Journal of Educational Psychology Copyright 1987 by the American Psychological Association, Inc.

1987. Vol. [Link]. 2. i?9-!88 9O22-O663/87/500.75

Cultural Variations in Family Beliefs About Children's

Performance in Mathematics: Comparisons Among
People's Republic of China, Chinese-American, and
Caucasian-American Families
Robert D. Hess Chang Chih-Mei
Stanford University Palo Alto, California

Teresa M. McDevitt
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

University of Northern Colorado

This document is copyrighted by the American Psychological Association or one of its allied publishers.

We examined beliefs about children's performance in mathematics through interviews with

mothers and their sixth-grade children in the People's Republic of China (PRC) and in Chinese-
American and Caucasian-American groups in the United States. Explanations for relatively high
and low performance were indicated by attributions to ability, effort, training at home, training
at school, and luck. We also asked mothers about specific instances of unusually high or low
achievement. The groups showed different patterns of attributions. Mothers in the PRC viewed
lack of effort as the major cause of low performance. The Chinese Americans also viewed lack
of effort as important but assigned considerable responsibility to other sources. The Caucasian
Americans distributed responsibility more evenly across the options. PRC mothers offered partial
reinforcement to children who brought home a good grade; American mothers, both Chinese
and Caucasian, were Hkely to offer unqualified praise. National differences in performance may
occur in part because of such cultural variations in beliefs.

Students from the United States do not perform as well on members' beliefs about children's performance in mathemat-
tests of mathematics and science as do students from Asian ics and mothers' reports of their behavioral responses to
countries, particularly Japan, the People's Republic of China, specific occasions of high and low performance by their
and Taiwan (Husen, 1967; International Association for the children. Although all the institutions of a society express its
Evaluation of Educational Achievement, 1985; Stanley, cultural features, the family has a unique role in representing
Huang, & Zu, 1986; Stigler, Lee, Lucker, & Stevenson, 1982; culture to the young and thus provides a useful point for
Wong, 1980). Although reports of such national differences national comparison. The family members participating in
have been available for almost 20 years, knowledge of the this study were mothers and children from three cultural
factors responsible for these discrepancies is still incomplete. groups—Chinese families from the People's Republic of
National differences have been attributed to such divergences China (PRC) and Chinese-American and Caucasian-Ameri-
in national educational systems as the time students spend in can families from the San Francisco Bay area.
school (Kirst, 1981; Stigler etal., 1982), time on task while in
the classroom (Stigler et al., 1982), and comprehensiveness of
the mathematics curriculum (Stigler et al., 1982). Asian- Role of Family Beliefs in Achievement in School
American students in the United States, however, also per-
form at higher levels than Caucasian students on tests of The logic of this study assumes that family beliefs about
academic achievement and other measures of educational achievement influence both parents' behavior and children's
attainment (California State Department of Education As- performance. A considerable body of research and theory
sessment Program, 1981; College Board, 1982; Vernon, 1982). supports this view. Studies of family characteristics and aca-
This suggests that national differences are not solely due to demic attainment show parents* attitudes toward achievement
differences in educational systems but may also reflect cultur- to be related to children's performance in Western cultures
ally transmitted values, beliefs, and behaviors. (Hess & Holloway, 1984; Marjoribanks, 1980) and in China
(Huang & Wen, 1984). Presumably, these associations occur
In this article, we examine cultural differences in motiva-
because parents' beliefs about the importance of achievement
tional aspects of school achievement, using data on family
influence parents' daily interactions with their children. Par-
ents' conceptions of the malleability of the child's character
This study was conducted with funds provided by the Spencer have been observed to affect their responses to specific acts
Foundation and the National Science Foundation (Grant NSF BNS (Sameroff& Feil 1981), Furthermore, parents' attributions
91-07542). Staff for the study of Caucasian Americans included Susan
and perceptions of the extent to which children are willful,
Holloway, William Arsenic, and Peggy Estrada.
Correspondence concerning this article should be addressed to causal agents in regulating their behavior may determine how
Robert D. Hess, School of Education, Stanford University, Stanford, parents respond to children's actions, especially actions con-
California 94305. strued negatively (Dix & Grusec, 1984). One way that parents

179
 180 R. HESS. C. CHIH MEI, AND T. McDEVITT

influence children's achievement is by promoting the value persistence (e.g., "Being diligent in study means devoting one's
of effort. Weiner (1980) demonstrated the efficacy of attribu- effort to it for a long time," Confucius, Zi Zhang chapter),
tions to lack of effort for unsuccessful performance. By attrib- and (d) the efficacy of effort (e.g., "One will inevitably succeed
uting relative failure to lack of effort, individuals may expect in one's study if one is diligent and takes delight in study,"
that they can do better on subsequent tasks by working harder. Confucius, Xu Er chapter). In Huang's view, the enemies of
When parents hold these beliefs, they may transmit them to study are laziness, giving up quickly, and overconfidence.
their children and encourage children to persist when con- Scholarly values appeared in Chinese culture well before
fronting obstacles. Parents may express in various ways the the 19th-century immigration of Chinese to California. It is
belief that children bear responsibility for performance be- thus reasonable to assume that early immigrants brought with
cause they can control the degree of effort directed toward a them the cultural view that scholarship and effort are the
task. routes to social and economic advancement. The upward
mobility of Chinese in the United States is consistent with the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

assumption that such values were present in families of early

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Role of Cultural Differences in Motivation in migrants (Sucheng, 1981;Vernon, 1982). Given the cohesion
of the Chinese family system, we would expect these values
Educational Achievement to be preserved over several generations, although they might
Considerable evidence suggests that Americans emphasize decline somewhat over time.
achievement in schools less than the Chinese do. Historical Achievement and individual effort have also been promi-
materials and cultural folklore document the value that the nent themes in the cultural and economic development of
Chinese place on scholarly endeavor and toil. The Chinese the United States. Some observers suggest, however, that the
have long respected academic accomplishments, and with value placed on achievement and hard work has declined in
some brief interruptions, this regard persists to the present the United States in the past half century (Spence, 1985). For
day (Wan, 1985; Xiangrong, 1985). Intellectual facility is example, in observational studies, the discrepancy between
appreciated in its own right, but it is also valued as a path to actual time spent on work and employees' reports of time at
employment and status. In China, the examination system work increased between 1965 and 1975 (Stafford & Duncan,
has provided a route to social, political, and economic mo- 1979, cited in Yankelovich & Immerwahr, 1983); according
bility for more than a thousand years (Ridley, 1973). This to Cherrington's (1980) observations, only about half of work-
respect for scholarly attainment parallels a high regard for ers' time was spent in actual job-related activities. About half
effort: Hard work offers the major route to accomplishment the workers surveyed by Yankelovich and Immerwahr (1983)
and competence (Chang, 1985). Historical accounts of prep- said that they put in no more effort than what was required.
aration for examinations reveal exceedingly high levels of Also, the heroes of American culture, although praised for
motivation (Ridley, 1973). The ancient teachers and scholars individualism, strength of character, and accomplishment,
urged on young students by saying, "If you are not diligent in include fewer in whom effort and scholarly achievement are
study when your hair is black, it will be too late to sigh about closely linked. Thus, despite a traditional regard for achieve-
study when your hair is white." In modern times, parents, ment and success in American culture, effort seems to be less
teachers, and adults refer to folklore containing vivid exem- salient and less oriented toward scholarly achievement in the
plars of the value of effort. Among the many legends well- United States than in Chinese culture. We expected, then,
known to the Chinese, three can be used to illustrate this that the families from the PRC would attribute children's
respect for effort: Kuang Heng, a boy in a poor family that performance in mathematics to effort and individual respon-
could not afford fuel for lights at night, bored a small hole sibility more than would either group from the United States.
through a wall that his home had in common with a more We also expected that within the United States, the Chinese-
affluent neighbor's house to find light by which to study; Che American families would emphasize these qualities more than
Yin made light for his studies by carrying fireflies; and Sun would the Caucasian-American families.
Kang studied by the light of the moon reflected off the snow
(Huang, 1969). These and similar legends that extol the virtue
of effort in pursuit of learning are familiar throughout China. Overview of the Study
Beliefs about the efficacy of effort also appear in writings The design of the study permitted an examination of both
about educational practice. A volume describing educational national and cultural differences in family beliefs. The na-
principles (Huang, 1969) taken from Confucius and other tional comparisons are indicated by contrasts between fami-
revered scholars includes several injunctions and approba- lies from the PRC and those from the United States. The
tions about effort. These statements, quoted by Huang (1969), existence of such distinctions between the PRC families and
offer advice on several aspects of scholarly activity—(a) the the United States families would argue for national differences
role of mental concentration {e.g., "Study as if you could not in beliefs about motivation. These distinctions might arise
attain your aim and were afraid you should lose it," Confu- from both institutional policies and cultural practices.
cius, Tai Bo chapter), (b) the primacy of effort (e.g., "Talent Cultural comparisons come from contrasts between Cau-
and will come first in study; will is the teacher of study and casian-American families and both the PRC and Chinese-
talent is the follower of study. If a person has no talent, it American groups. Divergences between the Caucasian-Amer-
[achievement] is possible. But if he has no will, it is not worth ican families and the American-Chinese and PRC families
talking about study," Xu Gan, Zhong Lun), (c) the need for would make a case for cultural distinctions. A difference
 CULTURAL VARIATIONS IN FAMILY BELIEFS 181

between Caucasian-American and Chinese-American groups The backgrounds and experience of families in the United States
would be particularly important because the students from and China are obviously not comparable. Parents in the PRC who
these families participate in the same political and educational participated in the study were in school during the time of the Cultural
systems, ruling out the effect of institutional differences. Revolution, when official attitudes toward academic education were
less favorable than previously or at the present time. Also, as is
The data gathered in the interviews included causal expla-
common in cross-cultural studies, demographic characteristics that
nations by mothers and children for why the children did as seem, on the surface, to be similar across societies may, on inspection,
well as they did and why they did not do better. In this turn out to be noncomparable. For example, equating socioeconomic
comparison of the beliefs of families from the PRC, Chinese- or educational status was, in a real sense, impossible. Occupational
American families, and Caucasian-American families, we ex- roles that involve comparable tasks in the two countries may earn
pected that Chinese in the PRC would emphasize effort, very different incomes on a relative as well as an absolute basis and
especially its absence, as a cause of relatively low performance, may invoke quite different levels of prestige. A given number of years
more than either of the other two groups would. We also of schooling involves very different things in the two countries in
terms of content presented in courses that deal with social or political
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

expected that traditional values about scholarship and

This document is copyrighted by the American Psychological Association or one of its allied publishers.

achievement would persist to some degree in the Chinese- subjects or level of mastery in more comparable subjects, such as
American families and that the Caucasian-American mothers mathematics and science.
and children would place less emphasis on effort and individ- These constraints made it difficult to follow some of the usual
standards for equating research groups and devising similar data-
ual responsibility than the other two groups would. We based
gathering procedures. Nevertheless, the opportunity to gather data on
our analyses on the entire set of attributions presented to the families in the PRC was rare, and this fact gives the data a certain
families, however, to obtain a more complete understanding interest and value. At the least, comparative data offer a descriptive
of the patterns of attributions held by the three groups. view of families in the two countries, yielding evidence and hypotheses
As a supplement to the analyses of attributions, we obtained about the differences in achievement that can be confirmed or altered
reports from mothers about their responses to instances of by future studies.
children's relatively low and relatively high performance. We Although these circumstances suggest caution in using inferential
expected that the families from the PRC would treat low statistics, we include the results of such statistics to inform readers
performance more seriously than the other groups would. If who may find them of interest.
the Chinese parents believe, as we proposed, that their chil-
dren's failure to achieve at higher levels is due to an internal Subjects
factor over which the children have control (i.e., effort), they For this study we drew from three cultural groups—families from
may show anger and criticism at children's low performance the PRC and Chinese-American and Caucasian-American families.
(Dix & Grusec, 1984). Such feelings are more likely to spur The original Caucasian sample consisted of 67 native-born American
parental intervention than are feelings of resignation or pity, mothers from a range of socioeconomic backgrounds and their 4-
which may spring from a belief that failure was caused by an year-old firstborn children. We recruited these families from pre-
external source (the school) or an internal, unchangeable schools and daycare centers in the San Francisco Bay area. The
source (natural ability) beyond the child's control. A combi- median education of parents was about 2 years of college. When the
nation of explanations that focuses on internal and control- children were in sixth grade, we conducted a follow-up study. At this
lable factors offers a particularly powerful basis for parental time, we were able to locate and contact 48 families from the original
sample; 47 agreed to participate. Of these, 23 had girls, and 24 had
response and initiative.
boys. Only data from the follow-up phase are reported in this article.
We recruited 51 Chinese-American mothers from a range of socio-
economic backgrounds and their sixth-grade children in the San
Method Francisco Bay area, mainly through schools. We considered families
to be Chinese if they had ancestral origins in the PRC, Taiwan, or
We gathered the data on which this article is based under unusual Hong Kong and both parents were Chinese. For most of the families
circumstances. Responses from Caucasian mothers in the United (38), the children were third-generation Chinese: Their grandparents
States were obtained as part of an earlier study of socialization of were horn in a Chinese country, and their parents were born in the
school readiness in the United States. When an unexpected oppor- United States. For the remaining families (13), the children were
tunity to gather information about family beliefs in the PRC arose at second-generation Chinese: Their parents were born in China but
a later time, we obtained funds to extend a part of the study to had moved to the United States at least 15 years prior to the interview.
include mothers from Chinese-American and PRC families. The median education of parents was 3.8 years of college, slightly
The study thus exploited an unexpected opportunity. We devel- higher than that of the Caucasian mothers. There were 28 girls and
oped it, however, under constraints, some of which follow from the 23 boys in the group. The sample was not restricted to firstborn
different climates for research in the two countries. Procedures for children.
obtaining permission to conduct a study of this kind in the PRC The research group from the PRC consisted of 47 native-born
involve conversations with several levels of government. Schools and mothers from a range of socioeconomic backgrounds. Their children
families are not accustomed to participating in research. There was were in the sixth grade (or the equivalent) and had taken the entrance
not a body of data on beliefs and attitudes of families in contemporary examination to middle school. We recruited them in Beijing primarily
China to use in constructing research instruments. For example, the through the schools. There were 21 boys and 26 girls in the research
interviews in the PRC elicited parental beliefs that suggested addi- group. As with the Chinese-American group, the PRC sample was
tional comparisons. This led us to reinterview families in the United not restricted to firstborn children. The median education of parents
States on specific points. Unfortunately, many of the Caucasian- was 12 years.
American families had moved, and only a small subsample from We recognize, of course, that a small number of families cannot
which to collect data was available. represent a nation of roughly a billion people and small samples
 182 R. HESS, C. CHIH MEl, AND T. McDEVITT

cannot represent the range of attitudes and responses in the United 1 to 8; for relative falure, it was 1 to 10. This indicated that the
States. It seems very unlikely, however, that large studies using constraint on allocating chips to ability and effort was negligible.
representative samples of families from the two nations and the three The mothers were also asked about what they would say to their
cultural groups will be conducted. In the absence of other data, child if he or she brought home an unusually good (and unusually
studies of this type provide initial glimpses into motivational processes bad) grade in mathematics. These questions were part of the interview
in the two countries. The validity of such preliminary inquiries will in the PRC. In the United States, they were asked as a telephone
gain or lose support as other investigations are conducted. follow-up to the original interview. Of the original 51 Chinese-
American mothers, 41 could be later located and contacted; these
mothers responded to the questions. The Caucasian-American fami-
Materials and Procedure lies had been interviewed three years earlier, and only a small part {n
= 18) of the Caucasian-American group could be reached for the
In both the United States and the PRC, we interviewed mothers follow-up questions.
and children separately. With minor modifications, we asked mothers The child interview was identical to the mother interview except
and children the same questions about the children's performance.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

for minor changes required to make the questions relevant to the

For the interview in the PRC, we used back-translation to ensure
This document is copyrighted by the American Psychological Association or one of its allied publishers.

child's own performance.

similarity of questions in the two countries.
We conducted the maternal interview as follows. First, we asked
mothers to judge how well the child had performed in mathematics
during the sixth grade. The mother was to indicate the child's level
Results
of performance on a 6-point scale ranging from not doing as well as The three groups in this study differed in the causal expla-
most (1) to doing the very best in the class (6). The interviewer noted
nations that they gave for children's performance in mathe-
the ranking the mother gave and, for the relative-success condition,
asked why the child performed at that level {e.g., "You have indicated matics. The mothers of the groups also differed in the re-
that your child is doing better than many in the class. Why do you sponses that they said they would give to specific instances of
think she is doing that well?"). The interviewer then showed the high and low performance by their children (Figures 1 -4 and
motherfivecards, each labeled with one of the following attributional Table 1).
phrases: In explaining why their children "did not do better than
they did," all three groups put more weight on effort than on
My child has natural ability for math. other causes, but the salience of effort varied substantially
My child tries hard in math. among the groups, as did the degree to which other sources
My child has had good training at school in math.
My child has had good training at home in math. played a role. For the PRC families, lack of effort was the
My child has been lucky in math. predominant cause for low performance; other factors shared
little responsibility. The Chinese-American group showed a
The interviewer then asked the mothers to distribute 10 plastic token similar pattern, with somewhat less weight assigned to effort
chips across the five cards to indicate the importance of each as a and more weight assigned to lack of natural ability, poor
reason for the child's having done as well as he or she did. This school training, and poor home training. Although the Cau-
procedure resulted in five attribution variables, each with possible casian-American group gave more weight to lack of effort
ranges of 0 to 10. We selected these specific attributions on the basis than to other sources, they distributed blame more evenly
of previous research (Holloway & Hess, 1982).
across all five sources than did the Chinese groups. Lack of
An analogous procedure was used for questions about relatively
ability and poor training at school—both factors over which
low performance. In this condition, the question was "You have
indicated that your child wasn't at the (very) top of the class. Why parents and children have little or no control—were seen as
do you think he/she isn't doing (even) better in math?" (If the mother almost as important as effort.
reported that the child was at the top of the class, the interviewer The tendency to blame the schools for children's poor
suggested that the child might not be the best in the school district performance also varied considerably among the three groups.
and asked why the child was not doing even better.) The purpose of Mothers from the Caucasian-American group put only
these questions was to examine the factors that mothers believed slightly less responsibility on schools than on the child's
limited the children's performance. The attributional phrases were ability. A complementary contrast shows in the mothers'
worded to reflect low performance (e.g., "My child does not try hard
views of the role of training at home. Both Chinese groups
in math").
Because allocating chips to one source necessarily affected the put more chips on "lacks training at home" than on "lacks
number of chips available to assign to other sources, the procedure training at school." For the Caucasian group, the emphasis
of limiting the number of chips to 10 constrained the range of scores was reversed.
for individual attributions. We examined the data to see if there was The three groups differed also in the degree to which they
adequate variance across the five options by inspecting variability of selected causes under the control of the students or their
two attributions that tended to receive high scores, ability and effort. families. This contrast is seen by combining two options,
If the combined value of these two sources was very high and thus effort and training at home. The tendency of the three groups
similar across mothers, we would obtain no additional information of mothers to place responsibility on causes that can be
by considering both of these attributions because when the value of manipulated was greatest in the PRC group (7.58 chips of a
one was set, the other would also be known.
possible 10), next in the Chinese-American group (6.35 of
To examine this possibility, we created additive composites of
ability and effort attributions (separately for the two conditions) for 10), and least in the Caucasian-American group (4.74 of 10).
the Caucasian group and examined the resulting distributions of The Caucasian mothers gave more weight (5.26 of 10) to
scores for the 47 mother-child pairs. We found a considerable range causes not under their control (child's ability, training at
and distribution among the scores. The range for relative success was school, and luck).
 CULTURAL VARIATIONS IN FAMILY BELIEFS 183

7 T

• PRC CHINESE
W CHINESE AMERICANS
I I CAUCASIAN AMERICANS

Mean
Attribution
Scores
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Ability Effort School Training Home Training Luck

Figure I. Mothers' attributions for relative success.

Comparable data for the children showed a similar pattern were more likely to see luck as playing a role, particularly in
but with less emphasis on controllable causes. The students the failure condition.
from the PRC were most likely to select controllable causes Cultural differences were also evident in explanations for
(7.49 of 10); American students, both Chinese and Caucasian, why children "did as well as they did." Mothers and children
were less likely to blame causes over which families have some from the PRC gave most credit, by far, to the schools. The
control (4.86 of 10 and 3.38 of 10, respectively). In general, Chinese Americans viewed the home as the most important
children also gave more emphasis than did mothers to con- of the five explanations but gave almost as much credit to
trollable causes in attributions for success. American children ability, effort, and school training. The Caucasian-American

7i
• PRC CHINESE
6 •
S3 CHINESE AMERICANS
111 CAUCASIAN AMERICANS
5 •

Mean
Attribution
Scores
3

2 •

02

Lack of Ability Lack of Effort Poor School Poor Home Bad Luck
Training Training

Figure 2. Mothers* attributions for relative failure.

 184 R. HESS, C. CHIH MEI, AND T. McDEVITT

• PRC CHINESE
6 M CHINESE AMERICANS
M CAUCASIAN AMERICANS
5 •

4 •• 3.4
Mean
(1.5)
Attribution 2.8 2.8 2.7 2.7
Scores
0.3) 0.4) 2.3
2.0 0.1)
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

1.8
This document is copyrighted by the American Psychological Association or one of its allied publishers.

(1.5)

0 6

Ability Effort School Training Home Training Luck

Figure 3. Children's attributions for relative success.

mothers and children again tended to spread responsibility them what they might say to their children if the children
rather evenly across four of the five options. really did bring home a grade that was higher or lower than
usual. In response to the question about what mothers would
do when the children performed especially well, three cate-
Maternal Responses to Specific Instances of Children's gories of responses were coded. Two raters who coded re-
Good and Poor Performance sponses independently obtained 90% agreement. In most
cases, mothers offered only one response; in cases in which
In addition to asking mothers to explain why their children they offered more than one type of response, only their first
might perform at a level higher or lower than usual, we asked response was coded. The three categories were as follows: (a)

• PRC CHINESE
M CHINESE AMERICANS
I I CAUCASIAN AMERICANS

Mean
Attribution
Scores

Lack of Ability Lack of Effort Poor School Poor Home Bad Luck
Training Training

Figure 4. Children's attributions for relative failure.

 CULTURAL VARIATIONS IN FAMILY BELIEFS 185

Table 1
Strategies Used by Mothers in Response to Specific Instances of Children's Relatively High and Relatively Low Performance
People's
Republic of Caucasian
China Chinese Americans Americans
(N = 47) (AT =41) (JV= 18)
Response type % No. % No. % No.
Relatively high performance
Provides rewards without
setting higher standards 19 9 73 30 83 15
Sets higher standards with
or without praise 66 31 24 10 6 1
Other 15 7 .02 1 11 2
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Relatively low performance

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Punishes or threatens
punishment 21 10 .02 1 0 0
Determines reason 32 15 61 25 67 12
Mentions need for more
care or effort 28 13 29 12 22 4
Other 19 9 .07 3 11 2

provides rewards, verbal or tangible, without setting higher Examples of responses from Chinese-American mothers con-
standards (e.g., "Good job!" and "I'd give him something he firm this finding:
likes to eat"), (b) sets higher standards, with or without praise "I would say 'Good job,' like everybody else would say."
(e.g., "You did well but maybe next time you can do even "I would encourage her and praise her and get her some-
better" and "You should be getting even higher grades"), and thing that she wants to make her happy."
(c) other—emphasizes value of effort or working carefully, Mothers from the PRC were less free with their praise. They
attempts to determine reason for good performance, or pro- more often reported attempting to set even higher standards.
vides some other response (e.g., "You're studying hard, and Two examples illustrate the responses of the mothers from
it's paying off," "This time you checked all of your answers— Beijing:
that's why you did so well," and "Why do you think you did "That's nothing to be proud about. With your ability, you
so well this time?"). should get 100 every time. If you don't, it is because you are
In response to the question about what mothers would do careless."
when the children performed poorly, raters coded four cate- "Don't get too excited. Compare yourself with other class-
gories. They obtained 81% agreement. Again, in most cases, mates; you still have a big gap!"
mothers offered only one response; when they offered more Caucasian-American mothers' responses, although not in-
than one type of response, only their first response was coded. cluded in the tests of significance, were similar to the Chinese-
The four categories were as follows: (a) punishes or threatens American mothers' responses in their use of praise:
punishment (e.g., 'Td show him I was angry, or Fd spank "I would be pleased and would tell him that I was pleased."
him"), (b) determines reasons for low performance (e.g., "I'd "I would be visibly impressed! Knowing her, I would feel
find out exactly what she didn't understand" and "Fd ask if that she had done exceptionally well, and I would be quite
he didn't try hard"), (c) urges more care or effort (e.g., "You pleased with her."
need to try harder" and "You were careless. Next time be The mothers in the three cultural groups did not differ as
more careful!"), and (d) other—comforts child, describes long- much in responses to relatively low performance. Both the
term consequences, ignores the information, or provides other Chinese-American mothers and Caucasian-American moth-
response (e.g., "Fd tell her that it's all right, that she shouldn't ers often attempted to determine the reason for the failure of
get upset," "How will you get into middle-school if you do so their children's relatively low performance more so than did
poorly?", and "I wouldn't say anything"). the PRC mothers. The following are representative responses
Mothers in the three cultural groups appeared to use specific of the Chinese-American mothers:
categories to various degrees. We did not include Caucasian- "Your grade seems to be lower than usual. Is there some
American mothers in the analyses because of the small sample reason for that we should know about?"
size, but their responses are included with those of the other "I will ask about the reason and check to see what she is
two groups in Table 1. The analyses indicated that the PRC doing everyday, putting enough time in study or wasting her
and Chinese-American mothers responded in different ways time in playing too much. Or, she does not understand the
for both relatively high performance, * 2 (2, N = 88) = 26.28, classwork."
p < .0001, and relatively low performance, x2(3, N = 88) = The responses from the Caucasian-American mothers were
12.55,p< .01. similar:
Differences between the PRC and Chinese-American moth- "I would want to know if he were struggling with that
ers in responses to relatively high performance were striking. particular concept. We would want to help him through that
Chinese-American mothers more often reported that they since he hadn't apparently gotten that in the classroom envi-
would reward their children than did the PRC mothers. ronment."
 186 R. HESS, C. CHIH MET, AND T. McDEVITT

"We would want to find out why she was slipping and get Scheffe contrasts supported the existence of cultural differ-
her back to her previous level." ences that are visible in the graphs. The PRC Chinese saw
Only about one third of the PRC mothers offered this school training as more influential than did either the Chinese
response. One mother put it in these words: Americans or the Caucasian Americans. The Caucasian
"You should Find the reason why you made mistakes. You Americans viewed school training as having greater responsi-
should correct your mistakes and remember this lesson." bility than did the Chinese Americans. In contrast, the
About one fifth of the mothers from the PRC said that they Chinese Americans viewed home training as exerting greater
would express anger or punish the child in some way. Some influence than did the PRC Chinese or the Caucasian Amer-
examples follow: icans, and the Caucasian Americans viewed it as being more
"I would be angry. I would criticize him." important than did the PRC Chinese. The significant effect
"You are the kind of person who won't try harder without on luck appeared to be due to the tendency of the Caucasian
being punished." Americans to assign greater weight to it than did the PRC
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

"Some people get things wrong because they think they are Chinese.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

smart. If you want to be a success, being smart is not enough. Boys and their mothers' beliefs about the children's achieve-
Sometimes it is even useless." ment seemed to differ from girls and their mothers', although
The PRC mothers' reports of anger at their children's lack the pattern of child gender differences was not consistent
of effort are consistent with Weiner's (1980) formulations of across the three cultural groups. The gender difference was
the link between emotions and attributions. It is as if the PRC indicated by a significant interaction between gender of child
mothers believed that the child could do better if he or she and attribution, F(4, 136) = 2.92, p < .05. However, both the
tried harder; it was appropriate to place blame on the child. interactions between cultural group and attribution and be-
The level of emotion may also reflect the degree of importance tween gender of child and attribution were qualified by a
that these mothers give to achievement. higher order interaction among attribution, cultural group,
In summary, the interviews suggest that mothers in the two and gender of child, F<8, 272) = 2.44, p < ,05. This higher
American groups tended to reward the successes of their order interaction indicated that the differences in patterns of
children and to use a more diagnostic approach to children's attributions among the cultural groups were not constant
problems at school. The PRC mothers tended to respond to across the two sexes. Likewise, gender differences in patterns
successes by raising the standards; to failure, they showed of attributions were not constant across the three cultural
considerable feeling, often punishing the children or urging groups.
them to work more carefully and diligently. Mothers and children also differed from each other, F(4,
136) = 10.59, p < .001, but the interaction between informant
A Statistical Perspective (mothers vs. children) and attribution was qualified by an
interaction among cultural group, informant, and attribution,
We present the results of inferential statistics with the F(8, 272) = 1.98, p < .05. We computed additional MANOVAS
caution raised earlier: The three samples were small and were to see whether the three cultural groups differed when mothers
not randomly selected from larger populations. With this and children were taken separately. They did. The patterns of
reservation in mind, our analyses of the attributions com- attributions differed across the cultural groups for both moth-
menced with multivariate analyses of variance (MANOVAS). ers, F(8, 278) - 8.52, p < .0001, and children, F(8, 278) =
Child gender and cultural group were between-subjects vari- 3,86, p<. 001.
ables, and generation of informant (i.e., mother and child) Finally, the remaining interactions among attribution, cul-
and attribution (i.e., ability, effort, home training, school tural group, gender, and informant and among attribution,
training, and luck) were within-subject variables. We com- gender and informant were not significant.
puted analyses separately on explanations for why the children
did as well as they did and why they did not do better.
Why Children Did Not Do Better
Why Children Did as Well as They Did For explanations of why the children did not do better,
participants also viewed individual attributions as exerting
The combined groups distinguished among the five options different amounts of influence, F(4, 136) « 108.13, p< .001.
in explaining the reasons for children's success in mathemat- For the entire group, lack of effort was seen as the most
ics, F{A, 136) = 648.50, p < .001. For the entire group, school important reason for not doing better. It was followed by lack
training was seen as especially important. The second most of ability, poor home training, and poor school training.
important explanation was effort, followed by home training Again, luck was seen as having negligible influence. As with
and ability. Luck was seen as having negligible influence. the analysis on attributions about relative success, several
Several higher order interactions qualified this pattern. higher order interactions qualified this pattern.
The three cultural groups differed in their patterns of attri- The three cultural groups used different patterns of attri-
butions for relative success, F(8, 272) = 9.79, p < .001 (see butions, indicated by a significant interaction between cultural
Figures 1 and 3). The cultural groups did not differ on ability group and attribution when mothers and children were con-
or effort, but they did assign different weights to school sidered together, F(8t 272) = 10.88, p < .001 (see Figures 2
training, F{2, 142) = 29.57, p < .001; home training, F(2, and 4). The cultural groups did not differ on the importance
142) = 17.40, p < .001: and luck, F(2, 142) = 4.94, p < .01. of poor home training, but they did differ on lack of ability,
 CULTURAL VARIATIONS IN FAMILY BELIEFS 187

F<2, 142) = 8.97, p < .001; lack of effort, F{2, 142) - 39.25, PRC spend more time in school than do children in the
p < .001; poor school training, F(2, 142) = 6.78, p < .0 i; and United States. Beginning at the age of six years, the typical
bad luck, F(2, 142) = 20.05, p < .001. According to Scheffe child in the PRC spends from 7:30 a.m. to 3:00 p.m. at school,
contrasts, the Caucasian Americans saw lack of ability as a with 1 hr for lunch—6'/2 hr each day. In Grade 4, the school
more important reason than did the PRC Chinese. The PRC day is usually 1 hr longer. After school, many children engage
Chinese viewed lack of effort as more influential than did in study groups led by teachers. In addition, the school year
either the Chinese Americans or the Caucasian Americans. includes about 280 days, in contrast to the 180-200 days in
The Caucasian Americans assigned greater weight to poor the United States. Additionally, the school is charged with
school training than did the PRC Chinese. Finally, both the responsibility for moral and physical development as well as
Caucasian Americans and the Chinese Americans viewed bad for academic performance. Until recently, children were as-
luck as a stronger determinant than did the PRC Chinese. signed homework beginning in the first grade, and the teacher
Mothers and children also differed in their beliefs, F{4, 136) graded homework every day. Parents and teachers commu-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

— 9.44, p < .001, but the interaction between informant and nicate frequently, and the schools expect teachers to visit
This document is copyrighted by the American Psychological Association or one of its allied publishers.

attribution was qualified by a higher order interaction among families in their homes.
cultural group, informant, and attribution, F(8, 272) = 2.32, In some respects, the beliefs of the Chinese Americans
p < .05. We computed additional MANOVAS to see if the three appeared to show the influences of both their Chinese cultural
cultural groups differed when mothers and children were heritage and the American culture in which they resided. In
taken separately. The patterns of attributions differed across this regard, the Chinese Americans took an intermediate
cultural groups for both mothers, F{8y 278) = 4.27, p < .001, position on the importance of lack of effort as a source of low
and children, F(8, 278) = 8.36, p < .001. performance: They placed more emphasis on lack of effort
Finally, the patterns of attributions did not differ for boys than did the Caucasian Americans but less emphasis on it
and girls, and the remaining interactions—Cultural Group X than did the Chinese from the PRC. In other respects, how-
Gender x Attribution, Gender x Informant x Attribution, ever, the Chinese Americans differed in a way that did not
and Cultural Group x Gender x Informant X Attribution— reflect an intermediate position. For example, the Chinese
were not significant. Americans viewed the home as more responsible for the
children's relative success than did either of the other two
Discussion groups.
Also, patterns of attributions differed by generation. When
There were dramatic cultural differences both in mothers' mothers and children were examined separately, however, the
and children's attributions and in mothers1 responses to high cultural differences in patterns of attributions persisted; chil-
and low performance. The higher achievement of Chinese dren in the three cultural groups distinguished themselves in
students may occur in part because they believe that perform- ways similar to those of their parents. There was, however,
ance is linked to an internal, controllable source: Failure can some unevenness across groups in differences between moth-
be avoided and achievement assured by trying harder. ers and children. For example, compared with their mothers,
The tendency of the PRC group to see the school as Caucasian-American children more often chose bad luck as a
responsible for relative success may reflect their high regard reason for relatively low performance. For both American
for teachers and scholars. It may also indicate a culturally groups, there was a large generational difference in attribu-
based reluctance to claim credit for achievement of family tions to lack of effort: Mothers relied on this source more
members; individuals are responsible for avoiding failure but than did children. In the PRC, the generations were similar
share credit with external sources for success. The Chinese in their attitudes about effort.
Americans also referred to training, an external source, as a In responses to specific hypothetical occasions when their
factor in high performance, but instead of focusing on the children brought home test scores showing unusually high or
school, they gave credit to the home. It may be that the low performance, mothers from the PRC often responded by
Chinese in the United States are dissatisfied with the education increasing their pressure for high performance and raising
their children are receiving and attempt to supplement the their standards. They mixed praise or acknowledgment with
school's offerings. Or the Chinese-American mothers may cautions to avoid complacency and with admonitions to strive
perceive that their children study harder than classmates and to do even better. This technique, which seems comparable
believe that this is due to family discipline and encourage- to partial reinforcement, may raise the children's own internal
ment. standards. The Chinese-American mothers, like the Cauca-
In follow-up interviews, we asked the Chinese-American sian-American mothers, set higher standards less often and
mothers to help interpret this finding. They reported that it is praised the children for success more often. Furthermore, the
the duty of parents to instill in their children a good attitude PRC mothers more often mentioned that they would punish
toward the school. Some mothers also suggested that they the children for low grades than did the two groups of Amer-
lacked confidence in the school and believed that the home ican mothers. This tendency to treat children's low grades
needs to take an active part in children's education. It also rather severely may be partly due to these mothers' assuming
seems possible that mothers from the PRC had received little that the children were fully responsible for their low scores—
formal education themselves and felt less prepared to help they simply did not try hard enough.
their children in mathematics. Indeed, in the PRC, the school These comparisons of the beliefe of families from different
takes the major responsibility for the student. Children in the cultural backgrounds suggest that motivational forces may be
 R. HESS, C. CHIH MEI, AND T. McDEVITT

important sources of national and cultural differences in Kirst, M. W. (1981, June). Japanese education: Its implications for
achievement on tests of mathematical skills. The persistence economic competition in the 1980s. Phi Delta Kappan, pp. 707-
of beliefs within cultural groups, as reflected in responses of 708.
Chinese-American families about the utility of effort, indicates Marjoribanks, K. (1980). Ethnic families and children's achievements.
that cultural beliefs are stable and are transmitted through the Sydney, Australia: George Allen & Unwin.
Ridley, C. P. (1973). Educational theory and practice in later imperial
family. One implication of these findings is that cultural
China: The teaching of writing as a specific case. Unpublished
influences on achievement through affective and motivational doctoral dissertation, Stanford University.
sources deserve more extensive analyses. In particular, it Sameroff, A. J., & Fell, L. A. {1981, April). Parental perspectives in
would be worthwhile in future investigations to determine development. Paper presented al the meeting of the Society for
whether similar cultural differences in attributions and moti- Research in Child Development, Boston, MA.
vational strategies exist in other scholastic domains. Spence, J. T. (1985). Achievement American style: The rewards and
costs of individualism. American Psychologist, 40, 1285-1295.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Stafford, F., & Duncan, G. J. (1979). The use of time and technology
References
This document is copyrighted by the American Psychological Association or one of its allied publishers.

by households in the United States. Ann Arbor: University of

Michigan, Institute for Social Research.
California Slate Department of Education Assessment Program
Stanley, J. C , Huang, J., & Zu, X. (1986, summer). SAT-M scores
(1981). Student achievement in California schools: 1978-79 annual
of highly selected students in Shanghai tested when less than 13
report. Sacramento: California State Department of Education.
years old. College Board Review, No. 140, pp. 10-13,28-29.
Chang, C. (1985). Family influences on school achievement in China.
Stigler, J. W., Lee, S., Lucker, G. W., & Stevenson, H. W. (1982).
US.-China Friendship, 9(4), 20.
Curriculum and achievement in mathematics: A study of elemen-
Cherrington, D. C, (1980), The work ethic. New York: AMACOM.
tary school children in Japan, Taiwan, and the United States.
College Board. (1982). Profiles, college-bound seniors. 1981. New
Journal of Educational Psychology, 74, 315-322.
York: College Entrance Examination Board.
Sucheng, C. < 1981). Contemporary Asian immigration and its impact
Dix, T. H.f & Grusec, J. E. (1984). Parent attribution processes in
on undergraduate education at the University of California, Berke-
the socialization of children. In I. E. Sigel (Ed.), Parental belief
ley. Berkeley: University of California, Center for Studies in Higher
systems: Psychological consequences for children. Hilisdale, NJ:
Education.
Erlbaura.
Vernon, P. E. (1982, July). Abilities of North American Orientals: A
Hess, R. D., & Holloway, S. D. (1984). Family and school as educa-
study in acculturation. Paper presented at a meeting of the Inter-
tional institutions. In R. D. Parke (Ed.), Review of child develop-
national Association of Applied Psychology, Edinburgh, Scotland.
ment research. Vol. 7. The family (pp. 179-222). Chicago: Univer-
Wan, L. (1985). Changing educational theory and methods. Beijing
sity of Chicago Press.
Review, 28(24), 19-21.
Holloway, S. D., & Hess, R, D. (1982). Causal explanations for school
Weiner, B. (1980). Human motivation. New York: Holt, Rinehart &
performance: Contrasts between mothers and children. Journal of
Winston.
Applied Developmental Psychology, 3, 319-327.
Wong, M. G. (1980). Model students? Teachers' perceptions and
Huang, G. (1969). Comparative educational philosophy. Taipei, Tai-
expectations of their Asian and White students. Sociology of Edu-
wan; Po Mi Er Press.
cation, 53, 236-246.
Huang, G. F., & Wen, M. R (1984). The relationship between family
Xiangrong, X. (1985). Education given top priority. Beijing Review,
characteristics and students' achievement Jin Chinese]. Ada Shang-
28(36), 4-5.
hai Normal, 3, 136-141.
Yankelovich, D., & Immerwahr, J. (1983). New roles: Searching for
Husen, T. (1967). International study of achievement in mathematics:
self-fulfillment in a world turned upside down. New York: Random
A comparison of twelve countries. New York: Wiley.
House.
International Association for the Evaluation of Educational Achieve-
ment. (1985). Second study of mathematics: Summary report, Received February 18, 1986
United States. Urbana-Champaign, IL: U.S. National Coordinating Revision received December 5, 1986
Center. Accepted December 19, 1986 •