Journal of Pediatric Nursing (2016) xx, xxx–xxx

The Development and Psychometric Properties

of the Children's Sleep Assessment Questionnaire
in Taiwan
Hsiao-Ling Chuang RN, MS Candidate a,b,c , Ching-Pyng Kuo RN, PhD b,c ,
Cheng-Ching Liu RN, PhD d , Chia-Ying Li RN, BS c , Wen-Chun Liao RN, PhD b,e,⁎
a
Institute of Medicine, Chung Shan Medical University, Taichung, Taiwan
b
School of Nursing, Chung Shan Medical University, Taichung, Taiwan
c
Department of Nursing, Chung Shan Medical University Hospital, Taichung, Taiwan
d
College of Nursing, Michigan State University
e
School of Nursing, China Medical University, Taichung, Taiwan

Received 4 December 2015; revised 28 June 2016; accepted 20 July 2016

Key words:
Purpose: To develop and examine the validity and reliability of the Children's Sleep Assessment
Children's sleep
Questionnaire (CSAQ) for school-aged children in Taiwan.
assessment questionnaire;
Design and Methods: We used a cross-sectional study design with stratified random sampling. Pairs of
School-aged child;
children and parents were recruited from a school-based sample of third- and fourth-grade students,
Sleep problems
enrolling 362 child and parent pairs. The content validity, construct validity, convergent validity,
internal consistency, and inter-rater reliability of the CSAQ were assessed.
Results: The CSAQ comprised three parts: sleep hygiene, sleep quality, and sleep disturbance. Sleep
hygiene showed a moderate intra-class correlation coefficient (0.37–0.66) between children and parents.
Results of exploratory factor analysis suggested a four-factor structure model for sleep quality with
64.9% of variance and a two-factor structure for sleep disturbance with 57.7% of variance. These two
models also demonstrated good fit with the confirmatory factor analysis.
Conclusions: The CSAQ is a valid and reliable instrument for assessing sleep problems in school-aged children.
Practice Implications: Both clinicians and researchers can use the CSAQ to screen or elucidate the children’
sleep problems.
© 2016 Elsevier Inc. All rights reserved.

SLEEP PLAYS A pivotal role in the health and well-being Kristensen, 2014; Pesonen et al., 2010), cognitive performance
of children. Insufficient sleep can impede metabolic function (Ferri et al., 2010; Gruber et al., 2010; Wang, Wang, et al.,
and lead to obesity in school-aged children (Cespedes et al., 2013; Wang, Xu, et al., 2013), and school performance (Li et
2014; Spruyt, Molfese, & Gozal, 2011). Moreover, insufficient al., 2013). Therefore, children's sleep affects not just physical
sleep can result in excessive daytime sleepiness, impaired growth but also behavior, cognitive function, and school
attention, and impaired concentration, as well as problems with performance.
learning (Kopasz et al., 2010), behavioral and emotional Unfortunately, sleep problems are common among
regulation (Hansen, Skirbekk, Oerbeck, Wentzel-Larsen, & school-aged children in various cultures. The incidence of
sleep problems ranges from 23.8% to 43.0% among Chinese
⁎ Corresponding author: Wen-Chun Liao, RN., PhD. (Li et al., 2013; Wang, Wang, et al., 2013; Wang, Xu, et al.,
E-mail address: wcl@csmu.edu.tw. 2013), Finnish (Simola et al., 2012), and American

http://dx.doi.org/10.1016/j.pedn.2016.07.008
0882-5963/© 2016 Elsevier Inc. All rights reserved.
2 H.-L. Chuang et al.

school-aged children (Surani et al., 2015). Recent studies Children reported more caffeine consumption than parents
have also revealed that daytime sleepiness occurs in over did (Meltzer et al., 2013). Additionally, approximately 40%
62% of school-aged children (Li et al., 2013). This high of parents are not aware of children's difficulties in
prevalence indicates that sleep problems are an important sleep-onset latency, night waking, or poor sleep quality
modern social issue for child health. (Meltzer et al., 2013). Therefore, sleep hygiene and quality
Polysomnography is the gold standard of objective sleep reported by children may be more accurate than parental
biophysiological changes (Marcus et al., 1992; Uliel, reports. By contrast, status of sleep disturbances is best
Tauman, Greenfeld, & Sivan, 2004). Actigraphy is also an reported by the parent as an observer (Meltzer et al., 2013).
objective measure of rest/activity cycles (Waldon et al., For example, snorers may not know they are snoring during
2016). However, they both fail to qualify the subjective sleeping. Thus, assessments that rely solely on parent- or
experience of sleep quality or sleep hygiene, such as child-reported questionnaires may result in incomplete
difficulty waking in the morning or daytime sleepiness. information. Both perspectives are necessary for a complete
Moreover, needed sleep hours are different in different picture. However, currently available multidimensional
people. Thus, the individual's objective sleep quantity may assessment tools on children's sleep rely exclusively on
be good, but subjective experience may be bad. If we obtain either parent- or child-reported questionnaires. Only few
both objective and subjective data to assess sleep, we will not measures include both parents' and children's reports
only know the sleep biophysiological changes or rest/activity (Lewandowski, Toliver-Sokol, & Palermo, 2011; Spruyt &
cycles but also know the individual's subjective perception. Gozal, 2011). To our knowledge, the Children's Sleep Habits
Thus, both objective quantity and subjective quality are Questionnaire (CSHQ) (Owens, Spirito, & McGuinn, 2000)
important for sleep assessment. A self-report measure to is the most commonly used tool in child sleep science in
assess children's subjective experience of sleep problems is Chinese societies. However, CSHQ is a parent-reported
therefore needed. measure and does not examine children's sleep hygiene.
Sleep problems in children are multifaceted, covering Because of the need for a comprehensive self-report
multiple domains. The best predictor for sleep change, such measure of Taiwanese school-aged children's sleep, we
as reduced sleep length and circadian shift, is age developed the Children's Sleep Assessment Questionnaire
(Iglowstein, Hajnal, Molinari, Largo, & Jenni, 2006). The (CSAQ). The CSAQ includes three components – sleep
most marked reduction in sleep duration occurs when hygiene, sleep quality, and sleep disturbance – with reports
children attend primary school. Since bedtimes become from both parents and children. We proposed that the
increasingly delayed, yet morning rising times remain fixed CSAQ will provide a holistic picture of children's sleep.
(Li et al., 2013). In Chinese society, children's academic The purpose of this study was to describe the development
performance is emphasized, which often requires extra of the CSAQ and examine the CSAQ's reliability and
classes after school. Children spend more time on their validity in measuring sleep problems in Taiwanese school-
studies, adversely affecting their sleep–wake habits and aged children.
sleep duration (Li et al., 2014). These differences suggest
that sleep problems are driven not only by biological Methods
processes but are also influenced by social and cultural Developing Initial Items
factors (Li et al., 2010). Thus, sleep assessment, reflecting The CSAQ focuses on children's sleep problems. The
sleep's multidimensional nature, is necessary for initial scale and items were developed on the basis of
school-aged children. previous literature (Meltzer et al., 2013; Orgilés et al., 2013;
Based on existing literature (Meltzer et al., 2013; Orgilés, Owens et al., 2000; Yolton et al., 2010) and the researchers'
Owens, Espada, Piqueras, & Carballo, 2013; Yolton et al., clinical experience. The questionnaire comprised three parts,
2010), a multidimensional children's sleep assessment tool chosen for their ability to identify sleep problems or
should include three components: sleep hygiene, sleep insufficient sleep. Part one, “sleep hygiene” concerns social
quality, and sleep disturbance. Sleep hygiene refers to and cultural factors, such as lifestyle and environmental sleep
lifestyle and environmental factors that affect an individual's factors that influence children's sleep quality. Thus, sleep
sleep quality (Meltzer et al., 2013; Yolton et al., 2010). Sleep hygiene includes indices of diet, exercise, electronics use,
quality refers to quantitative sleep patterns, night waking, sleep routine, and the environment. Part two, “sleep quality”
waking in the morning, and daytime sleepiness (Drake et al., includes bedtimes, wake times, sleep latency, night waking
2003; Meltzer et al., 2013). Sleep disturbances refer to frequency in varying situations, difficulty waking in the
behaviors that occur only during sleep, such as snoring or morning, and daytime sleepiness in different situations (in
sleepwalking (Meltzer et al., 2013). Currently, no multidi- school, during short car rides, while doing homework, and
mensional questionnaire on sleep assessment exists for while watching television). Part 3, “sleep disturbance”
school-aged children in Taiwan. includes parasomnias and sleep-disordered breathing. The
Secondary, parents may be unaware of certain aspects of original scale comprised three parts with a total of 44 items:
children's sleep hygiene (e.g., exposure to second-hand 16 items for sleep hygiene, 20 items for sleep quality, and 8
smoke or daytime caffeinated beverage consumption). items for sleep disturbances. Literature showed that sleep
Children’s Sleep Assessment Questionnaire 3

hygiene and quality reported by children were more accurate total of three city-funded schools were included in this study
than parental reports, and sleep disturbance reported by and each had six grades with 700–800 students at each
parents were more accurate than children's reports (Meltzer school. School 1, located outside the city center, has been
et al., 2013). Therefore, this study designed that parts one established for 15 years. School 2, located in a new
and two were reported by children, while part three was city-center community, has been established for 16 years.
reported by parents. Moreover, parents also completed all the School 3, located in an old city-center community, has been
CSAQ items because parental information was used to assess established for 55 years. Moreover, the socioeconomic status
children's and parents' inter-rater reliability. of the students' parents was similar among the three schools.
A four-point Likert scale indicated frequency of occur- After receiving permission from school administrators, the
rence: 0 = “never” to the given situation, 1 = “less than once first author gave instructions to all the 3rd- and 4th-grade
a week,” 2 = “one to three times a week,” and 3 = “four to classes. The investigator explained the research purpose, its
seven times a week.” We used the summation for each part. process, and all scales to the children. Further, a formal letter of
Item 4, item 5, and item 9 were reverse items. Higher scores consent was sent to children and their parents, who were
indicate poorer sleep hygiene, poorer sleep quality, or more required to provide signed informed consent along with their
sleep disturbances, respectively. returned questionnaires. Parents signed a consent form while
their corresponding children signed an assent form. Each child
Content Validity received a box of 12 colored pencils to thank them for their
An expert panel examined the original items' content validity time, none for parents. Furthermore, children or parents who
index (CVI) (Lynn, 1986). Panel experts (one elementary school wanted to withdraw from the study could do so at any time by
teacher, one clinician specializing in children's sleep problems, calling the researchers. Participants had to fulfill the following
one pediatrician, and two experienced researchers specializing criteria: (a) able to read Chinese, (b) child aged N8 years, and
in child care) rated each CSAQ item based on relevance, clarity, (c) parent completing the questionnaire had to be the child's
and simplicity using the following scale: 1 (not relevant), 2 primary caregiver. Figure 1 displays a flowchart of the
(somewhat relevant), 3 (relevant), or 4 (very relevant). Only recruiting process. All third- and fourth-grade classes from
items scored 3 and 4 were considered relevant and thus used to participating elementary schools were recruited, and all
calculate CVI. Results from panel experts yielded CVIs of 0.97 children and their parents were asked to complete the
and 0.80–one each for the scale (S-CVI) and items (I-CVIs), questionnaire including the CSAQ and demographic data.
respectively. These values fulfilled the requisite criteria The children also complete the Pittsburgh Sleep Quality Index
(S-CVI≥0.90 and I-CVI ≥0.78) (Polit, Beck, & Owen, 2007). (PSQI) and the Epworth Sleepiness Scale (ESS) to assess
No items were deleted from the instrument, and minor revisions convergent validity between the CSAQ and the PSQI, and
regarding items' clarity or wording were suggested. Revisions between the CSAQ and the ESS. Based on the suggestion from
were incorporated into the instrument, and the revised items the literature (Costello & Osborne, 2005), five to ten subjects
were used in subsequent reliability and validity testing. per item were used for sample size estimation in this study.
This study was approved by the Institutional Review Board
(CSMUH, no. CS11213). Each child and parent participant
Establishing Psychometric Properties provided informed consent for the study. All were informed
After developing initial items and performing the content that participating in this study would not harm or discomfort
validity test, we conducted a cross-sectional survey for them in any way. They were also informed that everyone
CSAQ item analysis and psychometric testing. We used item would receive individual report about the assessment results as
analysis to exclude items with a standard deviation of 1.5, as part of the benefit for their participation. To ensure data
well as item–total correlations less than 0.30 or greater than confidentiality, all data were encoded, used only for this study,
0.70 (Chiou, 2010). We then tested the final scale's and not disclosed to other parties. The study was conducted
psychometric properties, including construct and convergent within the parameters of the Helsinki agreement.
validities, internal consistency, and intra-class correlation
coefficient (ICC). The Pittsburgh Sleep Quality Index (PSQI)
and the Epworth Sleepiness Scale (ESS) were used to test Measures
CSAQ's convergent validity. The questionnaire includes demographic data, CSAQ,
Pittsburgh Sleep Quality Index (PSQI) and Epworth
Sleepiness Scale (ESS). Demographic data includes the age
Design, Sample, and Data Collection
of child and parent, sex and grade of child, and education and
A cross-sectional design was used in this study. Children
marital status of the parent.
were recruited by stratified random sampling. Three
administrative areas in Taichung were selected from the
eight districts of Xitun, Beitun, Nantun, West District, North Pittsburgh Sleep Quality Index (PSQI)
District, Central District, East District, and South District The PSQI, used to measure sleep quality during the
(Figure 1). Considering time and money, one elementary previous month, is a self-administered questionnaire. It
school (of 11) was selected from each administrative area. A contains 19 questions yielding seven subscores, covering
4 H.-L. Chuang et al.

Eight administrative areas

Three administrative areas selected

Three of 11 elementary schools selected from three administrative areas

559 3rd-and 4th-Grade students

188 refused to participate in the study

371 completed

9 primary caregivers excluded because they

were not the students’ parents

362 Successful questionnaires

290 Mothers 72 Fathers

Figure 1 Flowchart of 3rd- and 4th-grade children and their parents as participants in each stage of the Children's Sleep Assessment
Questionnaire study in Taiwan.

subjective sleep quality, sleep latency, sleep duration, sleep Data Analysis
efficiency, sleep disturbance, hypnotic medication use, and The CSAQ was developed and tested for item analysis,
daytime dysfunction. Each component is scored 0–3, validity, and reliability. Item descriptive statistics and
yielding a global PSQI score of 0–21, with higher scores corrected item–total correlations were performed for item
indicating lower sleep quality. The PSQI's reliability and selection (n = 362 pairs of children and parents). Items must
validity study showed acceptable internal consistency meet a minimum criterion of an all-item mean+ 1.5 standard
(Cronbach's α = 0.83) and test–retest reliability with a deviation (Chiou, 2010). We considered those items with
coefficient of 0.85 (Buysse, Reynolds, Monk, Berman, & item–total correlations of less than 0.30 not to contribute
Kupfer, 1989). A modified Chinese version of the PSQI sufficiently to measuring the concept and considered those
(CPSQI) was used to assess school-aged children's sleep with item–total correlations greater than 0.70 as possibly
quality in Taiwan (Tan, 2004; Wu, 2009). Here, we used the redundant (Chiou, 2010). Thus, we deleted items with item–
modified CPSQI for CSAQ's convergent validity. total correlations of less than 0.30 and greater than 0.70.
After performing item analysis, revised items were subse-
Epworth Sleepiness Scale (ESS) quently tested for validity and reliability. Because there were
The ESS is a widely used scale that evaluates the degree multiple indicators for sleep hygiene that may be unrelated
of daytime sleepiness. Responders are instructed to rate, on a (e.g., frequency that child drinks soda may not relate to
scale of 0–3 (never =0, slight =1, moderate =2, and high =3), frequency of exercising), calculating internal consistency for
the likelihood of falling asleep in eight different situations sleep hygiene was not appropriate and was omitted. IBM SPSS
(Johns, 1991). The Chinese version of the ESS (CESS) was statistics version 20.0 (IBM Corporation, 2011) was used to
translated and validated by Chen et al. (2002). The scale analyze the data. The p level of significance was less than.05.
showed acceptable internal consistency (Cronbach's α =
0.81) and test–retest reliability with a coefficient of 0.74 in Validity testing
individuals who experienced symptoms of sleep-disordered Exploratory factor analysis (EFA) was conducted using
breathing (Chen et al., 2002). A modified CESS was used to principal component analysis to examine the CSAQ's
assess excessive daytime sleepiness of Chinese children. The construct validity of sleep quality and sleep disturbance.
scale showed acceptable internal consistency (Cronbach's Confirmatory factorial analysis (CFA) was conducted to
α = 0.85) (Chan et al., 2009). Here, the modified CESS was validate the structure obtained in the EFA. The x2/df ratio,
also used to assess CSAQ's convergent validity. root mean residual (RMR), root mean square error of
Children’s Sleep Assessment Questionnaire 5

Table 1 Item Analysis and Exploratory Factor Analysis for the CSAQ
Items Missing Mean (SD) Correlation a Factor
rates (%) loadings
Sleep hygiene
Diet index
Frequency of fast food consumption 0.0 0.57 (0.70) − −
Frequency of sweetened food consumption 0.0 1.44 (0.92) − −
Frequency of caffeinated beverage consumption 0.0 1.18 (1.00) − −
Exercise index
Frequency of sports per week 0.0 0.78 (0.75) − −
Exercise for more than 30 minutes each time 0.5 1.14 (0.98) − −
Strenuous exercise in the hours before sleep 0.3 0.32 (0.70) − −
Electronic use index
Watching TV or playing computer games within 4 hours of sleeping 0.0 1.95 (1.00) − −
Watching TV or playing computer games for more than 2 hours per day 0.8 1.18 (1.10) − −
Sleep routine index
Goes to bed at the same time every night 0.8 1.01 (1.09) − −
Goes to bed after 10 pm 0.8 1.64 (1.16) − −
Keeps a blanket or toy while sleeping 0.0 1.15 (1.33) − −
Environment index
Exposure to second-hand smoke 0.0 0.75 (1.07) − −
Sound present during sleep 0.3 1.00 (0.92) − −
Light present during sleep 0.0 0.84 (0.72) − −
Alteration of the setting would result in insufficient sleep 0.0 0.67 (0.71) − −
Room-sharing 0.0 2.23 (1.21) − −
Sleep quality
Night waking
Wakes at night 0.0 1.58 (1.02) .44 .90
Goes to the toilet at night 0.0 1.13 (0.98) .33 .78
Alarmed by nightmares 0.0 0.57 (0.80) .48 .73
Discomfort (e.g., cough and stuffy nose) during sleep 0.0 0.73 (0.93) .50 .69
Wakes because it is too cold or too hot 0.0 0.72 (0.92) .36 .68
Daytime sleepiness
Daytime sleepiness 0.0 0.83 (1.04) .52 .92
Sleepiness in class 0.0 0.22 (0.57) .30 .83
Sleepiness while doing homework 0.0 0.30 (0.65) .40 .76
Sleepiness while watching TV 0.0 0.24 (0.65) .43 .74
Sleepiness while riding in a car or on a motorcycle 0.0 0.58 (0.91) .50 .59
Quantitative sleep pattern
Sleep-onset latency 0.0 0.71 (0.76) .31 .95
Sleep efficiency 0.0 0.49 (0.74) .33 .94
Waking in the morning
Wakes in a bad mood 0.0 0.84 (1.02) .34 .85
Has difficulty getting out of bed 0.0 1.60 (1.09) .35 .84
Seems tired in the morning 0.3 1.80 (1.13) .37 .50
Sleep disturbances
Parasomnias
Talking during sleep 0.0 0.69 (0.75) .43 .69
Nightmares or screaming 0.0 0.20 (0.47) .33 .64
Struggles at bedtime 0.0 1.28 (1.05) .43 .67
Night waking 0.0 0.66 (0.79) .48 .70
Sleep breathing-related disorder
Snoring 0.0 0.81 (0.95) .38 .88
Stuffy nose or mouth breathing 0.0 0.93 (0.98) .51 .81
Note. CSAQ, Children's Sleep Assessment Questionnaire; SD, standard deviation.
a
Corrected item–total correlation.
6 H.-L. Chuang et al.

approximation (RMSEA), goodness-of-fit index (GFI), and (Figure 2B). Results of the Kaiser–Meyer–Olkin index (0.74
comparative fit index (CFI) were conducted. Moreover, partici- and 0.73, respectively) and Bartlett's Test of Sphericity
pants were randomly split into two groups. EFA was performed on (χ2 = 1601.96 and χ2 = 310.120; all p b 0.000) demon-
Group 1 (n = 181), and CFA on Group 2 (n = 181). strated adequacy to conduct EFA. Criteria of factor with
Convergent validity was further evaluated by the Pearson eigenvalues N1 were used to select factors. Items with a
correlation examining association among the three parts of factor loading of b0.40 were excluded from the scale (Chiou,
the CSAQ, CPSQI, and CESS (n = 362), based on the 2010). No items were deleted from this instrument. Table 1
hypothesis that CPSQI and CESS scores should correlate shows the factor loadings' results.
positively with the CSAQ's three parts. Four factors were extracted, and they explained 64.9% of
variance in sleep quality. The first factor includes 5 items and
describes night-waking frequency in different situations,
Reliability Testing
including going to the toilet, having a nightmare, being
Cronbach's α coefficients were used to measure internal
uncomfortable, or anything else defined as night waking.
consistency for CSAQ sleep quality and sleep disturbance
The daytime sleepiness factor with 5 items and describes
(n = 362). Inter-rater reliability was assessed for sleep
frequency of sleepiness in different situations during the day,
hygiene and sleep quality using the ICC between children
and parents (n = 362). such as in school classrooms, during short car rides, and
while doing homework or watching television. The third
factor, quantitative sleep pattern, refers to sleep-onset latency
and efficiency, consist of 2 items. The fourth factor, waking
Results
in the morning, refers to difficulties in waking, consist of 3
Participants
items. Two factors were extracted and explained 57.7% of
Of the 559 pairs of children and parents (father or mother)
variance in sleep disturbance. The first factor described
invited to participate, 371 (66.4%) pairs completed the study.
parasomnias with four items. The second factor referred to
A further nine pairs were excluded because primary caregivers
sleep-disordered breathing and comprised two items.
were not parents, but grandparents, aunts, or nursemaids.
CFA was computed to validate the factorial structure
Finally, 362 pairs of children and parents completed the
obtained in the EFA. Model fit was considered acceptable if
questionnaire and 188 participants refused to do so (Figure 1).
the x2/df ratio ranged from 1.0 to 3.0, if both the GFI and CFI
Among the 188 who refused, parents claimed they were too
were greater than 0.90, if RMR was less than 0.05, and the
busy to complete the questionnaire and children responded
RMSEA was less than 0.08 (Li, 2006). Results (Figure 2A)
with no interest in sleep problems at all.
demonstrated good fit for the model of sleep quality with the
Of the 362 participating children, 172 were boys and 190
data (x2/df = 1.81; RMR = 0.04; RMSEA =0.06; GFI =
were girls, with an average age of 9.9 years (SD = 0.82,
0.91; CFI = 0.94). In addition, results (Figure 2B) also
range = 8.0–12.0). More than half the children were
demonstrated good fit for the model of sleep disturbance
fourth-graders (62.7%). Of the parents, 290 were mothers
with the data (x2/df = 1.53; RMR = 0.03; RMSEA =0.05;
(average age 39 years), and 72 were fathers (average age 43
GFI = 0.98; CFI = 0.98).
years). Most parents were educated through senior high
school (46.5% mothers, 31.9% fathers) or college (42.9% Convergent Validity
mothers, 46.8% fathers), and most were married (88.1% Convergent validity was determined using correlation
mothers, 85.4% fathers). among the three parts of the CSAQ, the CPSQI, and the
CESS (Table 2). The three parts of CSAQ including Sleep
Item Analysis hygiene, sleep quality, and sleep disturbance, had low to
Descriptive statistical analysis and item–total correlations moderate correlation with both the CPSQI (r = 0.20–0.41)
of retained items are presented in Table 1. All item–scale and the CESS (r = 0.15–0.42), respectively. All correlation
correlations fulfilled the prerequisite criterion of all-item coefficients are statistically significant (p b 0.05).
mean + 1.5 standard deviation. Otherwise, correlations
ranged from 0.30 to 0.52, indicating adequate performance. Internal Consistency
However, from the original scale of the remaining items, Cronbach's α for sleep quality and for sleep disturbance
seven (nos. 28, 30, 32, 33, 37, 38, and 40) were eliminated were 0.80 and 0.68, respectively (Table 2). Removal of one
because of their low item–total correlation. The revised scale or more weakly correlated items did not significantly
included 37 items: 16 items for sleep hygiene, 15 items for improve the alpha value.
sleep quality, and 6 items for sleep disturbances. This revised
scale was used in the subsequent reliability and validity test. Inter-Rater Reliability
ICC results between children and parents for CSAQ sleep
Factor Analysis hygiene and sleep quality are presented in Table 2. Moderate
EFA was conducted to examine the CSAQ's construct agreement between children's and parents' reports was found
validity for sleep quality (Figure 2A) and sleep disturbance for diet, exercise, electronics use, sleep routine, and
Children’s Sleep Assessment Questionnaire 7

A .85
.92 .85
.58
.76
.56 .32
.58
.24
.51
.26
.34
.85
.92
.34
.58
.03
.65 .42
.53
.21 .28
.84
.01 .70

.38
.37 .62
1.41
1.19
.21
.22
.47
.80 .64
.73 .58

B .25
.06

.29
.54
.39
.16
.49
.24
.67

.61 .37

.50
.71

Figure 2 Path diagram for confirmatory factor analysis model with standardized values A, CFA model for Sleep quality of 3rd- and
4th-grade children in Taiwan Note. Model fit indices: x2/df = 1.81; RMR = 0.04; RMSEA =0.06; GFI = 0.91; CFI = 0.94). The numbers on
the double-headed arrow refer to the correlation between latent variables (nightwaking, sleepiness, quamtitativesleep, and morningwaking);
the numbers on the direction arrow refer to the standardized regression weights between the latent variables and the items; the numbers on the
right side of item boxes refer to the squared multiple correlations of the items. B, CFA model for sleep disturbance of 3rd- and 4th-grade
children in Taiwan Note. Model fit indices: x2/df = 1.53; RMR = 0.03; RMSEA =0.05; GFI = 0.98; CFI = 0.98. The numbers on the
double-headed arrow refer to the correlation between latent variables (parasomnias and breathing); the numbers on the direction arrow refer to
the standardized regression weights between latent variables and items; the numbers on the right side of item boxes refer to the squared
multiple correlations of the items.

environmental indices (0.37–0.66). A similar result was Results from content validity according to panel experts
obtained for sleep quality (ICC = 0.41). yielded a high CVI. In addition, missing rates of all items ranged
from 0.3% to 0.8% (Table 1). We did not find any reported
difficulties for children and parents in completing this measure.
Discussion Thus, we can recommend that the CSAQ is easy to use both for
Here, we report psychometric properties of a children and parents, and that children can provide information
sleep-screening questionnaire designed primarily to assess regarding their own sleep hygiene and sleep quality.
children's sleep hygiene, sleep quality, and sleep disturbance Results of EFA suggest a four-factor structure for sleep
in a community setting. Based on these results, the CSAQ was quality and a two-factor structure for sleep disturbance. All
found to be a reliable and valid measure of children's sleep, items are significantly loaded to their corresponding factor as
with acceptable content validity, construct validity, convergent designed. No item is cross-loaded to difference factors.
validity, internal consistency, and inter-rater reliability. Furthermore, the factor structure of sleep quality obtained in
8 H.-L. Chuang et al.

Table 2 Convergent Validity Test, Internal Consistency, and comparable subscales in the CSHQ (0.36–0.65) (Owens
ICCs for the CSAQ et al., 2000), the most popular scale in Chinese society.
Score CPQSI a CESS a α b ICC ICCs between children and parents were moderately
range correlated for sleep hygiene indices (0.37–0.66) and sleep
quality overall (0.38–0.78), comparable with those of the
Sleep hygiene 0–48 .22*** .23*** −
Children's Report of Sleep Patterns (ICCs =0.39–0.71)
Diet index 0–9 − − − 0.40
Exercise index 0–9 − − − 0.40 (Meltzer et al., 2013). Because literature showed poor
Electronic use index 0–6 − − − 0.37 agreement (ICCs =0.05) between children and parents on the
Sleep routine index 0–9 − − − 0.49 parasomnias subscale (Meltzer et al., 2013), we did not
Environment index 0–15 − − − 0.66 examine ICCs between children and parents for sleep
Sleep quality 0–45 .41*** .42*** .80 0.41 disturbance of parasomnias and sleep-disordered breathing
Night waking 0–15 .81 0.38 subscales. However, another study showed moderate
Daytime sleepiness 0–15 .82 0.45 agreement (ICC = 0.36–0.51) between children and parents
Quantitative sleep for both subscales (Owens et al., 2000). Therefore, we
0–6 .90 0.78
pattern suggested that future study may be needed to examine ICCs
Waking in the morning 0–9 .66 0.42
between children and parents for sleep disturbance in
Sleep disturbances 0–18 .20*** .15* .68 -
the CSAQ.
Parasomnias 0–12 .61 -
Sleep breathing-related This study has several strengths. First, although this study
0–9 .64 - only included 3rd & 4th graders, the children's age ranges
disorder
from 8–12 years old. Thus, it supports accuracy of
Note. CSAQ = Children's Sleep Assessment Questionnaire; CPSQI =
Chinese version of the Pittsburgh Sleep Quality Index; CESS = Chinese information provided by children aged 8–12 years regarding
version of the Epworth Sleepiness Scale; ICC = Intra-class correlation their own sleep hygiene and quality. Second, the CSAQ is
coefficient between children and parents. the only multidimensional measure for examining
*p b .05; **p b .01; ***p ≤ .001. sleep-related problems reported by children aged 8–12
a
Refer to Pearson correlation coefficient.
b years in Taiwan. Furthermore, CSAQ use was more likely to
Refer to the internal consistency of sleep quality and disturbance
modules. result in documentation of additional sleep-related informa-
tion, including sleep hygiene, insufficient sleep, and sleep
disturbance. Such information may be important in identi-
fying potential intervention to prevent children's future sleep
the EFA demonstrated good fit with the CFA. A similar problems. For example, discussing the importance of
result was also found in that the factor structure of sleep bedtime routines with children and parents, and providing
disturbance obtained in the EFA demonstrated good fit with basic information regarding good sleep hygiene may help
the CFA. Construct validity of sleep quality and disturbance parents to identify workable strategies for children. Finally, a
were supported. Thus, we suggest four subscales for sleep single-reporter sleep scale may fail to uncover all relevant
quality and two subscales for sleep disturbance. clinical information necessary to identify sleep problems.
The CSAQ had low-to-moderate positive correlation with The multi-reporter (child and parent) CSAQ can provide
both the CPSQI and CESS (r = 0.15–0.42). Previous studies comprehensive information that may otherwise be missed.
showed that poor sleepers reported poor sleep hygiene This study's limitations must be considered when
(Meltzer et al., 2013; Yolton et al., 2010), whereas this study evaluating the scale's suitability. First, no specific observa-
showed that children with poor sleep hygiene had poor tion period was included in participant instructions, which
nighttime sleep quality and more daytime sleepiness. A asked only for frequency of sleep behaviors over one week.
similar result was also found in this study that children with This may have resulted in relative over-reporting of transient
more sleep disturbances had poor nighttime sleep quality and sleep problems, indicating a need to establish a timeline for
more daytime sleepiness. In the CSAQ, sleep quality also assessing habitual sleep. Second, the CSAQ should be
had moderate positive correlation with both nighttime sleep compared to objective measures of sleep problems, such as
quality and daytime sleepiness. The CSAQ demonstrated actigraphy or polysomnography, to assess its validity against
acceptable convergent validity. those measures. Notably, the CSAQ was designed as a
The CSAQ's reliability was demonstrated through screening rather than a diagnostic instrument, with utility in
acceptable internal consistency for sleep quality and sleep both clinical and research settings. A cutoff point would be
disturbance. Internal consistency coefficients of sleep quality useful for clinicians and researchers to make a diagnosis
and sleep disturbance were near (0.68) or above (0.90) a during clinical interviews. In future, a diverse sample from
required criterion of (0.70). However, a Cronbach's α value community-based and clinical settings will be needed to
between 0.6 and 0.7 is acceptable for a preliminarily establish a cutoff score. Finally, we assessed only third- and
developed scale (Chiou, 2010). For a community-based fourth-graders from selected schools in Taichung, and this
population, performances of the total scale and subscales of may limit the findings' generalizability to other school-aged
sleep quality and sleep disturbance were greater than children and populations. More studies are needed to assess
Children’s Sleep Assessment Questionnaire 9

the questionnaire's generalizability to other grades and Gruber, R, Laviolette, R, Deluca, P, Monson, E, Cornish, K, & Carrier, J
children from other populations, including different cities (2010). Short sleep duration is associated with poor performance on IQ
measures in healthy school-age children. Sleep Medicine, 11, 289–294.
and rural areas. Hansen, BH, Skirbekk, B, Oerbeck, B, Wentzel-Larsen, T, & Kristensen, H
(2014). Associations between sleep problems and attentional and
Practice Implications behavioral functioning in children with anxiety disorders and ADHD.
The CSAQ is a multidimensional measure that can provide Behavioral Sleep Medicine, 12, 53–68. http://dx.doi.org/10.1080/
15402002.2013.764525.
comprehensive information to clinicians and researchers. In Iglowstein, I, Hajnal, LB, Molinari, L, Largo, RH, & Jenni, OG (2006).
particular, CSAQ is a multi-reporter measure that can provide Sleep behaviour in preterm children from birth to age 10 years: a
additional information which may not be captured if relying longitudinal study. Acta Paediatrica, 95, 1691–1693.
solely on children or parents. Further, an increasing score on Johns, MW (1991). A new method for measuring daytime sleepiness: the
the CSAQ would alert clinicians or parents to provide Epworth sleepiness scale. Sleep, 14, 540–545.
Kopasz, M, Loessl, B, Hornyak, M, Riemann, D, Nissen, C, Piosczyk, H, &
additional screening for underlying sleep problems. Voderholzer, U (2010). Sleep and memory in healthy children and
adolescents - a critical review. Sleep Medicine Reviews, 14, 167–177.
Conclusion Lewandowski, AS, Toliver-Sokol, M, & Palermo, TM (2011). Evidence-based
The results of the study suggest that the CSAQ is a review of subjective pediatric sleep measures. Journal of Pediatric
Psychology, 36, 780–793. http://dx.doi.org/10.1093/jpepsy/jsq119.
reliable and valid instrument for evaluating sleep problems in
Li, MN (2006). An introduction to AMOS and its uses in scale development:
school-aged children, with acceptable content validity, graphics & basic. Taipei, Taiwan: Psychology Press.
construct validity, convergent validity, internal consistency, Li, S, Arguelles, L, Jiang, F, Chen, W, Jin, X, Yan, C, ... Shen, X (2013).
and inter-rater reliability. Additional studies should address Sleep, school performance, and a school-based intervention among
the use of the CSAQ in other populations to provide further school-aged children: a sleep series study in China. PLoS ONE, 8,
e67928. http://dx.doi.org/10.1371/journal.pone.0067928.
evidence of its use in a variety of settings.
Li, S, Yang, Q, Chen, Z, Jin, X, Jiang, F, & Shen, X (2014). Homework
schedule: an important factor associated with shorter sleep duration
Acknowledgments among chinese school-aged children. Behavioral Sleep Medicine, 12,
389–397. http://dx.doi.org/10.1080/15402002.2013.821654.
No funding was secured for this study.
Li, S, Zhu, S, Jin, X, Yan, C, Wu, S, Jiang, F, & Shen, X (2010). Risk factors
associated with short sleep duration among Chinese school-aged children.
Sleep Medicine, 11, 907–916. http://dx.doi.org/10.1016/j.sleep.2010.03.018.
Appendix A. Supplementary Data Lynn, MR (1986). Determination and quantification of content validity.
Nursing Research, 35, 382–385.
Supplementary data to this article can be found online at
Marcus, CL, Omlin, KJ, Basinki, DJ, Bailey, SL, Rachal, AB, Von Pechmann,
http://dx.doi.org/10.1016/j.pedn.2016.07.008. WS, ... Ward, SL (1992). Normal polysomnographic values for children
and adolescents. The American Review of Respiratory Disease, 146,
References 1235–1239. http://dx.doi.org/10.1164/ajrccm/146.5_Pt_1.1235.
Buysse, DJ, Reynolds, CF, III, Monk, TH, Berman, SR, & Kupfer, DJ Meltzer, LJ, Avis, KT, Biggs, S, Reynolds, AC, Crabtree, VM, & Bevans,
(1989). The Pittsburgh sleep quality index: a new instrument for KB (2013). The Children's report of sleep patterns (CRSP): a self-report
psychiatric practice and research. Psychiatry Research, 28, 193–213. measure of sleep for school-aged children. Journal of Clinical Sleep
Cespedes, EM, Rifas-Shiman, SL, Redline, S, Gillman, MW, Pena, M-M, & Medicine, 9, 235–245. http://dx.doi.org/10.5664/jcsm.2486.
Taveras, EM (2014). Longitudinal associations of sleep curtailment with Orgilés, M, Owens, J, Espada, JP, Piqueras, JA, & Carballo, JL (2013).
metabolic risk in mid-childhood. Obesity, 22, 2586–2592. http://dx.doi. Spanish version of the sleep self-report (SSR): factorial structure and
org/10.1002/oby.20894. psychometric properties. Child: Care, Health and Development, 39,
Chan, EYT, Ng, DK, Chan, C-H, Kwok, K-L, Chow, P-Y, Cheung, JM, & Leung, 288–295. http://dx.doi.org/10.1111/j.1365–2214.2012.01389.x.
S-Y (2009). Modified Epworth sleepiness scale in Chinese children with Owens, JA, Spirito, A, & McGuinn, M (2000). The Children's sleep habits
obstructive sleep apnea: a retrospective study. Sleep & Breathing, 13, 59–63. questionnaire (CSHQ): psychometric properties of a survey instrument
Chen, NC, Johns, MW, Li, HY, Chu, CC, Liang, SC, Shu, YH, ... Wang, PC for school-aged children. Sleep, 23, 1043–1051.
(2002). Validation of a Chinese version of the Epworth sleepiness scale. Pesonen, AK, Raikkonen, K, Paavonen, EJ, Heinonen, K, Komsi, N, Lahti, J, ...
Quality of Life Research, 11, 817–821. Strandberg, T (2010). Sleep duration and regularity are associated with
Chiou, HJ (2010). Quantitative research and statistical analysis in social behavioral problems in 8-year-old children. International Journal of
& behavioral sciences (5th ed.). Taipei, Taiwan: Wu-Nan Book Inc. Behavioral Medicine, 17, 298–305.
Corporation, IBM (2011). IBM SPSS statistics for windows, version 20.0. Polit, DF, Beck, CT, & Owen, SV (2007). Is the CVI an acceptable indicator of
Armonk, NY: IBM Corp. content validity? Appraisal and recommendations. Research in Nursing &
Costello, AB, & Osborne, JW (2005). Best practices in exploratory factor Health, 30, 459–467. http://dx.doi.org/10.1002/nur.20199.
analysis: four recommendations for getting the most from your analysis. Simola, P, Laitalainen, E, Liukkonen, K, Virkkula, P, Kirjavainen, T, Pitkaranta,
Practical Assessment, Research & Evaluation, 10 (http://pareonline.net/ A, & Aronen, ET (2012). Sleep disturbances in a community sample from
pdf/v10n7.pdf). preschool to school age. Child: Care, Health & Development, 38, 572–580.
Drake, C, Nickel, C, Burduvali, E, Roth, T, Jefferson, C, & Pietro, B (2003). Spruyt, K, & Gozal, D (2011). Pediatric sleep questionnaires as diagnostic
The pediatric daytime sleepiness scale (PDSS): sleep habits and school or epidemiological tools: a review of currently available instruments.
outcomes in middle-school children. Sleep, 26, 455–458. Sleep Medicine Reviews, 15, 19–32. http://dx.doi.org/10.1016/j.smrv.
Ferri, R, Drago, V, Aricò, D, Bruni, O, Remington, RW, Stamatakis, K, & 2010.07.005.
Punjabi, NM (2010). The effects of experimental sleep fragmentation on Spruyt, K, Molfese, DL, & Gozal, D (2011). Sleep duration, sleep regularity,
cognitive processing. Sleep Medicine, 11, 378–385. http://dx.doi.org/ body weight, and metabolic homeostasis in school-aged children.
10.1016/j.sleep.2010.01.006. Pediatrics, 127, e345–e352. http://dx.doi.org/10.1542/peds.2010-0497.
10 H.-L. Chuang et al.

Surani, S, Hesselbacher, S, Sadasiva, S, Surani, Z, Surani, SS, Khimani, A, Wang, X-Y, Wang, X, Huang, C-Q, Guo, Z-Y, Qian, Y-F, Yang, Y, ...
& Subramanian, S (2015). Sleep habits of elementary and middle school Wang, L (2013a). Effects of sleep deprivation on the intelligence
children in South Texas. Sleep Disorder, 2015, 179103. http://dx.doi. structure of school-age children in Changsha, China. Zhongguo Dang
org/10.1155/2015/179103. Dai Er Ke Za Zhi, 15, 866–869.
Tan, LP (2004). The effects of background music on quality of sleep in Wang, G, Xu, G, Liu, Z, Lu, N, Ma, R, & Zhang, E (2013b). Sleep patterns
elementary school children. Journal of Music Therapy, 41, and sleep disturbances among Chinese school-aged children: prevalence
128–150. and associated factors. Sleep Medicine, 14, 45–52. http://dx.doi.org/10.
Uliel, S, Tauman, R, Greenfeld, M, & Sivan, Y (2004). Normal polysomno- 1016/j.sleep.2012.09.022.
graphic respiratory values in children and adolescents. Chest, 125, 872–878. Wu, HC (2009). The sleep quality in the transition from late childhood to
Waldon, J, Begum, E, Gendron, M, Rusak, B, Andreou, P, Rajda, M, & early adolescence: a longitudinal study. (Unpublished master's thesis)
Corkum, P (2016). Concordance of actigraphy with polysomnogra- Taichung,Taiwan: China Medical University.
phy in children with and without attention-deficit/hyperactivity Yolton, K, Xu, Y, Khoury, J, Succop, P, Lanphear, B, Beebe, DW, &
disorder. Journal of Sleep Research. http://dx.doi.org/10.1016/j. Owens, J (2010). Associations between secondhand smoke exposure
pedn.2016.07.008. and sleep patterns in children. Pediatrics, 125, e261–268.