| Home | E-Submission | Sitemap | Contact us |  
top_img
Sleep Med Res > Volume 17(3); 2026 > Article
Kitch, Thomas, Taylor, Weinstock, and Emert: Psychometric Validation and Evaluation of the Sleep Need Questionnaire

Abstract

Background and Objective

The Sleep Need Questionnaire (SNQ) is a widely used self-report measure used to detect self-perceived sleep need but has not yet been validated. Validation and reliability analyses were conducted on the SNQ in a sample of college students across multiple universities.

Methods

Total score and item-level convergent and discriminant validity were assessed utilizing various sleep and psychological measures. Construct validity and the structure of the SNQ were assessed using exploratory factor analysis (EFA) and item response theory via graded response modeling (GRM).

Results

The SNQ demonstrated moderate internal consistency (Cronbach’s α=0.74). Convergent analyses showed significant moderate to strong convergent correlations with other sleep-related constructs (r=0.46–0.74) and discriminant analyses showed mixed weak correlations with dissimilar constructs (r=0.04–0.10). The accuracy of the SNQ was examined through a receiver operating characteristic analysis with the Epworth Sleepiness Scale (ESS) as the criterion and indicated acceptable discrimination (area under the curve=0.73). EFA revealed a one-factor solution (λ=2.36) with a mixed range of significant inter-item correlations (r=0.165–0.833) and variable factor loadings (0.419–0.985). GRM revealed considerable imbalances in item functioning and potential redundancy.

Conclusions

Overall, the SNQ demonstrates moderate internal consistency and moderate to strong convergent and discriminant validity. Further analysis is needed to confirm and examine the unidimensional structure of the SNQ, as weak inter-item correlations, mixed factor loadings, and mixed item level behavior indicate structural and construct validity problems with the SNQ.

INTRODUCTION

Sleep disorders include some of the most widespread health concerns in the USA, with roughly 30%–36% of the population endorsing at least one symptom of insomnia disorder and roughly 25% endorsing general dissatisfaction with their sleep [1,2]. Previous literature has found that excessive daytime sleepiness may be more common in some sleep disorders (i.e., obstructive sleep apnea, narcolepsy) [3], than others (e.g., insomnia relationship is mixed) [4]. Although not always a common symptom with insomnia disorder, increased daytime sleepiness has been associated with common treatments for insomnia disorder (i.e., sleep restriction therapy) [5,6]. Examining the psychometric properties of measures that screen, diagnose, and aid in the treatment of sleep-related disorders is critical for informing effective clinical decision-making.
One of the most widely used measure for assessing excessive daytime sleepiness is the Epworth Sleepiness Scale (ESS) [7]. Though, the effective utility of the ESS is mixed, with references of high variability of scores across populations [8] and inconsistent associations with physiological sleep need (i.e., multiple sleep latency test) [9]. Concerns regarding the limitations of the ESS may have led to the development of alternative measures, such as the Sleep Need Questionnaire (SNQ), that seek to capture similar constructs. The SNQ is an unpublished 4-item single time-point self-report measure developed by Arthur J. Spielman, PhD, used to detect self perceived sleep need through assessing questions related to daytime/evening fatigue, feeling sleepy/drowsy, taking intentional or unintentional naps, and feelings related to getting an adequate amount of sleep. The SNQ, while not validated, has been used in combination with sleep efficiency within sleep restriction therapy to inform matrix-style decision points for altering a patient’s sleep window [10].
The goal of the present study is to address these gaps by examining the SNQ’s psychometric properties and to validate the SNQ. The analyses conducted in the present study are critical to understanding the accuracy of the SNQ compared to the ESS, establishing the SNQ as a valid and reliable measure of sleep need, and the potential utility of the SNQ in sleep treatments within clinical populations. To test the SNQ’s psychometric properties, the present study conducted reliability analyses using Cronbach’s alpha, convergent and discriminant validity testing using Pearson r correlations, and daytime impairment analyses comparing the accuracy of the SNQ with the ESS using a receiver operating characteristic (ROC) curve. Finally, exploratory factor analysis (EFA) and item response theory (IRT) via graded response modeling (GRM) were used to assess the factor structure and individual item functioning of the SNQ.

METHODS

Sample

Analyses in the present study were from a cross-sectional dataset on college student health behaviors collected online in the spring of 2023 across seven different universities across the USA. A core battery consisting of demographic and other central measures were completed by all participants. Participants were then subsequently randomly assigned to complete one additional battery (i.e., Survey Battery A or Survey Battery B) that contained half of the remaining measures. Out of the 1,197 participants in the overall dataset, 581 individuals completed the SNQ and were included in all the analyses conducted for the SNQ (n=581). The sample was 18 to 24 years old, predominantly female (74.7%) and Caucasian (81.8%).
Inter-item consistency was measured via Cronbach’s alpha analyses for all psychological measures used in the current study across convergent and discriminant validity testing. All available data for each measure were used to generate a Cronbach alpha for maximum generalizability and power (Table 1). Correlations were conducted with the subset of people who completed both measures in the pair (Tables 2 and 3).

Primary Outcome

Sleep Need Questionnaire

The SNQ is an unpublished 4-item single time-point self-report measure developed by Arthur J. Spielman, PhD, to assess self-perceived sleep need (Supplementary Material in the online-only Data Supplement). The SNQ is based on a 5-point Likert scale from 1 (never) to 5 (always) with a total score ranging from 4 to 20, with a higher score indicating increased sleep need (Supplementary Material in the online-only Data Supplement). In the current study (n=581), the SNQ demonstrated acceptable internal consistency reliability (Cronbach’s α=0.74 and McDonald’s ω=0.80), with average inter-item correlation (matrix r= 0.43) [11].

Convergent Validity

Epworth Sleepiness Scale

The ESS is an 8-item single time-point self-report measure used to detect daytime sleepiness [7]. The ESS is based on a 4-point Likert scale from 0 (would never nod off) to 3 (high chance of nodding off) for common dozing situations, with a total score ranging from 0 to 24. A score of greater than 10 raises concerns of significant daytime sleepiness (as compared to healthy controls), and a score of greater than 16 indicates a high level of daytime sleepiness that may be due to a sleep disorder (i.e., moderate to severe sleep apnea, idiopathic hypersomnia, narcolepsy). In this study, the ESS demonstrated an acceptable internal consistency (Cronbach’s α=0.78).

PROMIS Sleep Disturbance

The PROMIS Sleep Disturbance (PSD) is an 8-item single-time-point self-report measure used to detect sleep disturbance over the last week [12]. The PSD is based on a 5-point Likert scale from 1 to 5 (anchors vary) with a total score ranging from 8 to 40, with higher scores indicating increased sleep disturbance. Responders rate how frequently they struggle with various components of sleep disturbance, such as satisfaction, quality, and sleep onset. In this study, the PSD demonstrated an excellent internal consistency (Cronbach’s α=0.92).

PROMIS Sleep Related Impairment

The PROMIS Sleep Related Impairment (PSRI) is an 8-item single time-point self-report measure used to detect sleep-related impairment over the last week [12]. The PSRI is based on a 5-point Likert scale from 1 (not at all) to 5 (very much) with a total score ranging from 8 to 40. Responders rate how frequently they struggle with various components of sleep-related impairment, including tiredness, alertness, and mood due to poor sleep. An increasing score indicates an increasing severity in sleep-related impairment. In this study, the PSRI demonstrated an excellent internal consistency (Cronbach’s α=0.91).

Insomnia Severity Index

The Insomnia Severity Index (ISI) is a 7-item single time-report measure used to assess subjective insomnia symptom severity (e.g., over the last two weeks) [13]. The ISI is based on a 5-point Likert scale (anchors vary) and has a total score ranging from 0 to 28 with higher scores representing greater insomnia severity. Responders rate their current severity of several domains that are thought to encapsulate insomnia, including difficulty falling/staying asleep, satisfaction with sleep, daily hindrance due to sleep, and worry around sleep over the past week (altered for purposes of the current study). In this study, the ISI demonstrated good internal consistency (Cronbach’s α=0.88).

Sleep Condition Indicator

The Sleep Condition Indicator (SCI) is an 8-item single time-report measure used to assess perceived insomnia symptoms and aid in insomnia disorder diagnosis [14]. The SCI is based on a 5-point Likert scale (reverse scored, i.e., 4–0; anchors vary) and has a total score ranging from 0 to 32, with lower scores indicating less severe insomnia symptoms. For the ease of scoring and correlation interpretation, this study reverse-coded this measure from its published version so that higher scores indicated worse insomnia severity. In this study, the SCI demonstrated a good internal consistency (Cronbach’s α=0.87).

Discriminant Validity

Alcohol Use Disorders Identification Test

The Alcohol Use Disorders Identification Test (AUDIT) is a 10-item single time-point self-report measure used to assess harmful levels of alcohol consumption [15]. The AUDIT is based on a 5-point Likert scale from 0 to 4 (anchors vary) with a total score ranging from 0 to 40. An increasing score indicates an increasing severity in risky alcohol consumption. In this study, the AUDIT demonstrated a good internal consistency (Cronbach’s α=0.85).

Cannabis Use Disorder Identification Test

The Cannabis Use Disorder Identification Test (CUDIT) is a 10-item single time-point self-report measure used to assess harmful levels of cannabis consumption [16]. The CUDIT is based on a 5-point Likert scale from 0 to 4 (anchors vary) and has a total score ranging from 0 to 40. An increasing score indicates an increasing severity in risky cannabis consumption. In this study, the CUDIT demonstrated a good internal consistency (Cronbach’s α=0.80).

Brief Sensation Seeking Scale

The Brief Sensation Seeking Scale (BSSS-4) is a 4-item single time-point self-report measure used to assess sensation seeking [17]. The BSSS-4 is based on a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree) and has a total score ranging from 5 to 20. An increasing score indicates an increase in sensation-seeking. In this study, the BSSS-4 demonstrated an acceptable internal consistency (Cronbach’s α=0.79).

RESULTS

Daytime Sleepiness Accuracy

The utility of the SNQ as a standalone tool, or as a decisional aid, is unknown at the time of the present study. An ROC curve analysis [18] using SPSS Version 29.0.2.0 (IBM Corp.) was conducted to gain preliminary insight into the SNQ’s ability to differentiate daytime sleepiness measured by the ESS. A score of 11 was used as the binary criterion for daytime sleepiness as constructed by the ESS [7]. Youden’s Index was generated to determine the optimal daytime impairment clinical cutoff score of the SNQ utilizing the following formula for all possible SNQ scores: J=(sensitivity+specificity)-1. The maximum of Youden’s Index across all scores of the SNQ indicates the score on the SNQ at which both sensitivity and specificity are maximized [19].
Results indicated the SNQ demonstrated acceptable discriminatory ability, area under the curve (AUC)=0.73, confidence interval=[0.68–0.77], p<0.001. Youden’s Index was highest at a cutoff score of 12 (J=0.345), with a sensitivity of 0.587, and a specificity of 0.758; a cutoff score of 11 yielded a similar Youden’s Index (J=0.342) with a sensitivity of 0.748 and a specificity of 0.594, potentially offering a better cutoff score for screening and detecting daytime sleepiness (Table 4).

Exploratory Factor Analysis

The lack of a priori knowledge or theory regarding the latent structure of the SNQ in previous literature at the time of the present study indicated the need for an EFA [20]. The goal of EFA was to explore the underlying factor structure of the SNQ and determine the degree to which the individual items in the SNQ properly load onto its underlying factor structure. SPSS was used in the present study to perform the EFA. A Kaiser–Meyer–Olkin (KMO) and Bartlett’s Test of Sphericity was conducted to determine if the SNQ was suitable for factor analysis [21,22].
Results of the KMO measure indicated mediocre sampling accuracy, KMO=0.65, and Bartlett’s test of sphericity was significant; χ2(6)=920.05, p<0.001. These results support moderate to acceptable sampling accuracy and factorability of the data. Results of inter-item correlations ranged from 0.165 to 0.833 (Table 5). All correlations were significant (p<0.001), with Items 1 and 2 demonstrating a large correlation (r=0.833); medium correlations were observed between Item 1 and 3 (r=0.339) and between Item 2 and 3 (r=0.435); a small correlation between Item 3 and 4 was found (r=0.165) [23]. The average inter-item correlation computed was r=0.43.
EFA analysis indicated a single factor with an eigenvalue greater than 1 (λ=2.36) accounting for 58.88% of the variance (2.36 units of the total 4 units of standardized variance). The remaining eigenvalues were less than 1 (Supplementary Table 1 in the online-only Data Supplement). Communalities indicated strong extraction for Item 1 (0.715) and Item 2 (0.969), but low extraction for Item 3 (0.192) and Item 4 (0.176) (Supplementary Table 2 in the online-only Data Supplement). This suggests Items 1 and 2 share considerable common variance with the construct, but Items 3 and 4 do not. Factor loadings ranged from 0.419 to 0.985 (Supplementary Table 3 in the online-only Data Supplement).

Item Response Theory

IRT is a family of statistical psychometric models that examine how each item in a measure provides varying degrees of information across the latent-trait continuum [24]. GRM is an IRT approach allowing for the modeling and analysis of polytomous response items [25]. GRM was selected for the present study due to the SNQ’s Likert response structure. Analyses were conducted in R (Version 4.5.1; https://www.r-project.org/) using the mirt package (Version 1.45.1) [26]. To our knowledge, this is the first application of GRM to the SNQ, providing novel insight into item functioning and measurement precision.
A unidimensional GRM was used due to the EFA eigenvalue calculations indicating a single factor structure. The model estimated discrimination (a) and difficulty (bm) parameters for each item along the latent trait (θ) continuum (Supplementary Table 4 in the online-only Data Supplement). The discrimination (a) of an item indicates how well responses to that item are able to differentiate between individuals at each level of the latent trait. The difficulties (b1, b2, b3, b4) of each item correspond to the demarcations between each response category (i.e., between never and rarely, between rarely and sometimes, etc.) and represent the points along the latent trait continuum at which an individual has a 50% probability of choosing at least that response category or higher. Individual item reliability across the latent trait continuum was assessed using item information functions (IIF), and the overall reliability of the SNQ was assessed using the test information function. Item probability functions were also used to visually represent how each item category responded across the latent trait continuum. Local dependence (LD) was evaluated using the standardized residual correlation and LD matrix.
Items 1 and 2 exhibited strong LD (16,384.57) and Item 4 exhibited poor fit to the model (S-X2=49.87, p<0.001). Given these issues, the following results should be interpreted with caution. Parameter estimates for Item 4 are not discussed due to poor model fit. GRM analyses indicated significant imbalances in functioning across the four items of the SNQ. Item 3 exhibits moderate discrimination parameters, while Item 1 exhibits a very high discrimination parameter and Item 2 exhibits a near perfect discrimination parameter [27]. Each of Items 1 and 2 corresponding difficulty parameters (bm) have overlapping 95% confidence intervals, indicating that Items 1 and 2 response categories measure essentially the same range and levels of the latent trait. Item 2’s much larger discrimination parameter demonstrates that it is much better at distinguishing between individuals at those levels (i.e., difficulties) of the latent trait. Item 3’s b4 of 3.51 (standard error=0.37) indicates that the highest response category (“always”) for item 3 measures the highest level of the latent trait across all four items.
The IIFs indicated that Item 2 overwhelmingly contributed to the total test information at its difficulty thresholds. Item 1 functions very similarly to Item 2 in terms of its range of information across the latent trait continuum, but contributes only an average of 2.67% of the total test information at its corresponding difficulty thresholds compared to Item 2’s average of 96.91%. Item 3 contributes negligibly to total test information around Items 1 and 2 difficulties (average of 0.19%) but provides a much higher proportion of total test information at the higher (42.61% at θ=+3) and lower range of the latent trait continuum (33.91% at θ=−3).

DISCUSSION

Findings

Although the SNQ displayed moderate to acceptable internal consistency reliability (Cronbach’s α=0.74) [11] and strong convergent and discriminant validity, there are issues when other analyses are interpreted together.
The SNQ demonstrated strong convergence with the PSRI and the ISI, potentially raising concerns about the overlap across constructs. Although there is no explicit threshold at which convergent validity becomes redundancy, correlations of this size are expected, as insomnia symptoms theoretically precede daytime sleep-related impairment and sleep need (increasing the magnitude of association). The SNQ demonstrated mixed item-level validity with a large discrepancy between inter-item correlation ranges. Item 3 (naps/dozing) displayed an inter-item correlation range of r=[0.165, 0.435], and Item 4 (adequate duration) displayed an inter-item correlation range of r=[0.165, 0.409]. It is important to note that within that range, the lowest correlation is between Items 3 and 4 (r=0.165), denoting item validity concerns as this relationship is much weaker than any other inter-item correlation. Additionally, Items 1 and 2 are almost identical in semantic construction for a general population and correlate very strongly (r=0.833). This result suggests these items may be too similar in construction, leading to redundancy in the measure, which is problematic, especially in low-item-count measures [11]. Given that the SNQ is only a 4-item measure, one might hypothesize that if Items 3 and 4 do not correlate moderately with Items 1 and 2, then they might correlate moderately with each other or load onto a separate factor within the measure. As noted in the results of the EFA analysis, this loading was not observed, as a one factor solution overwhelmingly prevailed.
Insight into the utility of the SNQ’s ability to detect daytime sleepiness per the ESS was examined, yielding the highest Youden’s Index (J=0.345) at a cutoff score of ≥12. At a cutoff score of ≥11 (J=0.342), the measure offered better sensitivity and worse specificity (Table 4). Total AUC demonstrated acceptable discriminatory utility for the SNQ (AUC=0.73). However, this must be interpreted with caution as the Youden Index demonstrates a clear imbalance between sensitivity and specificity. This imbalance most likely restricts the SNQ’s utility in clinical settings. Yet a cutoff score of 11 potentially gives it utility as a screener, as casting a wide net may be appropriate for brief screening where high sensitivity is prioritized [28].
Currently, the Department of Veterans Affairs is utilizing the SNQ as a tool to make treatment decisions for sleep restriction in CBT-I protocols in combination with sleep efficiency (SE; i.e., total sleep time divided by total time in bed [TIB]) as an indicator for whether to increase, decrease, or maintain the current TIB [10]. Within this protocol, SE and total SNQ score are assessed. If SE ≥85%, an SNQ score of ≤9 indicates no change in TIB, a score of 10–12 indicates a 15-minute increase in TIB, and a score of ≥13 indicates a 30-minute increase in TIB. However, if SE is <80%, an SNQ score of ≤9 indicates a 15-minute reduction in TIB [10], implementing SNQ cutoff scores of 9, 10, and 13, respectively. Given our findings, the cut scores used in this CBT-I protocol may need to be reevaluated. However, additional research is needed, particularly within a clinical sample.
EFA yielded several key implications of the underlying structure of the SNQ. The analysis indicated a one factor model (λ= 2.36) accounting for roughly 59% of the total shared variance across items. Given that the SNQ is only a 4-item measure, it is not surprising that the underlying structure is unidimensional. However, the pattern of loadings and communalities suggests that the underlying factor is only weakly defined by the 4 items. Items 1 and 2 demonstrated excellent extraction (h2=0.715 and 0.969) and strong factor loading (0.846 and 0.985), which suggests that both items represent the underlying factor well. However, when considering how strongly the items correlate (r=0.833) and their very similar question wording, it is possible that they are too similar in construction. This conclusion is further supported by the GRM analysis indicating a strong LD and indistinguishable difficulty parameters among Items 1 and 2. Given these results, it is possible that the SNQ would benefit from the exclusion of Item 1 entirely.
Furthermore, Items 3 and 4 demonstrated low communalities (h2=0.192 and 0.176) and weak factor loadings (0.439 and 0.419). This raises concerns about the conceptual coherence of how sleep need is measured in the SNQ. Additionally, the moderate discrimination parameter of Item 3 and poor model fit of Item 4 in the GRM analysis suggest that these items contribute negligibly to the total information provided by the SNQ across the latent trait continuum. However, Item 3’s high difficulty parameter (b4) suggests that it may be more sensitive to higher levels of the latent trait, and thus it may be a more reliable indicator of an individual’s sleep need in a clinical population (Supplementary Table 4 in the online-only Data Supplement).
The inter-item correlations, unidimensional structure, communalities, and loadings together suggest that the SNQ may have a single-core factor in which only 1–2 items contribute meaningfully to the factor identified. Further, the observed LD of the highest precision items (i.e., Items 1 and 2) in the GRM analysis strongly suggests that the latent factor structure of the SNQ is overwhelmingly driven by a single item (i.e., Item 2). Practical implications of these EFA and IRT results potentially weaken the SNQ’s structural validity and indicate the need for a confirmatory factor analysis (CFA) in an independent sample to verify these findings.

Limitations

The cross-sectional nature of the dataset is a limitation of the present study because the temporal stability of the SNQ could not be assessed. Establishing test–retest reliability is a critical component in psychometric validations, though the present study could not examine the SNQ’s consistency and reliability over time. Future studies should incorporate repeated within-participant measures of the SNQ to assess the measure’s psychometric reliability longitudinally.
Additionally, the lack of additional data sources restricts the comprehensiveness of the present analysis. The SNQ measures subjective sleep need, and more in-depth methods used to assess sleep need, such as multiple sleep latency testing, daily sleep diaries, longitudinal modelling, and ecological momentary assessment (EMA) were not used in the present study. Further exploration of these more complex methodologies to capture the construct of sleep need more comprehensively would strengthen the conceptual validation of the SNQ.
Although college students make up a large portion of the USA population, the current sample limits the generalizability of the present results. For example, mixed findings around the clinical cutoff score of the ISI have been found in different samples; a cutoff score of greater than or equal to 8 in a sample of college students [29] and a cutoff score of greater than or equal to 11 in a clinical population [30]. The sample homogeneity across various demographic factors additionally limits the generalizability and validity of these analyses.

Future Directions

More research needs to be done on the mapping of sleep need as a construct to excessive daytime sleepiness. Future studies should investigate whether the trade-offs between the clinical cutoff scores of 11 and 12 found in the present study are a relevant indicator of modifying TIB in sleep restriction therapy [10]. With a goal of contributing to the broader psychometric validation of the SNQ, future studies should compare the SNQ to multiple sleep latency testing, daily sleep diaries, use EMA, and CFA to support the present study’s EFA results. Because EFA is inherently data-driven rather than theory-driven, CFA is also necessary to support the underlying latent structure of the SNQ found in the present study [31,32]. Additional investigation into the generalizability of the structure and validation of the SNQ should also be explored in clinical populations.

Conclusion

To our knowledge, the current study conducts the first psychometric evaluation of the SNQ, presenting preliminary evidence for the measure’s reliability, convergent validity, discriminant validity, underlying latent structure, and item-level behavior. Strong convergent and discriminant validity suggest that the SNQ converges strongly with measures examining similar constructs and discriminates measures examining different constructs. EFA analyses supported a one-factor solution but should be interpreted with caution due to weak inter-item correlations, weak to moderate factor loadings, and redundancy between items. Item level behavior displayed redundancy through LD between items as well as a questionable factor structure through observed imbalances in item information contributions to total test information across the latent trait continuum. Collectively, the present findings provide a preliminary essential step towards a comprehensive validation of the SNQ, however further analysis of temporal stability, confirmation of the underlying latent structure, and further research around clinical cutoffs are needed.

Supplementary Materials

The online-only Data Supplement is available with this article at https://doi.org/10.17241/smr.2026.03699.

Supplementary Table 1.

Initial eigenvalues and variance explained
smr-2026-03699-Supplementary-Table-1.pdf

Supplementary Table 2.

Item communalities from the exploratory factor analysis of the Sleep Need Questionnaire
smr-2026-03699-Supplementary-Table-2.pdf

Supplementary Table 3.

Factor loadings from the exploratory factor analysis of the Sleep Need Questionnaire
smr-2026-03699-Supplementary-Table-3.pdf

Supplementary Table 4.

Graded response model parameter estimates
smr-2026-03699-Supplementary-Table-4.pdf

Availability of Data and Material

The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.

NOTES

Ethics Statement
Informed consent was obtained from all participants. Institutional Review Board approval was obtained from Saint Louis University (protocol #30467) with corresponding institutional authorization agreements with the remaining participating universities.
Author Contributions
Conceptualization: Austin Kitch, Daniel J. Taylor, Sarah E. Emert. Data curation: Austin Kitch, Daniel J. Taylor, Jeremiah Weinstock, Sarah E. Emert. Formal analysis: Austin Kitch, Jacob Thomas. Investigation: Austin Kitch, Jacob Thomas. Methodology: all authors. Project administration: Austin Kitch, Daniel J. Taylor, Jeremiah Weinstock, Sarah E. Emert. Resources: Sarah E. Emert. Software: Austin Kitch, Jacob Thomas. Supervision: Daniel J. Taylor, Jeremiah Weinstock, Sarah E. Emert. Validation: all authors. Visualization: Austin Kitch, Jacob Thomas. Writing—original draft: Austin Kitch, Jacob Thomas. Writing—review & editing: Daniel J. Taylor, Jeremiah Weinstock, Sarah E. Emert.
Conflicts of Interest
The authors have no potential conflicts of interest to disclose.
Funding Statement
None
Acknowledgements
None

REFERENCES

1. Morin CM, Jarrin DC. Epidemiology of insomnia: prevalence, course, risk factors, and public health burden. Sleep Med Clin 2022;17:173-91.
pmid
2. Morin CM, LeBlanc M, Daley M, Gregoire JP, Mérette C. Epidemiology of insomnia: prevalence, self-help treatments, consultations, and determinants of help-seeking behaviors. Sleep Med 2006;7:123-30.
crossref pmid
3. Ulander M, Hedner J, Stillberg G, Sunnergren O, Grote L. Correlates of excessive daytime sleepiness in obstructive sleep apnea: results from the nationwide SESAR cohort including 34,684 patients. J Sleep Res 2022;31:e13690.
pmid pmc
4. Fasiello E, Mombelli S, Sforza M, Zucconi M, Casoni F, Chadia K, et al. Challenging subjective excessive daytime sleepiness as an insomnia symptom: a retrospective study. J Sleep Res 2024;33:e14118.
crossref pmid
5. Vgontzas AN, Zoumakis E, Bixler EO, Lin HM, Follett H, Kales A, et al. Adverse effects of modest sleep restriction on sleepiness, performance, and inflammatory cytokines. J Clin Endocrinol Metab 2004;89:2119-26.
crossref pmid
6. Spielman AJ, Saskin P, Thorpy MJ. Treatment of chronic insomnia by restriction of time in bed. Sleep 1987;10:45-56.
pmid
7. Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 1991;14:540-5.
crossref pmid
8. Scharf MT. Reliability and efficacy of the Epworth sleepiness scale: is there still a place for it? Nat Sci Sleep 2022;14:2151-6.
crossref pmid pmc
9. Fong SY, Ho CK, Wing YK. Comparing MSLT and ESS in the measurement of excessive daytime sleepiness in obstructive sleep apnoea syndrome. J Psychosom Res 2005;58:55-60.
crossref pmid
10. Manber R, Friedman L, Siebern AT, Carney C, Edinger J, Epstein D, et al. Cognitive behavioral therapy for insomnia in veterans: therapist manual. Washington, DC: U.S. Department of Veterans Affairs 2014.

11. Clark LA, Watson D. Constructing validity: basic issues in objective scale development. Psychol Assess 1995;7:309-19.
crossref
12. Buysse DJ, Yu L, Moul DE, Germain A, Stover A, Dodds NE, et al. Development and validation of patient-reported outcome measures for sleep disturbance and sleep-related impairments. Sleep 2010;33:781-92.
crossref pmid pmc
13. Morin CM. Insomnia: psychological assessment and management. New York: Guilford Press 1993.

14. Espie CA, Kyle SD, Hames P, Gardani M, Fleming L, Cape J. The sleep condition indicator: a clinical screening tool to evaluate insomnia disorder. BMJ Open 2014;4:PMID: e004183.
crossref
15. Saunders JB, Aasland OG, Babor TF, de la Fuente JR, Grant M. Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption--II. Addiction 1993;88:791-804.
crossref pmid
16. Adamson SJ, Sellman JD. A prototype screening instrument for cannabis use disorder: the Cannabis Use Disorders Identification Test (CUDIT) in an alcohol-dependent clinical sample. Drug Alcohol Rev 2003;22:309-15.
crossref pmid
17. Stephenson MT, Hoyle RH, Palmgreen P, Slater MD. Brief measures of sensation seeking for screening and large-scale surveys. Drug Alcohol Depend 2003;72:279-86.
crossref pmid
18. Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clinical Chemistry 1993;39:561-77.
crossref pmid
19. Youden WJ. Index for rating diagnostic tests. Cancer 1950;3:32-5.
crossref pmid
20. Suhr D. Exploratory or confirmatory factor analysis? Proceedings of the 31st Annual SAS Users Group International Conference. 2006; Cary, NC. Cary, NC: SAS Institute Inc. Paper; 200-31.

21. Kaiser HF. An index of factorial simplicity. Psychometrika 1974;39:31-6.
crossref
22. Bartlett MS. Tests of significance in factor analysis. Br J Psychol 1950;3:77-85.
crossref
23. Cohen J. The concepts of power analysis. In: Cohen J. Statistical power analysis for the behavioral sciences (revised edition) New York: Academic Press; 1977;75-7.
crossref
24. Chen Y, Li X, Liu J, Ying Z. Item response theory—a statistical framework for educational and psychological measurement. Statist Sci 2025;40:167-94.
crossref
25. Samejima F. Graded response model. In: van der Linden WJ, Hambleton RK. Handbook of Modern Item Response Theory New York, NY: Springer; 1997;85-100.
crossref
26. Chalmers RP. mirt: a multidimensional item response theory package for the R environment. J Stat Softw 2012;48:1-29.

27. Baker FB. The basics of item response theory. 2nd ed. Washington, DC: ERIC Clearinghouse on Assessment and Evaluation 2001.

28. Trevethan R. Sensitivity, specificity, and predictive values: foundations, pliabilities, and pitfalls in research and practice. Front Public Health 2017;5:307.
crossref pmid pmc
29. Emert SE, Dietch JR, Bramoweth AD, Kelly K, Taylor DJ. Psychometric evaluation of the Insomnia Severity Index in U.S. college students. Behav Ther 2024;55:990-1003.
crossref pmid pmc
30. Bastien CH, Vallières A, Morin CM. Validation of the Insomnia Severity Index as an outcome measure for insomnia research. Sleep Med 2001;2:297-307.
crossref pmid
31. Hurley AE, Scandura TA, Schriesheim CA, Brannick MT, Seers A, Vandenberg RJ, et al. Exploratory and confirmatory factor analysis: guidelines, issues, and alternatives. J Organ Behav 1997;18:667-83.
crossref
32. Widaman KF. Exploratory factor analysis and confirmatory factor analysis. In: Cooper H, Camic PM, Long DL, Panter AT, Rindskopf D, Sher KJ. APA handbook of research methods in psychology. Vol. 3. Data analysis and research publication Washington, DC: American Psychological Association; 2021;361-89.
crossref

Table 1
Demographic and clinical characteristics of the study sample
Value
Age (yr) 19.09±1.15
Sex
 Male 147 (25.3)
 Female 434 (74.7)
Race
 American Indian/Alaska Native 15 (2.6)
 Asian 42 (7.2)
 Black/African American 57 (9.8)
 Caucasian 475 (81.8)
 Middle Eastern 12 (2.1)
 Other 12 (2.1)
Ethnicity
 Hispanic/Latinx 73 (12.6)
 Not Hispanic/Latinx 508 (87.4)
SNQ (n=581) 11.35±3.12
ESS (n=581) 7.99±4.39
PSD (n=1,185) 21.70±7.58
PSRI (n=1,182) 21.32±7.15
ISI (n=580) 8.03±5.75
SCI (n=1,183) 10.98±7.09
AUDIT (n=878) 6.77±4.86
CUDIT (n=383) 6.98±4.86
BSSS-4 (n=1,186) 13.01±3.73

Values are presented as mean±standard deviation or number (%). When selecting race, multiple answers were permitted.

SNQ, Sleep Needs Questionnaire; ESS, Epworth Sleepiness Scale; PSD, PROMIS Sleep Disturbance; PSRI, PROMIS Sleep Related Impairment; ISI, Insomnia Severity Index; SCI, Sleep Condition Indicator; AUDIT, Alcohol Use Disorders Identification Test; CUDIT, Cannabis Use Disorder Identification Test; BSSS-4, 4-item Brief Sensation Seeking Scale.

Table 2
Correlation matrix of convergent validity measures
SNQ M±SD ESS (n=581) PSD (n=581) PSRI (n=581) ISI (n=580) SCI (n=581)
Item 1: tired/fatigue 2.95±1.00 0.31** 0.53** 0.65** 0.62** 0.59**
Item 2: sleepy/drowsy 2.92±1.01 0.41** 0.52** 0.68** 0.64** 0.61**
Item 3: naps/dozing 2.60±1.10 0.46** 0.22** 0.42** 0.33** 0.29**
Item 4: adequate duration 2.89±1.04 0.20** 0.57** 0.50** 0.57** 0.54**
Total 11.35±3.12 0.46** 0.60** 0.74** 0.71** 0.67**

Correlation analyses included all participants with complete data on both variables, with total n for each analysis listed in the corresponding column.

** p<0.01 (2-tailed).

SNQ, Sleep Need Questionnaire; M, mean; SD, standard deviation; ESS, Epworth Sleepiness Scale; PSD, PROMIS Sleep Disturbance; PSRI, PROMIS Sleep Related Impairment; ISI, Insomnia Severity Index; SCI, Sleep Condition Indicator.

Table 3
Correlation matrix of discriminant validity measures
SNQ M±SD AUDIT (n=428) CUDIT (n=231) BSSS-4 (n=581)
Item 1: tired/fatigue 2.95±1.00 0.04 0.13 0.07
Item 2: sleepy/drowsy 2.91±1.01 0.06 0.14* 0.07
Item 3: naps/dozing 2.60±1.10 0.10* 0.06 −0.05
Item 4: adequate duration 2.88±1.05 0.09 −0.02 0.03
Total 11.35±3.12 0.10* 0.10 0.04

Correlation analyses included all participants with complete data on both variables, with total n for each analysis listed in the corresponding column.

* p<0.05 level (2-tailed).

SNQ, Sleep Need Questionnaire; M, mean; SD, standard deviation; AUDIT, Alcohol Use Disorders Identification Test; CUDIT, Cannabis Use Disorders Identification Test; BSSS-4, 4-item Brief Sensation Seeking Scale.

Table 4
Diagnosis utility determinant measures
SNQ cutoff Sensitivity Specificity Youden’s Index (J)
≥8 0.935 0.239 0.175
≥9 0.910 0.354 0.264
≥10 0.832 0.455 0.288
≥11 0.748 0.594 0.342
≥12 0.587 0.758 0.345
≥13 0.471 0.850 0.321
≥14 0.323 0.910 0.224

SNQ, Sleep Need Questionnaire.

Table 5
Inter-item correlation matrix
SNQ Item 1 Item 2 Item 3 Item 4
Item 1: tired/fatigue 1.00
Item 2: sleepy/drowsy 0.833*** 1.00
Item 3: naps/dozing 0.339*** 0.435*** 1.00
Item 4: adequate duration 0.392*** 0.409*** 0.165*** 1.00

*** p<0.001 (1-tailed).

SNQ, Sleep Need Questionnaire.