INTRODUCTION
During the COVID-19 pandemic, many high school students’ daily lives were disrupted. The disruptions during the school year in Korea were evident as the start of the school year was delayed by five weeks and a shift to online learning occurred [1]. This sudden change in the school environment affected students’ emotional states and routines [2]. Throughout 2021, the Centers for Disease Control and Prevention conducted the adolescent behaviors and experiences survey, revealing that since the beginning of the pandemic, 66% of students found it more challenging to complete their academic work and experienced difficulties in accessing education and healthcare [3]. Various psychological disturbances emerged as the COVID-19 pandemic caused significant changes in adolescents’ daily lives, including pressure from adjusting to online schooling, social isolation from peers, and spending excessive time with their immediate families, potentially leading to heightened tensions and increased conflicts [4,5].
Furthermore, many adolescents struggled to maintain regular eating and sleeping habits due to decreased activity and social interactions [6]. Many students reported reduced sleep quality and bedtime procrastination, with high school students particularly susceptible to excessive screen time during the pandemic due to quarantines, infections, and online classes, making them more vulnerable to sleep disturbances and further bedtime procrastination.
Bedtime procrastination is characterized as the intentional delay of bedtime without external justifications [7]. This behavior in adolescents can lead to or exacerbate sleep disturbances [8]. During the COVID-19 pandemic, bedtime procrastination became more severe and prevalent among youth [9]. The lack of a structured schedule and online learning contributed to this issue. Moreover, staying up late to watch videos or browse social media on the phone often leads to additional procrastination at bedtime, which results in poorer sleep quality and can negatively impact both physical and psychiatric health [10]. Engaging in smartphone use, video games, and online streaming before or in bed acts as an aftermath of disrupted routines, increased screen time, and heightened stress, anxiety, and feelings of isolation, with limited opportunities for physical activity due to lockdowns or social distancing [11].
During the COVID-19 pandemic, high school students experienced elevated levels of depression and anxiety. In a survey conducted two years after the onset of COVID-19, a significant number of high school students reported moderate to severe anxiety symptoms. High anxiety levels were associated with missing social events, spending time with friends at school, and students with high levels of anxiety believed these circumstances impacted their maturity and personal growth [12]. Another study also reported increased symptoms of depression, and a multivariable linear regression suggested that greater concerns about COVID-19 in school settings were significantly associated with higher depression levels [13]. The uncertainty surrounding the virus and fear of infection may have caused anxiety, leading many students to experience delayed sleep. Not only did the pandemic itself provoke viral anxiety, but the stress of disrupted daily routines and changes in classroom environments also fostered depression and distress. Preoccupation with the coronavirus is driven by anxiety, depression, and reassurance-seeking behaviors, with mediation analysis supporting the cognitive behavioral model of hypochondriasis, which can potentially lead to sleep disorders including bedtime delays and poor sleep quality [14]. A previous study on medical school students also reveals that viral anxiety and depression may affect individuals’ obsession with COVID-19, and reassurance-seeking behaviors may mediate this [15]. Similarly, anxiety and depression related to the pandemic may impact sleep hygiene among Korean high school students, ultimately leading to bedtime procrastination.
Bedtime procrastination, a concept introduced in 2014 by Kroese et al. [8], is defined as “failing to go to bed at the intended time, while no external circumstances prevent a person from doing so.” The Bedtime Procrastination Scale (BPS) was developed to assess individuals’ behaviors related to delaying sleep, such as staying up late for activities like watching TV, using a nine-item self-report questionnaire. BPS is significantly correlated with low self-regulation and is associated with poor sleep quality, psychiatric health issues, and diminished overall wellbeing [16]. One study validated the Korean version of BPS, confirming its reliability and relevance for use in the Korean general population. The results indicated that BPS strongly correlates with insomnia, perceived stress, and depression, making it an effective predictor of sleep and mental health disorders in clinical settings [17]. Subsequent studies with the Korean BPS showed that individuals with high bedtime procrastination levels exhibited significantly more depression, anxiety, and insomnia, and spent considerable time on media and leisure activities on smartphones before sleeping [18]. These findings suggest that BPS is negatively correlated with sleep and mood and should be addressed as a serious behavioral issue.
In a randomized controlled trial involving young adults in Korea, a behavioral intervention targeting BPS demonstrated promising results in reducing sleep interfering behaviors [19]. The study also seeks to validate the Korean version of BPS for high school students, since previous validations focused solely on the adult population. Given the developmental significance and unique characteristics of adolescence, such as high digital device usage, erratic sleep patterns, academic and peer pressures, it is crucial to confirm that the BPS effectively evaluates sleep behaviors in high school students. This validation will enhance our understandings and enable more targeted interventions to mitigate sleep disturbances, thereby improving the well-being of adolescents during demanding times like the COVID-19 pandemic.
METHODS
Participants and Procedure
We conducted an online survey involving 300 high school students from October 18–24, 2021, using the services of the professional survey company EMBRAIN (www.embrain.com). Participants responded anonymously as no identifiable personal information was collected. The Institutional Review Board of Asan Medical Center approved the study protocol (2021-1361) and waived the requirement for written informed consent. The survey commenced once a participant’s parent consented for their child to participate and selected “yes” to the agreement questions at the start of the survey. Demographic information collected included age, sex, grade, school type, and living area. The survey also contained questions about COVID-19, such as experiences of being quarantined, infection status, and vaccination.
Measures
Bedtime Procrastination Scale
The BPS is a self-report rating scale specifically designed to measure the extent of an individual’s bedtime procrastination [8]. It consists of 9 items scored on a five-point scale from 1 (never) to 5 (always), with items 2, 3, 7, and 9 being scored in reverse. The Korean version of the BPS [17], previously validated in young adult populations, was utilized.
Patient Health Questionnaire-9 items
The Patient Health Questionnaire-9 items (PHQ-9) is a selfreport scale designed to assess depression severity, which was validated in both adults and adolescents [20,21]. It comprises 9 items rated on a Likert scale from 0 (never ever) to 3 (almost every day). For this study, the Korean version of the PHQ-9 was employed.
Stress and Anxiety to Viral Epidemics-6 items
The Stress and Anxiety to Viral Epidemics-6 items (SAVE-6) is a self-report rating scale, designed to assess one’s anxiety towards viruses during a pandemic [22]. The SAVE-6, derived from the original SAVE-9 scale, was developed specifically for healthcare workers [23]. Each of the 6 items on the SAVE-6 can be rated on a five-point Likert scale ranging from 0 (never) to 4 (always). In this study, we used the original SAVE-6 scale validated among high school students [24].
Statistical Analysis
In this study, we evaluated the reliability and validity of the BPS. We assessed sampling adequacy and data suitability using the Kaiser-Meyer-Olkin (KMO) value and Bartlett’s test of sphericity before conducting the factor analysis. The factor structure was investigated through confirmatory factor analysis (CFA), with satisfactory model fit criteria defined by a standardized root mean square residual (SRMR) ≤0.05, root mean square error of approximation (RMSEA) ≤0.10, and comparative fit index (CFI) and Tucker-Lewis index (TLI) values ≥0.90 [25,26]. Additionally, the Graded Response Model (GRM) and Rasch analysis were applied to explore the psychometric properties of the BPS. Internal consistency reliability was assessed using Cronbach’s alpha and McDonald’s omega. We also tested the convergent validity of the BPS with the existing PHQ-9 and SAVE-6 scales using Pearson’s correlation coefficients. For statistical analysis, we utilized jamovi version 1.6.23 (https://www.jamovi.org/; Sydney, Australia).
RESULTS
Table 1 presents the demographic characteristics of the participants.
Validity and Reliability of the Korean Version of the BPS
Sample adequacy and data suitability were evaluated using the KMO measure (0.86) and Bartlett’s test of sphericity (p<0.001). The CFA performed with the DWLS methods indicated a good model fit (χ2/df=0.597, CFI=1.000, TLI=1.012, RMSEA=0.000, SRMR=0.040) (Tables 2 and 3). Fig. 1 illustrates the CFA factor structure with factor loadings and error variances. The multigroup CFA, including configural, metric, and scalar invariance, confirmed that the Korean version of the BPS accurately measures the same construct across sex and depression levels (Supplementary Table 1 in the online-only Data Supplement).
Graded Response Model
Supplementary Table 2 (in the online-only Data Supplement) presents the GRM outputs. After controlling for the false discovery rate (FDR), all the p-values of the S-χ2 for both factors were nonsignificant at a p-value of 0.01, except for item 3 in factor II. However, all RMSEA values were below 0.080. According to these RMSEA values, all items belong to the scale. For factor I, the slope/discrimination parameters (α) ranged from 2.231 to 4.772 (mean=3.115). All items in factor I had very high slopes. For factor II, the slope/discrimination parameters (α) ranged from 1.276 to 2.024 (mean=1.661). Item 3 exhibited a moderate slope, item 9 a high slope, and items 2 and 7 very high slope coefficients. These items provide valuable information and efficiently discriminate among individuals assessed using the Korean version of the BPS. The threshold coefficients (b) for factor I indicate that a higher latent trait or theta is necessary to endorse Likert-type response options “almost always” in all items. For factor II, the threshold coefficients (b) indicate that a higher latent trait or theta is necessary to endorse Likert-type response options from “frequently” to “almost always” for all items, except item 3, which is the most challenging item in this factor. The item characteristic curves in Fig. 2 display this information. The scale information curves (Fig. 3) show that factor I provides more information on bedtime procrastination than factor II. Several peaks in the curve for factor I could result from the polytomous nature of the data.
Rasch Model
Supplementary Table 3 (in the online-only Data Supplement) presents the outputs of the Rasch model for the Korean version of the BPS. The infit and outfit mean squares of all items in both factors are within the recommended range (0.50–1.50). In factor I, item 1 is the least difficult, and item 5 is the most difficult. In factor II, item 9 is the least difficult, and item 3 is the most difficult. All subscales have an acceptable item separation index (≥2) and item and person reliability (≥0.7) (Table 3). Factor I has an acceptable person separation index, but factor II does not.
Reliability and Evidence Based on Relations to Other Variables
The Korean version of the BPS demonstrates good internal consistency with support from Cronbach’s alpha, McDonald’s omega, and the split-half reliability for each factor (Table 3). The BPS score was significantly correlated with the PHQ-9 score (r=0.31, p<0.01), yet it showed no correlation with the SAVE-6 score (r=-0.05, p=0.56).
DISCUSSION
In this study, we aimed to explore the reliability and validity of the BPS in high school students. We observed that the BPS is a reliable and valid rating scale for measuring high school students’ bedtime procrastination during the pandemic. Furthermore, bedtime procrastination was positively associated with depression during the pandemic.
We aimed to explore the reliability and validity of the BPS among high school students particularly during the COVID-19 pandemic. The CFA demonstrated a good model fit for the single-factor model of the BPS in this study. The original study on the BPS proposed a single-factor model [8]. In a Korean validation study involving young adults, a single-factor model was proposed, yet both single- and two-factor models proved reliable and valid [17]. In this study, we examined the validity of the BPS as a two-factor model. Consistent with previous findings, the BPS among high school students displayed good internal consistency reliability [17,27].
Another significant strength of this study is that the psychometric properties of the BPS were assessed using the modern test theory approach [28]. This is the first study to assess the psychometric properties of this scale using this approach. The results exhibited good slope and threshold parameters, which are difficulty indices. Similar to internal consistency reliability, the BPS among high school students had good item response theory reliability, item reliability, and person reliability.
We also observed that depression was significantly associated with bedtime procrastination among high school students. Previous studies primarily involving adults reported similar findings [18]. While the link between bedtime procrastination and depression among adolescents remains understudied, the association between depressive symptoms and delays in sleep time has been reported [29]. The results of this study contribute to the literature on the association between depression and bedtime procrastination during the pandemic.
Contrary to expectations, the BPS was not significantly associated with SAVE-6. During the pandemic, many students experienced viral anxiety, leading to an expectation of reluctance to attend school, which could manifest as bedtime procrastination. Further studies will examine the negative correlation between BPS and SAVE-6 to understand the impact of viral anxiety on bedtime procrastination.
This study has several limitations. First, it was conducted through an online survey using a platform that may introduce sampling bias. An anonymous online survey could also influence the reliability of participants’ responses. Despite these issues, we opted for an online survey over face-to-face interviews to prevent viral transmission. Second, the research relied on self-report questionnaires without collecting objective measures such as actual bedtime or wake-up times, thus we could not verify participants’ bedtime procrastination against objective sleep indices.
In conclusion, the BPS was found to be reliable and valid in assessing bedtime procrastination in high school students, which was significantly related to depression during the pandemic. The BPS scale is suitable for measuring bedtime procrastination in this demographic within the clinical context.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Supplement1
Print 
