:::

青少年時期之認知自主性: 美國與台灣高中生之泛文化比較

Cognitive Autonomy in Adolescence 7
Research Articles
Cognitive Autonomy in Adolescence:
A Cross-Cultural Comparison of
American and Taiwanese
High School Students
Troy E. Beckert✽, Chien-Ti Lee✽✽, J. Mitch Vaterlaus✽✽✽
ABSTRACT
Most often, researchers of adolescent autonomy focus on behavioral
and emotional aspects of autonomy within western cultures. The
current study provides insight on cognitive autonomy from two cultural
perspectives. Adolescents from the United States (n = 330) and
Taiwan (n = 376) completed the Cognitive Autonomy and Self-Evaluation
(CASE) inventory, designed to assess five domains of cognitive
autonomy. Latent Class Analysis (LCA) resulted in a two-class model
of cognitive autonomy development across both culture and gender.
Taiwanese females and males were less likely to self-rate high in cognitive
autonomy areas including voicing opinions and self-assessing
whereas American youth had lower probabilities to be highly autonomous
in evaluating thinking. Adolescents from both cultures self-
  ✽  Department of Family, Consumer, and Human Development, Utah State University
2905 Old Main Hill, Logan, Utah 84322 – 2905. E-mail: troy.beckert@usu.edu
 ✽✽  Department of Community and Family Medicine, Duke Medical Center
✽✽✽  Department of Family, Consumer and Human Development, Utah State University
8 調查研究—方法與應用/第28 期
rated lower in comparative validation. Gender differences were
observed among Taiwanese youth but not Americans. We conclude
with a discussion of implications toward peer pressure and risky
behaviors.
Keywords: cognitive autonomy, adolescence, cultural differences,
latent class analysis
青少年時期之認知自主性:
美國與台灣高中生之泛文化比較
Troy E. Beckert, Chien-Ti Lee, J. Mitch Vaterlaus
大多數青少年自主性的研究者,只單一地將其研究焦點放在西方
文化脈絡底下,專注在行為與情緒的自主面向上。本研究提供了從兩
個文化的比較觀點,進行認知自主研究的獨特見解。本研究的樣本
中,有來自美國的330 位青少年,以及來自台灣的376 位青少年,共
同填寫一份名為「認知自主與自我評量」的問卷;該問卷是設計用以
評估「認知自主」的五個面向。根據潛在類別分析(Latent Class
Analysis)的結果,本研究發現:無論是美國或台灣的資料,或是男
性或女性的資料,都符合兩個類別的認知自主發展模式。台灣的青少
年,在認知自主的其中兩個面向:即「意見表達」和「自我評價」上,
無論男女,都較不可能給自己較高的評分;反觀,美國的青少年,較
不可能在「評斷性思考」上,給自己較高的自主性評分。而這兩個不
同文化的青少年,在「比較性的確認」上,自我評分都偏低。性別差
異的情況在台灣青少年的資料中顯著,但在美國青少年的資料中則無
此現象。最後,本論文以本研究之發現在「同儕壓力」與「風險行為」
方面的可能應用,進行討論。
關鍵字:認知自主性、青少年時期、文化差異、潛在類別分析
Cognitive Autonomy in Adolescence 9
I. Introduction
Adolescents achieve a healthy psychosocial development when they
become autonomous from parents and other adults (Freud, 1958; McElhaney
and Allen, 2001; Noom, Deković, and Meeus, 2001; Yeh, Liu,
Huang, and Yang, 2007) and gain a sense of personal identity (Bukowski
and Newcomb, 1983; Erikson, 1963; Meeus, Iedema, Maassen, and Engels,
2005). Although autonomy and identity are separate constructs within psychosocial
development, they relate closely to each other. From most theoretical
perspectives, autonomy is an important factor in the development of
identity (Erikson, 1963; Meeus et al., 2005; Spear and Kulbok, 2004).
Adolescent autonomy manifests as increased self-reliance, evidenced by
distinguished ideas from authority figures, organized personal experiences,
regulated personal behavior, guided individual life-goals, and independent
decisions based on their own rationales and experiences without undue
parental emotional support (Yeh et al., 2007). From a psychodynamic perspective,
autonomy is essential to the development of a strong ego; thus,
individuals avoid psychopathology by becoming autonomous (see Freud,
1958). Some scholars suggest that the psychoanalytic view over emphasizes
the distancing aspect of adolescent autonomy between adolescents
and their parents. These researchers posit that even though adolescents
decrease in closeness and become individualized from parental ties, most
adolescents still maintain strong positive relationships with parents (Allen,
Hauser, Bell, and O’Connor, 1994; Noom, Daković, and Meeus, 1999).
Accordingly, detachment may not serve as a valid indicator for capturing
adolescent autonomy.
10 調查研究—方法與應用/第28 期
Although researchers have tried many approaches to conceptualize and
operationalize adolescent autonomy, a clear definition of autonomy has yet
to emerge (Noom et al., 2001). Nonetheless, across research studies, autonomy
is commonly conceptualized into three domains including: behavioral,
emotional, and cognitive autonomy (Beckert, in press; Cicchetti and
Rogosch, 2002; Noom et al., 2001; Spear and Kulbok, 2004; Yeh et al.,
2007; Yeh and Yang, 2006). Operationally, behavioral autonomy includes
adolescents’ ability to act in an age appropriate manner (Allen, Marsh,
McFarland, McElhaney, Land, Jodl, et al., 2002; Anderson, Worthington,
Anderson, and Jennings, 1994; Cicchetti and Rogosch, 2002). Emotional
autonomy includes adolescents’ feelings of confidence to define goals independent
of the wishes of their parents and peers toward achieving interpersonal
competence (Anderson et al., 1994; Noom et al., 2001). Cognitive
autonomy represents an adolescent’s ability to think independently (Beckert,
2007; Lee, Beckert, and Goodrich, 2010). Until recently, scholars
tended to rely on theoretically rather than empirically definitions of cognitive
autonomy. One way to measure adolescent independent thought is to
appraise adolescents’ capacity to evaluate their own thoughts, voice opinions,
make decisions, self assess, and capitalize on comparative validations
(Beckert, 2007).
Behavioral and emotional autonomy are developmental tasks generally
introduced in early childhood. Most scholarly interest concerning adolescent
autonomy continues to focus on behavioral and emotional independence,
while cognitive autonomy has received less attention (Beckert, in
press). Of late however, professional interest in cognitive autonomy has
increased. One important reason is that adolescents are physically maturing
at earlier ages (Herman-Giddens, 2006). Because of early physical maturaCognitive
Autonomy in Adolescence 11
tion, adults place adolescents in adult situations at younger ages (Elkind,
2001). These situations require more advanced cognitive skills, such as
decision-making, than previously expected from younger adolescents. In
addition, neurological evidence suggests that cognitive development, which
includes cognitive autonomy, occurs much later in adolescence and into
early adulthood (Nelson, Thomas, and de Haan, 2006).
Scholars emphasize the importance of both culture and gender with
respect to most facets of adolescent development. These issues are also
central in examining components of autonomy in young people. Researching
and understanding cognitive autonomy in adolescents as it relates differently
by culture and gender may lead to new interventions in cognitive
development that could potentially accelerate independent thought and subsequent
identity exploration processes in adolescents (Lee, et al., 2010).
This could be especially helpful as adolescents encounter high-risk adult
decisions (e.g., drugs, alcohol, and sexual activity) at younger ages.
II. Literature Review
A. Cultural Contexts of Autonomy Development
Both popular media and developmental academic circles describe the
mainstream culture in the U.S. as more individualistically oriented while
they portray Asian or Chinese families, like those in Taiwan, as more collectivistically
oriented (Fuligni, Tseng, and Lam, 1999; Lee, et al., 2010).
Individualism is a cultural value affiliation that relaxes social bonds where
an individual’s needs are of more primary concern than the needs of others.
Collectivism, on the other hand, affiliates the interconnectedness
between individuals where the larger social group (e.g., family, commu12
 調查研究—方法與應用/第28 期
nity, and society) encompasses an individual’s needs. Individuals in collectivistic
societies are more disposed to give priority to family goals over
their own personal aspirations (Fuligni, Yip, and Tseng, 2002). Thus,
researchers have commonly sought to capture social (family) obligations/
duties and interdependence when comparing individualism and collectivism
(Fuligni, et al., 1999; Fuligni, et al., 2002; Phinney, Ong, and Madden,
2000). Some researchers suspect that teenagers from collectivistic societies
may be limited in the level of autonomy achieved when compared to
youth in individualistic contexts because collectivistic adolescents rely on
an external locus of control (Hui and Yee, 1994). While there are limited
and inconsistent findings in the literature to support this hypothesis, it is
worthwhile to notice that the validity of cultural comparisons is augmented
when both cultures have similar understandings of the constructs in question,
even if they might give different emphasis to the implications of the
constructs.
Taiwan has maintained a traditional collectivistic Chinese culture for
hundreds of years. This tradition emphasizes interdependence with hierarchical
authority (Lee, 1996). Taiwanese society, in general, expects youth
to respect elders and authorities within the family, at school, and in the
workplace. Young people are encouraged to be obedient and conforming.
Therefore, to a large measure, their elders’ expectations determine youth
decisions and behaviors (Yeh and Yang, 2006). There has been some evidence,
as communication and technology have evolved, that Taiwanese
youth, and society as a whole, are shifting away from collectivism.
Through frequent contact and identification with western culture, via Internet
and mass media, the values of many Taiwanese youth have rapidly
grown more individualistically oriented (Chattopadhyay and Marsh, 1999;
Cognitive Autonomy in Adolescence 13
Deaton and Paxson, 2000). This value shift might make Taiwanese adolescents
more likely to achieve an independent identity and become more
autonomous as what is defined in Western literature and which implies Taiwanese
youth become similar to U.S. teenagers in autonomy development.
Unfortunately, scholars know little concerning adolescent autonomy within
eastern cultures.
B. Cross-Cultural Comparisons of Adolescent Autonomy
Researchers in western cultures have conducted most of the investigations
on adolescent autonomy. Erikson (1963) indicated that the timing of
expectations for developmental task completion (e.g., when the adolescent
needs to achieve autonomy and what level of autonomy is acceptable)
might be culturally dependent. Erikson theorized that, although all cultures
deal with similar human developmental issues (e.g., identity formation and
autonomy) in the socialization process, the paths to desired developmental
outcomes might differ across cultures. Some empirical evidence supports
this theory. Feldman and Rosenthal (1991) found that the age expectations
of achieving behavioral autonomy for Chinese adolescents from Hong
Kong occurred later than it did for teenagers in both Australia and the
United States. Sheldon and his colleagues (2004) also indicated that late
adolescents from rural Taiwan had lower scores of self-concordance (i.e.,
people can express their opinions without being restricted by external controls)
than late adolescents from urban cities in the United States, China,
and Korea.
However, Xia et al. (2004) showed that Mainland Chinese adolescents,
from a midsized city, had similar developmental patterns of autonomy, in
terms of decision-making, as American youth. Accordingly, the develop14
 調查研究—方法與應用/第28 期
mental outcomes of adolescent participants of Chinese descent varied
across studies. These discrepancies are more likely to be the result of methodological
differences in the measurement of the different aspects of autonomy
(e.g., behavioral, emotional, and cognitive) and regional sampling differences
of Chinese youth. A better way to discover and understand differences
in cognitive autonomy between western and eastern cultures may
start by comparing adolescents from the United States with youth from
urban Taiwan where Western perspectives such as individualism are widely
recognized and therefore enhance participants’ understanding and accurate
responses that best presented their developmental status.
C. Gender Differences
Observations of differences in psychosocial development have not
been limited to cultural variations. Researchers have found that gender
plays a role in developing autonomy in adolescence. Zimmer-Gembeck
and Collins (2003) indicated that society generally has delayed expectations
for female behavioral autonomy compared to males, but we know little
about the relationship of gender to cognitive autonomy. As Gilligan
(1977) argued, most cultures socialize females to be caring and attentive to
others’ needs whereas males are not held to the same expectations.
Kashima and his colleagues (1995) found that females were more likely to
be interdependent rather than independent regardless of cultural or ethnic
differences. In addition, it is believed that Chinese females may exhibit
lower autonomy because they are socialized to be submissive and to surrender
their wills to (male) authority. For example, one traditional teaching
posits that “a woman has three pathways to follow. . . . In her youth
she must follow and obey her father. In her adulthood, she must follow
Cognitive Autonomy in Adolescence 15
her husband. In her later years she must follow her oldest son.” (Shon and
Ya, 1982: 211–212). Gender, as a factor, may moderate adolescent psychosocial
development. Further exploration, especially in Taiwan, is warranted
because little is known of gender differences in Taiwanese adolescent
autonomy levels.
D. The Present Study
Overall, cognitive autonomy is an important but under-researched construct
in adolescent psychosocial development. Most of studies in adolescent
autonomy, conducted primarily in western cultures, focus on behavioral
and emotional aspects. Adolescents across cultures could have similar
developmental tasks, such as autonomy achievement, but they may have
different expectations from their cultures concerning the timing of these
developmental tasks. These different expectations may be linked to a cultural
preference toward individualism or collectivism. It may also be linked
to cultural expectations related to gender. This study seeks to identify differential
patterns in adolescents’ self-perceptions of cognitive autonomy
across two cultural groups. There are two objectives for this study. First,
we seek to describe the self-identified levels of cognitive autonomy from
male and female adolescents within the United States and Taiwan. Second,
we attempt to compare the patterns of cognitive autonomy across domains
between these groups. Although this study is exploratory, based on the theory
used to anchor our argument and the previously reviewed literature, we
put forth two general hypotheses.
General Hypothesis 1. Taiwanese adolescents have different patterns
of autonomy (the combinations of attained cognitive autonomy status
across five domains) than adolescents from the United States.
16 調查研究—方法與應用/第28 期
General Hypotheses 2. Cross-cultural gender differences in patterns of
cognitive autonomy will be identified with male respondents scoring themselves
higher in all areas of cognitive autonomy.
III. Methods
A. Sample
For the purposes of this study, we recruited high school youth from
both Taiwan and the United States. Both samples were non-probability
samples but, because of cultural differences in research protocol, we
recruited each sample in a slightly different manner. In each case, methodological
procedures, including sample selection, followed IRB guidelines
and approval.
We identified Taiwanese adolescents for participation by using a modified
stratified cluster sampling technique from both high schools and vocational
schools in Taipei City, Taiwan. Taipei City is the capital of Taiwan
and is its largest metropolitan area. Students self-select into their high
school or vocational school mostly based on test scores and not geographic
location. The sampling procedure was a three-step process. First, we
obtained permission from school district administrators in Taipei. With
their permission, we identified target schools, based on similarities in
school characteristics in terms of their rank and sent letters seeking local
administrator’s permission to proceed. Next, participating local administrators
identified teachers of tenth, eleventh, and twelfth grades to invite for
participation. Finally, the cooperating teachers invited the students in their
classes to complete self-report questionnaires at school. As is common in
studies with Taiwanese youth, compliance rate was very high (98%). The
Cognitive Autonomy in Adolescence 17
final Taiwanese sample consisted of 376 adolescents; 20.6% of them were
vocational high school students. The mean age of the sample was 16.83
(SD = .80) years. As expected, the sample was predominately the same ethnicity
with 98.4% of participants indicating a background associated with
Han ancestry (71.5% Holo, 6.2% Hakka, and 20.6% Mainlanders). There
were slightly more female (51.1%) than male (48.9%) participants in the
sample. Based on the education level of fathers that was provided by the
students, 22.3% completed a 9th grade education or less, 28.7% had some
high school education, 39.9% earned a college degree, and 9.0% went to
graduate school.
The American sample of adolescents came from a convenience sample
of public and private high schools in the western United States. The
cooperating schools’ principal sent an informational letter supporting and
describing the research to parents of tenth, eleventh, and twelfth grade students.
The informational letter also informed parents that the survey was
voluntary and their student’s identity would remain confidential. The
American sample consisted of 330 adolescents. Participation rates were
also high for the American sample (96%). The mean age of the participants
was 16.19 (SD = .80) years. Among the valid responses (n = 299),
The majority of the sample were from public schools (54%) and reported
themselves as being Caucasian (67.9%), followed by Hispanics (12.7%),
Other (5.4%), Asian (4.3%), and Black (2.7%), and none of the participants
in the US self-identified as first generation of Asian American. The
final sample was comprised of 55.5% male participants and 44.5% female
participants.
18 調查研究—方法與應用/第28 期
B. Instrumentation
The Cognitive Autonomy and Self-Evaluation (CASE) inventory
assesses five domains of adolescent cognitive autonomy (Beckert, 2007).
The CASE inventory required translation from English into Mandarin Chinese
in order to be used with the Taiwanese sample. The translation procedure
included three adolescent development researchers, fluent in both
English and Mandarin Chinese, separately translating the items from English
to Mandarin. These researchers identified discrepancies and collectively
resolved problem areas to produce a final translation. Next, five
monolingual Chinese speakers evaluated their understanding of the statements
in the translated inventory. We compared feedback from monolingual
speakers and from adolescents who participated in a field test to assess
culture applicability. We made modifications based on the feedback to the
final translation toward increasing ease and clarity for Taiwanese adolescents.
Table 1 contains all the items of the CASE inventory in both English
and Chinese.
The CASE inventory is a 27-item 5-point Likert scale, scored from 1
to 5, consisting of five, empirically derived subscales (Beckert, 2007).
Subscales were designed to assesses implications toward evaluative thinking
that include thinking about the consequences of decisions, looking at
situations from other’s perspectives, weighing possible risks, evaluating
daily actions, considering alternative decisions, thinking about effect of
actions, weighing the long term effect of decisions, and evaluating
thoughts. A young person’s inclination to voice opinions is measured on
the inventory by the adolescent’s willingness to speak up in class discussions,
to share views when disagreements arise, to stand up for what the
Cognitive Autonomy in Adolescence 19
Table 1 English and Chinese Version of the Cognitive Autonomy and
Self Evaluation (CASE) Inventory
CASE Inventory
I. Evaluating Thinking
1. I think about the consequences of my decision
在我做決定前,我會先考慮做這個決定的後果
2. I look at every situation from other people’s perspectives before making
my own judgment
在我做決定之前,我會站在其他人的角度來思考
3. I think of all possible risks before acting on a situation
在我做事情(採取行動)前,我想過各種可能的風險
4. I like to evaluate my daily actions
我喜歡檢視並評定自己每日的所作所為
5. I think about how my actions will affect others
我會思考考慮到自己的行為對他人可能產生的影響
6. I think about how my actions will affect me in the long run
我會思考自己的行為對自己未來的可能影響
7. I like to evaluate my thoughts
我喜歡檢視、評估自己的想法是否成熟、有智慧
II. Voicing Opinions
8. If I have something to add to a class discussion I speak up
在課堂中,如果我對討論的議題有看法時,我就會舉手發言
9. When I disagree with others I share my views
當我不同意他人的想法時,我會說出自己的想法
10. I stand up for what I think is right regardless of the situation
不論任何情況下,我都會支持我所認為對的事情
11. I feel that my opinions are valuable enough to share
我認為我自己的意見是很值得與他人分享的
12. At school I keep my opinions to myself (reverse coded)
在學校的時候,我會保留自己的意見(想法),不願提出
III. Decision Making
13. There are consequences to my decisions
20 調查研究—方法與應用/第28 期
我知道任何一項決定,都將會有其好的或者不好的結果
14. I can tell that my way of thinking has improved with age
我知道我的思考能力已經隨著年紀而成長
15. I think more about my future today than I did when I was younger
比起小時候,我現在比較常思考自己的未來
16. My decision making ability has improved with age
我做決定的能力╱對事情判斷的能力隨年紀而成長
17. I am good at evaluating my feelings
我很擅長察覺、評估自己的感受
18. I consider alternatives before making decisions
做決定前,我會先考慮其他各種可能的方案
19. I am better at decision making than my friends
我比我的同學╱儕還擅長於做決定
IV. Self-Assessing
20. I am good at identifying my own strengths
我擅於認定、辨別自己的優點╱長處╱強項
21. I am the best judge of talents
我清楚了解自己的天賦、長處
22. I am best at identifying my abilities
我是最瞭解自己能力的人
V. Comparative Validation
23. I need family members to approve my decision (reverse coded)
我需要我的家人支持我所做的任何決定
24. I need my views to match those of my parents (reverse coded)
我需要我的想法與我父母的想法一致,才會使我心裡踏實
25. It is important to me that my friends approve of my decisions (reverse
coded)
對我來說,我朋友能夠認同並支持我做的決定是很重要的
26. I need my views to match those of my friends (reverse coded)
我需要使我自己的想法與我朋友們的想法一致,才會使我心裡踏實
27. I care about what others think of me (reverse coded)
我在乎別人怎麼看待我
Cognitive Autonomy in Adolescence 21
adolescent thinks is right, valuing their own opinion, and speaking out in
other school situations. Decision making, as measured on the inventory,
entails a recognition that there are consequences to decisions, that their
way of thinking and decision making have improved with age, that they
think more about the future than previously, and that they are better at
decision making than their friends. Self-assessing includes an ability to
identify one’s own strengths, abilities, and talents. Lastly, comparative validation,
on the inventory includes needs to have views match those of parents,
friends, and others, and to have family members and friends approve
decisions (Beckert, 2012). Response options are “1 = Never”, “2 = Seldom”,
“3 = Sometimes”, “4 = Often”, and “5 = Always”. Overall, Cronbach’s
alphas were relatively good across scores from both groups across
the five subscales. In the sample of adolescents from the United States,
the alpha coefficients ranged from .63 to .87. In the Taiwanese adolescent
sample, results were similar with alpha coefficients ranging from .63 to
.86. The convergent validity of cognitive autonomy among Taiwanese
youth were shown by the positive correlated results between CASE and
emotional and behavioral autonomy, assessed by Noom et al. (2001) Integrative
Model of Adolescent Autonomy Questionnaire (AAQ). Detail psychometric
properties of CASE among Taiwanese sample were presented in
Table 1. Table 2 provides the correlation matrix of the CASE Inventory
stratified by gender and cultural group. As expected, preliminary differences
can be identified between genders and cultures. Cognitive autonomy
in a particular area was considered present or established if it respondents
indicated “often” or “very often.” Accordingly, we dichotomized each cognitive
autonomy subscale into a composited mean score of less than 3.5 as
“Not yet established consistent autonomy” in that area vs. a mean score of
22 調查研究—方法與應用/第28 期
Table 2 Psychometric Properties of the CASE for Taiwanese Adolescents
#Items M SD α
Range Bivariate Correlation
Potential Actual
Emotional
Autonomy
Behavioral
Autonomy
Evaluative
thinking
8 3.66 .66 .86 1–5 1–5 .156 .270
Voicing opinions 5 3.20 .62 .63 1–5 1–5 .433 .320
Decision-making 6 3.86 .62 .76 1–5 1–5 .296 .352
Self-assessing 3 3.27 .89 .82 1–5 1–5 .258 .317
Comparative
validation
5 2.44 .70 .68 1–5 1–5 .233 .043
3.5 and above as “Established consistent autonomy” in that area.
C. Analytic Strategies
A preliminary analysis was used to explore adolescent characteristics
of cognitive autonomy on an aggregate level within each society. Because
cultural differences in desirable autonomy are inevitable, differential item
functioning of the scale of cognitive autonomy is expected as part of culture
difference and therefore was not of interest in this study. Therefore, examining
chi-square difference across cultures and genders serves as a preliminary
analysis to delineate the overall feature of the data. In addition, merely
comparing percentages of cognitive autonomy across cultures does not
yield integrated information, that is, whether there are heterogeneous latent
groups (classes) in terms of cognitive autonomy across cultures and genders.
Therefore, latent class analysis (LCA) was selected to assess the overCognitive
Autonomy in Adolescence 23
all patterns underling the domains of cognitive autonomy between cultures
and genders. This statistical procedure allowed for further validation concerning
whether adolescent samples within the same culture should be considered
either homogeneous or heterogeneous in cognitive autonomy. It
also allowed for examinations between identified latent classes across cultures
and genders. We used the statistical analysis software package Mplus
6.1 for a series of latent class analyses with maximum likelihood estimation.
When distinct subgroups (classes) were identified within each cultural
group for each gender, post-hoc group comparisons were used to identify
patterned statistical significance across culture and gender. Specifically,
these comparisons help explain the existence of the latent classes; whether
one class in one cultural/gender group would closely resemble one in their
counter group. Details of this procedure can be found elsewhere (Castle,
Sham, Wessely, and Murray, 1994; Finch and Bronk, 2011). A significant
Satorra-Bentler (Satorra and Bentler, 2001) chi-square value indicates a statistical
difference in the tested parameter between constrained and unconstrained
models. Results from this statistical procedure were presented in
terms of probabilities and represent the likelihood of class membership.
IV. Results
A. Self-Identified Levels of Cognitive Autonomy
Table 3 contains percentages for each of the five subscales of cognitive
autonomy for both American and Taiwanese participants who established
consistent autonomy. Chi-square tests were conducted to evaluate pairwise
differences between genders within each culture group and between cultures
within each gender group. Differences between males and females from
24 調查研究—方法與應用/第28 期
Table 3 Cognitive Autonomy Correlation Matrix for American and
Taiwanese Adolescents
1 2 3 4 5
Americans
1. Evaluative Thinking .259** .647** .406** −.038
2. Voicing Opinions .234** .355** .324** .150*
3. Decision-Making .455** .453** .467** −.080
4. Self-Assessing .421** .272** .434** .038
5. Comparative Validation .123 .223** .157 .250**
Taiwanese
1. Evaluative Thinking .381** .596** .415** −.081
2. Voicing Opinions .267** .377** .363** −.121
3. Decision-Making .690** .379** .506** −.227**
4. Self-Assessing .407** .356** .521** −.015
5. Comparative Validation −.252** −.170* −.259** −.184*
Note. Data for females are presented below the diagonal where data for males are showed above the
diagonal. Comparative Validation was recoded, higher score means greater autonomy.
*p < .05 **p < .01.
the United States existed only for voicing opinions where a significantly
higher percentage ( p < .05) of males (57.9%) than females (46.9%) were
categorized as consistently autonomous in that regard. Taiwanese males
and females differed in percentages of evaluative thinking (males 69.6%,
females 58.3%), and comparative validation (males 14.1%, females 2.6%).
Differences between cultural groups existed in all but one subscale, decision
making, for both males and females. The subsequent Latent Class
Analysis provides additional insights to these differences.
Cognitive Autonomy in Adolescence 25
B. Latent Class Analysis
Latent Class Analysis (LCA) was first conducted for each gender
within the American and Taiwanese adolescent sample. The initial analysis
procedure yielded latent class models defining 1–3 latent classes. All the
models had non-significant Pearson and likelihood ratio Chi-Square values,
which indicated that the data fit the hypothetical models. The rejected
three-class model had the higher Bayesian information criterion (BIC)
value and non-significant Lo-Mendell-Rubin adjusted likelihood ratio test
(LMR) across subsamples, which implied that these three-class models did
not fit significantly better than two-class models whereas two-class models
fitted better than one-class models. Therefore, a two-class model was identified
as a final solution in terms of numbers of latent classes for each cultural
group with the most parsimony and acceptable model fit (for detail
information about model fit indicators, see Lanza, Flaherty, and Collins,
2003). The two classes are referred to as class 1 and class 2. Overall,
except for the domain of comparative validation, participants in the class 1
had improved chances of being highly cognitively autonomous and class 2
had a lower probability of cognitive autonomy. Heterogeneity was evident
between and within the samples using this procedure. Figure one showed
the percentage of participants in each LCA derived class. Preliminary chisquare
tests were conducted for assessing whether there was significant difference
in numbers of participants classified in either class between genders
and cultural group. he difference was found between genders among American
(χ2 = 7.744, df = 1, p = .006) but not among Taiwanese youth. These
result indicated that among American youth, more females than males were
classified into class 1.
26 調查研究—方法與應用/第28 期
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0.25
0.75
0.28
0.72
0.39
0.61
0.35
0.65
Class 1
Class 2
American
Males
Taiwanese
Males
American
Females
Taiwanese
Females
Figure 1 Percentages for each LCA classification of cognitive autonomy
stratified by culture and gender.
Following a previously established analytic approach (see Castle et al.,
1994), further comparisons of the latent class structure proceeded by constraining
the conditional probability of the resembling class of Taiwanese
males to be equal to U.S. males and Taiwanese females to be equal to U.S.
females. Similarly, corresponding classes of females from the U.S. was
constrained to be equal to U.S. males and females from Taiwan to be equal
to Taiwanese males. Accordingly, the following eight models were fitted.
Using a Chi-Square difference test based on loglikelihood values (LL),
scaling correction factors (for maximum likelihood estimation with robust
standard errors) were calculated for each pair of nested models.
M1: Two latent classes for U.S. males and Taiwanese males, totally
unconstrained.
Cognitive Autonomy in Adolescence 27
M2: Two latent classes for U.S. males and Taiwanese males, constrained
class 1 of U.S. males were parallel structure to the class 1 of Taiwanese
males, whereas class 2 of the U.S. males had similar pattern as class
2 of Taiwanese males.
M3: Two latent classes for U.S. females and Taiwanese females, totally
unconstrained.
M4: Two latent classes for U.S. females and Taiwanese females constrained
the conditional probability of being in class 1 in U.S. female sample
is the same as being class 1 in Taiwanese female group.
M5: Two latent classes for U.S. males and females, totally unconstrained.
M6: Two latent classes for U.S. males and females, constrained class
1 of U.S. males were parallel structure to the class 1 of U.S. females whereas
class 2 of the U.S. males had a similar pattern as class 2 of U.S. females.
M7: Two latent classes for Taiwanese males and females, totally
unconstrained.
M8: Two latent classes for Taiwanese males and females constrained
the conditional probability of being in class 1 to being the same across gender
that was the same for class 2.
Table 4 provides the model fit and the Chi-Square difference results.
Overall, from the evidence of significant Chi-Square differences and poor
fit of the constrained models (e.g., significant G2, larger BIC, and smaller
Entropy values), the non-constrained models fit significantly better than
the constrained models with one exception (M5 and M6 had similar model
fit; see Table 4). The latent structure between the U.S. and Taiwanese
males, U.S. and Taiwanese females, and Taiwanese genders were significantly
different from each other, informed by the significant Chi-square
28 調查研究—方法與應用/第28 期
Table 4 Percentage of Established Consistent Autonomy for Each
Domain of Cognitive Autonomy for American and Taiwanese
Adolescents
Evaluative
Thinking
Voicing
Opinions
Decision-
Making
Self-
Assessing
Comparative
Validation
Americans
Males 38.8 57.9 85.8 58.5 23.0
Females 38.1 46.9 81.6 51.0 22.4
χ2 .02 3.95 1.05 1.83 .01
p .896 .047 .306 .176 .914
Taiwanese
Males 69.6 32.1 81.5 43.5 14.1
Females 58.3 23.4 79.7 34.9 2.6
χ2 5.13 3.50 0.20 2.91 16.50
p .023 .062 .653 .088 < .001
Males
Americans 38.8 57.9 85.8 58.5 23.0
Taiwanese 69.6 32.1 91.5 43.5 14.1
χ2 34.99 24.79 1.22 8.25 4.73
p < .001 < .001 .269 .004 .020
Females
Americans 38.1 46.9 81.6 51.0 22.4
Taiwanese 58.3 23.4 79.7 34.9 2.6
χ2 13.64 20.60 .20 8.89 32.94
p < .001 < .001 .654 .003 < .001
Cognitive Autonomy in Adolescence 29
difference tests based on LL values and scaling correction factors (CF)
obtained with the MLR estimator.1 Therefore, the unconstrained models,
M1, M3, and M7 and the constrained model M6 were the final solutions.
The conditional probabilities and corresponding standard error of each
model are presented in Table 5. The difference between American and
Taiwanese youth was more obvious in males than with female adolescents.
American youth had a pattern of similar probabilities in high self-ratings
in voicing opinions. Taiwanese youth had higher probabilities to be high in
evaluative thinking when compared to American teenagers across class
membership and gender. Youth in both cultures, in general, had higher
probability to be high in decision-making but a lower probability in achieving
autonomy in comparative validation. In addition, although American
Table 5 Model Fit Indices and Chi-Square Differences between Pairs
of Models
Model G2 df BIC Entropy LL CF Parameters Δχ2 p
M1 32.96 40 2633.165 0.848 –1248.671 1.020 23
M2 131.05 50 2672.201 0.799 –1297.716 1.149 13 115.089 <.0001
M3 49.08 40 2338.742 0.827 –1102.372 1.022 23
M4 137.63 50 2369.032 0.827 –1146.647 1.097 13 95.782 <.0001
M5 52.365 40 2468.104 0.834 –1167.362 1.027 23
M6 60.069 50 2417.816 0.818 –1171.214 1.059 13 7.818 .6466
M7 29.677 40 2505.631 0.842 –1184.625 1.015 23
M8 60.293 50 2476.951 0.812 –1199.933 1.042 13 31.244 .0005
Note. G2 = Likelihood ration Chi-Square; LL = Loglikelihood; CF = Scaling correction factor.
1. The Chi-square difference was obtained by [−2 × (LLc − LLuc)] ÷ [(#Pc × CFc − #Pc × CFuc) ÷
(#Pc − #Puc)]. LL = Loglikelihood, CF = Scaling correction factor, C = Constrained model,
UC = Unconstrained model, #P = Number of Parameters.
30 調查研究—方法與應用/第28 期
teenagers also had lower probabilities of high autonomy in the comparative
validation domain, the conditional probabilities were still higher than Taiwanese
youth (see Figures 2 and 3).
Gender differences, in terms of latent structure existed within the Taiwanese
sample. Taiwanese females overall had a lower probability in
achieving evaluative thinking when compared to Taiwanese males. More
gender differences were revealed in class 1 than class 2. Males in class 1
had a better chance to be high in evaluative thinking and voicing opinions,
whereas females were better in self-assessing and comparative validation.
On the other hand, females in class 2 had a slightly higher probability in
voicing opinions but a lower probability in comparative validation than
males in the same class. (see Figures 4 and 5). Because different latent
structure between genders with the United State sample were not found, the
Americans_Class 1
Americans_Class 2
Taiwanese_Class 1
Taiwanese_Class 2
Comparative Validation
Self-Assessing
Decision-Making
Voicing Opinions
Evaluate Thinking
1
0.8
0.6
0.4
0.2
0
Figure 2 Conditional probability of being high autonomy in per latent
class for Taiwanese and American males.
Cognitive Autonomy in Adolescence 31
Americans_Class 1
Americans_Class 2
Taiwanese_Class 1
Taiwanese_Class 2
Comparative Validation
Self-Assessing
Decision-Making
Voicing Opinions
Evaluate Thinking
1.0
0.8
0.6
0.4
0.2
0.0
Figure 3 Conditional probability of being high autonomy in per latent
class for Taiwanese and American females.
1
0.8
0.6
0.4
0.2
0
Males_Class 1
Males_Class 2
Females_Class 1
Females_Class 2
Comparative Validation
Self-Assessing
Decision-Making
Voicing Opinions
Evaluate Thinking
Figure 4 Conditional probability of being high autonomy in per latent
class for American males and females.
32 調查研究—方法與應用/第28 期
Males_Class 1
Males_Class 2
Females_Class 1
Females_Class 2
1
0.8
0.6
0.4
0.2
0
Comparative Validation
Self-Assessing
Decision-Making
Voicing Opinions
Evaluate Thinking
Figure 5 Conditional probability of being high autonomy in per latent
class for Taiwanese males and females.
comparison of in numbers of youth classified in either class became meaningful.
As stated previously, female adolescents in the U.S. had a higher
percentage of being classified in class 1 than males, implying among
American youth, more females achieved their autonomous than the opposite
gender.
V. Discussion
The primary purpose of this study was to explore the self-identified
cognitive autonomy levels from male and female adolescents living in Taiwan
and the United States. Speculation continues to grow about the traditional
influence of collectivistic values on the development of young people
in westernized Asian areas such as Taiwan. To provide an initial, in-depth
Cognitive Autonomy in Adolescence 33
picture, LCA was used to compare patterns of cognitive autonomy between
cultures and genders. The use of this analysis provides information about
the five domains at the same time, transcending potentially misleading pairwise
comparisons. In addition to mean differences in cognitive autonomy
between genders within each culture, we found that teenagers from the
United States and Taiwan had meaningfully different latent structures of
conditional probability in cognitive autonomy even though both samples
exhibited two-class patterns of cognitive autonomy development, loosely
represented a high and low autonomy class. Accordingly, these findings
give credence to our general research hypothesis that patterns of cognitive
autonomy would vary in both cultures. These findings relate nicely to
Erikson’s proposition that all cultures experience similar human developmental
issues (e.g., autonomy and identity formation) in the socialization
process, though the ways to foster desired developmental outcomes might
differ among cultures (Erikson, 1963).
A. Descriptive Differences in Cognitive Autonomy
Females, in general, tended to be more dependent than males. Previous
research suggests that adolescent females tend to be more protected and
restricted during the socialization processes associated with transitioning to
adulthood (see Bumpus, Crouter, and McHale, 2001). This process might
work counter to females’ agency in terms of autonomy development. Additionally,
the influence of collectivistic culture may even exaggerate the differences
between genders by seemingly granting males more power and
authority to make strategic and final decisions (see Shon and Ya, 1982).
With recent trends in globalization, however, researchers speculate that Taiwanese
youth have become more westernized and, consequently, are more
34 調查研究—方法與應用/第28 期
likely to be influenced by individualistic ideas resulting in adolescents
maintaining collectivistic values in some areas of their lives while embracing
more individualistic notions in others (Chiu, 2001). In this study, male
participants had higher ratings in evaluative thinking, voicing opinions,
self-assessing, and comparative validations when compared to female adolescents.
This finding supported general impressions concerning the influence
of collectivistic culture on psychosocial development in patriarchal
Chinese societies (e.g., Taiwan). Male adolescents are more likely to be
autonomous than females within these societies.
B. Patterns in Cognitive Autonomy
Observing mean differences of cognitive autonomy outcomes across
cultures is often misleading in research because the lack of a golden measure
for social behaviors that is culturally invariant but also limited in scope
because it fails to summarize patterns across five dimensions of cognitive
autonomy within and between groups. In this study, the very low percentage
scores of both males and females from Taiwan in comparative validation
underscored this point. Although all the scales of the CASE inventory
were considered culturally applicable to Taiwanese youth, it was evident in
this study that comparative validation is valued differently between cultures.
Other studies using this instrument with diverse cultures have found
that comparative validation lags behind the other four scales in successful
completion (Beckert and Bundy, 2008). Indeed, comparative validation
appears to reach completion for most young people during emerging adulthood.
However, Taiwanese youth in this study were significantly behind
their US counterparts in comparative validation. Therefore, we employed
LCA to identify cultural implications for desirable autonomy performances
Cognitive Autonomy in Adolescence 35
among adolescents. The LCA findings draw attention to important differences
in patterns of cognitive autonomy across cultural groups while maintaining
gender difference considerations. The varied conditional probabilities
between cultural and gender groups might reflect the subtle differential
socialization and expectations toward both ideal and desirable cognitive
autonomy within each culture. Nonetheless, similar patterns in terms of
higher probability in achieving decision-making but lower chance in being
autonomous from comparative validation were also found across both cultural
and gender groups.
Cultural Differences. Cultural differences resulted from class membership
comparisons between these Taiwanese and American cultures.
Many teachings and cultural influences can sway young people to be less
likely to provide high self-ratings in areas of cognitive autonomy. Class 1
provided the most obvious differences. Taiwanese youth overall had lower
conditional probabilities of being highly autonomous in voicing opinions,
comparative validations, and self-assessing (especially males) when compared
to American adolescents. Taiwanese youth had higher self-ratings in
evaluative thinking than youth living in the United States. A similar pattern
also manifested between American and Taiwanese youth in class 2.
Taiwanese youth, in general, compared to North American adolescents
were less likely to share their opinions in public. There is a lasting influence
of cultural teachings in Taiwan, such as the old proverbial saying,
“silence is golden.” The emphasis of harmony and respect for hierarchical
authority in Taiwan might also lead these youth, regardless of class membership,
to lower conditional probabilities in achieving comparative validations.
In addition, Taiwanese culture traditionally teaches youth to be humble
and self-effacing in order to maintain harmony in interpersonal relation36
 調查研究—方法與應用/第28 期
ships. Therefore, Taiwanese youth might hesitate to acknowledge their talents
in personal communications or in public (see Markus and Kitayama,
1991). Consequently, self-assessment, as measured by the CASE would not
be culturally appreciated.
An underlying reason that Taiwanese adolescents had better conditional
probabilities of being highly autonomous in evaluative thinking
might be the influence of collectivistic affiliations and Confucius teachings.
Both would strongly recommend and counsel individuals to constantly
reflect on the ability of their intentions and corresponding plans/behaviors
to meet societal expectations and collectivistic interests (see Lam, 1997).
Yet, the same reason might serve as a two-edge sword impeding Taiwanese
youth’s autonomy in comparative validation.
It is worthwhile to note that adolescents in both samples had higher
chances in achieving autonomy in decision-making but low probabilities of
being highly autonomous in comparative validation. American and Taiwanese
youth acquire their decision-making skills through different socialization
processes. However, decision making abilities appear to be similar
across both cultures. Developing decision making abilities may be one area
in which adolescents manifest fewer difficulties. Difficulties in making
decisions, therefore, may serve as an early sign for parents or teachers to
identify youth (age 15 and above) who need assistance within this domain
of autonomy development. Not all young people mature at the same rate.
With partial maturation in manifesting cognitive capability, adolescents (in
general) are more likely to give in to peer pressure and/or subjectively
experiencing an imaginary audience (Elkind, 1985). Therefore, they tend
to be overly concerned about how they appear to others in an effort to
receive others’ approval. The obvious implication of this trajectory includes
Cognitive Autonomy in Adolescence 37
Table 6 Conditional Probabilities and Standard Errors of Being Highly Autonomous in per Latent Class
for American and Taiwanese Male and Female Adolescents
Evaluate
Thinking
Voicing
Opinions
Decision-
Making
Self-Assessing
Comparative
Validation
p SE p SE p SE p SE p SE
Class 1
US
Males 0.533 0.066 0.677 0.053 0.982 0.018 0.734 0.048 0.258 0.041
Females 0.540 0.065 0.544 0.056 1.000 0.000 0.763 0.088 0.266 0.049
Taiwan
Males 0.906 0.033 0.469 0.053 0.961 0.020 0.591 0.054 0.155 0.034
Females 0.788 0.060 0.322 0.049 1.000 0.000 0.763 0.055 0.266 0.015
Class 2
US
Males 0.000 0.000 0.319 0.085 0.525 0.123 0.186 0.113 0.152 0.077
Females 0.107 0.059 0.341 0.092 0.500 0.123 0.075 0.051 0.153 0.056
Taiwan
Males 0.242 0.096 0.000 0.000 0.500 0.087 0.098 0.050 0.113 0.046
Females 0.208 0.067 0.074 0.040 0.424 0.112 0.084 0.049 0.041 0.029
38 調查研究—方法與應用/第28 期
the potential for illicit drug use, unprotected sexual activity, and other negative
risk taking behaviors. Early intervention with decision making skills
might be logical first step in avoiding some of these pitfalls.
Gender Differences. There were some differences in probabilities for
cognitive autonomy between genders within each cultural group. Overall,
gender role expectations could explain the observed differences between
males and females across cultures. It was not surprising that gender differences
within Taiwanese participants were more obvious than American
sample. This difference makes sense when considering the expectation of
respect for hierarchical authority that is more pronounced for females than
males in Taiwan. The self-reflected inability to voice opinions for both
classes of Taiwanese females speaks to this point. Additionally, class 2 Taiwanese
females had much lower probability in comparative validations.
The findings in this study could provide additional support to the idea that
socialization processes influencing females tend to be more interdependent
(Kashima et al., 1995).
There were no structural differences in Class 1 between U.S males and
females, and therefore it should be noted that females in the United States
were more likely than males in the United States to be classified in Class 1.
This finding is yet another indication that females in western societies are
increasingly encouraged to establish a sense of self and develop autonomously
across the domains of cognitive autonomy. While this is encouraging
for females, a new trend seems to be developing in that males in western
cultures are postponing important life transitions including issues of
cognitive autonomy.
Cognitive Autonomy in Adolescence 39
C. Limitations
There are some perceivable limitations to this study. First, although
the sample sizes for this study were sufficient for the advanced categorical
variable analysis, a larger sample would be desirable. Additionally, adolescents
in this study were from limited areas in both Taiwan and the United
States and do not represent the entire population of adolescents in either
country. Finally, applying any western latent construct and an instrument
developed in the United States with an Eastern (Taiwanese) sample might
restrict comparisons especially because validity information was limited.
This was especially true for the subscale of comparative validation. More
research is needed to validate the applicability of this construct within Taiwanese
culture.
D. Implications and Future Research
The use of LCA strengthened the analysis of these data. The LCA and
the resulting post-hoc model comparisons provided further validation,
beyond mean differences, by capitalizing on factorial analysis of variance.
Knowing that cultures differ, in terms of the differential patterns of participants,
may be beneficial in future research because each culture group may
emphasize certain domains of cognitive autonomy over the others. Given
that each cultural group may have differential developmental patterns, comparisons
and interpretations of mean differences might be more appropriate
and less misleading within a cultural group rather than between cultural
groups.
Gender differences that went beyond mean differences were also found
in latent structures. Differences between genders in cognitive autonomy
40 調查研究—方法與應用/第28 期
were expected because of gender role expectations across societies and cultures.
However, equality, in terms of treatments and standards between
genders, should take the effect of differential gender traits into consideration.
In other words, equality is not easily accomplished by any arbitrary
standards to foster androgyny. Rather, the process of promoting gender
equality should develop in an environment that provides “goodness of fit”
for both genders in order to help them reach their potential in the development
of cognitive autonomy.
This study was a good first step in examining Erickson’s idea that adolescents
in all cultures will go through identity and autonomy development
but that cultural variations should be expected. Fostering balanced cognitive
autonomy, while considering cultural expectations, can promote cognitive
autonomy across cultures. In addition, adolescents in general, had low
achievement in the comparative validation domain; this is especially concerning
because of adolescents’ vulnerability to negative peer pressure and
risky behaviors. Accordingly, a supportive “goodness of fit” environment
could provide the culturally appropriate assistance that adolescents need to
develop further within this domain. Such an environment, considering cultural
expectations, may include opportunities for adolescents to think aloud,
listen to inductive reasoning strategies, share their experiences and thoughts
with others, and receive assistance from adults.
In the future, researchers could consider applying a longitudinal
design, recruiting representative samples, and assessing convergent validity
by adding other measurements of similar constructs (e.g., behavioral and
emotional autonomy) to the study from both Taiwan and the United States.
This would provide insights to the timing of achieving autonomy and patterns
of developmental trajectories in cognitive autonomy across cultures.
Cognitive Autonomy in Adolescence 41
Results from this study provide a promising first step toward understanding
further the implications and utility of using western models with eastern
populations.
REFERENCES
Allen, J. P., P. Marsh, C. McFarland, K. B. McElhaney, D. J. Land, K. M. Jodl, et al.
2002 “Attachment and Autonomy as Predictors of the Development of Social
Skills and Delinquency during Midadolescence,” Journal of Consulting and
Clinical Psychology 70: 56–66.
Allen, J. P., S. T. Hauser, K. L. Bell, and T. G. O’Connor
1994 “Longitudinal Assessment of Autonomy and Relatedness in Adolescent-Family
Interactions as Predictors of Adolescent Ego Development and Selfesteem,”
Child Development 65(1): 179–194.
Anderson, R. A., L. Worthington, W. T. Anderson, and G. Jennings
1994 “The Development of an Autonomy Scale,” Contemporary Family Therapy
16(4): 329–347.
Beckert, T.
2007 “Cognitive Autonomy and Self Evaluation in Adolescence: A Conceptual
Investigation and Instrument Development,” North American Journal of Psychology
9: 579–594.
2012 “Autonomy and Its Assessment,” pp. 241–247 in R. J. Levesque (ed.), Encyclopedia
of Adolescence. New York, NY: Springer.
Becket, T. and R. Bundy
2008 “Measuring Cognitive Autonomy in Residential Treatment Center Youth,”
Poster Session Presented at the Society for Research on Adolescence, Chicago,
IL.
Bukowski, W. M. and A. F. Newcomb
1983 “The Association between Peer Experiences and Identity Formation in Early
Adolescence,” Journal of Early Adolescence 3(3): 265–274.
Bumpus, M. F., A. C. Crouter, and S. M. McHale
2001 “Parental Autonomy Granting during Adolescence: Exploring Gender Differences
in Context,” Developmental Psychology 37(2): 163–173.
42 調查研究—方法與應用/第28 期
Castle, D. J., P. C. Sham, S. S. Wessely, and R. M. Murray
1994 “The Subtyping of Schizophrenia in Men and Women: A Latent Class Analysis,”
Psychological Medicine 24: 41–51.
Chattopadhway, A., and R. Marsh
1999 “Changes in Living Arrangement and Familial Support for the Elderly in Taiwan:
1963–1991,” Journal of Comparative Family Studies 30: 523–537.
Chiu, C.
2001 “Normative Expectations of Social Behavior and Concern for Members of
the Collective in Chinese Society,” The Journal of Psychology 124: 103–111.
Cicchetti, D. and F. A. Rogosch
2002 “A Developmental Psychopathology Perspective on Adolescence,” Journal
of Consulting and Clinical Psychology 70: 6–20.
Deaton, A. and C. Paxson
2000 “Growth, Demographic Structure, and National Saving in Taiwan,” Population
and Development Review 26: 141–173.
Elkind, D.
1985 “Egocentrism Redux,” Developmental Review 5: 218–226.
2001 The Hurried Child (3rd ed.). Cambridge, MA: Da Capo.
Erikson, E. H.
1963 Childhood and Society (2nd ed.). New York: W. W. Norton.
Feldman, S. S. and D. A. Rosenthal
1991 “Age Expectations of Behavioral Autonomy in Hong Kong, Australian and
American Youth,” International Journal of Psychology 26: 1–23.
Finch, W. H. and K. C. Bronk
2011 “Conducting Confirmatory Latent Class Analysis Using Mplus,” Structural
Equation Modeling 18: 132–151.
Freud, A.
1958 “Adolescence,” Psychoanalytic Study of the Child 13: 255–278.
Fuligni, A. J., T. Yip, and V. Tseng
2002 “The Impact of Family Obligations on Daily Activities and Psychological
Well-being of Chinese American Adolescents,” Child Development 73(1):
302–314.
Fuligni, A. J., V. Tseng, and M. Lam
1999 “Attitudes toward Family Obligations among American Adolescents with
Asian, Latin American, and European Backgrounds,” Child Development
Cognitive Autonomy in Adolescence 43
70(4): 1030–1044.
Gilligan, C.
1977 “In a Different Voice: Women’s Conceptions of Self and of Morality,” Harvard
Educational Review 47(4): 481–517.
Herman-Giddens, M. E.
2006 “Recent Data on Pubertal Milestones in United States Children: The Secular
Trend toward Earlier Development,” International Journal of Andrology
29(1): 241–246.
Hui, C. H. and C. Yee
1994 “The Shortened Individualism-collectivism Scale: Its Relationship to Demographic
and Work-Related Variables,” Journal of Research in Personality 28:
409–424.
Kashima, Y., S. Yamaguchi, U. Kim, S. Choi, M. J. Gelfand, and M. Yuki
1995 “Culture, Gender, and Self: A Perspective from Individualism-Collectivism
Research,” Journal of Personality and Social Psychology 69: 925–937.
Lam, C. M.
1997 “A Cultural Perspective on the Study of Chinese Adolescent Development,”
Child and Adolescent Social Work Journal 14: 95–113.
Lanza, S. T., B. P. Flaherty, and L. M. Collins
2003 “Latent Class and Latent Transition Analysis,” pp. 663–685 in J. A. Schinka
and W. F. Velicer (eds.), Handbook of Psychology: Vol. 2. Research Methods
in Psychology. Hoboken, NJ: Wiley.
Lee, C., T. Beckert, T. Goodrich
2010 “The Relationship between Individualistic, Collectivistic, and Transitional
Cultural Value Orientation on Adolescents’ Autonomy and Identity Status,”
Journal of Youth and Adolescence 39(8): 882–893.
Lee, E.
1996 “Chinese Families,” pp. 249–267 in M. McGoldrick, J. Giordano, and J. K.
Pearce (eds.), Ethnicity and Family Therapy (2nd ed.). New York, NY: The
Guilford Press.
Markus, H. R. and S. Kitayama
1991 “Culture and the Self. Implications for Cognition, Emotion, and Motivation,”
Psychological Review 2: 224–253.
McElhaney. K. B. and J. P. Allen
2001 “Autonomy and Adolescent Social Functioning: The Moderating Effect of
44 調查研究—方法與應用/第28 期
Risk,” Child Development 72(1): 220–235.
Meeus, M., J. Iedema, G. Maassen, and R. Engels
2005 “Separation-Individuation Revisited: On the Interplay of Parent-Adolescent
Relations, Identity and Emotional Adjustment in Adolescence,” Journal of
Adolescence 28: 89–106.
Nelson, C., K. Thomas, and M. de Hann
2006 “Neural Bases of Cognitive Development,” pp. 3-57 in W. Damon and R.
Lerner (eds.), Handbook of Child Psychology (6th ed.). New York: Wiley.
Noom, M. J., M. Deković, and W. H. J. Meeus
1999 “Autonomy, Attachment, and Psychological Adjustment during Adolescence:
A Double-Edge Sword?” Journal of Adolescence 22: 771–783.
Noom, M. J., M. Deković, and W. Meeus
2001 “Conceptual Analysis and Measurement of Adolescent Autonomy,” Journal
of Youth and Adolescents 30(5): 577–595.
Phinney, J. S., A. Ong, and T. Madden
2000 “Cultural Values and Intergenerational Values Discrepancies in Immigrant
and Non-Immigrant Families,” Child Development 71(2): 528–539.
Satorra, A. and P. M. Bentler
2001 “A Scaled Difference Chi-square Test Statistic for Moment Structure Analysis,”
Psychometrika 66(4): 507–514.
Sheldon, K. M., A. J. Elliot, R. M. Ryan, V. Chirkov, Y. Kim, C. Wu, M. Demir, and Z. Sun
2004 “Self-Concordance and Subjective Well-Being in Four Cultures,” Journal of
Cross-cultural Psychology 35: 209–223.
Shon, S. and D. Ya
1982 “Asian Families,” pp. 208–228 in M. McGoldrick, J. Pearce, and J. Giordano
(eds.), Ethnicity and Family Therapy. New York: The Guilford Press.
Spear, H. J. and P. Kulbok
2004 “Autonomy and Adolescence: A Concept Analysis,” Public Health Nursing
21(2): 144–152.
Xia, Y. P., X. Xie, Z. Zhou, J. DeFrain, W. H. Meredith, and R. Combs
2004 “Chinese Adolescents’ Decision-Making, Parent Adolescent Communication
and Relationships,” Marriage and Family Review 36(1): 119–145.
Yeh, K., Y. Liu, H. Huang, and Y. Yang
2007 “Individuating and Relating Autonomy,” pp. 123–146 in J. Liu, C. Ward, A.
Bernardo, M. Karasawa, and R. Fischer (eds.), Casting the Individual in
Cognitive Autonomy in Adolescence 45
Societal and Cultural Contexts (2nd ed.). Korea: Kyoyook-Kwahak-Sa Publishing.
Yeh, K. and Y. Yang
2006 “Construct Validation of Individuating and Relating Autonomy Orientations
in Culturally Chinese Adolescents,” Asian Journal of Social Psychology 9:
148–160.
Zimmer-Gembeck, M. J. and W. A. Collins
2003 “Autonomy Development during Adolescence,” pp. 175–204 in G. R. Adams
and M. D. Berzonsky (eds.), Blackwell Handbook of Adolescence. Malden,
MA: Blackwell Publishing.

AD1AD2AD3
cron web_use_log