Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Feb 1.
Published in final edited form as: J Sleep Res. 2021 Jun 16;31(1):e13420. doi: 10.1111/jsr.13420

The Theory of Planned Behavior and Sleep Opportunity: An Ecological Momentary Assessment

Michael P Mead 1,2, Leah A Irish 1,3
PMCID: PMC8674382  NIHMSID: NIHMS1748642  PMID: 34137110

Abstract

Many American adolescents and adults report that they are not meeting sleep duration recommendations. Although insufficient sleep duration can occur due to factors outside an individual’s direct control, many individuals choose to restrict their own sleep. The Theory of Planned Behavior offers a framework to study this phenomenon. Recent research efforts have used the Theory of Planned Behavior to study sleep and have demonstrated success predicting sleep-related intentions and behavior but have failed to consider volitional sleep behavior or consider between- and within-person differences. This study used an intensive longitudinal design to test how constructs of the Theory of Planned Behavior relate to nightly sleep opportunity. Healthy college students (N=79) participated in a weeklong study in which they completed 4 ecological momentary assessment signals per day that measured their attitudes, subjective norms, perceived behavioral control, and intentions relating to their nocturnal sleep opportunity. Participants wore an actiwatch each night of the study to measure their sleep opportunity. Mixed linear models found that both intentions and perceived behavioral control were significant predictors of subsequent sleep opportunity, and that perceived behavior control was the strongest within-day predictor of intentions. Results demonstrate that within-person changes in perceived behavioral control and intentions predict subsequent sleep opportunity and provide insight into the potential refinement of sleep promotion efforts.

Keywords: Sleep, Health Behavior Theory, Intensive Longitudinal Design

Introduction

Insufficient sleep is common among American adolescents (Wheaton et al., 2018), college students (Lund et al., 2010), and adults (Liu et al., 2016), which is highly significant due to sleep’s robust associations with health outcomes (Itani et al., 2017). However, the source of the U.S. sleep deficit is not entirely understood. It is estimated that 10% of adults in the U.S. have a clinical sleep disorder (Ram et al., 2010), and 7.7% of college students have insomnia (Schlarb, Seirawan, & Kumar, 2010), demonstrating that many individuals may be choosing to restrict their sleep. Whereas sleep duration is influenced by a broad range of behavioral, environmental, and physiological factors (Klerman & Dijk, 2005; Irish et al., 2015), the amount of time allowed for sleep at night (i.e., sleep opportunity) is more directly within behavioral control, and therefore represents an important target for sleep health promotion efforts. It is not well understood why or how individuals make decisions about establishing healthy sleep opportunities.

The Theory of Planned Behavior (TPB) offers an established theoretical framework to study sleep opportunity. The TPB has been used for decades to study why people participate, or do not participate, in various health behaviors. The TPB states that volitional behavior results from intentions to perform that behavior, and intentions are influenced by attitudes, subjective norms, and perceived behavioral control (PBC) of that behavior (Ajzen, 1991; Fishbein & Ajzen, 2011). If an individual has a favorable attitude towards a behavior, believes that others think they should engage in a behavior, and has a high perception of control over the behavior, then they are likely to form an intention to perform the behavior.

A recent review illustrated the utility of the TPB in the study of sleep (Mead & Irish, 2020). To date, several studies have used the TPB to predict sleep-related behavior, such as coffee consumption or keeping a consistent sleep schedule (Kor & Mullan, 2011; Lao, Tao, & Wu, 2016; Tagler, Stanko, & Forbey, 2017; Stromg et al., 2018), and an additional 4 studies have specifically examined sleep duration intentions and behavior (Knowlden, Sharma, & Bernard, 2012; Stanko, 2013; Robbins & Niederdeppe, 2015; Tagler, Stanko, & Forbey, 2017). Knowlden and colleagues (2012) found that more positive attitudes, subjective norms, and PBC were all associated with stronger intentions to achieve a healthy sleep duration. Moreover, higher PBC and greater intentions both significantly predicted longer sleep duration. While this study found that all components of the model were significant predictors, this has not been demonstrated elsewhere. Stanko (2013) found that only subjective norms and PBC were positively associated with intentions to obtain 7–8 hours of sleep. Further, intentions did not significantly predict sleep duration, but PBC was a significant predictor. In another study of college students, more positive attitudes and subjective norms, but not PBC, predicted stronger intentions to obtain a healthy sleep duration (Robbins & Niederdeppe, 2013). Similar to Stanko 2013), intentions were not a significant predictor of sleep duration, but PBC significantly predicted sleep duration (Robbins & Niederdeppe, 2013). Lastly, Tagler and colleagues (2017) found that more positive attitudes, subjective norms, and PBC all predicted stronger intentions to obtain a healthy sleep duration, yet only intentions predicted sleep duration behavior. These studies demonstrate that the attitudes, subjective norms, and PBC significantly predict sleep intentions, and that intentions and PBC significantly predict sleep behavior.

Taken together, these studies demonstrate that the TPB is an effective framework for predicting sleep duration. However, there are two key limitations in the literature. First, the current research focusing on sleep duration does not focus on volitional behavior. Most studies measure intentions and behavior for sleep duration (i.e., time spent asleep), and not sleep opportunity, with only one exception (Tagler, Stanko, & Forbey, 2017). Although the terms sleep opportunity and sleep duration are often used interchangeably they are conceptually and practically distinct constructs (Hirshkowitz et al., 2015). Specific definitions of the behavior’s action, target, context, and time are crucial components of the TPB (Fishbein & Ajzen, 2011). The second major limitation of the current literature investigating sleep through the TPB is its exclusive focus on between-person effects, which fails to account for the intra-individual variability of attitudes, subjective norms, PBC, intentions, and behavior. Moreover, the findings from between-person methods do not always translate to within-person processes (Curran & Bauer, 2011). In addition, we are more likely to be influenced by proximal behaviors and cognitions (Dickerson, Klingman, & Jungquist, 2016). While a recent study demonstrated between- and within-day patterns of intra-individual variability, it has not yet been tested how fluctuations in these constructs relate to sleep (Mead & Irish, 2021).

Examining the intra-individual variability in sleep attitudes, subjective norms, PBC, intentions, and behavior can provide important insight into why people may choose to restrict their sleep opportunity and how sleep health promotion efforts may be improved or refined. The current study builds upon the current literature by using an intensive longitudinal design and had two aims. The first aim was to test whether sleep opportunity attitudes, subjective norms, and PBC predict intentions, and tested both between- and within-person effects of these predictors. Testing a between-person effect examined whether on days when an individual differs from the sample grand mean predicted subsequent intentions. Testing a within-day, within-person effect examined whether at signals when individuals deviate from their daily trend predicted subsequent intentions. It was hypothesized that on days with more positive attitudes, subjective norms, and PBC (i.e., between-person effect), participants would have greater intentions towards sleep opportunity at each signal (afternoon, evening, bedtime). It was further hypothesized that at signals when an individual has more positive than their daily typical attitudes, subjective norms, and PBC (i.e., within-person effect), this predicted greater intentions at the next signal. The second was to test whether sleep opportunity intentions and PBC predict behavior, and tested both between- and within-person effects of these predictors. Testing between-person effects examined whether individual mean differences from the sample mean predicted subsequent sleep opportunity. In contrast, within-person effects tested whether deviation from an individual’s typical trend predicted subsequent sleep opportunity It was hypothesized that participants who generally had greater intentions and PBC than the sample average would obtain a longer sleep opportunity each night (i.e., between-subjects effect), and on days in which intentions or PBC were greater than an individual’s typical trend, sleep opportunity would be longer that same night (i.e., within-subjects effect). This project was pre-registered at https://osf.io/2q6s8.

Methods

Participants

Participants were recruited from a midwestern university undergraduate research pool. Individuals were eligible to participate if they were between the ages of 18–25 and had a mobile phone with text message and internet capabilities. This age range was selected to reduce confounding effect of age on sleep opportunity and because the TPB measure used in this study was validated for participants of this age range. Additional exclusion criteria were selected to minimize medical and behavioral confounds likely to affect sleep-related behaviors. Individuals were ineligible for participation if they had an atypical sleep schedule (e.g., shift work), were currently being treated for a sleep disorder, were binge drinkers (more than 5 drinks in a single episode more than once per month), smoked cigarettes or e-cigarettes (more than once cigarette in the last month), used sleep aids or medications that can disrupt sleep, or had a medical condition that can influence sleep (e.g., chronic migraines, rheumatoid arthritis, hypertension).

Procedure

Inclusion and exclusion criteria were assessed with an online screening questionnaire. Eligible individuals received an email which briefly described the study’s purpose, procedures, and compensation. Individuals who wished to participate signed up for an appointment via SONA, an online study management system.

All study materials and procedures were approved by the institution IRB, reference number SM20024. Three days before the lab session, participants received an email with instructions to begin the study and a digital consent form. After consenting to participate, participants completed bedtime and morning assessments each day before their time 1 session to reduce reactivity to study protocols during the one-week-in-home assessment period. During the time 1 session, participants completed a demographic questionnaire, received instructions for the in-home assessment period, and were given an actiwatch which they wore on their non-dominant wrist to measure sleep opportunity for 7 consecutive nights. Concurrently, an ecological momentary assessment (EMA) protocol (Schwarz, 2007) was used to assess TPB constructs 4 times per day (morning, afternoon, evening, and bedtime). The bedtime and wake time signals were completed via sleep diary by participants before they tried to start sleeping at night and after they were done trying to sleep in the morning, respectively. The afternoon and evening EMA signals were initiated by a text message and contained a link to the survey. All afternoon signals were sent between 12:00PM-2:00PM, and all evening signals were sent between 5:00PM-7:00PM. To maximize adherence, participants selected a fixed time within each of these timeframes to receive their text messages. Participants who completed at least 95% of these assessments were eligible for one of two $50 cash prizes. Daily adherence checks were also conducted, and participants with missing signals were contacted via email with a reminder to adhere to the study protocol. Following the weeklong in-home assessment, participants came back to the research lab to return the actiwatch device. All participants received credit toward their psychology course as compensation for their participation along with a copy of their sleep report.

Measures

Sleep opportunity.

Actigraphy was used to measure objective sleep opportunity (i.e., the amount of time a person makes available for sleep) each night of the study period. Each participant wore an actiwatch (Philips Respironics, Bend OR) on their non-dominant wrist and were instructed to wear the watch continuously for 7 consecutive days and nights. An actiwatch is a wrist-worn accelerometer which infers sleep and wake states from movement-based algorithms, and is an effective tool for objective sleep measurement, especially when measuring sleep over an extended period (Ancoli-Israel et al., 2003). Actigraphy provides multiple measures of sleep duration and continuity, but the variable of interest for the current study was sleep opportunity (i.e., the number of minutes that participants allow themselves to sleep at night). Actiwatch’s algorithms compute a rest period, indicating when participants try to fall asleep at night and wake up in the morning. However, several steps were taken to improve the validity of this measurement. First, participants were instructed to press a marker button each night when they try to begin sleeping, and each morning when they have woken up for the final time and do not intend to try to get back to sleep. This places a marker in the sleep data and assists in accurately defining the rest interval. In addition, each morning participants self-reported, using a modified version of the Pittsburgh Sleep Diary (Monk et al., 1994), what time they tried to start sleeping and what time they have woken up for the final time and do not intend to try to get back to sleep. The device algorithm, marker button, and morning sleep diaries were used together to identify participants’ sleep opportunity each night of the study.

Constructs of the TPB.

Attitudes, subjective norms, PBC, and behavioral intentions to allow adequate sleep opportunity were adapted from Robbins & Niederdeppe (2015). The constructed measure consists of 14 items that assess attitudes, subjective norms, PBC, and intentions to sleep 8–9 hours most nights of the week. Responses for each item were on a 7 point scale. In order to reduce possible restriction of range in repeated assessment, the current study used an 11 point scale. As noted by Ajzen (2002), adding additional response items to an established measure does not reduce its reliability. One item for each TPB construct was included in the EMA signals to reduce participant burden. Attitudes were assessed with the following item: “Overall, I think allowing myself the time to sleep 8–9 hours tonight is” 1 (good) – 11 (bad), reverse scored. Subjective norms were assessed with the following item: “People who are important to me think that I should allow myself the time to sleep 8–9 hours tonight” 1 (strongly disagree) – 11 (strongly agree). PBC was assessed with the following item: “For me to allow myself the time to sleep for between 8–9 hours tonight is” 1 (easy) – 11 (difficult), reverse scored. Lastly, intentions were assessed with the following item: “I intend to allow myself the time to sleep for between 8–9 hours tonight” 1 (definitely don’t) – 11 (definitely do).

Data Analysis

Before testing study aims, multilevel model assumptions and data missingness were evaluated. Skewness and kurtosis were tested for sleep opportunity and it was normally distributed (skew= .03(.11), kurtosis= .46(.21)). Overall protocol adherence was 92%. Days on which participants missed more than 50% of EMA signals were removed from analyses, and participants missing more than 50% of EMA study days were removed from all analyses. Twenty-five total signals were removed from analyses due to being taken at incorrect times (e.g., taking the afternoon assessment right before bed). Two total study days from two participants were removed due to having less than 50% of EMA signals. Only one participant was removed from final analyses, and this was due to non-adherence to the EMA protocol. Out of 548 study days, there were 29 days with only 2 signals, and 92 days with 3 signals. Adherence to study protocol was not related to participant attitudes, subjective norms, PBC, intentions, or sleep opportunity (all p’s > .05) and no corrections to the models were needed. Before testing study hypotheses, unadjusted models were run to calculate the intraclass correlations, and it was revealed that 30.20% of the variability in sleep opportunity existed within days, and 30.74% of the variability existed between days. Age and gender were tested as possible covariates in each model, however neither were significant predictors in either model, and these variables were removed from analyses.

Aim 1: Predictors of sleep opportunity intentions

Within each study day, there were three lagged relationships that could be tested: 1) morning attitudes, subjective norms, and PBC predicting afternoon intentions 2) afternoon attitudes, subjective norms, and PBC predicting evening intentions 3) evening attitudes, subjective norms, and PBC predicting bedtime intentions. Study day was added as a random effect to account for the nesting of the signals. Centering procedures followed recommendations from Curran & Bauer (2011). Attitudes, subjective norms, and previous intentions were between-person centered on the grand mean intercept, and within-person centered on the within-day slope. PBC was between-person centered on the sample mean and within-person centered on the individual mean.

Mixed linear modeling was used to test, on a within-day level, whether attitudes, subjective norms, and PBC predict sleep opportunity intentions. The lagged between- and within-centered variables for attitudes, subjective norms, PBC, and previous intentions were entered simultaneously into the same model to predict future sleep opportunity intentions. Heterogenous autoregressive covariance structure was used to allow the variances at each timepoint to covary and the intercept in the full model was random to allow for mean differences between participant sleep opportunity.

Aim 2: Predictors of sleep opportunity behavior

Centering procedures followed recommendations from Curran & Bauer (2011). Intentions were between-person centered on the sample intercept mean, and within-person centered their individual slope. PBC was between-person centered on the sample mean and within-person centered on the participant’s weeklong mean.

Mixed linear modeling was used to test whether daily deviation from one’s typical intentions and PBC predicted sleep opportunity each night. Heterogenous autoregressive covariance structure was used to allow the variances at each timepoint to covary and the intercept in the full model was random to allow for mean differences between participant sleep opportunity. The between- and within-person variables for intentions and PBC were entered as fixed effects into the final model.

Power Analyses

Prior to data collection, two power analyses were conducted using Monte Carlo simulation in Mplus version 8.1 to determine whether a sample size of 60, with 85% adherence, would provide adequate power to test study aims 1 and 2. For study aim 1, there is adequate power (.87) to detect a small effect size (.1). For study aim 2, there is adequate power (.83) to detect a medium effect size (d=.4).

Results

Descriptives

Participants were majority female (58.2%) and white (83.5%), with an average age of 19.01(1.16) years (see Table 1). On average, participants had a sleep opportunity of just under 8 hours (SD= 94.44 minutes), positive attitudes (M= 9.45, SD= 2.24), positive subjective norms (M= 9.01, SD= 2.10), moderate PBC (M= 7.16, SD= 3.19), and moderate intentions (M= 7.43, SD= 3.04) to allow themselves the opportunity to sleep at least 8 hours.

Table 1.

Demographic and health characteristics (N=79)

Gender, n(%)
 Male 33(41.8%)
 Female 46(58.2%)
Race, n(%)
 White 66(83.5%)
 Black 4(5.1%)
 Asian 9(11.4%%)
Age, mean(SD), years 19.01(1.16)
Sleep opportunity, mean(SD), minutes 479.34(94.44)
*Attitudes, mean(SD) 9.45(2.24)
*Perceived norms, mean(SD) 9.01(2.10)
*Perceived behavioral control, mean(SD) 7.16(3.19)
*Intentions, mean(SD) 7.43(3.04)
Open in a new tab
*

11 point scale with higher values indicating more positive values

Aim 1

Between-person Effects

Results of the mixed linear model revealed that after controlling for previous intentions, there was only one significant between-person effect (see Table 2). PBC significantly predicted intentions (b= .43, p<.001), in that on days when PBC was higher than the sample average, this predicted greater intentions at each time point on that day. The between-person effects of attitudes (b= −.02, p=.61) and subjective norms (b=.07, p= .07) were not significant.

Table 2.

Attitudes, perceived norms, and perceived behavioral control predicting future sleep opportunity intentions

b SE t p
Previous intentions BP .32 .03 9.73 <.001
Previous intentions WP −.75 .08 −10.36 <.001
Attitudes BP −.02 .04 −.52 .61
Attitudes WP −.02 .09 −.17 .87
Perceived norms BP .07 .04 1.81 .07
Perceived norms WP −.02 .10 −.15 .89
Perceived behavioral control BP .43 .05 7.69 <.001
Perceived behavioral control WP .25 .03 7.79 <.001
Open in a new tab
*

BP= between-person, WP= within-person

Within-person Effects

The mixed linear model revealed one significant within-person effect after controlling for previous intentions (see Table 2). PBC significantly predicted future intentions (b= .25, p<.001), in that signals when participants indicated greater PBC than their trend for that day predicted greater intentions at the next signal. Attitudes (b= −.02, p= .87) and subjective norms (b= −.02, p= .89) were not significant predictors.

Aim 2

Between-Person Effects

Results of the mixed linear model (see Table 3) revealed that the between-person effects of PBC (b= 7.43, p= .01) were a significant predictor, in that participants who were generally higher in PBC had longer sleep opportunity each night. Intentions were not a significant predictor of sleep opportunity (b= 1.49, p= .63).

Table 3.

Sleep intentions and perceived behavioral control predicting sleep behavior

b SE t p
Intentions BP 1.49 3.11 .48 .63
Intentions WP 20.50 3.29 6.25 <.001
Perceived behavioral control BP 7.43 2.91 2.56 .01
Perceived behavioral control WP 13.32 2.78 4.78 <.001
Open in a new tab
*

BP= between-person, WP= within-person

Within-person Effects

The mixed linear model revealed that both within-person effects were significant (see Table 3). The within-person effect of intentions was significant (b= 20.49, p<.001), in that days with higher than usual intentions to obtain a sleep opportunity of at least 8 hours were associated with longer sleep opportunity that night. In addition, the within-person effect of PBC was significant (b= 13.32, p<.001). Specifically, days in which individuals had greater than their usual PBC predicted longer sleep opportunity that night.

Discussion

Sleep is important for both physical and mental health, yet many Americans choose to restrict their sleep at night. Little is known about behavioral sleep restriction, and the current study sought to learn more about why people may choose to do so.

Study aim 1 was to test, on a daily level, whether attitudes, subjective norms, and PBC predicted sleep opportunity intentions. It was hypothesized that, after controlling for previous time point intentions, all three would significantly predict intentions, but that PBC would be the strongest predictor. Both the between- and within-person effects of previous intentions predicted future intentions. However, the within-person effect was negative, indicating that higher intentions at a given time point predicted lower future intentions. This is likely explained by the fact that intentions significantly decreased throughout the day. Neither the between- or within-person effects of attitudes or subjective norms on intentions were significant, and it is possible that these null results are explained by ceiling effects and restricted range. Lastly, the between- and within-person effects of PBC on subsequent intentions were both significant, such that on days in which their PBC was higher than the sample average, intentions increased. Moreover, intentions were significantly higher following signals with higher than their typical trend of PBC for that day. Taken together, results suggest that within a day, PBC is the only construct to significantly predict future intentions. This contradicts previous literature that has utilized cross-sectional designs, which have shown attitudes, subjective norms, and PBC to all be significant predictors of sleep duration intentions (Knowlden, Sharma, Bernard 2012; Stanko, 2013; Robbins & Niederdeppe, 2015; Strong et al., 2017; Branscum, Fay, & Senkowski, 2018). This discrepant finding could be due to the use of a daily framework, which would suggest that momentary attitudes and subjective norms are not significant predictors of future intentions, while PBC remains a significant predictor of intentions regardless of context. Thus, PBC is related to sleep opportunity intentions at both a trait and state level and may be the most robust predictor of intentions.

Study aim 2 tested whether daily deviation from one’s typical intentions and PBC predicted sleep opportunity each subsequent night. Results supported the hypothesis that on days with greater intentions and PBC there would be longer sleep opportunity that same night, but between-person effects of intentions were not significant which suggest that mean level sleep opportunity intentions do not translate into a longer sleep opportunity each night. PBC showed significant between- and within-person effects. Participants who had greater average PBC, compared to the sample average, had significantly higher sleep opportunity. In addition, days on which participants had higher PBC than their typical day had significantly higher sleep opportunity. Thus, both mean level and momentary aspects of PBC may play a role in nightly sleep opportunity. An exploratory analysis was conducted to test whether the between- or within-person effects of intentions and PBC were stronger predictors of sleep opportunity. The model was re-run after standardizing the between- and within-person variables of intentions and PBC. The within-person effect (β= 23.87, p<. 001) of intentions was a stronger predictor of sleep opportunity than the between-person effect (β= 3.47, p= .64). The between-person effect (β= 18.97, p=.01) of PBC was a slightly stronger predictor of sleep opportunity than the within-person effect (β= 17.98, p<.001). In general, these findings illustrate that both intentions and PBC are significant predictors of sleep opportunity, and collectively, it is the within-person variability of these factors that play a significant role in sleep opportunity. This finding that both intention and PBC predict subsequent sleep behavior contradict the cross-sectional literature. Some studies have identified PBC as the significant predictor (Stanko, 2013; Robbins & Niederdeppe, 2015), while others found only intentions to be significant (Tagler, Stanko, & Forbey, 2017; Strong et al., 2018). These mixed findings could be due to testing the relationships between intentions and PBC and sleep exclusively from a between-person perspective.

There are several areas for future research. In addition to sleep duration, previous research has used the TPB to predict sleep hygiene (Kor & Mullan, 2011; Tagler, Stanko, & Forbey, 2017; Strong et al., 2018), sleep patterns (Lao, Tao, & Wu, 2016), and late night electronic use (Zhao et al., 2019). Given that findings may differ between cross-sectional and intensive longitudinal designs, further research is needed to test these daily associations. A limitation of the TPB is the intention-behavior gap, which occurs when intentions do not always translate to behavior (Sheeran & Webb, 2016). Rebar and colleagues (2020) posit that this may also be true for sleep-related behaviors. For instance, individuals may have strong intentions to obtain a sleep opportunity of at least 8 hours, but then as they approach bedtime, they avoid going to bed (e.g., nightly routine consists of reading in bed for an hour). Thus, automatic processes (e.g., habits) may explain the intention-behavior gap. In the current study, intentions were a significant predictor of behavior, but accounting for automatic processes may improve this framework. Future research efforts might consider the use of dual process theories (Strack & Deutsch, 2004), which account for both reflective and automatic processes. These may be particularly effective frameworks for studying sleep-related behaviors (Rebar et al., 2020). Accordingly, intensive longitudinal designs are an effective framework for testing dual process theories (Dunton et al., 2019), and the current study provides the first example of how to implement this methodology in the context of health behavior theory and sleep health. However, there is very limited data on the use of dual process theories in the study of sleep health, and expectancy-value theories (e.g., The Theory of Planned Behavior) have shown promise (Mead & Irish, 2020). As noted by Mead and Irish (2019), theory evaluation and comparison is an important next step in the study of health behavior theory and sleep health. Lastly, results of this study have implications for the development of theoretically based interventions to improve sleep opportunity. For example, dynamic interventions, such as just in time adaptive interventions (Nahum-Shani et al., 2014), targeting participants on days of low intentions may be a more effective approach than a single time point intervention (e.g., 1 hour session detailing sleep hygiene). In contrast, interventions targeting both trait and state aspects of PBC may improve sleep intentions and behavior. While promising, theoretically based interventions to improve sleep are rare. To date, no studies have specifically targeted sleep duration or opportunity using the TPB as a framework. However, one recent study (Lin et al., 2018) utilized a theoretically based intervention targeting sleep hygiene and found that the intervention had significantly greater improvements in sleep hygiene compared to the control group (education material related to sleep hygiene). Indeed, tailored (Krebs, Prochaska, & Rossi, 2010) and theoretically based (Glanz & Bishop, 2010) interventions may be more effective in changing health behavior, and recent advancements in eHealth and mHealth interventions (Morh et al., 2014) provide a platform for implementing these interventions.

The present study has some limitations. Given the use of a convenience sample, study participants were a mostly homogeneous group of healthy (e.g., excluded binge drinkers and smokers) college students and these results may not generalize to other samples. Sleep restriction is not exclusive to college students, and future research should extend this framework to other samples. As noted by Mead and Irish (2019), one barrier to using the TPB as a framework in other populations is that elicitation studies are needed to develop TPB measures for different populations. Thus, existing measures based on elicitation studies in college samples may not adequately capture attitudes, subjective norms, and PBC in different samples. For example, schoolwork and environmental determinants related to dorm living may be influential barriers to sleep opportunity unique to college students. This study was likely limited by its measurement of attitudes and subjective norms, given the lack of variability. This study adapted a measure constructed by Robbins and Niederdeppe (2015), and only one item from each subscale was chosen. It is possible that the other items for attitudes (e.g., Overall, I think allowing myself the time to sleep 8–9 hours tonight is: unpleasant, pleasant) and subjective norms (e.g., I feel social pressure to allow myself the time to sleep 8–9 hours tonight: strongly disagree, strongly agree) could capture more intra-individual variability. In addition, each construct of the TPB was assessed with a single item, and single item measures tend to be more unreliable than multi-item measures. Future studies should utilize the full measure to increase the reliability of attitudes, subjective norms, PBC, and intentions. Lastly, this study was likely limited by participant reactivity (Robbins & Kubiak, 2014), and fact, self-monitoring is a strategy for sleep health improvement (Adachi et al., 2008). In order to minimize possible reactivity, participants completed a three day “practice” period before starting the week long assessment, but future research should consider a longer study period which may not only benefit possible reactivity, but could also provide more reliable estimates of variability (Rowe et al., 2008; Wang & Grimm, 2012). Relatedly, the fixed EMA schedule could lead to anticipatory effects and reduced ecological validity, and future studies could send random signals during specified time frames (Palmier‐Claus et al., 2011). Despite these limitations, the study had many strengths including its intensive longitudinal design, advanced analytical approach, and use of objective sleep monitoring (Lauderdale et al., 2008).

The present study provides insight into the processes underlying the within-person dynamics of sleep health and could inform the development and improvement of sleep health interventions. While limited, the recent use of the TPB has improved our understanding in factors that influence sleep. This study advanced this work by characterizing the intra-individual variability in sleep opportunity attitudes, subjective norms, PBC, and intentions, demonstrated that daily intentions and PBC predict sleep behavior, and that within a given day, PBC significantly predicts future intentions. In addition, this study highlights the need for more use of intensive longitudinal designs in studying sleep. In conclusion, this study addressed significant gaps in our understanding of how sleep health intentions predict behavior and has the potential to influence and inform the development of sleep health intervention strategies.

Acknowledgments

Financial Disclosure: Research reported in this publication was supported by National Heart, Lung, and Blood Institute of the National Institutes of Health under award number T32HL007909.

Footnotes

Non-financial Disclosure: none

References

  1. Adachi Y, Sato C, Kunitsuka K, Hayama J, & Doi Y (2008). A brief behavior therapy administered by correspondence improves sleep and sleep-related behavior in poor sleepers. Sleep and Biological Rhythms, 6(1), 16–21. [Google Scholar]
  2. Ajzen I (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211. [Google Scholar]
  3. Ajzen I (2002). Perceived behavioral control, self‐efficacy, locus of control, and the theory of planned behavior. Journal of Applied Social Psychology, 32(4), 665–683. [Google Scholar]
  4. Ancoli-Israel S, Cole R, Alessi C, Chambers M, Moorcroft W, & Pollak CP (2003). The role of actigraphy in the study of sleep and circadian rhythms. Sleep, 26(3), 342–392. [DOI] [PubMed] [Google Scholar]
  5. Branscum P, Fay KQ, & Senkowski V (2018). Do different factors predict the adoption and maintenance of healthy sleep behaviors? A reasoned action approach. Translational Behavioral Medicine, 1–9. [DOI] [PubMed] [Google Scholar]
  6. Curran PJ, & Bauer DJ (2011). The disaggregation of within-person and between-person effects in longitudinal models of change. Annual Review of Psychology, 62, 583–619. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Dickerson SS, Klingman KJ, & Jungquist CR (2016). Common meanings of good and bad sleep in a healthy population sample. Sleep Health, 2(3), 253–259. [DOI] [PubMed] [Google Scholar]
  8. Dunton GF, Rothman AJ, Leventhal AM, & Intille SS (2019). How intensive longitudinal data can stimulate advances in health behavior maintenance theories and interventions. Translational Behavioral Medicine. 10.1093/tbm/ibz165 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Fishbein M, & Ajzen I (2011). Predicting and changing behavior: The reasoned action approach: New York: Psychology Press. [Google Scholar]
  10. Glanz K, & Bishop DB (2010). The role of behavioral science theory in development and implementation of public health interventions. Annual Review of Public Health, 31, 399–418. [DOI] [PubMed] [Google Scholar]
  11. Hirshkowitz M, Whiton K, Albert SM, Alessi C, Bruni O, DonCarlos L, … & Hillard PJA (2015). National Sleep Foundation’s sleep time duration recommendations: Methodology and results summary. Sleep Health, 1(1), 40–43. [DOI] [PubMed] [Google Scholar]
  12. Irish LA, Kline CE, Gunn HE, Buysse DJ, & Hall MH (2015). The role of sleep hygiene in promoting public health: A review of empirical evidence. Sleep Medicine Reviews, 22, 23–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Itani O, Jike M, Watanabe N, & Kaneita Y (2017). Short sleep duration and health outcomes: A systematic review, meta-analysis, and meta-regression. Sleep Medicine, 32, 246–256. [DOI] [PubMed] [Google Scholar]
  14. Klerman EB, & Dijk DJ (2005). Interindividual variation in sleep duration and its association with sleep debt in young adults. Sleep, 28(10), 1253–1259. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Knowlden AP, Sharma M, & Bernard AL (2012). A theory of planned behavior research model for predicting the sleep intentions and behaviors of undergraduate college students. The Journal of Primary Prevention, 33(1), 19–31. [DOI] [PubMed] [Google Scholar]
  16. Kor K, & Mullan BA (2011). Sleep hygiene behaviours: An application of the theory of planned behaviour and the investigation of perceived autonomy support, past behaviour and response inhibition. Psychology & Health, 26(9), 1208–1224. [DOI] [PubMed] [Google Scholar]
  17. Krebs P, Prochaska JO, & Rossi JS (2010). A meta-analysis of computer-tailored interventions for health behavior change. Preventive Medicine, 51(3–4), 214–221. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Lao HCF, Tao VYK, & Wu AMS (2016). Theory of planned behaviour and healthy sleep of college students. Australian Journal of Psychology, 68(1), 20–28. [Google Scholar]
  19. Lauderdale DS, Knutson KL, Yan LL, Liu K, & Rathouz PJ (2008). Self-reported and measured sleep duration: How similar are they? Epidemiology, 19(6), 838. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Lin CY, Strong C, Scott AJ, Broström A, Pakpour AH, & Webb TL (2018). A cluster randomized controlled trial of a theory-based sleep hygiene intervention for adolescents. Sleep, 41(11). [DOI] [PubMed] [Google Scholar]
  21. Liu Y, Wheaton AG, Chapman DP, Cunningham TJ, Lu H, & Croft JB (2016). Prevalence of healthy sleep duration among adults—United States, 2014. Morbidity and Mortality Weekly Report, 65(6), 137–141. [DOI] [PubMed] [Google Scholar]
  22. Lund HG, Reider BD, Whiting AB, & Prichard JR (2010). Sleep patterns and predictors of disturbed sleep in a large population of college students. Journal of Adolescent Health, 46(2), 124–132. [DOI] [PubMed] [Google Scholar]
  23. Mead MP, & Irish LA (2020). Application of health behaviour theory to sleep health improvement. Journal of Sleep Research, 29(5), 10.1111/jsr.12950. [DOI] [PubMed] [Google Scholar]
  24. Mead MP, & Irish LA (2020). Re: Accounting for automatic processes in sleep health. Journal of Sleep Research. 10.1111/jsr.12992 [DOI] [PubMed] [Google Scholar]
  25. Mead MP & Irish LA. (2021). Intra-individual variability of sleep opportunity attitudes, perceived norms, perceived behavioral control, and intentions. Annals of Behavioral Medicine. 10.1093/abm/kaaa112 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Mohr DC, Schueller SM, Montague E, Burns MN, & Rashidi P (2014). The behavioral intervention technology model: an integrated conceptual and technological framework for eHealth and mHealth interventions. Journal of Medical Internet Research, 16(6), e146. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Monk TH, Reynolds Iii CF, Kupfer DJ, Buysse DJ, Coble PA, Hayes AJ, … Ritenour AM (1994). The Pittsburgh sleep diary. Journal of Sleep Research, 3(2), 111–120. [PubMed] [Google Scholar]
  28. Nahum-Shani I, Smith SN, Tewari A, Witkiewitz K, Collins LM, Spring B, & Murphy S (2014). Just in time adaptive interventions (jitais): An organizing framework for ongoing health behavior support. Methodology Center Technical Report, 2014, 14–126. [Google Scholar]
  29. Palmier‐Claus JE, Myin‐Germeys I, Barkus E, Bentley L, Udachina A, Delespaul PAEG, … & Dunn G (2011). Experience sampling research in individuals with mental illness: Reflections and guidance. Acta Psychiatrica Scandinavica, 123(1), 12–20. [DOI] [PubMed] [Google Scholar]
  30. Ram S, Seirawan H, Kumar SKS, & Clark GT (2010). Prevalence and impact of sleep disorders and sleep habits in the United States. Sleep and Breathing, 14(1), 63–70. [DOI] [PubMed] [Google Scholar]
  31. Rebar AL, Reynolds AC, Ferguson SA, Gardner B, (2020). Accounting for automatic processes in sleep health. Journal of Sleep Research. 10.1111/jsr.12987 [DOI] [PubMed] [Google Scholar]
  32. Robbins ML, & Kubiak T (2014). Ecological momentary assessment in behavioral medicine. The Handbook of Behavioral Medicine (429–446). John Wiley and Sons, Ltd. [Google Scholar]
  33. Robbins R, & Niederdeppe J (2015). Using the integrative model of behavioral prediction to identify promising message strategies to promote healthy sleep behavior among college students. Health Communication, 30(1), 26–38. [DOI] [PubMed] [Google Scholar]
  34. Rowe M, McCrae C, Campbell J, Horne C, Tiegs T, Lehman B, & Cheng J (2008). Actigraphy in older adults: comparison of means and variability of three different aggregates of measurement. Behavioral Sleep Medicine, 6(2), 127–145. [DOI] [PubMed] [Google Scholar]
  35. Stanko KA (2013). Predicting sleep duration in college students: A reasoned action approach (Thesis). 2013 Ball State University, Muncie, IN. Doctoral Dissertation. [Google Scholar]
  36. Strong C, Lin CY, Jalilolghadr S, Updegraff JA, Broström A, & Pakpour AH (2018). Sleep hygiene behaviours in Iranian adolescents: An application of the Theory of Planned Behavior. Journal of Sleep Research, 27(1), 23–31. [DOI] [PubMed] [Google Scholar]
  37. Schlarb AA, Kulessa D, & Gulewitsch MD (2012). Sleep characteristics, sleep problems, and associations of self-efficacy among German university students. Nature and Science of Sleep, 4, 1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Schwarz N (2007). Retrospective and concurrent self-reports: The rationale for real-time data capture. The Science of Real-Time Data Capture: Self-Reports in Health Research, 11, 26. [Google Scholar]
  39. Sheeran P, & Webb TL (2016). The intention–behavior gap. Social and Personality Psychology Compass, 10(9), 503–518. [Google Scholar]
  40. Strack F, & Deutsch R (2004). Reflective and impulsive determinants of social behavior. Personality and Social Psychology Review, 8(3), 220–247. [DOI] [PubMed] [Google Scholar]
  41. Tagler MJ, Stanko KA, & Forbey JD (2017). Predicting sleep hygiene: A reasoned action approach. Journal of Applied Social Psychology, 47(1), 3–12. [Google Scholar]
  42. Wang L, & Grimm KJ (2012). Investigating reliabilities of intraindividual variability indicators. Multivariate Behavioral Research, 47(5), 771–802. [DOI] [PubMed] [Google Scholar]
  43. Wheaton AG, Jones SE, Cooper AC, & Croft JB (2018). Short sleep duration among middle school and high school students—United States, 2015. Morbidity and Mortality Weekly Report, 67(3), 85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Zhao X, Feng X, Garg R, & Kelly KM (2019). Reducing late evening bedtime electronic device intentions and use among young adults. Sleep Health, 401–408. [DOI] [PubMed] [Google Scholar]

RESOURCES