• adolescent health;
  • quality of life;
  • questionnaire;
  • Taiwan;
  • validation


  1. Top of page
  2. Abstract

Abstract  This study was to evaluate the psychometric properties of the Taiwanese quality of life questionnaire for adolescents and the factors affecting the quality of life of Taiwanese adolescents. The survey involved 5538 junior high school students, aged 13–15 years. An initial 90-item questionnaire was shortened to 38 items by means of principal component analyses. Quality of life assessment involved seven factors: family, residential environment, personal competence, social relationships, physical appearance, psychological well-being, and pain. The rate of missing data was low. The Cronbach α coefficient remained above the 0.75 threshold criterion for the global scale and seven subdomain scales. A lower quality of life score was evident for female adolescents in higher grades in school, those living with a single parent or other relatives, and those living in rural areas. This 38-item questionnaire should serve as a reliable tool for future studies.


  1. Top of page
  2. Abstract

Adolescent health is increasingly being recognized as an important facet of overall public health.1 Accordingly, the interest in measuring the quality of life (QOL) in adolescents is increasing. Adopting QOL approaches might help to understand adolescents’ health and establish policy to promote their health in Taiwan. Such QOL evaluations have been constrained by the lack of a generic self-administered instrument that accurately reflects the Taiwanese adolescent's point of view. Developing a reliable and valid QOL questionnaire for adolescents in Taiwan is, therefore, an important and necessary step.

The Chinese version of the quality of life questionnaire for adolescents (QOLQA) was developed in China on the basis of the QOL project of the World Health Organization.2,3 The first quantifiable version – a 70-item questionnaire divided into five domains of physical, psychological, independence, social relationship, and environment – originated in Japan. When the questionnaire was tested in China, the Cronbach α was 0.92 for the entire 70 items and ranged from 0.70 to 0.87 for each of the five domains. In a pilot study, we determined that the Cronbach α ranged from 0.16 to 0.87 for these five domains, which are not satisfactory. Thus, the present study was undertaken to develop and test the reliability of the Taiwanese QOL questionnaire for adolescents (TQOLQA).

There is widespread agreement that QOL is a multidimensional  concept  that  encompasses psychological, physical, and social well-being. We planned this questionnaire to cover the important domains for adolescents, fitting the personal (physical and mental), interpersonal (family function, intimate friends, and social networks), external (income and housing), and global (macro environment) spheres that Lindstrom proposed.4 For example, it is well known that the family is highly valued in Chinese society. Indeed, several studies have shown that a single-parent family is a strong risk factor for adolescent childbearing in Taiwan.5,6 Therefore, the QOL assessment also considered the influence of single-parent versus two-parent families.


  1. Top of page
  2. Abstract


This study was conducted at three public junior high schools in Taiwan. Two of the schools were in urban areas (where the majority of Taiwanese citizens dwell), and the remaining school was in a rural area. The choice of junior high schools reflected the fact that over 90% of 13–15-year-old teenagers from diverse economic backgrounds attend these schools, according to the Ministry of Education in Taiwan. The remainder of this age group attends private schools. The Ministry of Education also prescribes specific standards for curriculum, instructional material and equipment for all junior high schools under its jurisdiction. The choice of public junior high schools, thus, provided the greatest sampling diversity and a relatively homogeneous school environment.

Development of a new questionnaire: item generation and reduction

The original 70 items in the Chinese version of the QOLQA were pooled with 20 new items in the first-phase questionnaire. The majority of the new items addressed the relationships of the adolescents with parents, teachers and classroom learning. These choices reflected expressed concerns of Taiwanese adolescents.

Next, the 90 items were reduced to a more workable number of questionnaire items. To accomplish this, a random sample of 515 adolescents from the above three schools were selected. The age and gender of this selected population did not differ from the whole population. Principal component analyses with varimax rotation was used to determine the underlying factor structure of the initial 90-item questionnaire. Only those factors with an eigenvalue > 1.5 upon the application of Kaiser's ‘eigenvalues greater than 1’ rule and scree plot test7 were retained. Items with a loading of < 0.5 on any of the factors were omitted. After confirming the domain content, each domain was named according to the meaning of its constitutive items. The final version of TQOLQA (Table 1) was comprised of 38 items.

Table 1. List of the 38 items of the Taiwanese Quality Of Life Questionnaire For Adolescents
  1. PC, personal competence; PW, psychological well-being; PA, physical appearance; RE, residential environment; SR, social relationship.

 1Do you worry about pain or discomfort?Pain 1
 2Do you have any difficulty in managing or coping with pain or discomfort?Pain 2
 3Does your pain or discomfort interfere with things you need to do?Pain 3
 4Can you concentrate?PC 1
 5Do you feel inferior because of your appearance?PA 1
 6Do you feel uneasy about any part of your body or physical appearance?PA 2
 7Are you upset?PW 1
 8Does the feeling of depression or sorrow interfere with your daily activities?PW 2
 9Do you worry about depression?PW 3
10Do you have any difficulty in performing your daily activities?PW 4
11Do you feel safe in your daily life?RE 1
12Do you feel safe and protected in your home?RE 2
13Is your home cozy?RE 3
14Do you like where you live?RE 4
15Is your living environment healthy?RE 5
16Can you accept your physical appearance?PA 3
17Can you finish your daily affairs?PC 2
18Are your friends reliable when you need them?SR1
19Does your home fulfill your needs?RE6
20Does your family have enough money?RE7
21Are you satisfied with your quality of life?RE8
22Do you satisfy your abilities?PC 3
23Are you satisfied with your physical appearance?PA 4
24Are you satisfied with the support you receive from friends?SR 2
25Are you satisfied with the help and support you receive from your family?Family 1
26Do you have friends with whom you can have fun and talk to about anything?SR 3
27Do your friends consider you someone whom they can count on for support?SR 4
28Do you have a friend who understands you well?SR 5
29Are you satisfied with your memory?PC 4
30Do you think you have a good relationship with your parents?Family 2
31Are you satisfied with your family atmosphere?Family 3
32Do you think it is important for you to have an intact family?Family 4
33Are you satisfied with your parents’ attitude towards you?Family 5
34Are you satisfied with your relationship with your parents?Family 6
35Do your parents affirm your efforts?Family 7
36Do you think you can finish your school work?PC 5
37Can you learn normally?PC 6
38How do you rate your learning ability?PC 7

The Taiwanese Quality of Life Questionnaire for Adolescents

This new questionnaire is rated on a 5-point Likert scale (‘very much’, ‘quite a bit’, ‘average’, ‘not much’, and ‘not at all’) ranging from 1 to 5 for each item. After the raw scores were converted for the reverse questions, higher QOLOA scores reflected a better QOL.

The questionnaire enquired as to the respondent's experiences over the preceding 2 weeks. The items were grouped into dimensions after principal component analysis with varimax rotation. Each scale is then standardized so that they each range from zero (lowest level of functioning) to 100 (highest level).

Survey procedures

A letter describing the study objectives and methods was mailed to the school principals of the selected schools for their approval. Students in these schools (n = 5538) then answered a self-administered QOLQA with basic demographic data. The students also reported their illnesses, if any. The students were classified as having a chronic health problem (e.g. asthma, epilepsy), an acute health problem (acute respiratory tract infection), or no health difficulty.


Statistical analysis was performed using the Statistical Products of Service Solution (SPSS) software, version 11 (SPSS 11.0 for Windows; SPSS, Inc., Chicago, IL). Differences between mean values or the percentages for two or three groups (three schools, sex, grade, health status and living environment) were evaluated using the Student t-test, a one-way anova with Bonferroni correction, or the χ2 test.

Principal component analysis with varimax rotation was used to examine the factor structure of the questionnaire. Reliability was assessed by using the Cronbach α coefficient as a measure of internal consistency. We hypothesized that the subjects who were ill or living with single parent or other relatives had a lower QOL.

A P-value < 0.05 was indicative of a statically significant difference, and the significance of post-hoc comparisons after the Bonferroni modification was accepted at the P < 0.006 level.


  1. Top of page
  2. Abstract

Descriptive statistics for the students

In total, 5538 students in the seventh to ninth grades completed the questionnaire. The response rate was 89.4%.

Table 2 shows the descriptive statistics of the study population in the three sampled schools. The rural school had significantly fewer male students, a higher percentage of students living with a single parent or other relatives (mostly grandparents), and more students with acute illness (P < 0.001).

Table 2. Descriptive statistics of the study population
 Total No. (n = 5538)Urban Junior High School A (n = 1002)Urban Junior High School B (n = 3081)Rural Junior High School C (n = 1455)χ2P-value
  • χ2 test.

Sex     7.40.025
Grade    54.40.000
Living with    48.80.000
 Two parents88.3%91.5%89.3%83.8%  
 Single parent 9.2% 7.0% 8.7%11.8%  
 Other relatives 2.5% 1.5% 2.0% 4.4%  
Sickness    91.20.000
 Acute disease15.0%15.2%11.5%22.4%  
 Chronic disease 4.9% 4.5% 5.2% 4.6%  

Factor analysis of study measures

Principal-component factor analysis with varimax rotation of the final 38-item TQOLQA, revealed seven factors that accounted for 63.5% of the variance (Table 3). The seven factors were family, residential environment, personal competence, social relationships, physical appearance, psychological well-being, and pain. The family domain involved the subjects’ relationship with their parents, their family atmosphere, and the support they received from their family. The residential environment domain consisted of the subjects’ living conditions and material well-being. The personal competence domain addressed the confidence the subject has in his or her academic performance and other abilities. The social relationship domain consists of friendship and social support. The physical appearance domain concerned the subject's feelings about his or her appearance. The psychological well-being domain consisted of emotional symptoms. The pain domain addressed the feeling of pain.

Table 3. Factor analysis after varimax rotation of the final Taiwanese Quality Of Life Questionnaire For Adolescents
  1. Factor loading < 0.10 not reported.

  2. PC, personal competence; PW, psychological well-being; PA, physical appearance; RE, residential environment; SR, social relationship.

Family 1 0.630.3250.2010.2360.162  
Family 2 0.8180.2180.141    
Family 3 0.8070.2870.1360.1130.1010.111 
Family 4 0.6170.2160.1470.108   
Family 5 0.7990.2160.1680.115   
Family 6 0.8350.2280.1340.105   
Family 7 0.6080.1980.3740.144   
RE 1 0.1450.5580.2220.1590.1330.2050.103
RE 2 0.1380.7090.1760.131   
RE 3 0.2500.7900.1640.120   
RE 4 0.3120.7720.1400.106   
RE 5 0.1990.5690.2090.132   
RE 6 0.2990.7300.1700.1710.123  
RE 7 0.2850.5310.2020.1650.148  
RE 8 0.3680.5790.2190.2210.1420.121 
PC 1 0.1220.1910.6430.133   
PC 2 0.1180.3110.5750.1520.141  
PC 3 0.1800.1800.5200.1970.4090.134 
PC 4 0.167 0.5840.1270.1840.140 
PC 5 0.1790.2160.7780.108   
PC 6 0.2070.2490.7760.134   
PC 7 0.1860.1410.7750.1230.139  
SR 1 0.1110.2800.1640.700   
SR 2 0.2280.1980.1520.7330.128  
SR 3 0.1390.1210.1450.732   
SR 4 0.1130.1140.748   
SR 5 0.1350.1120.1450.770   
PA 1    0.8110.234 
PA 2 0.113   0.7260.3220.146
PA 3 0.3000.3240.1710.640  
PA 4 0.16720.2180.2830.1970.740  
PW 1 0.130   0.2730.7100.132
PW 2 0.135   0.8070.212
PW 3    0.1370.8290.173
PW 4 0.1140.1900.1000.1010.5350.201
Pain 1     0.1740.816
Pain 2     0.1920.823
Pain 3     0.3290.733
Variance explained (%)

Item-level analysis

Table 4 shows the psychometric results in our study population. The rate of missing data ranged from 0.2% to 1.2%. The rate on most surveys (81.1%) was less than 0.5%. The most frequently missing item was question 2: ‘Do you have any difficulty in managing or coping with pain or discomfort?’ Results in all domains and the total questionnaire demonstrated high internal consistency.

Table 4. Psychometric results from the Taiwanese Quality Of Life Questionnaire For Adolescents in the study population
Dimension (Numbers of item)Missing Data Rate (%)MeanSDHigh (%)Low (%)Reliability
  • Range, 1–5;

  • Cronbach α.

  • PC, personal competence; PW, psychological well-being; PA, physical appearance; RE, residential environment; SR, social relationship.

Family (7)0.464.520.50.3 4.20.91
RE (8)0.460.717.90.1 2.50.89
PC (7)0.455.516.90.2 0.90.87
SR (5)0.460.018.10.6 2.00.78
PA (4)0.366.418.30.4 5.70.82
PW (4)0.470.916.40.1 6.80.79
Pain (3)1.073.316.

The effect of sex, age, health status and living environment

Table 5 presents the seven subdomain scores across the different health status scores. The TQOLQA scores by sex, grade, living environment, and family structure are reported in Table 6.

Table 5. Global and domains of Taiwanese Quality Of Life Questionnaire For Adolescents by health status
Health StatusFamilyREPCSRPAPWPain
  • Differs from the healthy group;

  • ‡Differs from the acute-illness group;

  • §

    §Differs from the chronic-illness group.

  • PC, personal competence; PW, psychological well-being; PA, physical appearance; RE, residential environment; SR, social relationship.

Healthy (n = 4281)65.0 ± 20.361.3 ± 17.856.1 ± 16.860.4 ± 18.067.2 ± 17.971.8 ± 16.0§74.2 ± 15.7§
Acute illness (n = 792)63.0 ± 21.158.7 ± 18.253.1 ± 16.859.1 ± 18.463.5 ± 19.667.4 ± 17.669.6 ± 16.9
Chronic illness (n = 258)62.4 ± 22.259.5 ± 18.354.4 ± 17.957.6 ± 19.263.7 ± 20.067.8 ± 17.070.7 ± 17.8
Table 6. Global and domains’ scores of the Taiwanese Quality Of Life Questionnaire For Adolescents by sex, grade and living environment
  • *

    P < 0.006.

  • Differs from grade 7;

  • ‡Differs from grade 8;

  • §

    §Differs from grade 9;

  • ¶Differs from subjects who lived with two parents;

  • †Differs from subjects who lived with single parent;

  • ‡‡

    ‡‡Differs from subjects who lived with other relatives.

  • PC, personal competence; PW, psychological well-being; PA, physical appearance; RE, residential environment; SR, social relationship.

 Male (n = 2924)63.9 ± 20.561.7 ± 18.3*55.6 ± 17.558.6 ± 18.4*69.8 ± 17.8*74.1 ± 16.7*76.5 ± 16.2*
 Female (n = 2481)65.3 ± 20.659.6 ± 17.4*55.3 ± 16.161.7 ± 17.7*67.5 ± 18.0*67.0 ± 15.1*69.4 ± 15.2*
 t-value 2.5−4.2−0.7 6.4−14.9−16.5−16.7
 7 (n = 1711)68.8 ± 20.0§63.3 ± 18.3§58.4 ± 17.3§61.9 ± 17.9§69.0 ± 18.3§71.8 ± 16.3§72.8 ± 16.2
 8 (n = 1842)64.6 ± 20.9§61.3 ± 18.1§56.2 ± 16.9§60.6 ± 18.2§66.3 ± 18.3§72.2 ± 16.0§74.6 ± 15.8§
 9 (n = 1852)60.5 ± 20.157.8 ± 16.952.0 ± 15.857.6 ± 18.064.3 ± 18.068.7 ± 16.872.3 ± 16.3
 Urban (n = 3987)65.8 ± 20.6*62.4 ± 17.9*56.8 ± 17.3*61.1 ± 18.6*67.8 ± 18.3*71.8 ± 16.3*73.8 ± 16.2*
 Rural (n = 1418)60.8 ± 20.1*56.0 ± 17.0*51.6 ± 14.9*56.9 ± 16.5*62.7 ± 17.8*68.3 ± 16.4*71.6 ± 15.9*
 t-value 8.1 9.3 7.0 4.4
Living with
 Two parents (n = 4743)65.9 ± 20.261.6 ± 17.856.0 ± 16.860.4 ± 17.966.7 ± 18.071.2 ± 16.373.4 ± 16.0
 Single parent (n = 492)53.8 ± 20.554.9 ± 17.652.2 ± 16.957.5 ± 19.464.7 ± 20.068.4 ± 17.271.9 ± 16.8
 Other relatives (n = 137)56.7 ± 21.256.0 ± 18.848.9 ± 15.656.9 ± 20.762.5 ± 20.268.4 ± 17.973.1 ± 16.6
 F-value93.737.122.4 7.9 6.0 8.6 2.0

The results showed the expected direction of scores for the health status and family structure. The subgroups analysis showed students who lived in a single parent household had significantly lower scores in family and residential environment subdomains. These findings were consistent in all three schools. Students with acute or chronic illness had significantly reduced QOL scores in the psychological well-being and pain subdomains. The seven subdomain scores in the male adolescents were significantly higher than those of female adolescents, except for scores in the family and personal competence domains. In the subgroup analysis of the gender differences in these three schools, male adolescents consistently displayed significantly higher scores of psychological well-being and pain subdomains. Rural students, as well as those students in the higher grades, had lower scores in the seven subdomains. In subgroup analysis of grade effect between urban and rural schools, the urban students who were in the higher grades had lower scores in all subdomains. Nevertheless, the rural students who were in the higher grades had lower scores in the seven subdomains except social relationship and physical appearance.


  1. Top of page
  2. Abstract

The present study was conducted to examine the validity and the overall psychometric performance of the TQOLQA. The outcomes of these assessments are favorable, and indicate that the evaluation may be productively used in QOL studies.

This questionnaire was developed with the aim of producing an instrument that was sufficiently short and simple in format to be feasible for use in a wide range of health assessments. The observation that few questions were left unanswered attests to the acceptance of the questionnaire's structure and length.

In the present study there was a continuous decrease in QOL scores with grade among Taiwanese adolescents. This mirrors the findings with Japanese, British, Nordic, and Canadian adolescents, but differs from the pattern produced with Chinese adolescents.3,8–10

In Taiwan, the ninth grade is an important turning point for adolescents. Most ninth grade students must take an entrance examination to attend senior high school or other occupational schools. Much parental attention and focus is paid to this examination. As a result, the affected adolescents experience a great deal of stress. This might be at the root of the observed relationship between age and QOL in the present survey.

We found that female adolescents aged 13–15 years reported significantly lower QOL, especially in the psychological well-being and pain subdoamins. There are significant gender differences in self-esteem as a consequence of different patterns of social roles and interpersonal experience that characterize men and women from their earliest years.11 These differences might further reflect their subjective perception of QOL. Previous studies also demonstrated that the perception of pain  differed  between  the  genders.12 The  results  of our studies are in line with the previous studies in adults.11,12

In studies by Apajasalo et al.,13,14 healthy female adolescents aged 12–15 years reported a significantly lower status on the dimensions of vitality, sleeping, physical appearance, and depression, but those aged 8–11 years did not. In the surveys conducted in China, Japan, and the United Kingdom, the scores did not significantly differ by sex.3,8 Nevertheless, Nordic girls aged 12–17 years had a tendency to have a higher QOL.9 Despite the statistically significant differences, the absolute difference in the scores in these studies (including this one) was small, and the clinical importance is questionable. The other explanation for this controversy is the cultural effect. Although the population in China and Taiwan are both ethnic Chinese, the two sides across Taiwan Strait have had no official contact since 1949. Private exchanges between the two sides began in 1987 after the Taiwanese government began allowing citizens to visit their relatives in China, and have increased rapidly since then. Nevertheless, both societies differ due to the long-term separation.

The present study has documented lower QOL scores in urban and rural adolescents. This contrasts with the results of a survey of 6–14-year-old children in the United Kingdom.8 However, our results are consistent with several reports conducted in Taiwan which documented the disadvantage of rural communities for adolescents which included a greater prevalence of depression, teenage pregnancies, single-parent or ‘broken families’, and substance abuse in rural areas versus urban communities.15–18 More job opportunities in cities attract a lot of people from rural areas. Therefore, young and middle-aged rural parents are working in cities and leaving their children in rural areas with their grandparents. In addition, the families with better socioeconomic status choose to migrate to urban areas. In our survey, we also found that a higher percentage of  rural  adolescents  lived  with  a  single  parent  or  with a  foster  parent.  Both  the  adverse  rural  environment in Taiwan and the selective migration of families of higher socioeconomic status to the cities might explain this urban–rural difference.

This study had three possible methodological limitations. First, some causal items that are not highly correlated with other items might be omitted in factor analysis.19,20 Nevertheless, only three items in our questionnaire related to symptoms and almost all of them are indicator variables. Thus, this bias might be minimal. Second, very few of the students in our sample population were afflicted with a severe disease. Of course, the reported health status was self-assessed and so is subjective. While the differences between the healthy, acute and chronic groups were small; the observed relationship between health status and reported QOL was not unexpected. A final limitation concerns the selection of only two urban junior high schools and a single rural junior high school among the 708 such schools in Taiwan in 2001. However, because of the high questionnaire response rate and the homogeneity of the junior high school system in Taiwan, we believe that this student population might provide a representative sample of 13–15-year-old Taiwanese adolescents.


  1. Top of page
  2. Abstract

In conclusion, the present study demonstrates that the psychometric testing with the TQOLQA survey satisfies most conventional psychometric criteria. On the basis of these experiences, the TQOLQA is a promising tool for the evaluation of QOL of adolescents. Further studies will be necessary to validate this potential.


  1. Top of page
  2. Abstract

The authors wish to thank all of the teachers and students at Yun-Lin, Shih-Pai and Dai-Zh Junior High Schools for their great help, and Dr Xiangdong Wang for providing the original Chinese version of the QOL questionnaire for adolescents. This study was supported in part by grants from the National Science Council (NSC-90–2314-B-010–026, NSC-91–2314-B-075–042).


  1. Top of page
  2. Abstract
  • 1
    Raphael D, Brown I, Rukholm E et al. Adolescent health: moving from prevention to promotion through a quality of life approach. Can. J. Public Health 1996; 87: 8183.
  • 2
    Ma H, Wang XD, Shihfuku N et al. Validity and reliability of QOL questionnaire for adolescents. Chinese Mental Health J. 1999; 13: 1819 (in Chinese).
  • 3
    Wang XD, Matsuda N, Ma H et al. Comparative study of quality of life between the Chinese and Japanese adolescent populations. Psychiat. Clin. Neuro. 2000; 54: 147152.
  • 4
    Lindstrom B. Quality of life: a model for evaluating health for all. Conceptual considerations and policy implications. Soz Praventivmed 1992; 37: 301306.
  • 5
    Lee MC. Family and adolescent childbearing. J. Adol. Health 2001; 28: 307312.
  • 6
    Wu LL, Martinson BB. Family structure and the risk of a premarital birth. Am. Sociol. Rev. 1993; 58: 210232.
  • 7
    Hays RD, Alonso J, Coons SJ. Possibilities for summarizing health-related quality of life when using a profile instrument. In: StaquetMJ, HaysRD, FayersPM (eds). Quality of Life Assessment in Clinical Trials. New York, Oxford University Press, 1998; 206207.
  • 8
    Collier J, MacKinlay D, Phillips D. Norm values for the generic children's quality of life measure (GCQ) from a large school-based sample. Qual. Life Res. 2000; 9: 617623.
  • 9
    Berntsson L, Kohler L. Quality of life among children aged 2–17 years in the five Nordic countries. Eur. J. Public Health 2001; 11: 437445.
  • 10
    Raphael D, Rukholm E, Brown I et al. The quality of life profile-adolescent version: background, description, and initial validation. J. Adol. Health 1996; 19: 366375.
  • 11
    Josephs RA, Markus HR, Tafarodi RW. Gender and self-esteem. J. Pers. Soc. Psychol. 1992; 63: 391402.
  • 12
    Robinson ME. Altering gender role expectations: effects on pain tolerance, pain threshold, and pain ratings. J. Pain 2003; 4: 284288.
  • 13
    Apajasalo M, Rautonen J, Holmberg C et al. Quality of life in pre-adolescence: a 17-dimensional health-related measure (17D). Qual. Life Res. 1996; 5: 532538.
  • 14
    Apajasalo M, Sintonen H, Holmberg C et al. Quality of life in early adolescence: a sixteen-dimensional health-related measure (16D). Qual. Life Res. 1996; 5: 205211.
  • 15
    Cheng TA. Urbanization and minor psychiatric morbidity: a community study in Taiwan. Soc. Psychiatry Psychiatr. Epidemiol. 1989; 24: 309316.
  • 16
    Ching MY, Chan KW, Cheng ATA. Substance use disorders among adolescents in Taiwan: prevalence, sociodemographic correlates and psychiatric co-morbidity. Psychol. Med. 1999; 29: 13871396.
  • 17
    Li YT, Yin CS, Chan CC. Psychosocial risk factors of teenage pregnancy in eastern Taiwan. Zhonghua Yi Xue Za Zhi (Taipei) 1999; 62: 425430.
  • 18
    Cheng TA. Urbanization and minor psychiatric morbidity: a community study in Taiwan. J. Affect Disord. 1989; 17: 137152.
  • 19
    Fayers PM, Hand DJ. Causal variables, indicator variables and measurement scales: an example from quality of life. J.R. Statist. Soc. A 2002; 165: 233261.
  • 20
    Fayers PM, Hand DJ. Factor analysis, causal indicators and quality of life. Qual. Life Res. 1997; 6: 139150.