobsson & Lindberg, 1983; Lothe et al., 1982), changing formulas or

feeding techniques (Jakobsson & Lindberg, 1978; McGee, 1950; Stewart

et al., 1954; White, 1936), and introducing thickened cereal feedings

(White, 1929). Pharmacological therapy has included enemas and sup-

positories (Holmes, 1969), antispasmodics and sedatives for the in-

fant (Carey, 1968; Harley, 1969; Holmes, 1969; Illingworth, 1966;

Taylor, 1957; White, 1929, 1979), oral progesterone homones (Clark et

al., 1963), and whiskey (McGee, 1950). Although there is little evi-

dence that pharmacological agents are effective in treating colic

(O'Donovan & Bradstock, 1979), this fo.'m of therapy is still widely

used among medical practitioners. Researchers and authors who have

thought that colic is related to environmental tension and psychologi-

cal variables in the external environment have encouraged the applica-

tion of treatments including sedatives for the parents (Bakwin, 1956;

Harley, 1969), education and reassurance for the parents, and treatment

of anxieties (Brazelton, 1962; Carey, 1968; Harley; 1969, Holmes,

1969; Rambar, 1956; Taylor, 1957), and providing a calm, quiet environ-

ment (Neff, 1940). Additional therapies have included pacifiers

(Anderson, 1983; Levine & Bell, 1950; Meyer, 1958; Spitz, 1951),

placing the baby in a prone or semi-inclined position (Holmes, 1969;

Snow, 1937), and application of heat or pressure to the abdomen

(Holmes, 1969; Meyer, 1958; Wessel et al., 1954). Paradise (1966),

believing central nervous system immaturity to be a central cause of

colic, suggested that a colicky baby can be soothed by sounds and

vibrations which act to interrupt certain afferent proprioceptive

stimuli. Even folk wisdom has its set of soothing techniques for the

colicky infant, including taking the infant for car rides on bumpy

roads and administering colic water, a mixture of water, sugar, and

alcohol. The reader is referred to Asnes and Mones's (1982) and

illingworth's (1954) extensive reviews of the literature for more

thorough reviews of the etiology and treatment of colic.

Although many techniques have been offered for the management of

colic, none has been found to give complete or consistent relief.

Meyer's (1958) comment, that the lack of facts concerning the etiology

of colic is reflected in the absence of a valued therapeutic approach,

could still be made today.

Colic Research

Perusal of the literature indicates that much of what has been

written about infantile colic has been inspired by clinical experience

rather than systematic research. Subjective impression and opinion

concerning aspects of colic abound. Much of the research that has been

undertaken has focused on the relationship between colic and cow's

milk allergy, immaturity of the gastrointestinal tract, maternal

anxiety, and infant temperament. A review of this research is presented


Cow's milk allergy. There is no doubt that allergy is responsible

for colic symptoms in some babies. The number of cases to which this

etiology can be applied, however, is debtable.

Intolerance of some infants to milk-based formulas has been rec-

ognized for years (Breslow, 1956; Harris et al., 1977; Lothe et al.,

1982; White, 1929). In a double-blind study of the role of cow's milk

in infantile colic in formula-fed infants, Lothe et al. (1982) found

colic to be cow's-milk-dependent in approximately two-thirds of the

infants fed cow's milk formula. This research, however, has been

criticized by LeBlanc (1983) who claimed that the researchers failed

to subject their data to appropriate tests for statistical signifi-

cance and that the number of infants helped by switching formulas was

actually much lower than that suggested. Lothe et al. responded to

this criticism by agreeing that future studies should not use soy

formula as a placebo because soy, like milk, has been shown to be an

allergen for many infants.

Breast-fed infants may not oe escaping the allergic problem of

bottle-fed infants, as cow's milk proteins have been found in human

breast milk (see Asnes & Mones, 1982). Some researchers (Jakobsson &

Lindberg, 1978, 1983) have suggested that cow's milk products ingested

by the lactating mothers of colicky infants could be responsible for

colic symptoms in their infants. In a double-blind crossover study

involving 66 mothers of breast-fed infants with colic, Jakobsson and

Lindberg (1983) found infantile colic to be related to cow's milk con-

sumption by the mother in one-third of infants. They suggested a diet

free of cow's milk products as a first trial in the treatment of

colic in breast-fed infants.

In another double-blind crossover study with 20 breast-fed infants

with colic, Evans, Allardyce, and Fergusson (1981) were unable to

uncover a relationship between colic and maternal ingestion of cow's

milk or presence of cow's milk antigens in breast milk. The relation-

ship between colic in breast-fed infants and the ingestion of milk

products by lactating mothers demands further investigation.

Gas and intestinal dysfunction. Because colicky babies frequently

appear to have abdominal pain, intestinal gas and immaturity of the

intestinal tract have been popular explanations for the colic syndrome.

Many authors (Harley, 1969; Meyer, 1958; Illingworth, 1954), however,

have rejected excess flatus as an etiological factor, as x-ray

studies have usually failed to indicate excess gas in the intestines

of colicky babies.

In Illingworth's (1954) x-ray study, neither of the two colicky

babies showed a large amount of gas in the intestine during a colic

attack, but all three controls did. The small sample size, however,

makes these results tentative. Stewart et al. (1954) identified ex-

cessive gas in the gastrointestinal tracts of excessively crying

babies; however, these results are not as informative as Illingworth's

in that the authors did not indicate whether the infants were sympto-

matic at the time of the study.

Harley (1969) stated that "routine x-rays of infants frequently

reveal marked gastric distention without indications of discomfort"

(p. 139). Through use of x-rays, Harley found a normal amount of gas

in the abdomens of crying babies. A second film, taken moments after

the crying episode had ended, showed an excess of gas in the abdomen,

yet the infant fell into a comfortable sleep. As a result of such

comparisons, Harley concluded that "the distention is obviously the

result of rather than the cause of crying" (p. 139).

Lack of a clear relationship between gas pain and colic does not

preclude an association between colic and some type of immaturity of

the gastrointestinal tract. Although Illingworth (1954) was unable to

demonstrate excess gas in the intestines of colicky babies, he still

felt that colic could result from spasms or kinks of the colon produc-

ing localized obstruction. Brennemann (1940) also suggested that the

seeming intestinal discomfort of colicky babies resulted, not from

excess flatus, but from ineffective intestinal peristalsis.

In a more recent study (Clark et al., 1963), significant amounts

of pregnanediol monoglucurnide, a metabolite of progesterone, were

found in the urine of noncolicky infants, but only trace amounts were

found in the urine of eight colicky infants. When the colicky infants

were given oral progesterone, they experienced a remission of colic

symptoms. Clark et al. concluded that -olic could be a manifestation

of a transient progesterone deficiency state. They postulated that

progesterone withdrawal causes uterine contractions in the pregnant

woman and, therefore, colicky infants could be experiencing intestinal

spasms as a result of progesterone deficiency. While the results of

this study seem important, no other research has been undertaken to

pursue investigation of this hypothesis.

Postpartum maternal anxiety. Although most authors and research-

ers have agreed that colic most likely stems from multiple causes, the

emphasis on emotional factors in one or both parents has been made in

almost every organized study of its occurrence (Brazelton, 1962;

Breslow, 1956; Carey, 1968; Stewart et al., 1954; Taylor, 1957;

Wessel et al., 1954).

they were tense, anxious women who worried excessively about their

babies" (p. 459). Again, however, there was no organized assessment

of emotions prior to the onset of colic.

Infant temperament and postpartum maternal anxiety. Brazelton

(1962) studied 80 mothers and their infants and found that "normal

crying" during the second week of life lasted a median of 1-3/4 hours

daily and gradually increased to a median of 2-3/4 hours at 6 weeks,

reaching a plateau for 2 weeks before diminishing. He claimed that

normal crying could be compounded by environmental tension and presented

the following report as typical of those given by parents of heavy


The young mother had believed the hospital nurses' re-
ports that her baby was quiet in the nursery, and she
was surprised by the amount of crying he did when they
arrived home. She immediately felt guilty and respon-
sible for this "change" in his behavior. She was breast-
feeding him and blamed herself for lack of milk, and
then for "bad milk." Her husband was also upset by the
infant's crying, alternated between worrying about the
possibility that something was wrong with the baby and
being irritated with him for upsetting their homelife.
The grandmother, wno had come to help, felt helpless
and concerned about her daughter's exhausted condition.
Her ideas and criticisms thus added to everyone's
mounting tension. The young mother did "everything
she knew to do" when the infant cried. She rapidly ran
the gamut of picking him up, holding and rocking him,
walking around with him, patting him furiously, feeding
him every hour, etc. Her tension and hostility to this
situation mounted with her fatigue and realization of
her helplessness to change it. The baby's crying, which
had begun at rather reliably cyclic intervals, began
to invade all periods, especially at night. The mother
began to lose her milk with this strain, began to be
less effective in handling the baby, was depressed and
frightened by her negative feelings about the baby, and
wished for someone who would take over. The father,
who had other, professional demands on him, was frightened
by the situation and tried to stay away from home during
the fussy periods. The grandmother felt unsure of herself
and was not able to take over effectively in these cry-
ing periods when the mother needed the help most. It

became obvious to them all that the tensions transmitted
themselves to the infant and increased his own reasons
for crying. (pp. 586-587)

Six infants in this study were classified as heavy fussers and could

be somewhat characterized by this report. Unfortunately, Brazelton did

not comment on the emotional climate in the families of the other 74

infants. In addition to this limitation, the initial selection of

subjects could be a critical factor in the author's results. The

sample was not random in that mothers were chosen "because they

presented a 'normal' positive approach to mothering their new babies.

. . They seemed to be relaxed, competent mothers" (p. 580). In addi-

tion, mothers were required to maintain daily records of their infants'

crying for 12 weeks. Brazelton discussed the possibility of this

being a selective factor in that "the willingness to cooperate in

this type of study denoted a certain intellectual curiosity and/or

intensity of the physician-mother relationship" (p. 580).

It would have been interesting if some measure of intelligence or

educational level of the mothers had been reported, for Paradise (1966)

has concluded that colic is related to maternal education and superior

intelligence. If this is a valid conclusion, it is possible that

Brazelton's conclusion that "normal" crying ranges from 1-1/2 to 2-3/4

hours daily during the first 6 weeks of life could be faulty. Perhaps,

the sample was biased in favor of heavier fussers to begin with and,

therefore, what Brazelton interpreted as normal crying was really

colicky crying.

Whether or not "normal crying" was indeed being investigated, the

question still remains as to whether "heavy fussers" cry more than

"light fussers" prior to the compounding influence of environmental

tension, and, if so, why. Brazelton hypothesized that innate differ-

ences in the ability to assimilate tension could be responsible for

this individual difference. It was found that heavy fussers demon-

strated more total oral activity than light fussers, suggesting a

vigorous congenital activity type that could be highly sensitive to

accumulated tension.

Wessel et al. (1954) found that maternal emotional factors con-

tributed to colic but added that the degree of colic was probably

influenced by constitutional factors within the infant. However, as

in previously cited studies, the assessment of emotional factors was

of questionable reliability and validity. Paradise (1966) has cited

several limitations to these researchers' assessment of family tension.

Evaluation of family tension [was] based on retrospective
review of records; bias [was] introduced both in initial
selection of mothers for study (Yale Rooming-In Project)
and in later elimination of those who failed to return
questionnaire; colic itself might have (1) engendered
family tension, or (2) prompted search for areas of
family tension which, among contented infants, might
have remained unexplained or undocumented. (p. 124)

In his study of 103 newborns of white, middle-class mothers,

Carey (1968) found that 12.6% of his entire sample of infants were

colicky. He employed a semistructured interview protocol to obtain a

profile of maternal anxiety. The protocol consisted of six items which

were designed to measure anxiety related to the mother's own rearing

experience, her feelings from previous experience bearing and raising

other children, her distress about her pregnancy itself, her anxiety

about expected family supports at home, other maternal anxiety factors,

and concerns about the baby. Each item was rated from 0 to 2 indicat-

ing little or no concern to great concern. Carey claimed that the

rating scale disclosed that postpartum distress resulted from situa-

tional stress (state anxiety) as well as maternal personality problems

(trait anxiety). However, it is difficult to see how the scale taps

both state and trait anxiety upon perusal of the items.

Based on this method of assessment, 3.2% of colicky babies had

mothers who experienced no anxiety, while 27.5% of the colicky babies

had mothers who experienced anxiety. This difference was highly sig-
nificant (X =13.4, p=0.01). Carey presented no data on the reliabil-

ity or validity of this rating scale for his subjects and admitted that

it had yet to be used by others to test its reliability. The sensitiv-

ity of the scale is highly suspect in that two-thirds of the mothers

received a 0 anxiety rating. However, aside from the question of

reliability of the scale, results may have been confounded by experi-

menter bias in that the investigator was aware of maternal anxiety

ratings when he made the diagnosis of colic.

Although Carey's results indicated that a larger percentage of

colicky babies than noncolicky babies had anxious mothers, he cautioned

that anxiety was not the sole ingredient in the causation of colic in

that most anxious mothers did not have colicky babies, while a few

nonanxious mothers did. Such results led him to conclude that

determining each infant's primary reactive pattern or
temperamental characteristics might show that those
infants who are generally more easily and intensely
reactive to environmental stimuli are more susceptible
to the disorganizing effects of maternal anxiety.
(p. 593)

Maternal support system. The support a woman receives during and

immediately alter her pregnancy may be a factor worth considering in

research investigating the relationship between maternal psychoemo-

tional factors and infantile colic. Social support has been found to

be helpful in mediating stressful situations in the woman who is preg-

nant for the first time (Nuckolls, Cassel, & Kaplan, 1972). Pleshette

et al. (1956) demonstrated a need for emotional support for primiparas,

especially in the first trimester of pregnancy, and Williamson and

Egeland (1981) emphasized the importance of social support for couples

during late pregnancy and the early postpartum period.

Few studies have focused directly on the effects of the maternal

support system on the course and outcome of pregnancy. In his study

of maternal anxiety and infantile colic, Carey (1968) found the preg-

nant woman's expectations of support after pregnancy to be negatively

associated with the emergence of infantile colic. He concluded, "Since

three out of five mothers expecting trouble in their family supports

had colicky babies, this form of anxiety appears to have the greatest

predictive value" (p. 593).

No other research has been conducted that addresses the relation-

ship between maternal support, during pregnancy or the postpartum

period, and the emergence of infantile colic. However, in discussing

the transition to parenthood, several authors (Hrobsky, 1977; Rossi,

1968; Shereshefsky & Yarrow, 1973) emphasized the importance of support

for the continuity of harmonious family life. Hrobsky (1977) wrote of

this transition period as a time of considerable anxiety and relearning

and noted that the family's support system acted as a "safety valve"

which helped stabilize the system by

offering validation of parents' perceptions and allaying
feelings of alienation experienced by new parents by
reinforcing the universality of their reactions .
releasing parents from some responsibilities and the
intensity of the home situation . functioning as
emotional sounding boards . providing concrete in-
formation on the care of the child . enhancing

parents' enjoyment of their new adventures by simply
sharing their experiences with them. (p. 464)

Paternal anxiety. To date, no study has examined the relationship

between paternal psychosocial or emotional experiences and infantile

colic. The literature focusing on the relationship between anxiety

and infantile colic indicates that, while the father is often a par-

ticularly salient individual in the infant's life (Lamb & Lamb, 1976),

exclusive attention has been given to the mother as the primary source

of environmental tension. Since the transition to parenthood has been

found to be a period of stress (Hrobsky, 1977) for both parents, the

effects of paternal anxiety on the infant should not be ignored.

Several authors and researchers (Bernstein & Cyr, 1957; Boehm,

1930; Hobbs, 1963, 1966; Lacoursiere, 1972; Zilborg, 1931) have sug-

gested that impending and new fatherhood are accompanied by increased

psychological stress for the father. The interruption of routine

habits and increased financial responsibility are major contributors

to this increased stress (Fein, 1976; Hobbs, 1963, 1966; Lacoursiere,

1972). In prebirth interviews, Fein (1976) found expectant fathers to

be concerned about labor and delivery, parenting, the emotional support

they would receive after birth, and changes in their marriages and


While minor emotional problems associated with fatherhood are

fairly common (Schaefer, 1965), psychiatric disorders associated with

prospective fatherhood have also been documented (Earls, 1976;

Lacoursiere, 1972; Zilborg, 1931). Such findings suggest that if

environmental tension is an etiological factor in the occurrence of

colic, the father's anxiety as well as that of the mother should be

a variable worthy of investigation.

Studies not supporting the relationship between colic and

maternal postpartum anxiety. In their classic studies, Illingworth

(1954) and Paradise (1966) attempted to dispel what they viewed as

the mythology surrounding infantile colic. Both researchers found

colic to be unrelated to sex, weight gain, type of feeding, and family

history of allergic or gastrointestinal disorders. In addition,

Paradise found that it was also unrelated to family economic class,

maternal age, birth order, and maternal emotional factors.

Although Paradise's research is valuable in that he attempted an

organized assessment of maternal emotional factors prior to the onset

of colic, it has been criticized for its use of the MMPI, and instrument

designed to measure abnormal personality traits. Paradise himself

suggested that the MMPI may not have been measuring emotional factors

of etiological importance. The instrument does measure anxiety but not

state anxiety, which might be the variable of most interest in a study

of pregnant women and colicky infants. In support of this thesis is

the fact that Paradise's own clinical assessment of maternal anxiety

was not consistent with estimations of anxiety measured by the MMPI.

Only 1 out of 10 women Paradise thought to be anxious received an

elevated anxiety score on the MMPI. In addition, of the six women

receiving elevated anxiety scores on the MMPI, only one was found to

be anxious in the clinical interview.

On the basis of his findings refuting the suggested etiologies

of most authors, Paradise concurred with the theory of Spock (1944)

and Lipton et al. (1960) that colic is a function of central nervous

system immaturity. He reasoned as follows:

infant, and Part III focuses on temperament, with an emphasis on its

role in the colic syndrome.

Part II: Prenatal Anxiety


Although the literature is replete with studies of the relation-

ship between prenatal anxiety and various maternal, pregnancy-related,

and neonatal factors, only two studies have investigated the relation-

ship between prenatal anxiety and infantile colic. Lakin (1957) found

a relationship between maternal anxiety during pregnancy and the devel-

opment of colic in the neonate; however, his evaluation of maternal

anxiety occurred after the onset of colic. Paradise (1966) found that

the incidence of colic was significantly higher in mothers who gave

histories of heightened emotional tension, or depression, or both,

during pregnancy. Although the assessment of these emotions occurred

on the second or third postpartum day, it was retrospective and,

therefore, subject to unreliability.

Prenatal Anxiety and Maternal Factors

Prenatal anxiety has been related to maternal factors such as age,

general education, length of marriage or relationship, and income.

Glazer (1980) found a negative relationship between anxiety and these

factors. Burstein, Kinch, and Stern (1974) found that anxiety in preg-

nancy decreased with age and number of previous pregnancies.

Prenatal Anxiety, Obstetrical Complications, and Childbirth

Prenatal anxiety has also been associated with pregnancy-related

factors such as obstetrical complications and childbirth abnormalities.

Upon administering a measure of trait anxiety to 146 women in their

third trimester of pregnancy, Crandon (1979a,b) found that the incidence

of pre-eclampsia, forceps delivery, prolonged and precipitate labor,

postpartum haemorrhage, manual removal of the placenta, and clinical

fetal distress were all significantly higher among the anxious women.

These women were also more likely to have infants with lower apgar

scores, a measure of the neonate's physical status minutes after birth.

Lederman, Lederman, Work, and McCann (1981) also found associations

between anxiety in labor and lower apgar scores. Davids and DeVault

(1962) administered a battery of psychological tests to 50 clinic

patients in their third trimester of pregnancy and found that highly

anxious women and their babies were more likely to experience preg-

nancy, labor and delivery complications, and childbirth abnormalities.

The reader is referred to McDonald (1968) for an extensive review of

the literature pertaining to the role of emotional factors in obstetric


Some researchers (Beck et al. 1980; Burstein et al., 1974; Farber

et al., 1981; Gorsuch & Key, 1974; Jones, 1978) have found no relation-

ship between trait anxiety and obstetrical complications and the condi-

tion of the fetus at birth. Grimm (1961) found no relationship between

psychological tension in pregnancy and pregnancy complications and

physical status of the newborn. However, he suggested that an associa-

tion might have emerged had he investigated tension throughout pregnancy

and not just during the last half of the third trimester.

Thinking that trait anxiety did not discriminate among exact

levels of felt anxiety over a short period, Gorsuch and Key (1974)

administered the State-Trait Anxiety Inventory (STAI) to 113 low SES

clinic patients several times throughout their pregnancy. They found

that while trait anxiety was not associated with pregnancy complica-

tions and childbirth abnormalities in their study, measures of state

anxiety, obtained early in pregnancy, were. The following explanation

was offered for their inability to replicate the results of previous


Lack of replication of former findings probably lies
in the more explicit definition of trait anxiety used
in the current study. The previous measures were trait
anxiety scales given in the latter half of pregnancy;
these measures probably measured some of the changes in
state anxiety that occurred during pregnancy itself.
However, the trait measures reflected not only trait
anxiety before pregnancy but also state anxiety during
the pregnancy; our data suggest it is only the latter
component of their scores--and particularly the states
of anxiety during the first trimester--that are related
to the problems of pregnancy. Therefore, it seems a
woman's having characteristically high anxiety before
pregnancy is not a factor in obstetric abnormalities.
(p. 361)

While the results of studies investigating the relationship between

prenatal anxiety and childbirth complications and abnormalities have

varied, most studies investigating severe or prolonged prenatal stress

have demonstrated associations with major morphological anomalies such

as cleft palate and Down's Syndrome as well as with many minor physical

anomalies such as hernias (Sontag, 1941). Drillien's study (cited in

Stott, 1973), associating prenatal stress with cleft lip in the newborn,

suggested that "severe emotional stress during the pregnancy is more

damaging to the child than physical illness" (p. 782). The results of

Stott's (1973) study of prenatal anxiety indicated that "serious

continuous inter-person tension during pregnancy [is] followed by high

child morbidity with what looks like a one-to-one relationship"

(p. 777). While extreme and prolonged levels of prenatal maternal

anxiety have been associated with obstetrical complications and child-

birth abnormalities, the effects of lesser degrees and durations of

anxiety need further clarification.

Maternal Blood Pressure and Pregnancy Outcomes

The effects of stress and anxiety on blood pressure have been well

documented (see Isselbacher, Adams, Braunwald, Petersdorf, & Wilson,

1980). In turn, high maternal blood pressure has been found to cause

morphological changes in the placenta and decreased interoplacental

blood flow resulting in a less than optimal environment for the develop-

ing fetus (Chisholm et al., 1978). Page and Christianson (1976) found

that higher maternal blood pressure in the fifth and sixth months of

pregnancy was related to perinatal mortality and intrauterine growth


Within normotensive samples, variations in blood pressure have

been associated with neonatal behavior. Chisholm et al. (1978) ex-

plained that

the distinction between normotension and hypertension
is arbitrary as blood pressure in pregnancy follows a
normal distribution. Therefore, the causal mechanisms
known to operate in hypertension in pregnancy could
well be operating among normotensive mothers. (p. 175)

In cross-cultural studies with Navajo, Malay, Chinese, and Tamil

mothers and infants, normal variations in prenatal blood pressure have

been associated with neonatal irritability and liability of states, as

measured by the Brazelton Neonatal Behavioral Assessment Scale

(Chisholm et al., 1978). In their Navajo sample, Chisholm et al. found

the correlation between neonatal irritability and second trimester blood

pressure to be as high as .71 (p= .001). The correlation between second

trimester blood pressure and liability of states was also high at .73

(j= .0001). Korner et al. (1980) replicated this study with a sample

of 70 middle class Caucasian mothers, all of whom were normotensive

throughout pregnancy, and their infants. They found a lower, but still

significant, correlation of .36 between maternal blood pressure in the

third trimester and spontaneous crying, as measured by an electronic ac-

tivity monitor. These results suggest that measure of blood pressure

taken during the second and third trimesters of pregnancy are more pre-

dictive of neonatal irritability than those taken earlier in pregnancy.

In 1979, Woodson et al. conducted a follow-up study to the

cross-cultural research undertaken by Chisholm et al. (1978). They

were unable to replicate the association between second trimester

blood pressure and neonatal irritability in a sample of 87

primiparous, English women. An association was found, however,

between prenatal growth retardation and exposure to either

oxytocin-stimulated labor or higher maternal blood pressure

during spontaneous labor and lower intrapartum fetal heart rate.

Also, greater infant irritability during the first postpartum week

was associated with lower intrapartum fetal heart rate. In order to

explain these relationships, the authors hypothesized that intra-

partum hypoxia, or reduced oxygen transport to the fetus, was a mediator

of the relationship between increased pregnancy and labor blood pres-

sure and newborn irritability. They suggested that the relationship

between maternal blood pressure and newborn irritability reflected the

ability, or inability, of the fetus to withstand the effects of varying

degrees of hypoxia.

The association between maternal prenatal blood pressure and neo-

natal behavior could possibly have some bearing on research investigat-

ing the relationship between maternal prenatal anxiety and subsequent

infant behavior. Although increases in prenatal maternal blood pres-

sure could be due to several causes, increased emotional stress has

been found to be an influential variable. The combined use of valid and

reliable measures of prenatal anxiety and blood pressure data could

help to clarify the relationship between maternal prenatal anxiety and

infant irritability.

Maternal Prenatal Anxiety and Neonatal Behavior

While blood pressure is thought to be a physiological correlate of

stress, psychological measures of stress have also been associated with

neonatal behavior (Joffe, cited in Chisholm et al., 1978). Sontag

(1944) reported associations between anxiety during pregnancy and the

subsequent emotionality of offspring. Bakow et al. (cited in Farber

et al., 1981) found the infants of highly anxious mothers to be less

alert and responsive to stimulation, as measured by the Brazelton Scale,

than infants of nonanxious mothers.

Ottinger and Simmons (1964) administered the IPAT Self-Analysis

Form to expectant mothers during each trimester of pregnancy. They

found that the infants of highly anxious mothers cried more before,

but not after, feedings than did newborns of nonanxious mothers. Using

the same measure of anxiety, Farber et al. (1981) found anxiety to be

related to neonatal behavior and mother-infant interaction, but only

for female infants. Highly anxious mothers had less active and alert

female infants and were slightly less skilled in feeding and play

interaction and showed less positive affect toward their infants. How-

ever, these correlations were all found to be rather low. Davids

(cited in Farber et al., 1981) found children of highly anxious mothers

performed less well on Bayley mental and motor scales than did infants

of less anxious mothers. These associations, however, raise questions

concerning the hypothesized relationship between prenatal anxiety and

infantile colic, as clinical impression suggests that colicky babies

are precocious in mental and motor development. Or, perhaps, this im-

pression is erroneous. Another explanation is that severe or prolonged

anxiety may result in poor infant condition at birth and later devel-

opmental lags, while normal amounts of anxiety, at particular points in

pregnancy, may produce maternal hormones that are transmitted directly

to the fetus. In the latter case, the condition of hypoxic stress, or

reduced oxygen transport to the fetus, may be bypassed. Additional

research is needed to shed light on this question.

The above studies investigated prenatal psychological stress and

behavioral outcomes. The only researchers to examine the relationship

between maternal prenatal anxiety and the specific behavioral outcome

of infantile colic have been Paradise (1966) and Lakin (1957). Both

found that women demonstrating greater anxiety during pregnancy were

more likely than nonanxious mothers to have colicky babies. Replication

of these results, using more valid and reliable instrumentation, has not

been forthcoming.

Animal Studies

There is still much to be learned about the effects of prenatal

anxiety on individual differences in the human fetus and neonate. How-

ever, it is now known that the fetus can respond to agents that are

maternally induced or received directly from the environment, or both

(Sontag et al., 1969) For example, the fetus has been found to re-

spond directly to sound stimulation (Sontag et al., 1969). It has also

been demonstrated that maternal stress can influence fetal heart rate.

As early as 1950, Montague cited evidence indicating that

nervous changes in the mother may affect the fetus through
the neurohumoral systems, i.e., the system comprising
the interrelated nervous and endocrine systems acting
through the fluid medium of the blood. He reported that
there is good evidence that the mother's emotional states
are, at least in chemical form, transmitted to the fetus.
(cited in Davids & DeVault, 1962, p. 468)

Although the mechanisms governing the relationship between prenatal

maternal anxiety and neonatal behavior are still unclear, animal

studies have offered some interesting possibilities for the mechanisms

operating in infrahuman subjects.

It has been shown that behavioral stress in pregnant rats and

mice alters the behavior and emotionality of the offspring (Ottinger

et al., 1963; Thompson et al., 1962). In order to investigate whether

these changes resulted from secondary effects produced by an alteration

of the placental exchange system or whether the fetus was affected

directly by the passage of hormones across the placental barrier,

Selye (cited in Lieberman, 1963) injected chicken eggs with epinephrine,

a hormone known to be highly responsive to stress, and tested the

offspring when they hatched. Results indicating that the

epinephrine-injected group's social responses were more vigorous and

intense suggested that the hormone can act directly to alter the

behavior of the offspring.

The results of Morishima, Pederson, and Finster's (1978) study of

the influence of maternal psychological stress on rhesus monkey fetuses

led them to conclude that maternal hyperexcitability can be hazardous

to the fetus. Maternal agitation was induced by exposure to a bright

light, considered to produce a mild degree of anxiety. A decrease in

heart rate and arterial oxygenation was seen in all fetuses.

Lederman et al. (1981) cited research with primates and sheep which

also produced evidence for the deleterious effects of experimentally

induced maternal excitement on the fetal heart rate and health status.

The hormone, epinephrine, was again implicated in affecting maternal

blood flow to the fetus. Maternal epinephrine level and anxiety,

measured at the onset of labor, demonstrated significant correlations

of .57 and .33 respectively with fetal heart rate pattern ratings.

Animal studies have made valuable contributions toward the explana-

tion of possible mechanisms for the association between maternal pre-

natal and labor anxiety and offspring behavior. It is now thought that

the hormone epinephrine, a hormone highly responsive to stress, can

affect fetal and neonatal behavior directly or by altering the maternal

blood flow to the fetus. Future research is needed to determine

whether similar mechanisms are operating in humans.

Part III: Temperament


Part I of this review indicated that numerous authors have sug-

gested that maternal emotional and psychological factors play a crucial

role in the etiology or intensification of infantile colic. Several

of these researchers (Brazelton, 1962; Carey, 1968; Wessel et al.,

1954), however, have suggested that the influence of these factors is

probably indirect as a result of the mediating influence of infant

temperament. It has been suggested that infant temperamental character-

istics, such as low sensory threshold, may predispose an infant to the

colic syndrome. By virtue of their temperamental characteristics some

infants may react to the disorganizing effects of maternal anxiety and

other internal and external stimuli by developing colic. Other tempera-

mental characteristics have also been associated with the onset of colic.

Historical Perspective

Although the concept of temperamental differences among individu-

als has been with us since ancient times (Stevenson & Graham, 1982),

the construct of temperament has only recently been subjected to

empirical investigation. In 1937, Gesell conducted one of the first

studies reporting individual differences in infant behavior. Since

then, other authors (Birns, 1965; Birns, Barten, & Bridger, 1959;

Korner, 1971) have studied individual differences in neonatal response

to stimulation with the idea that early characteristics may be pre-

cursors of later personality.

The New York Longitudinal Study (NYLS) (Thomas, Chess, Birch,

Hertzig, & Korn, 1963), the first large-scale study of temperament,

stimulated a mass of research. In an effort to investigate differences

in personality patterns, data were collected over a period of several

years, via parent interview, on 141 middle- to upper-middle-class

subjects beginning during their first few months of life. Defining

temperament as "behavioral style" or the "how" of behavior rather than

the "what" or "why," nine categories of temperament were identified by

an inductive content analysis of parent-interview protocols. The nine

categories and their definitions, constituting the authors' opera-

tional definition of temperament, are as follows:

(1) Activity Level: the motor component present in a
given child's functioning and the diurnal proportion of
active and inactive periods. Protocol data on motility
during bathing, eating, playing, dressing and handling,
as well as information concerning the sleep-wake cycle,
reaching, crawling and walking, are used in scoring this
(2) Rhythmicity (Regularity): the predictability and/or
unpredictability in time of any function. It can be
analyzed in relation to the sleep-wake cycle, hunger,
feeding pattern and elimination schedule.
(3) Approach or Withdrawal: the nature of the initial
response to a new stimulus, be it a new food, new toy
or new person. Approach responses are positive, whether
displayed by mood expression (smiling, verbalizations,
etc.) or motor activity (swallowing a new food, reaching
for a new toy, active play, etc.). Withdrawal reactions
are negative, whether displayed by mood expression (cry-
ing, fussing, grimacing, verbalizations, etc.) or motor
activity (moving away, spitting new food out, pushing
away new toy, etc.).
(4) Adaptability: responses to new or altered situations.
One is not concerned with the nature of the initial re-
sponses, but with the ease with which they are modified
in desired direction.
(5) Threshold of Responsiveness: the intensity level of
stimulation that is necessary to evoke a discernible re-
sponse, irrespective of the specific form that the response
may take, or the sensory modality affected. The behaviors

utilized are those concerning reactions to sensory stimuli,
environmental objects, and social contacts.
(6) Intensity of Reaction: the energy level of response,
irrespective of its quality or direction.
(7) Quality of Mood: the amount of pleasant, joyful and
friendly behavior, as contrasted with unpleasant, crying
and unfriendly behavior.
(8) Distractibility: the effectiveness of extraneous
environmental stimuli in interfering with or in altering
the direction of the ongoing behavior.
(9) Attention Span and Persistence: two categories which
are related. Attention span concerns the length of time
a particular activity is pursued by the child. Persistence
refers to the continuation of an activity in the face of
obstacles to the maintenance of the activity direction.
(Thomas & Chess, 1977, pp. 21-22)

Through the use of quantitative and qualitative analyses of

data, Thomas and Chess identified three functionally significant

temperamental constellations. The first, the "easy child," described

40% of their sample. These children were characterized by regu-

larity, positive approach responses to new stimuli, high adaptability

to change, and mild or moderately intense mood which was preponder-

antly positive. The second, the "difficult child," described 10%

of the sample. These children were irregular in biological functions,

responded with negative withdrawal to new stimuli, were nonadaptable

or adapted slowly to change, and exhibited intense mood expressions

which were frequently negative. The third, the "slow-to-warm-up

child," described 15% of the sample. These children exhibited

negative responses of mild intensity to new stimuli with slow adaptabil-

ity after repeated contact. Although Thomas and Chess found the "dif-

ficult" pattern to be predictive of behavioral disorders and other

aspects of future development, they cautioned that all three patterns

represented variations of behavioral style within normal limits. It

is only through the interaction between temperament and environment,

abilities, and motivations that behavior can be understood and perhaps


Listed among the strengths of the NYLS have been its large sample

size, low attrition rate, its prospective nature, young subject ages at

the onset of the study, concern with reliability and validity of parent

ratings, the large number of data collection points, and length of

the follow up (Persson-Blennow & McNeil, 1979). The study was not,

however, without weaknesses in that the sample was not representative,

siblings were used, and data were pooled.

Subsequent researchers have seen a need to simplify the rather

elaborate data collection and analysis methods used in the NYLS. In

1970, Carey devised the Infant Temperament Questionnaire (ITQ), a

parent report measure of temperament, with the intention of using it to

identify and treat children in his pediatric practice who demonstrated

the "difficult" child behavior pattern. In more recent years, several

other parent report instruments have been constructed and used in the

investigation of infant and child temperament. The clinical utility of

these instruments has been defended by many researchers and clinicians

(Carey, 1970, 1972, 1982; Chess, 1966, Chess, Thomas, & Birch, 1959;

Mclnerny & Chamberlain, 1978; Thomas and Chess, 1977); however, many

theoretical, conceptual, and methodological problems plague the study

of temperament. Future research must be geared toward clarification of

the origins, stability, definition, measurement, and clinical applica-

tions of temperamental differences (Carey, 1930).

Definition of Temperament

The literature is filled with definitions of temperament, present-

ing both minor and major variations. Stevenson and Graham (1982) have

taken issue with the tendency to view temperament as behavioral style

without any reference to the content of behavior. They have pointed

out that the content of behavior often dictates style and, therefore,

should be considered. Most researchers, however, have concurred that

temperament is a component of personality that describes the "how" of

behavior rather than the "what" or "why." Whether or not to include

suggestions as to the origins of temperament in its definition, however,

seems to be a matter of personal preference. While some theorists

(Allport, 1961; Buss & Plomin, 1975) have emphasized the genetic com-

ponent of temperament, others (Thomas & Chess, 1977; Carey & McDevitt,

1978) have focused on early-appearing stylistic differences in behavior,

interacting with the environment with no implications as to etiology.

Recently, Rothbart and Derryberry (1981) have looked at tempera-

ment within a psychobiological, maturational, and socio-experiential

framework. They defined temperament as constitutional individual

differences in reactivity and self-regulation

with "constitutional" seen as the relatively enduring
biological makeup of the organism influenced over time
by heredity, maturation, and experience. By "reactivity"
we refer to the characteristics of the individual's re-
action to changes in the environment, as reflected in
somatic, endocrine, and autonomic nervous systems. By
"self-regulation" we mean the processes functioning to
modulate this reactivity, e.g., attentional and behavioral
patterns of approach and avoidance. (p. 37)

Bates (1980) pooled the most frequently cited defining properties

of temperament and arrived at a composite definition which described

temperament as having a constitutional basis, appearing in infancy and

showing some degree of continuity, being an objectively definable

characteristic of an individual, and being affected by the environ-

ment. The lack of agreement on a single conceptual definition of tem-

perament has not seemed to be problematic, in a practical sense, in

that most investigators continue to operationally define temperament

according to the measures they use to assess it. Berger (1982) cautioned

that the study of temperament, like the study of personality in general,

is "diversity heading for chaos" (p. 180). He claimed that the lack of

good theories is responsible for the lack of good definitions and

presented three options for averting "chaos" in the study of tempera-

ment. He suggested that either theories could be improved, or tempera-

ment could simply be accepted as a "fuzzy entity," a term with a soft

core and indefinite boundary. The third option vould be to fall back

on the same saying that users of intelligence tests have depended upon

for years, "temperament is what temperament tests measure." It

seems that this third option is the one that most researchers have

inadvertently chosen.

Origins of Temperament

Although the utility of the temperament concept does not depend

upon a delineation of its origins, some researchers have used the

classic method of twin studies to investigate the existence of a

genetic component to temperament. Rutter, Korn, and Birch (1963)

found evidence for genetic components of the temperamental characteris-

tics of activity, approach/withdrawal, and adaptability but none for

regularity. Data also suggested that the genetic influence was

stronger in the first year of life than in either of the subsequent

two years studied. Most researchers would probably agree with Rothbart's

(1981) interactionist view expressed by her statement that

temperament is assumed to have a constitutional basis,
with "constitutional" defined as the relatively endur-
ing biological makeup of the individual influenced over
time by the interaction of heredity, life experience,
and maturation. (p. 559)

Continuity and Stability

The average test-retest correlation for temperamental characteris-

tics found in the literature has been estimated to be above .80 (Lyon &

Plomin, 1981). Persson-Blennow and McNeil (1979) found stability cor-

relations for temperamental characteristics for 6-month-olds, 1-year-

olds, and 2-year-olds to range from .50 to .92 (N=14, 2-3 week interval)

for eight of nine temperament categories. These correlations are

similar to those found in most temperament studies, but a little lower

than those of Carey and McDevitt (1978). Stability has been found to

vary depending on the particular temperamental characteristics being

measured as well as the age period during which the assessment takes


Birns et al. (1969) found the traits of irritability, tension,

sensitivity, and soothability to be stable from birth to 4 months of

age. Stability was not demonstrated, however, for alertness, vigor,

and maturity level. Based on these results, the researchers concluded

that certain temperamental characteristics evidence themselves at birth

and that these individual differences can play a major role in the

personality of the infant.

While short-term stability correlations may be of a respectable

magnitude, the long-term stability of temperamental characteristics

has been more difficult to demonstrate. In their analysis of data from

the NYLS, Persson-Blennow and McNeil (1979) found temperamental charac-

teristics to be significantly stable only over relatively short inter-

vals, suggesting slow changes rather than long-term stability. In a

study of their own, these same authors (1932) found that over an 18-month

period, most children changed their temperament type. In his review of

studies, Rutter (1982) found considerable stability up to one year of

age, but near zero correlations from year one to year five.

Some researchers, however, have found some continuity in tempera-

ment patterns. Mclnerny and Chamberlain (1978) found that infants

assessed by Carey's ITQ to be of intermediate or difficult temperament

at 6 months of age were seen by their mothers as "difficult" at age 2.

Based on their very large sample of 1855 infants gathered from well-baby

clinics in Helsinki, Huttunen and Nyman (1982) found significant posi-

tive correlations for seven of nine temperament dimensions studied at

6 and 8 months and again at 5 years of age.

The lack of long-term continuity found in most studies, however,

may have many sources. Perhaps what changes is not temperament but

the behaviors which define any particular temperament category. There

is "some evidence that similarly named functions at two different ages

may be based upon quite different behaviors, and that similar behaviors

at two ages may reflect different underlying functions" (Rutter et al.,

1963, p. 167). In addition to the possibility of functional inequiva-

lence of behaviors is the suggestion that variability itself may be a

stable attribute of some individuals (Rutter et al., 1963; Stevenson &

Graham, 1982). Such an "instability factor" has yet to be included in

any measure of temperamental characteristics. As previously mentioned,

another source of instability of temperament is its interaction with

environmental influences, maturation, and abilities. As Thomas and

Chess (1977) insisted:

Continuity and predictability can thus not be assumed
for a specific attribute or pattern of the child,
whether it be temperament, intellectual functioning,
motivational attributes or psychodynamic defenses. What
is predictable is the process of organism-environment
interaction. Consistency in development will come from
continuity over time in the organism and significant
features of its environment. Discontinuity will result
from changes in one or the other which make for modifi-
cations and change in development. (p. 174)

Therefore, although some authors (Buss & Plomin, 1975) have made lifespan

stability a requisite for temperament, others (Rothbart & Derryberry,

1981; Thomas & Chess, 1977) have insisted that temperament develops

over time and is influenced by maturation in the context of experience.

This view suggests

that during periods when there are no major maturational
shifts or changes in environmental conditions, stabili-
ties in temperament will be found. We also expect that,
even with maturational or experiential transitions in
temperament, a child's previous temperamental character-
istics will constrain the changes in behavior that can
occur over time. Viewed in this way, temperament at
time 1 will delimit the range of possible changes by
time 2. As yet, we do not know the rules for these
changes, but their identification should be a primary
focus of future longitudinal study. (Rothbart &
Derryberry, 1981, pp. 64-65)

Predictive Value of Temperament

Although there remains some question as to how predictive tempera-

ment measures are of future measures of temperament, several studies

have found certain temperament attributes to have predictive value for

characteristics of the individual other than temperament itself. The

utility of the temperament concept can be seen in understanding how

children may react differently to the same environmental stimuli. For

example, many newborns are able to follow a self-demand feeding schedule

rather successfully. However, many infants exhibiting the temperamental

characteristic of "irregularity" have been found to have difficulty

forming regular patterns of response to basic biological functions and,

therefore, do better on a more structured feeding schedule (Chess et al.,

1959). Carey (1982) insisted that informing parents of their child's

temperamental profile and its possible consequences can have a major

impact on providing the child with an optimal environment for living.

Temperament and behavior disorders. It has been suggested that

temperament can be linked to individual differences in the manner in

which other people respond to the child, personality development, and

childhood behavior disorders (Bates, 1980; Carey, 1980; Rutter, 1982;

Thomas & Chess, 1977). Based on the results of their NYLS, Thomas and

Chess (1977) concluded that "temperamental characteristics play signifi-

cant roles in the genesis and evolution of behavior disorders in

children" (p. 46). Additional evidence on the relationship between

temperament and childhood behavior disorders has been sparse and incon-

clusive. It has been suggested, however, that if a relationship does

indeed exist it is probably mediated by other variables, such as adverse

parenting practices (Broussard & Hartner, 1971; Thomas & Chess, 1977).

Several clinicians (Carey, 1970, 1972; Mclnerny & Chamberlain,

1978; Thomas & Chess, 1977) have used temperament scales to identify

children fitting the profile of the "difficult" child in the hope of

preparing for or preventing future behavioral problems. However, other

authors (Bates, 1980; Hubert, Wachs, Peters-Martin, & Grandour, 1982),

while encouraging further research on difficult temperament, have

cautioned against its use in screening and intervention programs until

the concept has been given a stronger empirical base.

Danger in "difficult child" label. Whether or not clinicians

choose to use temperament measures for the purposes of intervention,

it would seem wise to refrain from labeling a child as "difficult," as

such a label may bring about a self-fulfilling prophecy. Some clini-

cians (Carey, 1982; Mclnerny & Chamberlain, 1978) have claimed that

educating parents about the temperament of their child is a valuable

way of ameliorating any uneasiness they may have about their child's

behavior. In their study, Mclnerny and Chamberlain (1978) found that

very few mothers thought their baby was difficult

even when the rating is clearly so. When the baby is
scored as "difficult," this is pointed out with the
assurance that it is the baby's native temperament and
not caused by anything the mother does or does not do.
(p. 233)

The wisdom of such a statement is questionable in that it devalues the

mother's opinion of her own child. It assumes that the mother's sub-

jective opinion concerning the "difficulty" of her child is less ac-

curate and valuable than the assessment provided by standardized instru-

ments. Such an assumption may be erroneous in that the "difficulty" any

child poses to a mother is obviously influenced by the mother's view of

the child's behavior.

There is evidence that mothers do not define infant difficultness

in the same manner as clinicians. While clinicians and researchers

have described the "difficult" infant as beinq arrhythmic, withdrawing,

having low adaptability, intense, and negative, mothers have been

found to be more bothered by negative mood and low distractibility or

soothability (Carey & McDevitt, 1978). The mothers in another sample

indicated that a fussy, hard-to-soothe baby was "difficult" (Bates,

Bennett Freeland, & Lounsbury, 1979).

Perhaps mothers' perceptions of their infants' "difficulty" are

more predictive of future development than the clinicians' and re-

searchers' diagnostic labels. Broussard and Hartner (1971) found

mothers' perceptions of their infants at 1 month of age to be predictive

of later emotional development. The value of not bypassing parents'

perceptions, even if this were possible, has already been emphasized.

As Thomas and Chess (1977) stated:

Temperament has its main impact on socially-relevant
outcomes through a process of transaction between the
child and social environment. It makes good sense
then to measure the temperament or difficultness as
it is perceived by important figures in the child's
life. (p. 316)

In summary, while a cluster of temperament characteristics defining

"difficult" temperament has been associated with later personality

development and behavior patterns, further research is needed before

such results should be applied to intervention programs for children

with behavior disorders. Similarly, additional research is needed to

shed light on relationships which have been found between temperament

and other characteristics of the individual, such as developmental

delay, school performance, the incidence of accidents and visits to

the pediatrician for illness, and infantile colic (Carey, 1980).

Temperament and Infantile Colic

Several clinicians and researchers in the area of infantile colic

have suggested that constitutional or temperamental differences may

play a role in the degree to which any infant exhibits colic symptoms.

When describing colicky infants, clinicians have frequently referred to

their excessive motor activity and tendency to respond acutely to

internal and external stimuli (Holmes, 1969; Levine, 1956; Meyer, 1958).

Meyer (1958) thought that colic resulted from a combination of tempera-

ment and a frustrating immaturity which limited the infant's activity.

He wrote:

They [colicky infants] deviate from expected or conven-
tional behavior because they vary (as do adults) in
temperament, reaction to sensory stimuli, and in motor
skills. (p. 629)

Benjamin (1961) suggested that infants with low sensory threshold are

prone to colic

during the third or fourth postnatal week, when all
babies go through a maturational spurt in their
sensory capacities but when they have not as yet
developed an adequate stimulus barrier. (cited in
Korner, 1971, p. 614)

Brazelton (1962) also hypothesized that heavy fussers may differ in-

nately in their ability to assimilate stimuli. Several authors

(Brazelton, 1962; Carey, 1968; Cobb, 1956; Wessel et al., 1954) have

suggested that differing thresholds for sensory stimulation may regu-

late the degree to which infants react to internal and environmental


Carey (1970) was the first to investigate the hypothesized rela-

tionship between temperament and infantile colic. Out of eight infants

diagnosed as having colic, five were labeled as "difficult" or

"intermediate high," while three were labeled as "intermediate low," and

none were labeled as "easy." This overrepresentation of colicky

babies in the more difficult categories was found to be significant
(X =4.1, p < .05). These results must be considered as tentative,

however, as the sample size was quite small. Carey cautioned against

viewing the colicky infant and difficult child as synonymous. Not all

colicky infants were "difficult," and 9 of 11 difficult babies were

not colicky.

In another study, Carey (1972) found 13 out of 200 babies to be

colicky (7%). After the colic ceased, four were rated as "difficult,"

four as "intermediate high," four as "intermediate low," and one as

"easy." Again the overrepresentation in the first two groups was found
to be significant (X =6.7, p< .01) as was the difference between the

number of colicky infants in the "difficult" group versus the "easy"
group (X =5.3, p< .05). A more interesting finding in this study,

however, was that 11 (or 85%) of the colicky babies had low sensory

thresholds (2 =5.8, p< .02), as measured by the ITQ.

Other temperamental characteristics have been related to colic.

Colicky babies have been found to be more difficult on the rhythmicity

factor (cited in Licamele et al., n.d.). More recently, Huttunen and

Nyman (1982) found that infants having acute colic spasms were more

"intense" and had more symptoms of "negative mood" in infancy. They

also found that "intensity" at the infant age correlated with "low

sensory threshold" at a later age (r= .24, p< .001). As a result of

this finding, they questioned, "Could a high score in intensity at

the infant age reflect a low sensory threshold to environmental

stimuli?" If the answer is "yes," high levels of intensity among

colicky infants could also be a measure of their sensitivity to environ-

mental stimuli.

In addition to viewing environmental tension and other external

stimuli as impinging upon the sensitive colicky baby, one might also

question the role of internal stimuli. Korner and Grobstein (1967)

have suggested that "internal stimuli impinge on the manner in which

external stimuli are dealt with" (p. 678). If colicky babies really do

have abdominal cramps, this pain or discomfort could be more acutely

experienced by an infant with a low sensory threshold.

Assessment of Temperament

Most research in the area of temperament has been based on parent

questionnaires derived from the NYLS. Carey's (1970) Infant Temperament

Questionnaire (ITQ), one of the earliest parent-report measures, was

devised as a screening instrument for use by pediatricians and other

clinicians dealing with children. Because of its increasing use in

research, Carey and McDevitt (1978) revised the ITQ in a successful

effort to improve its psychometric properties. The Revised Infant

Temperament Questionnaire (RITQ) has now been standardized on infants

from 4 to 8 months of age.

In 1979, Persson-Blennow and McNeil constructed a Swedish version

of Carey's RITQ. Standardized on 6-month-olds, research with the

instrument has basically confirmed the work of Carey.

In an effort to obtain a more valid measure of infant temperament,

Bates et al. (1979) constructed a short screening device entitled the

Infant Characteristics Questionnaire (ICQ). A factor analysis of the

ICQ resulted in the identification of the four factors of Fussy-Difficult,

Dull, Unadaptable, and Unpredictable. The ICQ has been found to have

a meaningful factor structure, acceptable test-retest reliability, and

a low but statistically significant degree of mother and researcher

observation agreement on the difficultness of 6-month-old infants. The

ICQ and Carey's RITQ show a moderate degree of overlap, but they are

not equivalent measures.

Rothbart's (1981) Infant Behavior Questionnaire (IBQ).was devised

to measure an individual's reactivity and self-regulation, in addition

to Thomas and Chess's nine temperament dimensions, and to identify

conceptually distinct dimensions so that correlations between dimensions

could be explained "without inflating them by using similar items on

scales with different names" (p. 571). After conceptual and item

analysis the six scales of Activity Level, Smiling and Laughter, Fear,

Distress to Limitations, Soothability, and Duration of Orienting were

developed with adequate conceptual and psychometric characteristics.

The IBQ has been used with infants as young as 3 months old.

These instruments are representative, but not exhaustive, of the

many measures constructed to assess temperament. See Hubert et al.

(1982) for a more extensive review and critique of temperament measures.

Problems in the Assessment of Temperament

Although adequate test-retest reliability over short intervals has

been demonstrated frequently for several measures of temperament, dif-

ficulties still remain in the measurement of temperamental attributes.

The probability of functional inequivalence of different temperament

categories suggests that "developmental changes may modify or alter the

particular manner in which a characteristic is manifest" (Rutter,

1982, p. 7).

Although some researchers have developed and used instruments

yielding global descriptions of temperament (Buss & Plomin, 1975; Lyon

& Plomin, 1981), most have felt that instruments tapping specific con-

crete behaviors in specific contexts would have a better chance of

showing favorable psychometric properties. However, this attempt to

gather information on behavioral style within specific contexts has

often resulted in the emergence of context-specific factors which do

not match the scale developer's conceptual categories (Huitt & Ashton,


Yarrow's (1963) concerns about the assessment of mother-child

interaction could just as easily apply to the measurement of temperament.

By requiring the mother to place her behavior on a scale
point of frequency and intensity, we may be creating
illusions about childrearing environments and losing
important variations. In reality, behaviors of a given
parent and child may not have a simple modal level of
occurrence. They may occur in patterns of intensity and
frequency which are distinct and different according to
the situation, the developmental level of the child, and
the psychological state of the mother. (p. 218)

Items on any questionnaire can only sample from the numerous possible

contexts of behavior. A child may not show "persistence" in any of

the situations tapped by a particular questionnaire; however, the child

may show amazing persistence in other situations. The results of the

questionnaire would be misleading in that they would rate the child as

having low persistence without informing one of the actual selectivity

of the persistence.

Another difficulty inherent in questionnaires using adverbs, such

as "hard," "actively," and the like, is that the rater is usually

unaware of what the "norm" is when rating any particular behavior.

Results, therefore, may indicate different frames of reference rather

than true individual differences. The validity of parent-report ques-

tionnaires has been difficult to establish and has been cited as one

of the main limitations of the assessment procedures currently in use.

Validity of Assessment Measures

Research investigating the validity of temperament assessment

procedures has evidenced conflicting results. The validity coefficients

for most standardized temperament questionnaires, however, have been

significant, but low. Possible explanations for this low validity

include criticism of the instruments themselves, invalid validity

criteria, and the influence of parent perceptions.

Some researchers (Field, Dempsey, Hallock, & Shuman, 1978; Field

& Greenberg, 1982) have found that parent and observer ratings of neo-

natal behaviors were highly correlated. Mothers rated their 3-day-old

infants on the Mother's Assessment of the Behavior of Her Infant (MABI),

a modified version of the Brazelton Neonatal Behavioral Assessment

Scale, while trained testers evaluated the same infants using the

original version of the Brazelton scale. Despite the mothers' lack of

training, they agreed with the testers on at least 80% of the MABI item

ratings (Field et al., 1978).

Broussard (1975) and Broussard and Hartner (1970) found mothers'

perceptions of their infants' difficultness, as related to crying,

spitting, feeding, elimination, sleeping, and predictability at 1 month

of age, but not at 3 days, to be related to the infants' subsequent

emotional development. Other researchers (Meares, Penman,

Milgrom-Friedman, & Baker, 1932), using Broussard's Neonatal Perception

Inventory (NPI), found that the mothers' initial perceptions of their

babies seemed to have little effect on their judgments of the infants'

personalities 48 hours later. However, perceptions at 1 month of age

seemed to be based on a combination of their infants' exhibited com-

petencies and their own characteristics. Mothers in this study appeared

to have fairly accurate perceptions of their 1-month-old infants'

physical capabilities, as those infants perceived to be "not difficult"

on the NPI had good state control, as assessed by the Brazelton.

Most researchers have found parent-report measures of infant

temperament to be only minimally correlated with observations of the

child by trained observers. The validity of these forms of assessment

has, therefore, been questioned. Bates (1930) reviewed a number of

studies and found significant, but low, parent-observer correlations

ranging from .20 to .50. Sameroff, Seifer, and Elias (1982) found

correlations between home and lab observations with parent response

on Carey's ITQ to range from .01 to .25.

Inadequate instrumentation. One of several explanations that

can be offered for these low correlations is that the assessment pro-

cedures are inadequate. Difficulties due to relativity of measures,

long-term stability, social context, and functional inequivalence of

different measures at different ages are just some of the possible

factors undermining the valid use of parent-report measures. In

defense of the temperament questionnaire, McNeil and Persson-Blennow

(1982) made the following statement:

In the first steps of constructing a questionnaire, it
seems difficult enough to choose a sufficient number of
behaviors, in situations for e.g. a six-month-old child,

which represent abstractly defined concepts that are to
become variables, and that keep mood separate from
approach, activity separate from intensity, attention-
persistence separate from distractibility, etc. The
items have to be expressed in simple, unambiguous
language and be clearly related to the mother's every-
day experience with the child, yet they should be (we
feel) at least somewhat independent of her management
of the child. The choice and development of such items
presents no easy task. The demands then placed on these
poor items are that: they should be answerable by all
parents and relevant to all children; they should have
balanced distributions, and discriminate between high
and low scorers on the variable; they should correlate
highly with the score for the total variable, and not
be thought by blind judges to represent other variables;
they should be answered in the same way by the parent
who is given the unexpected (and perhaps undesired) op-
portunity to complete the very same questionnaire twice
within a short period of time; they should be answered
similarly by the father whose primary experience with
the baby often consists of short contacts in the
morning and evening; they should reflect a behavioral
style which the baby (and mother) will show when a
scientist comes to observe them even on only one or two
occasions; and, preferably, they should predict some-
thing important or abnormal about the child in the
future. It is more than a rhetorical question to ask
how many existing research methods of any type and
field have even a majority of these desirable charac-
teristics. (pp. 30-31)

Invalid validity criteria. A second explanation for the failure

to establish higher validity correlations is that the validity criterion

most frequently used is itself subject to invalidity. Because parent-

parent agreement has usually been somewhat higher than parent-observer

agreement, some researchers have questioned the frequently held

assumption that observations necessarily yield valid data. "They do

so only to the extent that the phenomena to be observed have not been

destroyed, missed, or misinterpreted by the observer" (Yarrow, 1963,

p. 223). There is little question that the presence of an observer

can alter, and therefore "destroy," the behavior of those being observed

(Werry, 1982), just as there is little doubt that "observers have

perceptions too" (Carey, 1932) and phenomena, such as reliability drift,

can alter perceptions of what is being observed. The fact that ob-

servations are merely samples of behavior means that many behaviors in

many situations are being "missed." Sameroff et al. (1982) found that

agreement between two observations and interview ratings was higher

than agreement between the observations themselves.

Sameroff et al. (1982) found correlations between home observation

and lab observation to range from .04 to .38. Although these correla-

tions are somewhat higher than those ranging from .01 to .26 for the

parent-observer agreement, they are still unimpressive causing one to

question the use of observations as the sole validating criterion for

parent-report measures. Wilson's (1982) finding that questionnaire

ratings for the Infant/Toddler Temperament Questionnaire correlated

sigrificantly (.41-.52) with laboratory ratings at 6 to 12 months seems

to support this view.

Influence of parent perception. A third possible source for the

low parent-observer correlations is the influence of parent perceptions.

Designers of recent parent-questionnaires have attempted to circumvent

parent perceptions by writing items that tap the child's actual behavior

in concrete situations. However, because temperament measures are not

independent of the home environment (Rothbart, 1981), it has not been

surprising that characteristics of parents would influence their ratings

of their child's temperament. The fact that parent-report measures

have more frequently been filled out by mothers than by fathers explains

the focus on "maternal" rather than "paternal" influences on parent-

report ratings.

Maternal characteristics and infant temperament. Maternal charac-

teristics thought to be related to infant temperament have included

anxiety, certain demographic characteristics, prenatal and postnatal

adjustment, mother-infant interaction, and maternal perception of infant


Several authors have linked maternal anxiety with infant tempera-

mental characteristics. Maternal prenatal anxiety has been associated

with neonatal irritability, as measured by the Brazelton and electronic

activity monitors (Farber et al., 1981; Ottinger & Simmons, 1954).

Others (Sameroff et al., 1982; Vaughn, Deinard, & Egeland, 1980) have

found prenatal anxiety to be related to perceptions of infant tempera-

ment. Carey (1930) has suggested that the relationship between anxiety

and temperament may have many sources. Anxiety may (1) directly affect

the fetus, (2) influence the postnatal mother-infant interaction,

(3) influence reported perceptions of the infant, or (4) both maternal

anxiety and infant behavior may result from some third common biologic

or environmental factor. Research has not yet been able to establish

a causal relationship between anxiety and infant temperament.

The relationship between mother-infant interaction and infant

temperament is unclear in that most studies show conflicting results

(cited in Bates, 1980). Although specific temperamental characteris-

tics cannot be predicted on the basis of maternal-infant interaction

patterns, it is assumed that this interaction must influence the sub-

sequent personality development of the child (Ieares et al., 1932) for

infants with differing temperamental characteristics elicit different

responses from individuals in their environments (Thomas & Chess, 1977).

Some researchers have investigated the relationship between

temperament and several maternal characteristics including race,

mental health, socioeconomic status (SES), parity, and personality.

Sameroff et al. (1982) found that infants of lower socioeconomic status

and Black and mentally ill mothers were rated as more difficult on

Carey's RITQ. During their third trimester of pregnancy, Vaughn et al.

(1980) had 187 low SES, mostly single, primiparas complete a battery of

psychological tests, including the IPAT, Eroussard's NPI, and the

Schaefer and Manheimer Pregnancy Research Questionnaire. At 3 and 6

months after birth, infants were observed at home, and mothers completed

Carey's ITQ. Results indicated that differences existed between

mothers of "easy" and "difficult" babies with mothers who rated their

infants as "difficult" on the RITQ being more anxious, having less

desire for pregnancy, and lower maternal feelings, and believing pre-

natally that their infants would be "difficult." The authors concluded

that the RITQ data were related "more to maternal characteristics and

expectations concerning the behavior of infants than they are to actual

infant behaviors" (p. 513). This study has been criticized, however,

because no statistical measure of the strength of correlations between

maternal feelings and ITQ findings and no data on correlations of ITQ

data with appropriate observations of infant behavior were presented

(Carey, 1980). However, these results cannot be totally dismissed in

that other authors (Bates et al., 1979) have also found maternal charac-

teristics to be related to measures of infant temperament. Bates et al.

(1979) found that maternal personality factors, SES, and parity all

entered the regression equation before the single observed behavior of

fussiness, suggesting that maternal characteristics are more predictive

of infant temperament than fussiness ratings made by observers. This

finding has been cited by several authors as supporting the claim that

infant temperament measures are better measures of maternal than in-

fant characteristics. However, these same authors failed to report


when observer "ratings" rather than observed behavior
scores were entered into the regression equation, the
observers' ratings of fussy-difficult did make a posi-
tive contribution to the prediction of mothers' ratings
of difficultness. . This might be expected, since
the observers' ratings were made on a scale identical
to the mothers' ratings, and the scale may have tapped
aspects of the child's behavior not measured by the
observation codes. (Rothbart & Derryberry, 1981, p. 76)

Other researchers have found maternal perceptions of their infants'

difficultness, as measured by Broussard's NPI, to be correlated with

other maternal characteristics. Broussard and Hartner (1971) found

maternal perceptions of the 1-month-old, but not 3-day-old, infants to

be associated with depression, negative aspects of childrearing, and

irritability, as measured by Schaefer and Manheimer's Postnatal Research

Inventory. They also found this relationship to be independent of

parents' educational level, father's occupation, change in income since

delivery, prenatal or postpartum complications, the type of delivery,

family moves, and sex of the child. The finding that 54% of the women

shifted their perceptions from 3 days to 1 month has led the authors to

emphasize the need for adequate support systems for mothers during the

early postpartum period when perceptions are still in a "fluid state."

Meares et al. (1982) found that mothers who were flexible and

prepared for motherhood did not perceive their infants as likely to be

difficult in terms of crying, as measured by the NPI. However, based

on results of the baby's assessment by trained testers on the Brazelton

Scale, the authors concluded that the mothers' perceptions of their

infants were a function of the babies' exhibited capacities, as well as

the mothers' own characteristics. Although researchers differ in the

degree to which they believe maternal and infant characteristics influ-

ence maternal ratings of infant behavior, most would agree that both

are sources. Sameroff et al. (1982) cautioned that researchers using

temperament questionnaires "must entertain the possibility that these

measures might reflect children's behaviors, more likely reflect parental

characteristics and most likely reflect a complex combination of the two"

(p. 173).

Although research is still needed to clarify the influences on

questionnaire ratings, progress has been made in establishing their

validity. Carey and McDevitt (1978) have cited studies using their

RITQ that have verified that "difficult babies do cry more" (p. 738),

and, as recently as 1982, Carey stated that

evidence is accumulating that parental ratings of tempera-
ment are largely valid, if we accept as an adequate meas-
ure of parental validity an agreement with the brief
professional ratings (or are they perceptions?) of the
child that have generally been used so far. (p. 195)

The probability that parent questionnaires are measuring infant tempera-

ment is also increased by the consistent finding of significant, al-

though low, correlations between parents and observers (Bates et al.,

1979). To say that parent ratings are to some extent confirmed by the

way the observer perceives the child is not to deny the influence of

parent perception. However, this influence is probably unavoidable in

that the temperament scales presently in use "measure an infant's be-

havior as seen in a particular social system involving caregivers and

siblings" (Rothbart & Derryberry, 1981, p. 70). In addition, the

confounding of infant temperament and parent perception may not pre-

clude the usefulness of temperament questionnaires, for "in the long

run the parents' perceptions of their child may have a greater impact

on their child's development than the child's own temperament"

(Sameroff et al., 1982, p. 172).


Lack of agreement concerning the etiology and treatment of infantile

colic has contributed to the inconsistencies in its definition. As a

result, most authors have chosen to describe, rather than define, colic.

The one aspect of this syndrome that is usually agreed upon is that it

involves infant crying for which there is no known treatment that gives

consistent or complete relief.

While most authors agree that several factors are involved in the

etiology of colic, a large number of studies emphasized the role of

maternal anxiety, expressed either prenatally or postnatally, in the

etiology or intensification of colic. Important methodological flaws,

however, have limited the value of many of these studies. Not only

has the assessment of maternal emotions often been subjective, but it

has taken place after the onset of colic. It is not surprising to find

elevated anxiety scores in mothers living with colicky babies.

In his classic study, Paradise (1966) dispelled many of what he

considered to be the "myths" that have been perpetuated about colic.

Paradise found the incidence of colic to be unrelated to family economic

class, maternal age, birth order, sex, weight gain, type of feeding,

family history of allergic or gastrointestinal disorder, and postnatal

maternal emotional factors. Superior maternal intelligence, advanced

maternal education, and prenatal emotional factors were associated

with a higher than average incidence of infantile colic. Paradise's

study seems to indicate that if maternal anxiety contributes to infan-

tile colic, the influence is most likely to occur prenatally. Postnatal

anxiety probably acts to intensify, but not cause, colic.

The literature on prenatal emotional and psychological factors

and their relationship to infantile colic is sparse. Only Paradise

and Lakin found associations between these variables. The studies

exploring the relationship between prenatal anxiety and pregnancy com-

plications, birth abnormalities, and neonatal behavior are numerous and

emphasize the influence that such emotional factors can have on the

mother, fetus, and neonate. A few studies carried their investigations

into the behavior of the neonate. Ottinger and Simmons (1964) found

infants of highly anxious mothers to cry more before, but not after,

feedings, than infants of less anxious mothers. Chisholm et al. (1978)

found a significant correlation of .71 between maternal blood pressure,

a physiological correlate of stress, and newborn irritability. Other

studies have associated prenatal anxiety with state changes, activity

level, and infant-parent interaction. Future replications, however,

are needed to lend more credence to these results.

Animal studies have supported many of the findings of research

with human subjects. Emotionality and behavior of animal offspring

have been associated with prenatal maternal anxiety. Although the

mechanisms for this relationship in humans are still far from clear,

studies with animals suggest that alterations in the placental-

interchange, as well as direct action of maternal hormones, resulting

from maternal stress, on the fetus, are both legitimate hypotheses.

Results of hypertension research have demonstrated neonatal abnor-

malities via reduction of oxygen transport to the fetus. Such an ex-

planation for the emergence of infantile colic is doubtful, however, as

colicky babies are thought to exhibit both motor and mental precocity.

The more likely explanation is that, somehow, the fetus is directly

affected by the hormonal changes, produced by stress, within the mother.

Several investigators who have linked colic with maternal or family

tension have qualified this relationship by hypothesizing that the de-

gree to which an infant manifests colic symptoms may be determined by

constitutional or temperamental factors. Infants have been found to

respond differentially to varying types and degrees of stimulation.

Benjamin (1961) hypothesized that infants with low thresholds for

sensory stimulation may be prone to colic "during the third or fourth

postnatal week, when all babies go through a maturational spurt in

their sensory capacities but when they have not as yet developed an

adequate stimulus barrier" (cited in Korner, 1971, p. 614). A low

sensory threshold might not only sensitize the colicky infant to

internal discomforts, but it might also make environmental tension

and influences particularly disorganizing.

Other temperamental attributes may also be characteristic of the

colicky infant. Huttunen and Nyman (1982) found that infants having

acute colic spasms were more "intense" and had more symptoms of "nega-

tive mood" in infancy. Other researchers (cited in Licamele et al.,

n.d.) have found colicky babies to be less rhythmic than those not

exhibiting colic symptoms.

Meyer (1958) speculated that colic resulted from a combination of

temperament and a frustrating immaturity which limits the infant's motor

outlets. If his hypothesis is correct, it is conceivable that Roth-

bart's "Frustration to Limitations" temperament dimension may be im-

portant in describing the colicky infant. Clinical impressions of the

colicky baby as being precocious in motor and mental development seem

to suggest that this temperamental characteristic may be a relevant


The research linking prenatal maternal anxiety with neonatal ir-

ritability and activity could be seen as supporting the hypothesis that

anxiety is associated with infant temperament, as both of these behaviors

have been considered to reflect temperament. Prenatal maternal anxiety

has also been associated with infant temperament, as measured by

parent-report questionnaires (Sameroff et al., 1982; Vaughn et al.,

1980). However, the identification of specific temperamental character-

istics in infants has been difficult in that measurement scales tend to

reflect parent characteristics as well as infant behavior. This dis-

tortion of infant characteristics, however, may not only be unavoidable,

it may be necessary in that temperament itself is assumed to be insepa-

rable from environmental influences. "Perceptions of infant tempera-

ment may actually be more important to the child's future development

than the child's temperament itself" (Sameroff et al., 1982, p. 172).

The present study was intended to contribute to knowledge of the

influences of parental psychosocial and emotional factors and infant

temperament on infantile colic.



The literature review presented in Chapter II indicated that the

study of infantile colic has, for the most part, received inadequate

attention. Researchers who have investigated the problem have gener-

ally focused on only one dimension of the problem. Attention has

usually been focused on the relationship between infantile colic and

either maternal emotional factors, infant physiological factors, or

infant temperamental factors.

The present study was designed to study the interrelationships

among several demographic factors, parent psychosocial and emotional

factors, infant factors and infantile colic, measured during the sixth

week of life. The variables of interest were organized into three

categories: demographic and descriptive variables, parent psychosocial

and emotional variables, and infant variables. Specifically, the

present study addressed the following questions:

I. What is the relationship between certain demographic
and descriptive characteristics and infantile colic?

II. What is the relationship between infantile colic and
selected parent psychosocial and emotional factors
assessed prenatally and postnatally?

III. What is the relationship between selected infant be-
haviors observed during the first week of life and
infantile colic?

IV. What is the relationship between infantile colic
and infant temperament characteristics assessed
at 3 and 8 months of age?

The methodology employed to investigate these questions is dis-

cussed in this chapter. Information relevant to sample selection,

variables, research procedures, instrumentation, hypotheses, and methods

of statistical analysis is presented in the following sections.


The participants for this study consisted of 163 couples who were

recruited from prepared childbirth classes conducted through the

Birthplace, Alachua General Hospital, Shands Hospital, and private

obstetrical practices, all located in Gainesville, Florida. Although

emphasis on content varied depending upon the individual instructor,

classes generally dealt with the process of conception, reproduction,

implantation, fetal growth, and labor. In all classes, the focus was

placed on labor and delivery and on how both parents could take an

active and important role in the birth of their child.

Recruitment of subjects began in November of 1982 and continued

until March of 1983 when a sample size of 170 couples was achieved.

Five couples dropped out of the study for reasons including family ill-

ness and death, overwhelming responsibilities after the baby's birth,

and marital discord. Two families were eliminated from the study

because of the birth of twins. Approximately 80% of the invited sample

consented to participate in the study.

The average age for mothers and fathers was 26 years and 29 years,

respectively. The mean number of years completed in college was 2

years for mothers and 3 years for fathers. Most couples were White,

married, and expecting their first child. Demographic and descriptive

statistics for the sample are presented in Tables 3.1 to 3.6.

Table 3.1

Mean Age, Education, and Parity

for Parents

M N SD Min. Max.

Mother's Age 26.52 years 163 4.28 18 38

Father's Age 29.33 years 161 5.57 18 63

Education 14.21 (2 years college) 163 2.27 9 20

Education 15.02 (3 years college) 161 2.70 10 21

(including this 1.43 children 131 .77 1 5

Table 3.2

Frequency Distribution of Sample by Race

Race Frequency %

White 151 92.64

Black 9 5.52

Other 3 1.84

Table 3.3

Frequency Distribution of Sample by Marital Status

Marital Status


Married 156 96.30

Unmarried 3 1.35

Unknown 4 1.85

Table 3.4

Frequency Distribution of Sample by
Father's Occupation

Mother's and

Mother's Occupation

Frequency %

Father's Occupation

Frequency %

Blue Collar 12 7.45 66 41.51

White Collar 60 37.27 35 22.01

Semiprofessional 38 23.60 32 20.13

Professional 4 2.48 14 8.81

College Student 3 1.86 9 5.66

Homemaker 44 27.33 1 .63

Unemployed 0 0.00 2 1.26

Table 3.5

Frequency Distribution of Sample by Mother's Parity



1 (primipara) 92 70.23

2 27 20.61

3 8 6.11

4 3 2.29

5 1 0.76

Table 3.6

Frequency Distribution of Sample by Infant Gender

Gender Frequency %

Males 82 50.31

Females 74 45.40

Unknown 7 4.30


The present study investigated the relationship between variables

and subsets of variables and measures of infantile colic, collected at

6 weeks of age. Variables were organized into the three categories of

demographic, parent, and infant factors.

Demographic Variables

The demographic and descriptive variables of particular interest

included parent educational level, occupation, birth order, allergy,

and feeding method, all of which have been cited, in research or

theory, as having some association with colic. Other demographic and

descriptive characteristics which were also investigated included par-

ents' age, number of times pregnant, baby's sex, mother's drug use in

labor and delivery, father's work hours, baby's birthweight, baby's

gestational age, and gastrointestinal distress.

Parent Variables

Psychosocial and emotional experiences of parents during pregnancy

and the postpartum period were investigated through the global construct

of parent anxiety, as well as through more specific measures of mater-

nal prenatal emotional experiences and maternal postpartum fatigue and


Parent anxiety and environmental tension have frequently been

linked with the onset and severity of infantile colic. A differentia-

tion was made in the present study between state and trait anxiety as

it was hypothesized that state anxiety, as defined as a transitory

emotional state of tension and apprehension, might be a more appropri-

ate measure of anxiety assessed during the time-limited conditions of

pregnancy and colic than the more stable measure of anxiety proneness

(trait anxiety). Measures of state and trait anxiety were, therefore,

collected at different points during the study.

The influences of parental anxiety may be exerted prenatally,

through maternal hormonal changes affecting the developing fetus, or

postnatally via the communication of tension to the newborn infant.

To avoid methodological shortcomings of previous studies, the parent

variable of anxiety was assessed at several different junctures in the

study. Anxiety measures were collected during pregnancy and as soon

after birth as possible, in an effort to obtain data prior to the onset

of colic.

Because it was hypothesized that more specific measures of parent

psychosocial and emotional experiences might be more informative than

the global construct of anxiety, measures of maternal prenatal experi-

ences and postpartum fatigue and support were collected. Mother's

fears for herself and her infant, desire for pregnancy, dependency,

irritability, maternal feelings, and depression could possibly pinpoint

more specific sources of maternal prenatal anxiety.

Other possible sources of state anxiety could be the fatigue

experienced by the new mother as well as the existence or nonexistence

of an adequate support system. The mother's perceptions of support,

however, might be more crucial than the actual support provided in

terms of caring for the baby. Therefore, in addition to the general

variable of state-trait anxiety, more specific variables of maternal

prenatal experiences, postpartum fatigue, and support were investigated.

Infant Variables

While colic has often been related to factors in the infant's

environment, it has been postulated that the condition may stem from

factors within the infant. Because infantile colic is a time-limited

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs