Journal of American College Health

ISSN: 0744-8481 (Print) 1940-3208 (Online) Journal homepage: https://www.tandfonline.com/loi/vach20

Bedtime procrastination predicts the prevalence

and severity of poor sleep quality of Chinese
undergraduate students

Xiaohan Ma, Dexin Meng, Liwei Zhu, Huiying Xu, Jing Guo, Limin Yang, Lei Yu,
Yiming Fu & Li Mu

To cite this article: Xiaohan Ma, Dexin Meng, Liwei Zhu, Huiying Xu, Jing Guo, Limin Yang, Lei
Yu, Yiming Fu & Li Mu (2020): Bedtime procrastination predicts the prevalence and severity of
poor sleep quality of Chinese undergraduate students, Journal of American College Health, DOI:
10.1080/07448481.2020.1785474

To link to this article: https://doi.org/10.1080/07448481.2020.1785474

Published online: 15 Jul 2020.

Submit your article to this journal

Article views: 3

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=vach20
JOURNAL OF AMERICAN COLLEGE HEALTH
https://doi.org/10.1080/07448481.2020.1785474

MAJOR ARTICLE

Bedtime procrastination predicts the prevalence and severity of poor sleep

quality of Chinese undergraduate students
Xiaohan Ma, MEda, Dexin Meng, MD, PhDb, Liwei Zhu, BEda, Huiying Xu, MS, MDc, Jing Guo, BEa,
Limin Yang, MD, PhDd, Lei Yue, Yiming Fu, MEda, and Li Mu, MD, PhDa
a
Research Center of Brain and Cognitive Neuroscience, Liaoning Normal University, Dalian, China; bDepartment of Physiology, Jiamusi
University, Jiamusi, China; cSchool of Clinical Medicine, Changchun University of Chinese Medicine, Changchun, China; dSchool of Medicine,
Dalian University, Dalian, China; eDepartment of Physiology, College of Basic Medical Sciences, Jilin University, Changchun, China

ABSTRACT ARTICLE HISTORY

Objective: Bedtime procrastination is a type of unhealthy sleep behavior, but whether it affects Received 22 October 2019
sleep quality remains unknown. The current study aimed to examine the relationship between Revised 29 April 2020
bedtime procrastination and poor sleep quality among Chinese college students. Participants: A Accepted 12 June 2020
total of 1550 Chinese undergraduates participated and effectively completed the study. Methods:
KEYWORDS
This cross-sectional study collected information on bedtime procrastination and self-reported sleep Age; bedtime
quality via the Bedtime Procrastination Scale and the Pittsburgh Sleep Quality Index. Results: The procrastination; education;
prevalence of poor sleep quality was 39.42% among Chinese undergraduates. Age, education, and sleep quality;
bedtime procrastination independently predicted the prevalence of poor sleep quality. Bedtime undergraduate students
procrastination was a unique contributor to the severity of poor sleep quality. Conclusions: This
study highlighted the independent associations between bedtime procrastination and the preva-
lence and severity of self-reported poor sleep quality. These findings may be helpful in designing
effective measures to improve the sleep quality of college students.

Introduction after 9 p.m. and turning off the lights late,23 or sleeping at
midnight,20 frequently occur in the population of sleepers
Good sleep quality is a critical factor in the psychological and
with poor sleep quality, suggesting sleep patterns and habits,
physical health development of the youth.1,2 However, many
bedtime habits and preferences in particular, may be import-
studies have reported that poor sleep quality is common
ant factors that determine sleep quality.
among college students and seriously impacts their everyday
Bedtime procrastination, defined as “failing to go to bed at
lives;3,4 these include an increased feeling of fatigue,5 attenu-
the intended time while no external circumstances prevent a
ation of learning and memory,6 and difficulties in sustaining
attention and regulating emotions,7,8 which could further person from doing so”, is a type of health behavior procrastin-
affect a student’s overall academic performance and social ation that has been found in relation to sleep habits.24
life.9,10 The previous study has shown that about 50% of Bedtime routines (e.g., brushing teeth, setting the alarm, and
Chinese university students report that their hours of sleep is washing the face/removing make-up) and immersive activities
insufficient, and less than six hours of sleep is independently before sleeping (e.g., using media without self-control) are
and significantly associated with a poor self-reported health suggested to induce bedtime procrastination,25,26 making it
status.11 Therefore, it is crucial to understand what factors possible that bedtime procrastination contributes to the quality
contribute to sleep quality in order to avoid undesirable con- of sleep of young people. Specifically, physiological and psy-
sequences among Chinese college students. chological evidence has shown that chronotype and self-con-
It has been well-recognized that several physiological, trol may account for bedtime procrastination, which, in turn,
psychological and external environmental factors, such as may influence sleep quality. For instance, people with an even-
sex,12 body weight,13 chronotype,14 psychological distress,15 ing-type or self-control failure are likely to have more bedtime
academic stress,16 and social media,17 can affect the quality procrastination behaviors.27–29
of sleep of younger people. Recently, an increasing number In addition, it is noted that young adults, who are in the
of studies have focused on investigating the association special transition period from adolescence to adulthood,
between an individual’s sleep hygiene behaviors and sleep exhibit more evening chronotypes than other age groups
quality.18–21 For example, unhealthy nighttime habits, such due to their biological characteristics.30,31 Moreover, some
as using electronic media before bedtime,22 consuming caf- studies have also demonstrated that college students are eas-
feine and alcohol after 6 p.m.,18 doing too many activities ily addicted to the use of media or the Internet due to

CONTACT Li Mu [email protected] Research Center of Brain and Cognitive Neuroscience, Liaoning Normal University, No.850 Huanghe Road,
Shahekou District, Dalian, Liaoning Province, 116029, China.
Xiaohan Ma and Dexin Meng: These authors contributed equally to this work as the first author.
ß 2020 Taylor & Francis Group, LLC
2 X. MA ET AL.

theirlack of self-regulation skills.32,33 Thus, young adults education, height, and body weight. Body mass index (BMI)
may present more bedtime procrastination behaviors that was calculated as weight (kg) divided by the square of height
further result in subjective sleep complaints. Previous studies (m2).
have demonstrated that bedtime procrastination is an
important cause of sleep insufficiency of Polish adolescents34
and college students.35 However, it remains unknown Bedtime procrastination
Bedtime procrastination was assessed using the Bedtime
whether such observation also occurs in the population of
Chinese youth. Procrastination Scale (BPS),24 which has repeatedly been
It is well-known that sleep duration (sleep insufficiency) reported as having good internal and external validity and
is one dimension of sleep quality; the other dimensions acceptable test-retest reliability.24,27 The BPS contains nine
include sleep latency, daytime dysfunction, sleep efficiency, items involving nine different scenarios of bedtime procras-
sleep disturbances, and so on. There is no study that tination behaviors. Each item is answered on a 5-point scale
explores the relationship between bedtime procrastination ranging from 1 (never) to 5 (always) according to the fre-
and the other dimensions of sleep quality. Understanding quency of experiencing each scenario. Among these items,
the impact of bedtime procrastination on sleep quality will four (Items 2, 3, 7 and 9) are reverse-scored. The Chinese
greatly benefit the development of solutions or treatments to version of BPS (CBPS) was adapted to measure the behav-
improve sleep quality in college students. Therefore, the cur- iors of bedtime procrastination among Chinese students in
rent study focused on examining the relationship between the present study (Cronbach’s a ¼ 0.82) and in our previous
bedtime procrastination and sleep quality in university stu- study.36A principal component factor analysis was restricted
dents in China. In this population, we hypothesized that to extract one factor (eigenvalue ¼ 3.82) according to previ-
bedtime procrastination may correlate with sleep quality, as ous studies.24,35 The CBPS score, which is the average of the
increasing bedtime procrastination behaviors would lead to score of the nine items and ranges from 1 to 5, was used to
a decrease in sleep quality. Results from this study will pro- indicate the degree of bedtime procrastination, with a higher
vide evidence for formulating interventions for sleep-related score indicating more bedtime procrastination behaviors.
procrastination behaviors, and eventually will help improve
sleep quality in the younger Chinese population. Self-reported sleep quality
The Pittsburgh Sleep Quality Index (PSQI) is a self-rated
Methods questionnaire that assesses subjective sleep quality and dis-
turbances over a one-month period.37 The reliability and
Participants validity of the Chinese version of the PSQI (CPSQI) has
This study was conducted in Northeast China from April 2018 been confirmed and is widely used in epidemiology studies
to December 2019. A total of 1590 undergraduate students in China.38,39 It consists of 19 items with global scores rang-
from five public universities were enrolled in the study. They ing from 0 to 21, generating seven component scores: sub-
were randomly recruited by an advertisement for freshmen, jective sleep quality, sleep latency, sleep duration, habitual
sophomores, and juniors at one university (12.14%), by the sleep efficiency, sleep disturbances, use of sleeping medica-
disproportionate stratified random sampling method for fresh- tion, and daytime dysfunction. The scoring of each compo-
men at another university (8.62%), for freshmen, sophomores, nent is based on a 0  3 Likert scale, where 3 reflects the
and juniors at two more universities (39.06%/11.95%), and for negative extreme. A CPSQI global score of 5 is considered
sophomores and juniors at one college (28.24%). All partici- as the threshold in distinguishing good sleepers ( 5) from
pants were required to complete a series of questionnaires poor sleepers (> 5),38 and this threshold was also used in
about their demographics, bedtime procrastination, and self- this study. The Cronbach’s a used in the present sample was
reported sleep quality. Subjects were excluded if they had a 0.65, which is similar to previous reports in Chinese college
large amount of missing and/or multiple-choice responses (
students.40,41
12%), or regular responses with the same option (e.g.,
111111111) or loop/circular options (e.g., 123454321 or
Statistical analysis
123123123) in any questionnaire. Finally, 1550 respondents
were included in the statistical analyses with a 97.48% effective Continuous variables were expressed as the mean plus or
response rate. All participating students gave verbal or written minus the standard deviation, and categorical variables were
informed consent. This cross-sectional study was approved by expressed through frequency and proportion. Missing items
the Ethics Committee or its equivalent in all five universities in the CBPS and CPSQI were replaced by the mean of that
prior to data collection. item. However, none of the missing demographic variables
were replaced. To determine group differences, analysis of
variance (ANOVA) or chi-square was used to compare the
Measures
demographic variables, CBPS score, and CPSQI global score
Demographic characteristics in good and poor sleepers. Analysis of covariance
A Basic Information Questionnaire was used to collect (ANCOVA) was used to exclude the effect of covariates
demographic characteristics, including age, sex, years of such as age, sex, years of education, and BMI.
 JOURNAL OF AMERICAN COLLEGE HEALTH 3

Binary logistic regression analysis was conducted to from 16 to 23 years old) and years of education was on aver-
determine the related potential variables that were the stron- age 12.61 (SD ¼ 0.62 years, range from 12 to 14 years).
gest factors that showed a correlation with poor sleep qual- Nearly half of the population were freshmen (46.65%), with
ity, after controlling the demographic variables, including the remainder mostly sophomores (45.94%). The mean BMI
age, sex, years of education, and BMI. To further determine was 21.43 kg/m2 (SD ¼ 3.42 kg/m2).
the variables with the ability to identify poor sleepers, the As shown in Table 1, the average CPSQI global score was
receiver operating characteristic (ROC) curve was generated 5.13 (SD ¼ 2.76, 95% CI: 4.99  5.26). Based on the CPSQI
using MedCalc version 15.2.2 (MedCalc Software, global scores, 611 (39.42%) students were identified as poor
Mariakerke, Belgium). The area under the ROC curve sleepers with an average score of 7.83 (SD ¼ 2.09, 95% CI:
(AUC) was calculated, and the cutoff score of the optimal 7.66  7.99), whereas the rest of the 939 subjects (60.58%)
sensitivity and specificity was determined based on the were good sleepers with an average score of 3.37 (SD ¼ 1.38,
Youden index (Youden index ¼ sensitivity þ specificity  1). 95% CI: 3.28  3.46). However, no significant differences
In addition, the correlation of the CPSQI global scores between good sleepers and poor sleepers were found in
with demographic characteristics and CBPS scores was most demographic characteristics (e.g., age, sex, and BMI)
assessed using Pearson’s (for continuous variables) or (all p > 0.05), except for years of education where that of
Spearman’s (for categorical variables) correlation coefficients poor sleepers was significantly shorter than that of good
among poor sleepers. Multiple regression analysis through sleepers (F [1, 1548] ¼ 8.266, p ¼ 0.004, g2p ¼ 0.005). The
the stepwise method was further applied to analyze the rela- average CBPS scores were 3.16 (SD ¼ 0.76, 95% CI:
tionships among the severity of sleep quality, the degree of 3.10  3.22) in the poor sleeper group and 2.61 (SD ¼ 0.78,
bedtime procrastination, and demographic variables in poor 95% CI: 2.56  2.66) in the good sleeper group, suggesting
sleeper group. Demographic variables that have been repeat- that poor sleepers had more bedtime procrastination behav-
edly proven to have correlations with sleep quality, such as iors than good sleepers (F [1, 1548] ¼ 185.430, p < 0.001, g2p
age, sex, years of education, and BMI, were entered into the ¼ 0.107). Such difference remained significant in the covari-
multiple regression models as covariates. ance analysis with the demographic characteristics as covari-
Data were analyzed using SPSS version 16.0 for Windows ates (F [1, 1481] ¼ 169.523, p < 0.001, g2p ¼ 0.103).
(SPSS Inc., Chicago, IL, USA). Differences were considered
to be statistically significant at p < 0.05.
Factors that predicted the prevalence of poor
sleep quality
Results
Binary logistic regression analysis was conducted to assess
Demographic characteristics, bedtime procrastination, the independent factors that affect the prevalence of poor
and self-reported sleep quality of good sleepers and
sleep quality. As shown in Table 2, the risk of having self-
poor sleepers
reported poor sleep quality was about 2.5 times higher in
Table 1 outlines the differences in demographic characteris- students who had more bedtime procrastination behaviors
tics, bedtime procrastination, and self-reported sleep quality (OR ¼ 2.424, 95% CI ¼ 2.092  2.808, p < 0.001). Moreover,
of all participants, subjects with good and poor sleep quality. older students showed a higher risk of suffering from poor
Of the 1550 subjects, 459 (29.61%) were males, 1074 sleep quality than younger ones (OR ¼ 1.267, 95% CI ¼
(69.29%) were females, and 17 (1.10%) did not provide a 1.115  1.440, p < 0.001), and students with higher education
response to the question on their sex. Forty-three subjects showed a lower risk of poor sleep quality than those with
(2.77%) did not respond to age, weight, and height. The lower education (OR ¼ 0.729, 95% CI ¼ 0.594  0.893,
average age of the subjects was 19.30 (SD ¼ 0.98 years, range p ¼ 0.002). The relative abilities of using the CBPS score,

Table 1. Demographic characteristics, bedtime procrastination and self-reported sleep qualitya.

Poor sleepers Good sleepers
Variables Total CPSQI > 5 CPSQI  5 F/v2 p
Demographic characteristics
n (%) 1550 (100) 611 (39.42) 939 (60.58)
Age, Mean (SD), y 19.30 (0.98) 19.36 (0.96) 19.27 (0.99) 3.037 0.082
Sex, n (%)b 0.401 0.527
Male 459 (29.61) 175 (28.64) 284 (30.24)
Female 1074 (69.29) 428 (70.05) 646 (68.80)
Not filled 17 (1.10) 8 (1.31) 9 (0.96)
Years of Education, Mean (SD), y 12.61 (0.62) 12.55 (0.60) 12.64 (0.64) 8.266 0.004
12, n (%) 723 (46.65) 307 (50.25) 416 (44.30) 8.890 0.012
13, n (%) 712 (45.94) 271 (44.35) 441 (46.96)
14, n (%) 115 (7.42) 33 (5.40) 82 (8.73)
BMI, Mean (SD), kg/m2 21.43 (3.42) 21.47 (3.48) 21.40 (3.38) 0.127 0.721
CBPS Score, Mean (SD) 2.83 (0.82) 3.16 (0.76) 2.61 (0.78) 185.430 <0.001
CPSQI Global Score, Mean (SD) 5.13 (2.76) 7.83 (2.09) 3.37 (1.38) 2559.878 <0.001
a
Data are missing for some variables, because not all of the subjects had demographic data.
b
Not filled data were not included in Chi-square test. BMI ¼ Body Mass Index, CBPS ¼ the Chinese Version of Bedtime Procrastination Scale, CPSQI ¼ the Chinese
Version of Pittsburgh Sleep Quality Index, SD ¼ Standard Deviation.
4 X. MA ET AL.

Table 2. Logistic regression analysis predicting the prevalence of poor sleepers.

95% CI
Factors B SE Wald OR Lower Upper p
Age 0.237 0.065 13.220 1.267 1.115 1.440 < 0.001
Years of Education 0.317 0.104 9.322 0.729 0.594 0.893 0.002
CBPS Score 0.885 0.075 139.057 2.424 2.092 2.808 < 0.001
CBPS ¼ the Chinese Version of Bedtime Procrastination Scale, CI ¼ Confidence Interval, OR ¼ Odds Ratio, SE ¼ Standard Error.

age, and years of education to identify poor sleepers were

0.885, 0.237 and 0.317, respectively (Table 2). Therefore,
these results suggest that age, education, and the CBPS score
are independent predictors that can each be used in deter-
mining poor sleep quality.
The ROC analysis further showed that the CBPS score
was higher by 2.66 with a sensitivity of 76.92% and specifi-
city of 52.61% in discriminating poor sleepers (AUC ¼
0.692, 95% CI ¼ 0.668  0.715, Z ¼ 14.333, p < 0.001)
(Figure 1 and Table 3), whereas an age greater than 19 years
only showed a sensitivity of 40.13% and specificity of
64.44% in identifying poor sleepers (AUC ¼ 0.529, 95% CI
¼ 0.503  0.554, Z ¼ 1.990, p ¼ 0.047), and years of educa-
tion not exceeding 12 years had a low sensitivity of 50.25%
and specificity of 55.70% in discriminating poor sleepers
(AUC ¼ 0.536, 95% CI ¼ 0.511  0.561, Z ¼ 2.736,
p ¼ 0.006) (Figure 1 and Table 3). These results suggest that
the CBPS score is better than age (Z ¼ 8.069, p < 0.001) and
years of education (Z ¼ 8.672, p < 0.001) in identifying self-
reported poor sleep quality. Moreover, the combined score Figure 1. The receiver operating characteristic curve of the risk factors for iden-
of these three risk factors (AUC ¼ 0.702, 95% CI ¼ tifying poor sleep quality of Chinese undergraduate students. CBPS ¼ the
Chinese Version of Bedtime Procrastination Scale.
0.678  0.725, Z ¼ 14.989, p < 0.001) showed better diagnos-
tic reliability than age (Z ¼ 9.341, p < 0.001) or years of edu- Multiple linear regression analyses were further per-
cation alone (Z ¼ 9.821, p < 0.001) (Figure 1 and Table 3). formed to evaluate the CPSQI scores in relation to demo-
However, no significant difference was observed when com- graphic characteristics and the CBPS score of poor sleepers
paring the predictive power of the combined score to the (Table 5). The results showed that the CBPS score was an
CBPS score (Z ¼ 1.947, p ¼ 0.052). independent contributor to the CPSQI global score
(B ¼ 0.916, t ¼ 8.617, p < 0.001) and scores of subjective
sleep quality (B ¼ 0.259, t ¼ 7.460, p < 0.001), sleep duration
Factors associated with the severity of poor sleep (B ¼ 0.233, t ¼ 5.570, p < 0.001), and daytime dysfunction
quality in poor sleepers (B ¼ 0.204, t ¼ 5.028, p < 0.001) in poor sleepers. On the
The relationships between the CPSQI scores and demo- other hand, age was not an independent contributor to the
CPSQI global and component scores (all p > 0.05), while
graphic characteristics or the CBPS score were examined to
years of education was identified as a potential indicator of
find the factors that may contribute to the severity of poor
sleep duration (B ¼ 0.356, t ¼ 6.707, p < 0.001).
sleep quality of poor sleepers (Table 4). It was observed that
Subjective sleep quality was independently predicted by sex
the CBPS score positively correlated with the CPSQI global (B ¼ 0.135, t ¼ 2.339, p ¼ 0.020).
score (r ¼ 0.324, p < 0.001), suggesting that global sleep
quality decreases as the CBPS score increases. Moreover, the
results showed that higher CBPS scores were related to Discussion
poorer subjective sleep quality (r ¼ 0.303, p < 0.001), shorter
This cross-sectional study, to the best of our knowledge, is
sleep duration (r ¼ 0.249, p < 0.001), and more severe day- the first to evaluate the association between bedtime pro-
time dysfunction (r ¼ 0.202, p < 0.001). However, no signifi- crastination and sleep quality among Chinese undergraduate
cant correlation was seen between bedtime procrastination students. In this study, we found the high prevalence of self-
behaviors and other CPSQI components (all p > 0.05). In reported poor sleep quality of Chinese college students, and
addition, it was found that age and education were related identified that the CBPS score, years of education, and age
to sleep latency (r ¼ 0.090, p ¼ 0.029; r ¼ 0.105, p ¼ 0.009) are the predictive factors of poor sleep quality at the uni-
and sleep duration (r ¼ 0.160, p < 0.001; r ¼ 0.287, variate and multivariate levels, and illustrated the positive
p < 0.001). Age (r ¼ 0.092, p ¼ 0.025) and BMI (r ¼ 0.082, correlation between the severity of poor sleep quality and
p ¼ 0.047) were associated with sleep disturbances. the degree of bedtime procrastination of poor sleepers.
 JOURNAL OF AMERICAN COLLEGE HEALTH 5

Table 3. The ROC analyses of the optimized cutoff points for predicted factors in poor sleepers.
Factors AUC 95% CI Cutoff Sensitivity Specificity Z p
Age 0.529 0.503–0.554 >19 40.13 64.44 1.990 0.047
Years of Education 0.536 0.511–0.561 12 50.25 55.70 2.736 0.006
CBPS Score 0.692 0.668–0.715 >2.66 76.92 52.61 14.333 <0.001
Combined Score 0.702 0.678–0.725 – 80.61 49.23 14.989 <0.001
CBPS ¼ the Chinese Version of Bedtime Procrastination Scale, ROC ¼ Receiver Operating Characteristic, AUC ¼ Area Under the Curve,
CI ¼ Confidence Interval.

Table 4. Correlation between CPSQI scores and demographic characteristics or CBPS scores in poor sleepers.
Variables Mean SD CBPS Score Age Sexa Education BMI
CPSQI Global Score 7.83 2.09 0.324 0.037 0.028 0.073 0.004
Subjective Sleep Quality 1.41 0.67 0.303 0.026 0.064 0.028 0.055
Sleep Latency 1.59 0.86 0.044 0.090 0.037 0.105 0.049
Sleep Duration 1.69 0.81 0.249 0.160 0.016 0.287 0.068
Habitual Sleep Efficiency 0.63 0.85 0.079 0.024 0.047 0.038 0.009
Sleep Disturbances 1.18 0.56 0.007 0.092 0.037 0.066 0.082
Use of Sleeping Medication 0.09 0.40 0.017 0.030 0.044 0.044 0.006
Daytime Dysfunction 1.24 0.77 0.202 0.064 0.054 0.056 0.027
a
Data were analyzed using Spearman correlation analysis. BMI ¼ Body Mass Index, CBPS ¼ the Chinese Version of Bedtime Procrastination
Scale, CPSQI ¼ the Chinese Version of Pittsburgh Sleep Quality Index. SD ¼ Standard Deviation.
p < 0.05.
p < 0.01.
p < 0.001.

Table 5. Multiple linear regression indicating predictors of the severity of poor sleep quality in poor sleepers.
Global Score Subjective Sleep Quality Sleep Duration Daytime Dysfunction
Factors B t p B t p B t p B t p
Sexa – – – 0.135 2.339 0.020 – – – – – –
Years of Education – – – – – – 0.356 6.707 < 0.001 – – –
CBPS Score 0.916 8.617 <0.001 0.259 7.460 <0.001 0.233 5.570 < 0.001 0.204 5.028 <0.001
a
Sex: male ¼ 1 and female ¼ 2. CBPS ¼ the Chinese Version of Bedtime Procrastination Scale, CPSQI ¼ the Chinese Version of Pittsburgh Sleep Quality Index.

In this college-based survey of CPSQI, we reported that found more than 12 years of education (equivalent to higher
the prevalence of self-reported poor sleep quality is 39.42%, than high school education in China) has a protective effect
which is similar to previous studies assessing the prevalence on sleep quality. This phenomenon may be explained by the
of poor sleepers among university students in mainland relationship between academic and psychological stress and
China (41.4%, 171/413),21 Nepal (35.4%, 332/937),42 and the sleep quality.15,16 Freshmen with 12 years of education have
US (42.4%, 266/628).3 However, this prevalence is much to face the difficulty of adapting to a new life in college and
higher than two other studies also involving Chinese college the challenge of more professional curricula.41
students (9.8% and 17.7%).43,44 This discrepancy could be Our findings were also in agreement with some previous
partly attributed to the differences in the case definition of studies that suggest that age is associated with the preva-
poor sleep quality, regions, and/or sample size of the study lence of poor sleep quality.11,41,42,50 Interestingly, our obser-
population. For example, Feng et al. used a cutoff value of vation that older students predicted the prevalence of self-
seven in the CPSQI to identify poor sleepers;43 this is higher reported poor sleep quality in the undergraduate population
than the cutoff value of five that we used in this study. Wu was consistent with evidence on the effect of age on the
et al. included 4747 Chinese college students, far more than sleep quality of the healthy Chinese population.11 However,
our sample size, and their participants were recruited from it was inconsistent with a few studies about college students
universities in southern China.44 in Nigeria 15 and Saudi Arabia,20 which showed that age
Some demographic characteristics, such as age,42 sex,45 does not produce any significant variation in sleep quality.
education,46 and BMI,47,48 were assessed as potential predic- Regional social-cultural differences may be responsible for
tors of poor sleep quality due to their significant differences this discrepancy. Moreover, differences in the variable form
in sleep quality. However, we did not see any significant dif- of age may be used to explain this inconsistency. Age
ference between good and poor sleepers based on these vari- entered the statistical model as a continuous variable in this
ables, except for education, which was also identified as an and previous studies of the Chinese population,11 while it
independent predictor of poor sleepers in our multivariate was regarded as a dichotomous or polytomous variable in
model. We found that students with lower levels of educa- the studies from Nigeria15 and Saudi Arabia.20 Due to the
tion were more likely to suffer from poor sleep quality, limited range of age in college students, it seemed more rea-
which is similar to a previous study that suggested a 50% sonable to take age as a continuous variable to reflect the
reduction in the odds a person had poor sleep when they effect of age on sleep quality. Our ROC results showed that
were, at least, at college level education (compared to less an age higher than 19-year-old differentiated poor sleepers
than high school education).49 Our ROC results further further; and this confirms the rationality of age as a
6 X. MA ET AL.

continuous variable, which is more helpful in revealing the on the relationship between bedtime procrastination and
effect of age on sleep quality. In addition, some studies indi- sleep quality is still unclear, and further investigation is
cated that younger students tend to get poorer sleep.42 necessary to better understand the association between bed-
However, we found no correlation between age and the time procrastination and poor sleep quality.
severity of poor sleep quality among Chinese college stu- In addition, our results further demonstrated that the
dents, suggesting that other variables might contribute to CBPS was a moderate screening tool in discriminating poor
the severity of poor sleep quality. sleepers from good sleepers among Chinese undergraduate
It is worth noting that, in our study, bedtime procrastin- students. A previous study used 3.33, a 75% quartile on the
ation was considered a strong indicator in predicting not BPS score, as the cutoff value in distinguishing severe bed-
only the prevalence, but also the severity of self-reported time procrastination behaviors in a representative Dutch
poor sleep quality among students when the demographic sample.51 However, such cutoff value was much higher than
variables are comparably controlled. This supports our the optimal cutoff value of 2.66, making their identification
hypothesis that students with more bedtime procrastination of poor sleep quality less sensitive (40.13% in our sample).
behaviors are more likely to exhibit worse sleep quality, Thus, using the CBPS cutoff value of 2.66 in our study pro-
which is also consistent with previous findings.27,51 For vides higher sensitivity and specificity in identifying Chinese
example, Kroese et al. demonstrated that the impact of bed- youth with poor sleep quality. It seems like expanding bed-
time procrastination on poor sleep quality is due to insuffi- time procrastination to the intervention of sleep quality
cient sleep caused by reducing sleep hours.24,27 Our results could be more impactful from a public health perspective.
agreed with their finding after significant correlations were The strengths of our study include factors such as our
observed between bedtime procrastination and subjective large sample size, multi-center design, and multivariate ana-
sleep quality, bedtime procrastination and sleep duration, lysis. However, several limitations should be noted. First, the
and a surprising proportion that reported less than seven generalizability of our findings is limited to Northeast
hours of sleep among all samples (74.32%) and poor sleepers China, which might not be able to be extrapolated to other
(93.94%), respectively. Moreover, this study revealed that local regions of China or all over the China. Regional differ-
daytime dysfunction exhibited a close correlation with bed- ences in sleep quality have been reported. For example, the
time procrastination of poor sleepers, which is in agreement proportion of college students with poor sleep quality in
with previous studies, demonstrating that the bedtime pro- Hong Kong and Macao is significantly higher than that in
crastination phenomenon is associated with unhealthy out- mainland China.11 Regional differences in the economic
comes (e.g., fatigue) in the general population.24,27 development, social-cultural context, and climate may be the
Taken together, these results indicate that bedtime pro- reasons for such a discrepancy. Second, the study sample
crastination affects various aspects of sleep quality. This may was limited to Chinese undergraduate students. There was
be explained by evidence from studies related to physio- no data collection on older/younger people. Third, sleep
logical (circadian typology) and psychological (self-regula- measurements in this study were based on self-report ques-
tion) factors.27,29,52 Due to the influence of biological and tionnaires; thus, they lack objective evaluation. Moreover,
psychosocial factors, young adults tend to present circadian these instruments collected retrospective self-report data
phase delays that may result in more evening chronotypes over the last two weeks or one month. Therefore, the pre-
in this population.30,31 K€uhnel et al. observed the interaction sent study cannot provide day-specific or a more recent
between bedtime procrastination over the course of week- evaluation of bedtime procrastination and sleep quality.
days and the variability and alteration in sleep-wake sched- Fourth, a cutoff value of five in the CPSQI has been proven
ules.29 They found that evening-type subjects tended to to have a high diagnostic sensitivity (98%) but a low specifi-
report more bedtime procrastination behaviors than morn- city of 55% for self-reported poor sleep quality.38 Thus,
ing types, but day-specific bedtime procrastination declined recall bias and being misclassified as poor sleepers might be
over the course of the work week in later chronotypes.29 difficult to avoid. Fifth, the internal consistency of the PSQI
Thus, these data suggest that poor sleep quality is strongly in this study sample was low (Cronbach’s a ¼ 0.65).
influenced by bedtime procrastination, which is linked to Although it was similar to previous reports of Chinese col-
the individual chronotype of college students. lege students,40,41 this low consistency may affect the reli-
On the other hand, some studies revealed that lower lev- ability of the sleep quality assessment. More reliable tools
els of trait self-control was significantly associated with eve- should be sought in the future to assess the sleep quality of
ningness, which, in turn, predicted higher levels of bedtime college students. Last but not least, this study was performed
procrastination.28 Specifically, young adults with trait self- with a cross-sectional design that does not show the causal
regulatory resources depletion were likely to gain more relationship between sleep quality and bedtime procrastin-
bedtime procrastination behaviors after a stressful day,52 ation. A longitudinal study design, on the other hand, would
supporting the view that bedtime procrastination is a self- allow for an examination of the evolution of sleep quality
control failure.27,28 However, another study from K€ uhnel over bedtime procrastination or an exploration of the alter-
et al. showed the opposite trend; employees with less avail- native direction that poor sleep quality may contribute to
ability of momentary self-regulatory resources before going bedtime procrastination.
to bed showed lower bedtime procrastination.29 Therefore, In summary, our study found that poor sleep quality was
the exact mechanism that underlies the effect of self-control prevalent among younger Chinese students and was
 JOURNAL OF AMERICAN COLLEGE HEALTH 7

significantly associated with older age, shorter education 7. Louca M, Short MA. The effect of one night’s sleep deprivation
years, and more bedtime procrastination behaviors. A CBPS on adolescent neurobehavioral performance. Sleep. 2014;37(11):
1799–1807. doi:10.5665/sleep.4174.
score greater than 2.66 could be an independent indicator 8. Mauss IB, Troy AS, LeBourgeois MK. Poorer sleep quality is asso-
for the prevalence of poor sleep quality. In poor sleepers, ciated with lower emotion-regulation ability in a laboratory para-
the severity of sleep quality was positively correlated with digm. Cogn Emot. 2013;27(3):567–576. doi:10.1080/02699931.2012.
the degree of bedtime procrastination. These findings sug- 727783.
gest the importance and necessity of formulating effective 9. Mirghani HO, Mohammed OS, Almurtadha YM, Ahmed MS.
Good sleep quality is associated with better academic perform-
interventions for sleep-related procrastination behavior to ance among Sudanese medical students. BMC Res Notes. 2015;
help improve the sleep quality of Chinese college students. 8(1):706–710. doi:10.1186/s13104-015-1712-9.
Longitudinal studies are also needed to further elucidate any 10. Rezaei O, Mokhayeri Y, Haroni J, et al. Association between
causal relationship between bedtime procrastination and sleep quality and quality of life among students: a cross sectional
poor sleep quality, especially including the impact of poor study. Int J Adolesc Med Health. 2017;32(2):1–6. doi:10.1515/
ijamh-2017-0111.
sleep quality on bedtime procrastination using a day-spe- 11. Li L, Lok KI, Mei SL, et al. Sleep duration and self-rated health
cial CBPS. in Chinese university students. Sleep Breath. 2019;23(4):
1351–1356. doi:10.1007/s11325-019-01856-w.
12. Galland BC, Gray AR, Penno J, Smith C, Lobb C, Taylor RW.
Conflict of interest disclosure Gender differences in sleep hygiene practices and sleep quality in
New Zealand adolescents aged 15 to 17 years. Sleep Health. 2017;
The authors have no conflicts of interest to report. The 3(2):77–83. doi:10.1016/j.sleh.2017.02.001.
authors confirm that the research presented in this article 13. Kristicevic T, Stefan L, Sporis G. The associations between sleep
met the ethical guidelines, including adherence to the legal duration and sleep quality with body-mass index in a large sam-
requirements of the People’s Republic of China and received ple of young adults. Int J Environ Res Public Health. 2018;15(4):
758–767. doi:10.3390/ijerph15040758.
approval from the Ethics Committee or its equivalent of 14. Vollmer C, Jankowski KS, Dıaz-Morales JF, Itzek-Greulich H,
Liaoning Normal University, Jiamusi University, Changchun W€ ust-Ackermann P, Randler C. Morningness-eveningness corre-
University of Chinese Medicine, Dalian University, and lates with sleep time, quality, and hygiene in secondary school
Jilin University. students: a multilevel analysis. Sleep Med. 2017;30:151–159. doi:
10.1016/j.sleep.2016.09.022.
15. Seun-Fadipe CT, Mosaku KS. Sleep quality and psychological
Funding distress among undergraduate students of a Nigerian university.
Sleep Health. 2017;3(3):190–194. doi:10.1016/j.sleh.2017.02.004.
This work was supported by the National Natural Science Foundation 16. Alsaggaf MA, Wali SO, Merdad RA, Merdad LA. Sleep quantity,
of China (No. 31800924) and the Scientific Study Project for Education quality, and insomnia symptoms of medical students during clin-
Department of Liaoning Province (No. L201783641). ical years. Relationship with stress and academic performance.
Saudi Med J. 2016;37(2):173–182. doi:10.15537/smj.2016.2.14288.
17. Long Xu X, Zhu RZ, Sharma M, Zhao Y. The influence of social
media on sleep quality: a study of undergraduate students in
ORCID Chongqing, China. J Nurs Care. 2015;4(3):253–259. doi:10.4172/
2167-1168.1000253.
Li Mu http://orcid.org/0000-0003-2442-0221 18. Brick CA, Seely DL, Palermo TM. Association between sleep
hygiene and sleep quality in medical students. Behav Sleep Med.
2010;8(2):113–121. doi:10.1080/15402001003622925.
References 19. Lund HG, Reider BD, Whiting AB, Prichard JR. Sleep patterns
and predictors of disturbed sleep in a large population of college
1. Tarokh L, Saletin JM, Carskadon MA. Sleep in adolescence:
students. J Adolesc Health. 2010;46(2):124–132. doi:10.1016/j.
physiology, cognition and mental health. Neurosci Biobehav Rev.
jadohealth.2009.06.016.
2016;70:182–188. doi:10.1016/j.neubiorev.2016.08.008. 20. Siddiqui AF, Al-Musa H, Al-Amri H, Al-Qahtani A, Al-Shahrani
2. Magnavita N, Garbarino S. Sleep, health and wellness at work: a
M, Al-Qahtani M. Sleep patterns and predictors of poor sleep
scoping review. IJERPH. 2017;14(11):1347–1364. doi:10.3390/ quality among medical students in King Khalid University, Saudi
ijerph14111347. Arabia. Malays J Med Sci. 2016;23(6):94–102. doi:10.21315/
3. Mah CD, Kezirian EJ, Marcello BM, Dement WC. Poor sleep mjms2016.23.6.10.
quality and insufficient sleep of a collegiate student-athlete popu- 21. Li J, Zhou K, Li X, et al. Mediator effect of sleep hygiene practi-
lation. Sleep Health. 2018;4(3):251–257. doi:10.1016/j.sleh.2018. ces on relationships between sleep quality and other sleep-related
02.005. factors in Chinese mainland university students. Behav Sleep
4. James B, Omoaregba J, Igberase O. Prevalence and correlates of Med. 2016;14(1):85–99. doi:10.1080/15402002.2014.954116.
poor sleep quality among medical students at a Nigerian univer- 22. Levenson JC, Shensa A, Sidani JE, Colditz JB, Primack BA.
sity. Ann Nigerian Med. 2011;5(1):1–5. doi:10.4103/0331-3131. Social media use before bed and sleep disturbance among young
84218. adults in the united states: a nationally-representative study.
5. Herring MP, Monroe DC, Kline CE, O’Connor PJ, MacDonncha Sleep. 2017;40(9):1–7. doi:10.1093/sleep/zsx113.
C. Sleep quality moderates the association between physical 23. Bruni O, Sette S, Fontanesi L, Baiocco R, Laghi F, Baumgartner
activity frequency and feelings of energy and fatigue in adoles- E. Technology use and sleep quality in preadolescence and ado-
cents. Eur Child Adolesc Psychiatry. 2018;27(11):1425–1432. doi: lescence. J Clin Sleep Med. 2015;11(12):1433–1441. doi:10.5664/
10.1007/s00787-018-1134-z. jcsm.5282.
6. Gobin CM, Banks JB, Fins AI, Tartar JL. Poor sleep quality is 24. Kroese FM, de Ridder DT, Evers C, Adriaanse MA. Bedtime
associated with a negative cognitive bias and decreased sustained procrastination: introducing a new area of procrastination. Front
attention. J Sleep Res. 2015;24(5):535–542. doi:10.1111/jsr.12302. Psychol. 2014;5:611–618. doi:10.3389/fpsyg.2014.00611.
8 X. MA ET AL.

25. Nauts S, Kamphorst BA, Sutu AE, Poortvliet R, Anderson JH. 39. Luo J, Zhu G, Zhao Q, et al. Prevalence and risk factors of poor
Aversive bedtime routines as a precursor to bedtime procrastin- sleep quality among Chinese elderly in an urban community:
ation. European Health Psychol. 2016;18(2):80–85. results from the Shanghai aging study. PLoS One. 2013;8(11):
26. Exelmans L, Van den Bulck J. “Glued to the tube”: the interplay e81261. doi:10.1371/journal.pone.0081261.
between self-control, evening television viewing, and bedtime 40. Guo S, Sun W, Liu C, Wu S. Structural validity of the pittsburgh
procrastination. Commun Res. 2017:1–23. doi:10.1177/ sleep quality index in Chinese undergraduate students. Front
0093650216686877. Psychol. 2016;7:1126. doi:10.3389/fpsyg.2016.01126.
27. Kroese FM, Evers C, Adriaanse MA, de Ridder D. Bedtime pro- 41. Suen LK, Hon KL, Tam WW. Association between sleep behav-
crastination: a self-regulation perspective on sleep insufficiency ior and sleep-related factors among university students in Hong
in the general population. J Health Psychol. 2016;21(5):853–862. Kong. Chronobiol Int. 2008;25(5):760–775. doi:10.1080/
doi:10.1177/1359105314540014. 07420520802397186.
28. Kadzikowska-Wrzosek R. Self-regulation and bedtime procrastin- 42. Bhandari PM, Neupane D, Rijal S, Thapa K, Mishra SR, Poudyal
ation: The role of self-regulation skills and chronotype. AK. Sleep quality, internet addiction and depressive symptoms
Personality and Individual Differences. 2018;128:10–15. doi:10. among undergraduate students in Nepal. BMC Psychiatry. 2017;
1016/j.paid.2018.02.015. 17(1):106–113. doi:10.1186/s12888-017-1275-5.
29. K€uhnel J, Syrek CJ, Dreher A. Why don’t you go to bed on 43. Feng Q, Zhang QL, Du Y, Ye YL, He QQ. Associations of phys-
time? A daily diary study on the relationships between chrono- ical activity, screen time with depression, anxiety and sleep qual-
type, self-control resources and the phenomenon of bedtime pro- ity among Chinese college freshmen. PLoS One. 2014;9(6):
crastination. Front Psychol. 2018;9:77–87. doi:10.3389/fpsyg.2018. e100914. doi:10.1371/journal.pone.0100914.
00077. 44. Wu X, Tao S, Zhang Y, Zhang S, Tao F. Low physical activity
30. Dıaz-Morales J, Gutierrez Sorroche M. Morningness-eveningness and high screen time can increase the risks of mental health
in adolescents. Span J Psychol. 2008;11(1):201–206. doi:10.1017/ problems and poor sleep quality among Chinese college students.
s1138741600004248. PLoS One. 2015;10(3):e0119607. doi:10.1371/journal.pone.
0119607.
31. Delgado Prieto P, Diaz-Morales J, Escribano B, Collado Mateo
45. Fatima Y, Doi SAR, Najman JM, Mamun AA. Exploring gender
M, Randler C. Morningness-eveningness and health-related qual-
difference in sleep quality of young adults: findings from a large
ity of life among adolescents. Span J Psychol. 2012;15(2):613–623.
population study. Clin Med Res. 2016;14(3-4):138–144. doi:10.
doi:10.5209/rev_sjop.2012.v15.n2.38872.
3121/cmr.2016.1338.
32. Jiang Z, Zhao X. Self-control and problematic mobile phone use
46. Marques DR, Meia-Via AMS, da Silva CF, Gomes AA.
in Chinese college students: the mediating role of mobile phone
Associations between sleep quality and domains of quality of life
use patterns. BMC Psychiatry. 2016;16(1):416. doi:10.1186/
in a non-clinical sample: results from higher education students.
s12888-016-1131-z.
Sleep Health. 2017;3(5):348–356. doi:10.1016/j.sleh.2017.07.004.
33. Kim J, Hong H, Lee J, Hyun MH. Effects of time perspective
47. Vargas PA, Flores M, Robles E. Sleep quality and body mass
and self-control on procrastination and Internet addiction. J index in college students: the role of sleep disturbances. J Am
Behav Addict. 2017;6(2):229–236. doi:10.1556/2006.6.2017.017. Coll Health. 2014;62(8):534–541. doi:10.1080/07448481.2014.
34. Kadzikowska-Wrzosek R. Insufficient sleep among adolescents: 933344.
the role of bedtime procrastination, chronotype and autonomous 48. Park S, Jung J, Oh C, McIntyre R, Lee J. Association between
vs. controlled motivational regulations. Curr Psychol. 2020; 39(3): sleep duration, quality and body mass index in the Korean popu-
1031–1040. doi:10.1007/s12144-018-9825-7. lation. J Clin Sleep Med. 2018;14(08):1353–1360. doi:10.5664/
35. Herzog-Krzywoszanska R, Krzywoszanski L. Bedtime procrastin- jcsm.7272.
ation, sleep-related behaviors, and demographic factors in an 49. Patel N, Grandner M, Xie D, Branas C, Gooneratne N. “Sleep
online survey on a Polish sample. Front Neurosci. 2019;13:963. disparity” in the population: poor sleep quality is strongly associ-
doi:10.3389/fnins.2019.00963. ated with poverty and ethnicity. BMC Public Health. 2010;10:
36. Guo J, Meng D, Ma X, Zhu L, Yang L, Mu L. The impact of 475. doi:10.1186/1471-2458-10-475.
bedtime procrastination on depression symptoms in Chinese 50. Dregan A, Armstrong D. Adolescence sleep disturbances as pre-
medical students. Sleep Breath. 2020;1–9. doi:10.1007/s11325- dictors of adulthood sleep disturbances-a cohort study. J Adolesc
020-02079-0 Health. 2010;46(5):482–487. doi:10.1016/j.jadohealth.2009.11.197.
37. Buysse DJ, Reynolds CF, Monk TH, Berman SR, Kupfer DJ. The 51. Nauts S, Kamphorst BA, Stut W, de Ridder DTD, Anderson JH.
pittsburgh sleep quality index: a new instrument for psychiatric The explanations people give for going to bed late: a qualitative
practice and research. Psychiatry Res. 1989;28(2):193–213. doi:10. study of the varieties of bedtime procrastination. Behavioral
1016/0165-1781(89)90047-4. Sleep Medicine. 2019;17(6):753–762. doi:10.1080/15402002.2018.
38. Tsai P-S, Wang S-Y, Wang M-Y, et al. Psychometric evaluation 1491850.
of the Chinese version of the Pittsburgh Sleep Quality Index 52. Bernecker K, Job V. Too exhausted to go to bed: implicit theo-
(CPSQI) in primary insomnia and control subjects. Qual Life ries about willpower and stress predict bedtime procrastination.
Res. 2005;14(8):1943–1952. doi:10.1007/s11136-005-4346-x. Br J Psychol. 2020;111(1):126–147. doi:10.1111/bjop.12382.