Applied Developmental Psychology 24 (2003) 437 – 456

The effects of television advertising on materialism,

parent–child conflict, and unhappiness:
A review of research
Moniek Buijzen*, Patti M. Valkenburg
Amsterdam School of Communications Research ASCoR, University of Amsterdam, Kloveniersburgwal 48,
1012 CX, Amsterdam 1012, The Netherlands

Abstract

In this article, we introduce a model on the unintended effects of advertising. This model describes
the existing hypotheses about the impact of advertising on (a) materialism, (b) parent – child conflict,
and (c) unhappiness. The validity of each of these hypotheses was investigated using a vote-counting
analysis. Our analyses yielded a small to moderate effect size for the relation between advertising and
materialism as well as a small to moderate effect size for the relation between advertising and parent –
child conflict. However, support for the hypothesized relation between advertising and unhappiness
was not found. The outcome of our vote-counting analysis provided several working hypotheses for
further research as well as identified possible moderator variables that should be taken into account in
future research.
D 2003 Elsevier Inc. All rights reserved.

Keywords: Advertising; Materialism; Parent – child conflict; Unhappiness; Children; Review

1. Introduction

Television advertising aimed at children has been the subject of considerable public
concern, both in the United States and in Europe. In the United States, policies restricting
advertising aimed at children were formed as early as the 1970s. Two decades later,

* Corresponding author. Tel.: +31-20-525-2426; fax: +31-20-525-3681.

E-mail address: buijzen@pscw.uva.nl (M. Buijzen).

0193-3973/$ – see front matter D 2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0193-3973(03)00072-8
438 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456

several European countries implemented rules or regulatory policies concerning advertis-

ing directed at children. Belgium for example, issued the Five-minute rule in 1995, a
decree which states that children’s programs may not be immediately preceded or
followed by children’s advertisements. In the same period, Greece proclaimed a partial
ban on advertising, prohibiting toy advertising between 7 a.m. and 10 p.m. Finally, in
1995, Norway introduced a total ban on television advertising directed at children, a
policy also adopted by Sweden in 1996.
These restrictions on television advertising directed at children are based on either
ethical concerns among parents and policymakers about the unfairness of advertising to
children or on beliefs that television advertising has harmful effects on children.
Although the possible aversive effects of advertising have been debated since television
was introduced, there is still no consensus, however, about what these effects are, and
whether they are supported by empirical research. The main aim of this review study is
to identify and analyze the existing hypotheses regarding the negative effects of television
advertising and to examine the extent to which these hypotheses are supported by
empirical research findings. Our literature review focuses on three possible harmful
effects of television advertising: the effects on materialism, parent–child conflict, and
unhappiness.
Studies on the effects of advertising directed toward children are generally based on two
paradigms: the paradigm of the empowered child and the paradigm of the vulnerable child.
In the empowered child paradigm, children are viewed as skilled consumers capable of
critically processing commercial messages. Research based on this paradigm, mostly
marketing research, generally focuses on the intended effects of advertising. Intended
advertising effects refer to children’s brand awareness, brand attitudes, and purchase
intentions.
Advertising-effects research based on the vulnerable child paradigm assumes that
children lack the cognitive skills to protect themselves against advertising messages. It is
believed that children, more than adults, are susceptible to the seductive influences of
commercials. Research within the vulnerable child paradigm usually investigates the
extent to which advertising is harmful for children and addresses issues such as whether
children should be protected from these possibly harmful effects. This research focuses
on the unintended effects of advertising, for example, on family conflict and materialistic
values. In other words, unintended effects are the secondary, usually negative, effects of
children’s exposure to advertising.
Several earlier reviews have discussed the unintended effects of advertising (Atkin,
1980; Kunkel, 2001; Young, 1990). The present review differs from these reviews in two
respects. First, unlike earlier reviews, which were narrative in nature, we conducted a vote-
counting analysis. A vote-counting analysis is a formalized count of the number of studies
that either produce or fail to produce statistically significant findings in the hypothesized
direction. A vote-counting analysis is considered the most suitable and accepted method of
research synthesis when a literature is small and heterogeneous, as in the case of the
advertising-effects literature (Bushman, 1994; Cooper, 1990). Second, unlike earlier
narrative reviews, our vote-counting analysis provides a comprehensive review of the
 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456 439

research literature. Earlier narrative reviews on the unintended effects of advertising have
usually discussed a selection of the research, without reporting any criteria on how they
selected the empirical studies that were included in their reviews. Nonsystematic literature
searches can easily lead to biased samples and, as a result, to biased conclusions (Davies,
2000; Johnson & Eagly, 2000).
In preparing this vote-counting analysis, we collected all relevant references from the
standard computer-searchable databases (Econlit, ERIC, PsycInfo, Sociological Abstracts,
and Social Sciences Citation Index), including studies published up to the fall of 2002. In
addition, references from these publications were examined to trace articles, chapters, and
unpublished reports that were not recorded in these databases.

1.1. Hypotheses about the relations between advertising, materialism, parent–child conflict,
and unhappiness

In the past four decades, several hypotheses regarding the impact of advertising on
materialism, parent–child conflict, and unhappiness have been put forward in the literature.
For reasons of clarity, we have modeled the hypothesized relations between advertising and
these three unintended effects in Fig. 1. The model describes all the advertising-effects
hypotheses with regard to materialism, parent–child conflict, and unhappiness that have
received research attention.

1.1.1. Advertising enhances materialism (path 1)

Several authors suggest that advertising stimulates materialistic values in children
(Greenberg & Brand, 1993; Liebert, 1986; Pollay, 1986; Wulfemeyer & Mueller, 1992).
According to these authors, advertising enhances materialism because it is designed to
arouse desires for products that would not otherwise be salient. Advertising propagates the
ideology that possessions are important and that desirable qualities, such as beauty, success,

Fig. 1. Model for the three unintended effects of advertising. The arrows represent the hypotheses with regard to
the effect of advertising on materialism, parent – child conflict, and unhappiness that have received research
attention. The dashed square box represents a variable that has been investigated but not explicitly operationalized.
440 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456

and happiness can be obtained only by material possessions (Pollay, 1986; Wulfemeyer &
Mueller, 1992).

1.1.2. Advertising causes parent–child conflict (path 2)

A second hypothesis states that exposure to advertising leads to parent–child conflict
(see path 2; Atkin, 1980; Robertson, 1979). Although it is possible that advertising
exposure directly influences parent–child conflict, it seems more plausible that this
relation between exposure to advertising and parent –child conflict is mediated by
children’s purchase requests and parental denial of these requests. According to this
mediational hypothesis, advertising (a) leads to an increased number of requests for
advertised products on the part of children, which (b) leads to an increased number of
product denials (given that not all requests can be granted), which then (c) causes
conflict between parent and child. These mediated relations are represented by paths 2a,
2b, and 2c in Fig. 1.
The product denial variable is represented by a dashed square box in Fig. 1. This dashed
box indicates that studies investigating the relation between advertising exposure and parent–
child conflict only implicitly assume that this relation is mediated by parental denial.
However, none of these studies have actually operationalized and/or assessed product denial
as a mediating variable.

1.1.3. Advertising makes children unhappy (path 3)

Finally, several authors argue that advertising may cause unhappiness in children. Most
of these authors have investigated a direct relation between exposure to commercials and
unhappiness, represented by path 3a (e.g., Martin & Kennedy, 1993; Richins, 1991). One
explanation for this direct relation stems from social-comparison theory. Advertising is
assumed to depict a world with beautiful people and desirable products. When children
watch television advertising and compare their own situation with this idealized world in
the commercials, the discrepancy between the two worlds might cause unhappiness
(Atkin, 1980; Richins, 1991).
Another explanation for the direct advertising–unhappiness relationship has to do with
the way in which products and brands are depicted in commercials. By using special
camera and editing techniques (e.g., close-ups, moving images of toys that are unable to
move by themselves), commercials can create unrealistic expectations regarding the
performance and quality of products. It is assumed that young children are not yet capable
of seeing these techniques as representations that are not entirely consistent with reality.
After the purchase, when the product does not meet the child’s expectations, the child
becomes frustrated, dissatisfied, and unhappy (Atkin, 1980; Robertson, Rossiter, & Ward,
1985; Ward, Wackman, & Wartella, 1977).
Finally, there are hypotheses about one or more mediated paths between advertising
exposure and unhappiness. A first mediated hypothesis is that advertising exposure (a)
leads to materialism, which (b) leads to unhappiness because materialistic people are
assumed to be less happy (path 3b). For instance, it has been argued that materialistic
people consider objects as an important means to gain happiness (Sirgy, 1998). When the
 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456 441

products fail to yield the promised state of happiness, disappointment, and unhappiness
will follow (Richins, 1991). A second mediated hypothesis states that parental denial of
children’s purchase requests causes unhappiness (path 3c). It is assumed that greater
exposure to advertising causes children to make more purchase requests to their parents.
When children do not receive the requested products, they may become disappointed,
dissatisfied, and hence, unhappy (Atkin, 1980; Goldberg & Gorn, 1978; Sheikh &
Moleski, 1977).

1.2. Hypotheses about developmental differences in unintended advertising effects

It is generally assumed that younger children are more susceptible to advertising

effects than older children are (Gunter & Furnham, 1998; Young, 1990). It has been
shown that children in early childhood (ages two to seven) are more vulnerable to
persuasive information, because they have less experience and domain-specific knowledge
that they can use while processing commercials (Roedder, 1981). They are, therefore, less
able to come up with critical thoughts and counterarguments while watching commercials
(Brucks, Armstrong, & Goldberg, 1988). As children mature, they develop greater
cognitive skills, which result in a better understanding of the persuasive intent of
commercials. Consequently, in middle childhood (ages 8 to 12), children become
progressively more critical about, and thereby less susceptible to advertising messages.
Although these advertising processing theories suggest that younger children are more
susceptible to unintended advertising effects than older children are, no specific
hypotheses about developmental differences in advertising-induced materialism and
unhappiness have been formulated. However, in the case of parent–child conflict,
theories of parent–child interaction do suggest that younger children more often come
into conflict with their parents about advertised products than older children do. Various
explanations can be provided to support this hypothesis. First, younger children have
been shown to make more advertising-induced purchase requests than older children do,
which increases the chance of parent–child conflict (Robertson & Rossiter, 1974; Ward
& Wackman, 1972; Ward et al., 1977). Second, younger children more often have
difficulty delaying gratifications than older children have (Metcalfe & Mischel, 1999). If
young children see some item as attractive, they focus all their attention on the enticing
aspects of this stimulus and find it very difficult to resist, which may also increase the
chance of parent–child conflict.
A final explanation for a stronger effect of advertising on parent–child conflict among
younger children is children’s growing ability to apply sophisticated persuasion techni-
ques on their parents. Research has shown that young children quite often ask, whine,
and show anger to persuade their parents. Older children, in contrast, tend to use more
sophisticated persuasion techniques, such as negotiation, flattery, and white lies (Williams
& Burns, 2000). Such sophisticated persuasion strategies have been shown to lead to less
parent–child conflict than the persuasion strategies of younger children do (Atkin, 1978;
Kuczynski, Kochanska, Radke-Yarrow, & Girnius-Brown, 1987; Mangleburg, 1990;
Williams & Burns, 2000).
442 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456

2. Empirical evidence: Materialism

A total of eight studies examined the direct relation between television advertising and
materialism in children (path 1 in Fig. 1): Five correlational studies (Atkin, 1975a, 1975b;
Churchill & Moschis, 1979; Moschis & Churchill, 1978; Ward & Wackman, 1971), one causal-
correlational study (Moschis & Moore, 1982), one quasi-experiment (Greenberg & Brand,
1993), and one laboratory experiment (Goldberg & Gorn, 1978). Five of these eight studies
were conducted among adolescents and three among children in early and middle childhood.

2.1. Definition and operationalization of materialism

Materialism is considered as a preoccupation with money and possessions (Belk, 1985;

Rossiter, 1980) and the idea that personal wealth and material possessions are the key to success
and well-being (Fournier & Richins, 1991). Most correlational studies on the relation between
advertising and materialism have used Ward and Wackman’s (1971) operationalization of
materialism (Churchill & Moschis, 1979; Moschis & Churchill, 1978; Moschis & Moore,
1982). In these studies, a 4- or 5-item scale was used with items such as ‘‘It is really true that
money can buy happiness,’’ and ‘‘My dream in life is to be able to own expensive things.’’ In
Goldberg and Gorn’s (1978) experiment, materialism was operationalized as a preference for
toys over friends.

2.2. Correlational studies

The specific characteristics (sample size, age of children, control variables, and moderator
variables) of the correlational studies on the relation between television advertising and
materialism are presented in Table 1. The two right-hand columns in Table 1 present the
direction and strength of the observed relations. The coefficients (r, pr, b) are included in the
table to give an indication of the effect sizes that have been found.
Four studies reported zero-order correlations (Atkin, 1975a, 1975b; Moschis & Churchill,
1978; Moschis & Moore, 1982). In each of these studies, the reported correlations were positive
and significant, ranging from r = .13 (Moschis & Churchill, 1978) to r = .32 (Moschis & Moore,
1982). In three out of the four studies, the relation between viewing frequency and materialism
was still significant when controls were added for age and sex, socioeconomic status, school
performance, ethnicity, parent and peer communication about consumption, and birth order.
The significant correlation observed by Moschis and Churchill (1978) disappeared when they
simultaneously controlled for nine variables, including family communication about con-
sumption and motivations for watching commercials. Similarly, Ward and Wackman (1971) did
not find a significant correlation between television exposure and materialism when controlling
for motivation for watching advertising. However, it is questionable whether exposure and
motivation variables should be included simultaneously in a regression analysis. It is possible
that motivation measures are correlated highly with exposure measures. Highly correlated
predictors can lead to serious estimation problems and spurious relations (Lewis-Beck, 1980;
Stevens, 1996), and hence may lead to questionable results.
 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456 443

2.3. Quasi-experimental and experimental studies

Two quasi-experimental studies investigated the relation between advertising exposure and
materialism. Greenberg and Brand (1993) studied 782 adolescents and examined differences
in materialism between teenagers enrolled in schools receiving the commercial school
program Channel One and teenagers attending schools without this program. Teenagers
attending schools with Channel One were more materialistic than teenagers from schools
without the program, r = .11.1
Goldberg and Gorn (1978) exposed 231 four- to five-year-old children to a 10-min
preschool program. The children were randomly assigned to three groups. The first group
watched a program into which two identical toy commercials were edited. A second group
was exposed to the same program and commercials on two successive days. A control
group saw the program without the commercials. After the television program, all children
were shown two photographs. The first photograph showed a boy holding the advertised
toy. The second one pictured another boy, who was not holding a toy. The children were
told that the boy who was holding the advertised toy was a ‘‘not-so-nice boy.’’ The
children were then asked whether they would like to play with the not-so-nice boy with
the toy or with the nice boy without the toy. More than half of the children who were
exposed to the commercials chose to play with the not-so-nice boy with the advertised toy,
whereas in the control group only 30% of the children chose to play with the not-so-nice
boy with the toy.
In the same experiment, Goldberg and Gorn (1978) used a second operationalization of
materialism. The children were asked whether they would like to play with the advertised
toy or with their friends in the sandbox. More than half of the children who had watched
the commercials preferred the toy over their friends, whereas only 30% of the children in
the control group preferred the toy. However, it is doubtful whether this experiment has
really demonstrated that advertising makes children materialistic. It is also conceivable that
the attractive toy commercial temporarily aroused children’s curiosity, and that it is
curiosity (or some other factor) and not materialism that caused the children to choose a
toy over a friend.

2.4. Moderator variables

With the exception of Atkin (1975b), the majority of studies that investigated
developmental differences in the effect of advertising on materialism did not yield
significant results (Atkin, 1975a; Moschis & Moore, 1982; Ward & Wackman, 1971; see
Table 1). These studies also did not find a moderating effect for sex, SES, or peer
communication.

1
Greenberg and Brand (1993) have reported t values for the differences between the two experimental groups
(t [825] = 3.1, p < .005). The correlation size reported in the present article was computed with the formula
r = M (t2/t2 + df ); see Rosenthal (1991).
444 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456

Table 1
Correlational studies on the relation between advertising exposure and materialism
Study Age in N Measures Control Moderator Direction Strength
years of exposure variables variables of relation of relationa
Ward and 13 – 15; 1094 TV SES, IQ, 0 Age 0 (TV b= .01
Wackman 15 – 18 viewing money, family viewing ns
(1971) frequency communication, frequency)
motivations Younger b= .07
for watching children ns
commercials, Older children
magazine
readership
Atkin 9 – 11; 775 TV None and 0 Age + (TV r = .24
(1975a) 11 – 13 viewing age, sex, 0 Sex viewing pr = .18
frequency SES, school 0 SES frequency)
performance
Advertising 0 Age + (Advertising r = .22
viewing 0 Sex viewing pr = .13
frequency 0 SES frequency)
Atkin 4 – 6; 738 Saturday None and age, + Age + (Saturday r = .25
(1975b) 6 – 9; morning sex, ethnicity, + Sex morning pr = .10
9 – 12 viewing school viewing
frequency performance frequency)
+ (Only for
boys and 6+
children)
Moschis and 12 – 18 806 TV None and age, None + (TV viewing r = .13
Churchill viewing sex, SES, family frequency)
(1978) frequency communication, 0 (TV viewing b = .07
peer frequency) ns
communication,
motivations for
watching
commercials and
TV programs,
newspaper
readership,
consumer
education
Churchill and 12 – 18 806 TV Age, sex, SES, None + (TV viewing b = .11
Moschis viewing family frequency)
(1979) frequency communication,
peer
communication,
birth order
Moschis and 12 – 15; 211 Advertising None 0 Age Cross-sectional r = .32
Moore 15 – 18 viewing results:
(1982) frequency + (Advertising
viewing
frequency)
 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456 445

Table 1 (continued)
Study Age in N Measures Control Moderator Direction Strength
years of exposure variables variables of relation of relationa
Idem Idem Idem Idem Age, sex, SES, 0 Age Longitudinal pr = .24
ethnicity, family + Materialism results: pr = .22
communication, at year 0 + (Advertising
peer + Family viewing
communication, communication frequency;
materialism year 0. 0 Peer only for
communication adolescents
initially low
on materialism)
+ (Advertising
viewing
frequency;
only for
adolescents
whose parents
do not discuss
consumption
matters)
a
r = product-moment correlation coefficient; pr = partial correlation coefficient, b = standardized multiple
correlation coefficient.

In the causal-correlational study of adolescents by Moschis and Moore (1982), two

moderator effects were found: The effect for advertising on materialism held only for (a)
adolescents who initially scored low on materialism, and (b) adolescents from families
who do not discuss consumption matters. The first moderator effect was probably due to
a ceiling effect. After all, when children already score high on materialism at the outset of
the study, it is less likely that advertising can still contribute to this materialistic attitude.
The second moderator effect could be attributed to the mediating role of parents. A series
of studies have shown that parents can reduce or even counteract negative television
effects (Nathanson, 1999; Valkenburg, Krcmar, Peeters, & Marseille, 1999). Moschis and
Moore’s (1982) findings demonstrate that this also holds true for advertising effects on
materialism.

3. Empirical evidence: Parent–child conflict

Eleven studies have examined one or more causal paths linking advertising to parent–
child conflict (see path 2, 2a, 2b, and 2c in Fig. 1): Nine correlational studies (Atkin,
1975a, 1975b; Buijzen & Valkenburg, 2000; Galst & White, 1976; Isler, Popper, & Ward,
1987; Robertson & Rossiter, 1976, 1977; Robertson, Ward, Gatignon, & Klees, 1989;
Ward & Wackman, 1972) and two experiments (Goldberg & Gorn, 1978; Stoneman &
446 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456

Brody, 1981).2 All studies on the relation between advertising and parent–child conflict
were conducted among children 12 years old or younger.

3.1. Definition and operationalization of parent–child conflict

Parent–child conflict has been operationalized in two ways. In correlational research,

children or parents have been asked to indicate how often they had a conflict with the other
party about product purchases. In experimental research, diary studies, and observational
research, parent–child conflict has been indicated by children’s negative reactions to parental
denial of a purchase request.

3.2. Correlational studies

Eight studies investigated the relationship between advertising exposure and purchase
requests (path 2a; Atkin, 1975a, 1975b; Buijzen & Valkenburg, 2000; Galst & White, 1976;
Isler et al., 1987; Robertson & Rossiter, 1976, 1977; Robertson et al., 1989), four studies the
relation between purchase requests and parent–child conflict (path 2b/c; Atkin, 1975a, 1975b;
Robertson et al., 1989; Ward & Wackman, 1972), and two studies examined the direct
relation between exposure to advertising and parent–child conflict (path 2; Atkin, 1975a,
1975b).
The specific characteristics (sample size, age of children, control variables, and moderator
variables) of the correlational studies on the relations between television advertising and
parent–child conflict are presented in Table 2. The three right-hand columns of the table
represent the coefficients of the relations between exposure and purchase requests (path 2a),
between purchase requests and parent–child conflict (path 2b/c), and between advertising
exposure and parent–child conflict (path 2), respectively.
The correlational studies between advertising exposure and purchase requests show
without exception that advertising exposure is positively related to children’s purchase
requests. The zero-order correlations vary from r = .18 (Isler et al., 1987) to r = .41 (Atkin,
1975b). In the majority of studies the zero-order correlations were around r = .30.
The correlations between purchase requests and parent–child conflict (path 2b/c) were
investigated most often when possible third variables were controlled. These studies
yielded partial correlations varying from pr = .25 (Atkin, 1975b) to pr = .72 (Robertson
et al., 1989). Finally, the two studies that investigated the direct relation (path 2)
between advertising exposure and parent–child conflict reported relations of r = .08 and
r = .20.

2
Many studies refer to Sheikh and Moleski (1977) to show that advertising enhances parent – child conflict. In
our opinion, this study does not provide evidence that advertising causes conflict in the family, because (a) no
relation has been reported between exposure to advertising and conflict, and (b) no experimental comparison has
been made between children who were or were not exposed to advertising. Therefore, the study conducted by
Sheikh and Moleski (1977) has not been included in this portion of our review.
 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456 447

Table 2
Correlational studies on the relation between advertising exposure and parent – child conflict
Study Age in N Control Moderator Investigated relationa
years variables variables Advertising Purchase Advertising
exposure – requests – exposure –
purchase parent – child parent – child
requests conflict conflict
(path 2a) (path 2b/c) (path 2)
Atkin 9 – 11; 506 None and Age r = .32 pr = .46 r = .20
(1975a) 11 – 13 age, sex, Sex pr = .27 pr = .13
SES, school SES No moderator Moderator
performance effects effect sex:
boys r = .23,
girls r = .03
Atkin 4 – 6; 738 None and Age r = .41 pr = .25 r = .08
(1975b) 6 – 9; age, sex, Sex pr = .29 pr = .10
9 – 12 ethnicity, No moderator Moderator
school effects effect sex:
performance boys r = .14,
girls r = .07
Moderator
effect age:
4 – 6 years
pr = .02,
6 – 9 years
pr = .11,
9 – 11 years
pr = .28
Robertson 6 – 11 289 None None r = .24 – –
and Rossiter
(1976)
Galst and 4–6 41 None None r = .31 – –
White (1976)
Isler et al. 3 – 11 250 None None r = .18 – –
(1987)
Buijzen and 7 – 12 250 None and None r = .31, – –
Valkenburg age, sex pr = .29
(2000)
Robertson 3 – 10 267 Age, None b = .25 b = .72 –
et al. (1989) demanding
behavior,
communicating
behavior,
independent
behavior
Ward and 5 – 12 109 None None – r = .18 –
Wackman
(1972)
a
r = product-moment correlation coefficient; pr = partial correlation coefficient, b = standardized multiple
correlation coefficient.
448 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456

3.3. Experimental studies

Two experiments investigated the influence of advertising exposure on purchase

requests (path 2a; Stoneman & Brody, 1981) and parent–child conflict (path 2; Goldberg
& Gorn, 1978). Stoneman and Brody (1981) demonstrated that exposure to advertising
increased the number of children’s purchase requests. This study examined the
interaction between 36 mothers and their 3- to 5-year-old preschoolers in the store.
First, the mothers and children were shown a cartoon. Half of the mothers and children
saw just the cartoon, while the other half saw the cartoon with six commercials for
candy and chips edited into it. After exposure to the cartoon, the mothers went shopping
with their children. The children who had seen the candy and toy commercials made
approximately 50% more purchase requests than children who had not seen the
commercial. Moreover, children who had been exposed to the commercials asked for
the advertised products twice as often (M = 4.4) as the children who had not been
exposed (M = 1.9).
Goldberg and Gorn (1978) investigated the effect of advertising on parent–child conflict.
They showed 231 four- and five-year-olds a television program, either with or without
commercials. After exposure to the program, children were asked which product they
preferred, a tennis ball or the advertised toy. They were also told that their mother preferred
the tennis ball over the advertised toy. Children who had been exposed to the commercial
were significantly more often opposed to their mother’s wish (45.8%) than children who had
not seen the commercial (21.3%).
In the same experiment, Goldberg and Gorn (1978) used a second operationalization of
conflict. After exposure to the preschoolers’ program described previously, children were
told a story about a boy who asked his father for the advertised toy, but did not receive it.
Then the children were presented with pictures representing two hypothetical responses of
the boy. The first picture showed a child happily hugging his father, while the second
showed a child glumly walking away from his father. The children were asked how they
thought the boy in the story would react. The results showed that children who had seen
the commercial more often chose the picture of the boy walking away from his father,
although the differences between the two experimental groups were not statistically
significant.

3.4. Moderator variables

The moderating influence of age has been addressed in the correlational studies by Atkin
(1975a, 1975b. No moderating effect was found in the first study (Atkin, 1975a). In the
second study, the correlation between advertising exposure and parent–child conflict applied
only to 6- to 9-year-olds and 9- to 11-year-olds, and not to 4- to 6-year-olds. This greater
advertising effect on parent–child conflict for older rather than younger children was
surprising. After all, research on parent–child interaction has often found that younger
children have difficulty delaying gratifications relatively often (Metcalfe & Mischel, 1999),
and are less able to apply sophisticated persuasion strategies than older children (Kuczynski
 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456 449

et al., 1987), which are both characteristics that easily result in parent–child conflict
(McNeal, 1992).
The moderating effect of gender has also been established in two correlational studies
(Atkin, 1975a, 1975b). In both studies, the effect of advertising on parent–child conflict held
only for boys and not for girls. This moderating influence is consistent with general theories
on gender differences in parent–child interactions (e.g., Maccoby, 1990). It has been shown
that boys are on average less compliant than girls to the requests and demands of their parents
(Cowan & Avants, 1988). Boys also rely on forceful or demanding strategies more often
when trying to persuade their parents to comply with them, whereas girls are more likely to
rely on tact and polite suggestions (Cowan & Avants, 1988). These gender differences can
explain why the relation between advertising exposure and parent–child conflict holds
mainly for boys.

4. Empirical evidence: Unhappiness

4.1. Definition and operationalization of unhappiness

Unhappiness can occur as a short-term, temporary emotional state, but it can also be a
longer term, enduring state of mind. Long-term unhappiness has been investigated in studies
on the direct effect of advertising (path 3a), and in studies on the indirect effect of advertising,
mediated through materialism (path 3b). In these two types of research, unhappiness has been
conceptualized as dissatisfaction with one’s life situation or with oneself (e.g., Richins, 1991).
Short-term unhappiness has been investigated in studies into the consequences of parental
denial of children’s purchase requests (path 3c). In these studies, unhappiness has been
operationalized as crying and being sad or disappointed (Atkin, 1975b; Goldberg & Gorn,
1978).

4.2. Evidence of path 3a: The direct relation between advertising exposure and unhappiness

Two types of studies have addressed the direct relation between exposure to advertising
and unhappiness. The first type investigated whether people become unhappy because they
compare themselves unfavorably with the beautiful world of advertising. This type of
research has not been conducted with children in early or middle childhood. The existing
research has focused on the impact of beautiful models in print advertising on the self-
perceptions of female adolescents (Durkin & Paxton, 2002; Martin & Kennedy, 1993; Stice,
Spangler, & Agras, 2001), female graduate students (Richins, 1991), and male graduate
students (Gulas & McKeage, 2000).
These studies have produced mixed results. Three experiments (Durkin & Paxton, 2002;
Gulas & McKeage, 2000; Richins, 1991) found that exposure to print advertisements
containing idealized images of physical attractiveness and financial success negatively
affected males’ and females’ self-perceptions. These findings were corroborated in a
longitudinal experiment among 13- to 17-year-old girls (Stice et al., 2001), but the
450 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456

advertising exposure effect held only for girls that initially scored high on perceived
pressure to be thin and body dissatisfaction. However, in an experiment among 9- to 18-
year-old girls, Martin and Kennedy (1993) did not find an effect for advertising on the self-
perception of physical attractiveness. All studies were restricted to print advertising;
therefore, the question whether television advertising negatively affects children’s or
teenagers’ self-perception is still open.
The second type of research on the direct relation between advertising and unhappiness has
concentrated on the influence of commercials on disappointment with the advertised product
after purchase. Only one correlational study has investigated this relation (Robertson et al.,
1985). In this survey, among 253 six- to eleven-year-old boys, a small though significant
negative relation was found between exposure to advertising and satisfaction with the product
(r = .06). However, this finding applied only to the youngest children in the sample.

4.3. Evidence of path 3b: Does materialism lead to unhappiness?

One of the hypotheses modeled in Fig. 1 states that materialism mediates the relation
between advertising and unhappiness. Although there is empirical evidence to support the
relation between advertising exposure and materialism (path 1), the path between materialism
and unhappiness has never been investigated among children. There is, however, evidence for
this relation among adults. A meta-analysis on the relation between materialism and life
satisfaction has yielded a negative relation of r = .25 (Wright & Larsen, 1993), suggesting
that adults who are more materialistic are unhappier than adults who are less materialistic.
Although it is conceivable that this finding also applies to children, no empirical evidence
exists to confirm this proposition.

4.4. Evidence of path 3c: Does product denial lead to unhappiness?

Two correlational studies have investigated the relation between purchase requests and
unhappiness. Atkin (1975b) asked 211 mothers of 3- to 11-year-old children how often
their children became unhappy after they had denied a purchase request. He found a
significant positive correlation between advertising and unhappiness after denial of
requests for cereals (r = .14) and toys (r = .20). Both correlations remained positive
when controlling for age, sex, ethnicity, and school performance. A survey by Robertson
et al. (1985) among 253 six- to eleven-year-old boys yielded a smaller, though still
significant, correlation between advertising exposure and disappointment after denial of a
purchase request (r = .08).
In Goldberg and Gorn’s (1978) experiment, 4- and 5-year-olds were exposed to a
preschoolers’ television program. One group of children watched the program with
commercials and a second group watched the same program without commercials. After
the program, the children were asked how they thought a hypothetical child would react
when a request for the advertised product had been denied. More than half of the children
who had been exposed to the commercials thought that the child would be sad, compared to
35% of the control group.
 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456 451

5. Discussion

The main aim of this vote-counting analysis was to evaluate the current state of the
empirical research on three unintended advertising effects among children: materialism,
parent – child conflict, and unhappiness. The impact of advertising on children is a
controversial issue, which often provokes contradictory opinions. A systematic and
integrative review of the empirical research on this topic is important, because such a
review enables us to determine the extent to which the prevailing claims and assumptions
are supported by empirical research. In this discussion, we will first evaluate the validity of
each of the different hypotheses. In the final section, we will make some suggestions for
further research.

5.1. The validity of the hypotheses

5.1.1. Advertising enhances materialism

The majority of the correlational studies presented in Table 1 showed coefficients that
Cohen (1988) would identify as small to moderate effect sizes. Relations between advertising
exposure and materialism were established among children in early and middle childhood, as
well as adolescents. Most of these studies still showed a positive relation between advertising
and materialism when the influence of possible third variables, such as age, sex, socio-
economic status, and school performance was controlled.
These correlational findings were confirmed by the experimental studies by Green-
berg and Brand (1993) and Goldberg and Gorn (1978). Finally, the causal-correlational
study conducted by Moschis and Moore (1982) suggests that the influence of ad-
vertising applies only to children who initially scored low on materialism and children
from families who do not discuss consumption matters. However, the causal-correla-
tional and experimental studies only investigated adolescents and children in early
childhood. There is as yet no evidence for a causal relation among children in middle
childhood.
Although most research is correlational and does not permit causal conclusions, the
weight of evidence is in favor of the hypothesis that advertising makes children
materialistic. The causal-correlational and experimental studies that have been conducted
are in support of the hypothesis that the causal direction points from advertising to
materialism and not the other way around. Therefore, although more research is needed to
arrive at decisive conclusions, our vote-counting analysis clearly suggests that exposure to
television advertising stimulates materialistic values in children.

5.1.2. Advertising causes parent–child conflict

All correlational studies on one or more paths between advertising and parent–child
conflict yielded small to moderate positive correlations. These correlational results were
confirmed partially by the experiment by Goldberg and Gorn (1978). The two studies that
have investigated moderator variables suggest that the relation between advertising
exposure and parent–child conflict holds only for children older than six and for boys.
452 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456

However, because of the paucity of studies that have investigated moderator variables,
these results can, at most, be regarded as an indication for future research.
The results of the correlational and experimental studies are in agreement with the
hypothesis that advertising enhances parent–child conflict. However, the majority of the
studies are correlational. The sole experiment that has been conducted focused exclusively on
children in early childhood, and resulted in only a partial confirmation of the hypothesis.
Therefore, the scientifically conservative conclusion is that advertising enhances parent–child
conflict, but that decisive evidence is still lacking.

5.1.3. Advertising makes children unhappy

Hypotheses on the relation between advertising and unhappiness show that there are
four ways by which exposure to advertising can lead to unhappiness: (1) via social
comparison, (2) via materialism, (3) by leading to disappointment about the quality of
the advertised product, and (4) by leading to sadness after parental denial of purchase
requests.
With regard to the first two hypotheses, there is either a lack of studies that
investigate the hypothesis among children (in the case of materialism), or the empirical
evidence does not apply to television advertising (in the case of social comparison).
However, there is some evidence to support the latter two hypotheses. The only study
on the effect of advertising on unhappiness through disappointment about the advertised
product yielded a small positive relation. The three studies on the effects of advertising
on unhappiness caused by denial of purchase requests suggest a small to moderate effect
size.
In summary, the empirical research on the advertising–unhappiness relationship is not
developed to such an extent that it allows of causal conclusions. Our analysis does
suggest, however, that each of the four hypotheses on the advertising–unhappiness
relationship that we distinguished can be used as a viable working hypothesis in future
research.

5.2. Suggestions for further research

Traditionally, most studies on the unintended effects of advertising have focused on the
direct relation between the independent variable (advertising exposure) and one of the three
dependent variables (materialism, parent–child conflict, and unhappiness). None of these
studies has explored the mechanisms underlying the relations between advertising exposure
and each of the three dependent variables. Most research investigating the unintended
advertising effects is characterized by an input–output orientation, which has neglected the
specific ways in which advertising can lead to materialism, parent–child conflict, and
unhappiness. An important aim of our review was to reveal some of these underlying
mechanisms (or mediating variables).
Our vote-counting review has yielded a number of working hypotheses for future research,
and it identified several moderator variables, such as age, gender, socioeconomic status, and
parent–child communication. In most empirical studies, these moderator variables were
 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456 453

included on an exploratory basis, without guidance of specific theories on the nature and
direction of the moderating influence of these variables. In our view, a choice of moderator
variables can only be valuable when a priori hypotheses are formulated about how and why
certain moderators affect the relation between advertising exposure and materialism, parent–
child conflict, and unhappiness. Although it has not been decisively established how the
moderator variables identified in this review influence advertising effects, they may serve as a
basis or guideline for future research.
Although further research is needed to arrive at decisive conclusions, our review clearly
indicates that advertising aimed at children can have certain undesirable consequences for
the family, such as increased parent–child conflict, and materialistic and disappointed
children. Even though most Western countries have tightened their rules and regulations on
child-directed advertising, the lion’s share of the responsibility of dealing with negative
advertising effects is still shouldered by parents, who are usually the first to experience
inconvenience as a result of advertising.
This review study suggests that parents are able to counteract undesired effects of
advertising by interacting with their children about advertising and consumer matters.
However, it is advisable that those working in applied fields, such as marketers and
advertisers, also take on part of the responsibility of mitigating negative effects. Responsible
advertising aimed at children does not only serve an ethical purpose, but also a commercial
one. After all, it is progressively acknowledged among child marketers that parental approval
is a necessary condition for successful marketing and advertising towards children (Del
Vecchio, 1998; Guber & Berry, 1993; McNeal, 1992, 1999). Parents are eventually the ones
who will select and buy the advertised products for their children. It is no surprise, therefore,
that the most successful children’s products in the past decades were those that met the
approval of parents (McNeal, 1999).
An important way to avoid or reduce negative advertising effects is to prevent advertising
being misleading to children. One way to accomplish this is to test a new commercial among
children in order to examine whether the product is portrayed accurately and whether the
benefits communicated are not overpromised. Another way is to test the commercial among
parents to get their opinions about a certain advertising strategy or tactic. Not only can
parents indicate how other parents will react to the commercial, but they also have a keen
sense of what their children appreciate and understand (McNeal, 1992). In sum, testing
marketing and advertising efforts among children and parents can help applied professionals
to find ways of making child-directed advertising as effective, responsible, and publicly
acceptable as possible.

References

References marked with an asterisk indicate studies included in the vote-counting analysis

*Atkin, C. K. (1975a). Survey of pre-adolescent’s responses to television commercials: The effects of television
advertising on children (Report No. 6). East Lansing, MI: Michigan State University. ERIC Document
Reproduction Service No. ED116820.
454 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456

*Atkin, C. K. (1975b). Survey of children’s and mothers’ responses to television commercials. The effects of
television advertising on children (Report No. 8). East Lansing, MI: Michigan State University. ERIC Docu-
ment Reproduction Service No. ED123675.
Atkin, C. K. (1978). Observation of parent – child interaction in supermarket decision-making. Journal of Market-
ing, 42, 41 – 45.
Atkin, C. K. (1980). Effects of television advertising on children. In E. L. Palmer, & A. Dorr (Eds.), Children and
the faces of television: Teaching, violence, selling ( pp. 287 – 305). New York: Academic Press.
Belk, R. W. (1985). Materialism: Trait aspects of living in the material world. Journal of Consumer Research, 12,
265 – 280.
Brucks, M., Armstrong, G. M., & Goldberg, M. E. (1988). Children’s use of cognitive defenses against television
advertising: A cognitive response approach. Journal of Consumer Research, 14, 471 – 482.
*Buijzen, M., & Valkenburg, P. M. (2000). The impact of television advertising on children’s Christmas wishes.
Journal of Broadcasting & Electronic Media, 44, 456 – 470.
Bushman, B. J. (1994). Vote-counting procedures in meta-analysis. In H. M. Cooper, & L. V. Hedges (Eds.), The
handbook of research synthesis ( pp. 193 – 213). New York: Russell Sage Foundation.
*Churchill, G. A., & Moschis, G. P. (1979). Television and interpersonal influences on adolescent consumer
learning. Journal of Consumer Research, 6, 23 – 35.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Cooper, H. M. (1990). Integrating research: A guide for literature reviews (2nd ed.). Newbury Park, CA: Sage.
Cowan, G., & Avants, S. K. (1988). Children’s influence strategies: Structure, sex differences, and bilateral
mother – child influence. Child Development, 59, 1303 – 1313.
Davies, P. (2000). The relevance of systematic reviews to educational policy and practice. Oxford Review of
Education, 26, 365 – 378.
Del Vecchio, G. (1998). Creating ever-cool: A marketer’s guide to a kid’s heart. Gretna, LA: Pelican Publishing.
*Durkin, S. J., & Paxton, S. J. (2002). Predictors of vulnerability to reduced body image dissatisfaction and
psychological well being in response to exposure to idealized female images in adolescent girls. Journal of
Psychosomatic Research, 53, 995 – 1005.
Fournier, S., & Richins, M. L. (1991). Some theoretical and popular notions concerning materialism. Journal of
Social Behavior and Personality, 6, 141 – 403.
*Galst, J., & White, M. (1976). The unhealthy persuader: The reinforcing value of television and children’s
purchase—influencing attempts at the supermarket. Child Development, 47, 1089 – 1096.
*Goldberg, M. E., & Gorn, G. J. (1978). Some unintended consequences of TV advertising to children. Journal of
Consumer Research, 5(1), 22 – 29.
*Greenberg, B. S., & Brand, J. E. (1993). Television news and advertising in schools: The ‘‘Channel One’’
controversy. Journal of Communication, 43(1), 143 – 151.
Guber, S. S., & Berry, J. (1993). Marketing to and through kids. New York: McGraw-Hill.
*Gulas, C. S., & McKeage, K. (2000). Extending social comparison: An examination of the unintended con-
sequences of idealized advertising imagery. Journal of Advertising, 29(2), 17 – 28.
Gunter, B., & Furnham, A. (1998). Children as consumers: A psychological analysis of the young people’s market.
London: Routledge.
*Isler, L., Popper, E. T., & Ward, S. (1987). Children’s purchase requests and parental responses: Results from a
diary study. Journal of Advertising Research, 27(5), 29 – 39.
Johnson, B. T., & Eagly, A. H. (2000). Quantitative synthesis of social psychological research. In H. T. Reis, &
C. M. Judd (Eds.), Handbook of research methods in social psychology ( pp. 496 – 528). London: Cambridge
University Press.
Kuczynski, L., Kochanska, G., Radke-Yarrow, M., & Girnius-Brown, O. (1987). A developmental interpretation
of young children’s noncompliance. Developmental Psychology, 23, 799 – 806.
Kunkel, D. (2001). Children and television advertising. In D. Singer, & J. Singer (Eds.), Handbook of children and
the media ( pp. 375 – 394). Thousand Oaks, CA: Sage.
Lewis-Beck, M. S. (1980). Applied regression: An introduction. Newbury Park, CA: Sage.
 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456 455

Liebert, R. M. (1986). Effects of television on children and adolescents. Journal of Developmental and Behavioral
Pediatrics, 7, 43 – 48.
Maccoby, E. E. (1990). Gender and relationships. American Psychologist, 45, 513 – 520.
Mangleburg, T. F. (1990). Children’s influence in purchase decisions: A review and critique. Advances in Con-
sumer Research, 17, 813 – 825.
*Martin, M. C., & Kennedy, P. F. (1993). Advertising and social consequences for female preadolescents and
adolescents. Psychology & Marketing, 10, 513 – 530.
McNeal, J. U. (1992). Kids as customers: A handbook of marketing to children. New York: Lexington Books.
McNeal, J. U. (1999). The kids market: Myths and realities. New York: Paramount Market Publishing.
Metcalfe, J., & Mischel, W. (1999). A hot/cool-system analysis of delay of gratification: Dynamics of willpower.
Psychological Review, 106, 3 – 19.
*Moschis, G. P., & Churchill, G. A. (1978). Consumer socialization: A theoretical and empirical analysis. Journal
of Marketing Research, 15, 599 – 609.
*Moschis, G. P., & Moore, R. L. (1982). A longitudinal study of television advertising effects. Journal of
Consumer Research, 9, 279 – 286.
Nathanson, A. I. (1999). Identifying the relationship between parental mediation and children’s aggression.
Communication Research, 26, 124 – 144.
Pollay, R. W. (1986). The distorted mirror: Reflections on the unintended consequences of advertising. Journal of
Marketing, 50(2), 18 – 36.
*Richins, M. L. (1991). Social comparison and the idealized images of advertising. Journal of Consumer Re-
search, 18, 71 – 83.
Robertson, T. S. (1979). Parental mediation of television advertising effects. Journal of Communication, 29(1),
12 – 25.
Robertson, T. S., & Rossiter, J. R. (1974). Children and commercial persuasion: An attribution theory analysis.
Journal of Consumer Research, 1, 13 – 20.
*Robertson, T. S., & Rossiter, J. R. (1976). Short-run advertising effects on children: A field study. Journal of
Marketing Research, 13, 68 – 70.
*Robertson, T. S., & Rossiter, J. R. (1977). Children’s responsiveness to commercials. Journal of Communication,
27(1), 101 – 106.
*Robertson, T. S., Rossiter, J. R., & Ward, S. (1985). Consumer satisfaction among children. Advances in
Consumer Research, 12, 279 – 284.
*Robertson, T. S., Ward, S., Gatignon, H., & Klees, D. M. (1989). Advertising and children: A cross-cultural
study. Communication Research, 16, 459 – 485.
Roedder, D. L. (1981). Age differences in children’s responses to television advertising: An information-process-
ing approach. Journal of Consumer Research, 8(2), 144 – 153.
Rosenthal, R. (1991). Meta-analytic procedures for social research. Newbury Park, CA: Sage.
Rossiter, J. R. (1980). The effects of volume and repetition of television commercials. In R. P. Adler, G. S. Lesser,
L. K. Meringoff, T. S. Robertson, J. R. Rossiter, & S. Ward (Eds.), The effects of television advertising on
children ( pp. 153 – 183). Lexington, MA: Lexington Books.
Sheikh, A. A., & Moleski, L. M. (1977). Conflict in the family over commercials. Journal of Communication,
27(1), 152 – 157.
Sirgy, M. J. (1998). Materialism and quality of life. Social Indicators Research, 43, 227 – 260.
Stevens, J. (1996). Applied multivariate statistics for the social sciences (3rd ed.). Mahwah, NJ: Erlbaum.
*Stice, E., Spangler, D., & Agras, W. S. (2001). Exposure to media-portrayed thin-ideal images adversely affects
vulnerable girls: A longitudinal experiment. Journal of Social and Clinical Psychology, 20, 270 – 288.
*Stoneman, Z., & Brody, G. H. (1981). The indirect impact of child-oriented advertisements on mother – child
interactions. Journal of Applied Developmental Psychology, 2, 369 – 376.
Valkenburg, P. M., Krcmar, M., Peeters, A. L., & Marseille, N. M. (1999). Developing a scale to assess three styles
of television mediation: ‘‘Instructive mediation,’’ ‘‘restrictive mediation,’’ and ‘‘social coviewing’’. Journal of
Broadcasting & Electronic Media, 43, 52 – 66.
456 M. Buijzen, P.M. Valkenburg / Applied Developmental Psychology 24 (2003) 437–456

*Ward, S., & Wackman, D. (1971). Family and media influences on adolescent consumer learning. American
Behavioral Scientist, 14, 415 – 427.
*Ward, S., & Wackman, D. (1972). Television advertising and intrafamily influence: Children’s purchase attempts
and parental yielding. In E. A. Rubinstein, G. A. Comstock, & J. P. Murray (Eds.), Television in day-to-day
life: Patterns of use. Television and social behavior, vol. 4 ( pp. 516 – 525). Washington, DC: U.S. Department
of Health, Education, and Welfare.
Ward, S., Wackman, D. B., & Wartella, E. (1977). How children learn to buy: The development of consumer-
processing skills. Beverly Hills, CA: Sage.
Williams, L. A., & Burns, A. C. (2000). Exploring the dimensionality of children’s direct influence attempts.
Advances in Consumer Research, 27, 64 – 71.
Wright, N. D., & Larsen, V. (1993). Materialism and life satisfaction: A meta-analysis. Journal of Consumer
Satisfaction, Dissatisfaction, and Complaining Behavior, 6, 158 – 165.
Wulfemeyer, K. T., & Mueller, B. (1992). Channel One and commercials in classrooms: Advertising content
aimed at students. Journalism Quarterly, 69, 724 – 742.
Young, B. M. (1990). Television advertising and children. Oxford, UK: Clarendon Press.