Item response theory analysis of the Dysfunctional Beliefs and Attitudes about Sleep-16 (DBAS-16) scale in a university student sample

Abstract

Unhelpful beliefs about sleep have been shown to exacerbate distress associated with sleep-related difficulties. University students are particularly vulnerable to experiencing sleep-related problems. The Dysfunctional Beliefs and Attitudes about Sleep-16 (DBAS-16) scale is a widely used instrument that assesses for sleep-disruptive cognitions. Although psychometric support for the DBAS-16 is available, Item Response Theory (IRT) analysis is needed to examine its properties at the item level. Psychometric investigation in non-clinical samples can help identify people who may be at risk for developing sleep problems. We examined the DBAS-16 using IRT on a sample of 759 university students. Our results identified items and subscales that adequately/inadequately differentiated between students who held unhelpful beliefs about sleep and those who did not. The DBAS-16 is a valuable instrument to assess unhelpful beliefs about sleep. We outline recommendations to improve the discriminatory ability of the instrument. Future investigations should establish cross-validation with a clinical sample.

Full text

RESEARCH ARTICLE
Item response theory analysis of the
Dysfunctional Beliefs and Attitudes about
Sleep-16 (DBAS-16) scale in a university
student sample
Louise I. R. Castillo
1
, Thomas HadjistavropoulosID
1
*, L. Odell Tan
2
, Ying C. MacNab
3
1Department of Psychology and Centre on Aging and Health, University of Regina, Regina, Canada,
2Department of Psychiatry and Behavioural Neurosciences, McMaster University, Hamilton, Canada,
3School of Population and Public Health, University of British Columbia, Vancouver, Canada
*Thomas.Hadjistavropoulos@uregina.ca
Abstract
Unhelpful beliefs about sleep have been shown to exacerbate distress associated with
sleep-related difficulties. University students are particularly vulnerable to experiencing
sleep-related problems. The Dysfunctional Beliefs and Attitudes about Sleep-16 (DBAS-16)
scale is a widely used instrument that assesses for sleep-disruptive cognitions. Although
psychometric support for the DBAS-16 is available, Item Response Theory (IRT) analysis is
needed to examine its properties at the item level. Psychometric investigation in non-clinical
samples can help identify people who may be at risk for developing sleep problems. We
examined the DBAS-16 using IRT on a sample of 759 university students. Our results identi-
fied items and subscales that adequately/inadequately differentiated between students who
held unhelpful beliefs about sleep and those who did not. The DBAS-16 is a valuable instru-
ment to assess unhelpful beliefs about sleep. We outline recommendations to improve the
discriminatory ability of the instrument. Future investigations should establish cross-valida-
tion with a clinical sample.
Introduction
Sleep-related difficulties affect a large proportion of the population and are associated with
impairments in interpersonal, physical, and social functioning [1–4]. University students, in
particular, are vulnerable to experiencing sleep problems [5]. In a survey of university students
in the nited States, 43% reported difficulties staying asleep (e.g., waking up more than once in
the middle of the night) [6]. Similar results have been obtained in other countries [7]. This is
concerning given that sleep difficulties among university students are related to increased
mental health problems [8–10]. In addition, reduced duration and quality of sleep among uni-
versity students have been shown to negatively impact daily activities and academic perfor-
mance [11–13]. Even among students with generally healthy sleeping habits, reduced sleep
quality is associated with poorer mental health outcomes [14]. Given the high prevalence of
PLOS ONE
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 1 / 15
a1111111111
a1111111111
a1111111111
a1111111111
a1111111111
OPEN ACCESS
Citation: Castillo LIR, Hadjistavropoulos T, Tan LO,
MacNab YC (2023) Item response theory analysis
of the Dysfunctional Beliefs and Attitudes about
Sleep-16 (DBAS-16) scale in a university student
sample. PLoS ONE 18(2): e0281364. https://doi.
org/10.1371/journal.pone.0281364
Editor: Serena Scarpelli, Sapienza University of
Rome: Universita degli Studi di Roma La Sapienza,
ITALY
Received: December 16, 2021
Accepted: January 23, 2023
Published: February 2, 2023
Copyright: ©2023 Castillo et al. This is an open
access article distributed under the terms of the
Creative Commons Attribution License, which
permits unrestricted use, distribution, and
reproduction in any medium, provided the original
author and source are credited.
Data Availability Statement: The data set for this
investigation can be accessed through the
following repository: doi: 10.5683/SP3/BHAZW8.
Funding: The authors received no specific funding
for this work.
Competing interests: The authors have declared
that no competing interests exist.

sleep problems among university students, sleep-related measures have been developed for
and tested in this population [15,16]. In fact, use of a measure of attitudes about sleep should
not be restricted to clinical populations but may be useful for the identification of people who
may be at risk for developing insomnia later on. As the cognitive model of insomnia posits,
unhelpful beliefs about sleep exacerbate intrusive worries about lack of sleep which, in turn,
trigger physical arousal and interferes with ability to fall asleep [17]. As such, psychometric
analyses of scales concerning dysfunctional attitudes about sleep are important to validate in
student and other non-clinical populations so that those at risk for developing future sleep
problems can be identified and offered appropriate preventative interventions and guidance.
One of the widely used instrument used to examine problems and difficulties related to
sleep is the Dysfunctional Beliefs and Attitudes about Sleep (DBAS) scale. Morin [18] devel-
oped the 30-item scale to assess misconceptions about the causes of insomnia, misattributions
of the consequences of insomnia, and unrealistic sleep expectations. Morin et al. [19] later pre-
sented an abbreviated 16-item (DBAS-16) version of the original 30-item scale to ease the
administration and improve the utility of the scale. Since their development, the DBAS scales
have been widely used and tested in clinical and nonclinical samples [20–23]. Moreover, the
original and abbreviated scales have been translated into multiple languages [24–27]. The
DBAS-16 particularly targets the following sleep-disruptive cognitions: expectations about
sleep requirements, attributions of the causes and appraisals of the consequences of insomnia,
and issues of worry and helplessness about sleep [19]. The DBAS-16 has been used among uni-
versity students where elevated levels of unhelpful beliefs and attitudes about sleep were identi-
fied [28]. Although there are no cut-off scores in the current scoring guidelines of the measure
[19], previous research has identified that, using a 0–10 scale, total scores of 4 or higher may
indicate unhelpful beliefs and expectations about sleep [21,29,30]. The DBAS-16 has also
been shown to be valid in discriminating between patients with insomnia and controls [21,
31]. It is also sensitive to the effects of treatment [32–34].
The psychometric properties of the DBAS scales have been studied [18,19,24,27,35] and
concerns about several DBAS-16 items have been noted. For example, Items 10 (“I can’t ever
predict whether I’ll have a good or poor night’s sleep”), 11 (“I have little ability to manage the
negative consequences of disturbed sleep”), and 16 (“I avoid or cancel obligations (social, fam-
ily) after a poor night’s sleep”) have been shown to have low factor loadings under the worry/
helplessness about insomnia (Items 10 and 11) and perceived consequences of insomnia (Item
16) subscales but were retained due to their clinical relevance [19,25]. Moreover, the 2-item
sleep expectation subscale, which includes Item 1 (“I need 8 hours of sleep to feel refreshed
and function well during the day) and 2 (“When I don’t get the proper amount of sleep on a
given night, I need to catch up the next day by napping or the next night by sleeping longer”),
has been shown to have significantly low factor loadings but was retained due to its utility in
treatment planning for individuals with insomnia [19,25,36]. Furthermore, Carney and Edin-
ger [37] highlighted concerns about multiple items in the DBAS scales not adequately differen-
tiating between good and poor sleepers.
Commensurate with the need to evaluate treatments that address maladaptive cognitions in
individuals with sleep related difficulties is the need to develop and evaluate reliable and theo-
retically sound instruments capable of assessing treatment effectiveness. Considering that the
DBAS-16 developers called for an item response theory (IRT) analysis of the scale [19] as well
as the item specific concerns outlined in the literature, we set out to further investigate the psy-
chometric properties of the DBAS-16 in a student sample vulnerable to experiencing sleep
problems. Moreover, the DBAS-16 could help with the early identification of individuals with
sleep problems and, as such, would be important to study in non-clinical samples. IRT places
emphasis on the properties of each item of the scale as opposed to a test level focus (e.g., total
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 2 / 15

score) which is the basis of classical test theory [38]. We conducted an IRT analysis to examine
the probability of item endorsement as a function of degree of unhelpful beliefs and attitudes
about sleep by examining each subscale. In particular, the objective of our investigation was to
identify the subscales and items that best discriminate between university students with low vs.
high maladaptive beliefs about sleep.
Methods
All procedures performed in studies involving human participants were in accordance with
the ethical standards of the University of Regina Research Ethics Board (# 2015–093). Written
informed consent was obtained from all individual participants included in the study using an
electronic online form.
Participants
Data from this investigation were collected as part of a larger study of 765 university student
participants [15]. In the context of the larger study supporting the validity of the DBAS-16 in a
university sample, its scores were positively correlated with higher levels of daytime sleep
worry (Insomnia Daytime Worry Scale [39]; M = 14.83, SD = 11.34), neuroticism (Neuroti-
cism scale of the Big Five Inventory [40]; M = 3.18, SD = 0.75), depressive symptoms (Center
for Epidemiologic Studies Depression Scale–10 [41]; M = 9.94 SD = 5.30), worse sleep quality
(Pittsburgh Sleep Quality Index [42]; M = 7.58, SD = 3.14), and perception of sleep problems
(Insomnia Severity Index [43]; M = 9.79, SD = 5.52) [15].
Participants were recruited using a student-wide electronic mailing list of our institution
and were offered an opportunity to be entered in a draw for one of four $50 gift cards to either
a department store or a restaurant. There were no other specific inclusion or exclusion criteria.
Potential participants were provided with a link to the study through Qualtrics Survey Soft-
ware that led to an informed consent form, demographic information, and of the aforemen-
tioned questionnaires, including the DBAS-16. Six participants with incomplete DBAS-16
responses were excluded from the final sample. Furthermore, the proportion of missing data
for each item of the DBAS-16 was less than 5–10% (i.e., less than 0.3% in our sample) which is
consistent with published guidelines for deletion of cases (e.g., [44,45]). As such, a complete
case analysis was conducted. The omitted participants represented 0.8% of the overall sample.
After the deletion of the six incomplete participant data sets, a total of 759 participants were
included in this study. The mean age of our sample was 23.39 (SD = 6.89), and 77.3% were
female. Participants had completed an average of 14.83 years (SD = 3.84) of formal education.
Measures
Dysfunctional Beliefs and Attitudes about Sleep-16 (DBAS-16). The Dysfunctional
Beliefs and Attitudes About Sleep–16 (DBAS-16) [19] is a 16 item self-report questionnaire
designed to measure various maladaptive sleep-related beliefs and attitudes an individual may
hold. In this study, we used a Likert scale with scores ranging from 1 (Strongly disagree) to 10
(Strongly agree). This scale consists of four subscales which corresponds to the four-factor
model that has been identified for the scale: 1) sleep expectations; 2) worry/helplessness about
insomnia; 3) perceived consequences of insomnia and 4) medications. Scoring involves the fol-
lowing: 1) an average overall score (i.e., obtained by adding all the scores for all the items and
dividing by 16); and 2) an average subscale score (i.e., obtained by summing the score of each
subscale and dividing by the number of items per subscale) [19]. The degree of unhelpful
beliefs is reflected in the strength of endorsement for each item (i.e., a higher rating indicates
stronger agreement to an item). The DBAS-16 has shown adequate internal consistency in
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 3 / 15

clinical samples (a = 0.77; [19]) and among university students (a = 0.82; [29]). Validity in dis-
criminating between people with and without insomnia and has also been demonstrated [21,
31]. For our sample, the overall scale illustrated strong internal consistency (Cronbach’s
alpha = 0.861) and internal consistencies of as the subscales were as follows: 0.559 (expecta-
tions); 0.792 (worry/helplessness); 0.781 (consequences); and 0.523 (medications)
Analysis
Confirmatory factor analysis. To examine model fit and in accordance with the factors
previously identified for this scale, confirmatory factor analyses (CFA) were conducted using
AMOS (SPSS 26.0). Previous research has supported a four-factor structure model fit of this
scale [19].
Item response theory analysis. Item Response Theory (IRT) is grounded on the concept
that estimates of an individual’s latent ability or theta (θ) are dependent on how one responds
to the item and the properties (e.g., difficulty of item) inherent to the items in the instrument
[46]. IRT deviates from Classical Test Theory (CTT) in various ways, namely, CTT uses the
true score and the error to directly predict an individual’s total score with the assumption that
item properties are equivalent across an instrument and that item properties are not linked to
behavior [46]. In contrast, IRT identifies a person’s trait level based on individual response pat-
terns and locates both the individual and the scale item in the same continuum of an underly-
ing construct (e.g., trait level) [46]. The ability to examine how item properties influence trait
measurement has led to the varied use of IRT analysis in the development and refinement of
various measures [15,47].
The graded response model (GRM) [48] was formulated for the analysis of items consisting
of two or more ordered categorical responses (e.g., a Likert Scale), such as the DBAS-16 items
considered in this paper. In a GRM, each item is characterized by a slope parameter and a set
of between category threshold parameters b
1
,. . ., b
C-1
, where C is the number of item response
categories (e.g., for the DBAS-16 items, C = 10). The slope parameter reflects the strength of
the relationship of the item to the latent variable. Each of the threshold parameter quantifies
the trait level necessary to have 0.50 probability of choosing a response above a given score
(e.g., the level of unhelpful belief and attitude about sleep needed to have an equal probability
of choosing strongly disagree [1] or higher) [49]. In the present study, IRTPRO (version 4.20)
was used to conduct the analysis using the GRM with 9 threshold parameters for each of the
items.
Results
Table 1 outlines the descriptive statistics of the overall, item, and subscale scores for our
sample.
Confirmatory factor analysis
Initial results of the four-factor CFA revealed poor model fit: CFI = .876, RMSEA = .080, χ2
(98) = 573.75, p<.001. As such, items with low factor loadings (i.e., less than 0.55) were
dropped to improve model fit and gain model identification. After omitting Items 10 (0.43),
Item 16 (0.52), and Item 13 (0.32), the model fit for the four-factor structure improved signifi-
cantly: CFI = 0.918, RMSEA = 0.078, SRMR = 0.054l, χ2(58) = 326.44, p<.001. The final stan-
dardized factor loadings ranged from 0.56 (Item 3) to 0.80 (Item 15). The progression of item
removal and factor loadings are outlined in Table 2. The final items were included in the IRT
analysis.
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 4 / 15

Item response theory analysis
The perceived consequences of insomnia subscale included items 5, 7, 9, and 12. The model
fully converged at a statistically significant level: AIC = 12629.33, BIC = 12814.62, M
2
(482) =
673.92, RMSEA = 0.02, p <0.01. The worry/helplessness about insomnia subscale included
items 3, 4, 8, 11, and 14. The model fully converged at a statistically significant level:
AIC = 16790.96, BIC = 15022.56, M
2
(805) = 1099.14, RMSEA = 0.02, p <0.01. The expecta-
tions about sleep subscale included items 1 and 2. The model fully converged at:
AIC = 6436.78, BIC = 6529.42, M
2
(79) = 97.81, RMSEA = 0.02, p = 0.0743. The medication
subscale for this analysis included items 6 and 15. The model fit for the medication subscale
fully converged at: AIC = 5425.73, BIC = 5518.37, M
2
(79) = 77.73, RMSEA = 0.00, p = 0.5201.
The IRT analyses were conducted according to the final four-factor structure. For brevity,
Table 3 illustrates the item specific estimates (mean and standard errors) of the slope parame-
ter (a) and two of the nine category threshold parameters (i.e., item discrimination parameters
b
1
and b
9
). Items of higher slope parameters contribute more item information to the scale.
The slope parameter estimates ranged from 1.29 (Item 2) to 3.42 (Item 15). The threshold
parameter estimate reflects item difficulty or the extent to which a respondent with a given
latent trait has an equal probability of endorsing an item [46]. The b
1
parameter estimate indi-
cates the average latent ability threshold necessary for rating items as Strongly disagree or
higher. Across all the subscales, this ranged from -2.64 (Item 12) to 0.02 (Item 15). Similarly,
the b
9
parameter estimate indicates the average latent ability threshold necessary for rating
items as Strongly agree or lower response options. Estimates ranged from 0.80 (Item 1) to 3.18
Table 1. Item, subscale, and overall scores for the DBAS-16.
Mean Standard Deviation
Total Score 4.74 1.47
Expectations Subscale 6.76 2.26
Item 1 7.17 2.63
Item 2 6.34 2.80
Worry/Helplessness Subscale 4.25 1.86
Item 3 4.73 3.00
Item 4 3.62 2.59
Item 8 3.68 2.55
Item 10 5.63 2.88
Item 11 4.57 2.44
Item 14 3.25 2.43
Consequences Subscale 5.06 1.81
Item 5 6.37 2.53
Item 7 5.13 2.51
Item 9 4.42 2.47
Item 12 6.00 2.34
Item 16 3.40 2.58
Medication Subscale 3.82 1.79
Item 6 4.02 3.04
Item 13 4.92 2.21
Item 15 2.53 2.16
Note. DBAS-16 = Dysfunctional Beliefs and Attitudes about Sleep-16 Scale
https://doi.org/10.1371/journal.pone.0281364.t001
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 5 / 15

(Item 11). For Item 15, the standard error for the b
9
parameter was high (Mean = 2.33, stan-
dard error = 6.29). This is likely due to low frequencies of response score 9 (0.5%) and 10
(2.0%) in our data; these response scores have the lowest response frequencies among the 16
items. Moreover, 50.5% of respondents scored Item 15 as a 1 (strongly disagree) and the
majority of respondents disagreed.
Furthermore, IRTPRO calculates an item information function for each item. Item infor-
mation function reflects the level of precision or information that each item provides to the
overall scale and locates the precision across the underlying trait continuum [49]. As such, an
item that adequately discriminates among individuals along the trait continuum is expected to
provide the highest degree of information for the level of latent variable (i.e., those with
unhelpful beliefs and attitudes about sleep) that the item aims to examine (e.g., between the
range θ= 1.00 to θ= 3.00). Item information functions can therefore be used to identify items
that are most helpful to the overall scale [49]. The item characteristic curve and item informa-
tion function of each item were scoped to examine curve patterns and parameter estimates.
A summary of item information estimates is outlined in Table 4. Item 14 provided the most
information (1.01) among all the items for individuals on the extreme end of the spectrum (θ
= 3.00) (i.e., high levels of worry about sleep). For the expectations subscale, Item 1 provided
the most information about respondents with average levels of expectations about sleep. Item
4 in the worry/helplessness subscale provided the most information for individuals that
Table 2. Four-factor confirmatory factor analysis of the DBAS-16.
Subscale and item number Subscale loading
Original Final
Expectations
Item 1 0.58 0.58
Item 2 0.67 0.67
Worry/Helplessness
Item 3 0.62 0.56
Item 4 0.75 0.71
Item 8 0.66 0.68
Item 10
a
0.43
Item 11 0.57 0.57
Item 14 0.76 0.75
Consequences
Item 5 0.74 0.75
Item 7 0.61 0.62
Item 9 0.77 0.78
Item 12 0.60 0.61
Item 16
a
0.52
Medication
Item 6 0.58 0.58
Item 13
a
0.36
Item 15 0.75 0.80
Note.
a
= Item removed to improve subscale fit.
An additional correlation (0.31) was added between the errors of Items 3 and 4. DBAS-16 = Dysfunctional Beliefs
and Attitudes about Sleep-16 Scale.
https://doi.org/10.1371/journal.pone.0281364.t002
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 6 / 15

Table 3. Summary of item parameters estimates.
Subscale and item number Item parameters
a b1 b9
Expectations
Item 1 2.01 (0.54) -2.43 (0.32) 0.80 (0.11)
Item 2 1.29 (0.23) -2.50 (0.34) 1.57 (0.20)
Worry/Helplessness
Item 3 1.82 (0.13) -1.01 (0.08) 1.87 (0.12)
Item 4 2.87 (0.24) -0.62 (0.06) 2.17 (0.13)
Item 8 1.44 (0.11) -0.96 (0.09) 2.75 (0.20)
Item 11 1.16 (0.09) -2.05 (0.17) 3.18 (0.26)
Item 14 2.50 (0.20) -0.44 (0.06) 2.45 (0.15)
Consequences
Item 5 2.07 (0.16) -2.33 (0.15) 1.44 (0.09)
Item 7 1.63 (0.12) -1.94 (0.14) 2.58 (0.18)
Item 9 2.11 (0.16) -1.37 (0.08) 2.44 (0.15)
Item 12 1.74 (0.13) -2.64 (0.18) 2.00 (0.13)
Medication
Item 6 1.50 (0.19) -0.62 (0.11) 2.32 (0.25)
Item 15 3.42 (0.21) 0.02 (0.05) 2.33 (6.29)
Note.a(item discrimination estimate) = indicates the strength of the item to the latent variable; b1 (item difficulty
estimate) = indicates the average latent ability necessary to have a 0.50 probability of choosing 1 (strongly disagree)
or higher; b9 = indicates the average latent ability necessary to have a 0.50 probability of choosing 10 (strongly agree)
or a lower score. Values in parenthesis are item parameter standard error estimates.
https://doi.org/10.1371/journal.pone.0281364.t003
Table 4. Summary of item information estimates.
Subscale and item number Latent ability (theta; θ)
-3.00 -2.00 -1.00 0.00 1.00 2.00 3.00
Expectations
Item 1 0.76 1.29 1.32 1.30 1.05 0.31 0.05
Item 2 0.40 0.53 0.55 0.55 0.53 0.41 0.20
Worry/Helplessness
Item 3 0.08 0.41 0.96 1.08 1.08 0.91 0.34
Item 4 0.01 0.15 1.57 2.64 2.68 2.53 0.63
Item 8 0.10 0.32 0.59 0.68 0.69 0.67 0.55
Item 11 0.26 0.39 0.43 0.44 0.44 0.43 0.40
Item 14 0.01 0.12 1.00 1.98 2.03 1.99 1.01
Consequences
Item 5 0.69 1.33 1.39 1.38 1.35 0.79 0.16
Item 7 0.34 0.74 0.86 0.86 0.85 0.83 0.62
Item 9 0.14 0.75 1.38 1.44 1.43 1.40 0.82
Item 12 0.72 0.96 0.99 0.98 0.97 0.87 0.39
Medication
Item 6 0.06 0.22 0.55 0.72 0.74 0.70 0.45
Item 15 0.00 0.01 0.34 3.24 3.75 3.72 0.99
Note. Item information estimate indicates the degree of information or level of precision an item adds to the overall scale across the underlying latent variable (theta; θ).
https://doi.org/10.1371/journal.pone.0281364.t004
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 7 / 15

excessively worry about sleep (i.e., θ= 0.00 and θ= 2.00). For the consequences subscale, Item
9 provided the most information among those with high level of unhelpful beliefs about the
negative consequences of insomnia (i.e., θ= 0.00 and θ= 2.00). For the medication subscale,
Item 15 provided significant information about respondents who heavily attribute relief from
sleep-related distress to medications alone.
Of note, Items 6, 8, and 11 had characteristically flat and unique curves in comparison to all
the other items. Fig 1 illustrates this by contrasting their item characteristic curves with that of
the Item 15. Item 15 had the highest item discriminate estimate (i.e., the best item to discrimi-
nate from high vs low dysfunctional beliefs) and provided the most information across indi-
viduals with unhelpful beliefs (i.e., 0.99–3.75).
IRTPRO also calculates test information curves for each subscale. The item information
estimates for each item are summed together to create a total information function. As with
the item information estimate, the test information curves can provide information about the
scale as a function of the location on the trait continuum [49]. Fig 2 outlines the test informa-
tion curves for each subscale. The items in the expectation subscale exhibited low test informa-
tion estimates across the full range of the latent trait (θ= 3.00 and θ= 3.00). Similarly, the
medication subscale had low test information estimate in most of the negative range of the
latent trait (θ= 3.00 and θ= -0.50). Items in the consequences subscale provided the most
Fig 1. Item characteristic curves for dysfunctional Beliefs and Attitudes about Sleep (DBAS-16) scale items 6, 8,
11 and 15. The left y-axis represents the probability of choosing a response category and the right y-axis represent the
degree of information provided across the underlying trait continuum. The peak of each response function represents
the maximum probability of choosing that response category given a theta level (i.e., underlying trait). Item
information function is indicated with a dashed line. IRTPRO (Version 4.2) by default shows response alternatives
(strongly disagree [1] to strongly agree [10]) as 0 to 9.
https://doi.org/10.1371/journal.pone.0281364.g001
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 8 / 15

information about respondents with latent levels between approximately θ= -1.50 and θ=
1.50. The worry/helplessness subscale provided the most information about respondents with
latent levels approximately between θ= -0.50 and θ= 2.00 and the highest peak of total infor-
mation estimate (i.e., the information estimate ranged from 1.5 to 8.0) among all the subscales.
Discussion
To the best of our knowledge, this study represents the first IRT analysis of the DBAS-16 scale.
According to the results and in accordance with the objective of the study, we identified items
and subscales that adequately discriminated between those with low and high levels of unhelp-
ful beliefs and attitudes about sleep in a university student sample. The items identified may
elucidate understanding on the sleep-related cognitions germane to university students. Items
10 (“I can’t ever predict whether I’ll have a good or poor night’s sleep”), 13 (“I believe insomnia
is essentially the result of a chemical imbalance”) and 16 (“I avoid or cancel obligations [social,
family] after a poor night’s sleep”) were omitted from the IRT analysis due to low factor load-
ings. Low factor loadings for Items 10 and 16 are consistent with previous research in non-
clinical samples [25]. A proposed explanation for Item 16 has been the change in appraisal for
Fig 2. Test information function of the expectations, worry/helplessness, consequences, and medication subscales.
The peak of the total information curve represents the theta level (i.e., degree of unhelpful beliefs and attitudes about
sleep) at which the Dysfunctional Beliefs and Attitudes about Sleep (DBAS-16) Scale most accurately predicts degree of
dysfunctional beliefs and attitudes about sleep (i.e., worry, expectations, attributions, consequences). The peak of the
standard error curve represents the theta level at which the DBAS-16 has the most error in predicting the degree of
dysfunctional beliefs and attitudes about sleep.
https://doi.org/10.1371/journal.pone.0281364.g002
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 9 / 15

the potential consequences of canceling obligations due to insufficient sleep as a result of alter-
nate forms of communications (e.g., social media) [25]. That is, university students may not
view avoiding or canceling social obligations as a consequence of insufficient sleep because
they have access to technology as means for continuing communication.
The item characteristic curves for Items 6 (“To be alert and function well during the day, I
believe I would be better off taking a sleeping pill rather than having a poor night’s sleep”), 8
(“When I sleep poorly one night, I know it will disturb my sleep schedule for the whole
week”), and 11 (“I have little ability to manage the negative consequences of disturbed sleep”)
had characteristically flat curves in comparison to the rest of the items, particularly for average
to moderately high response levels. The lack of well-defined or unique peaks across each
response category (i.e., score ratings 2 to 9) suggest that these response options are not being
utilized as intended and therefore lack predictive value and discriminatory power. In addition,
there is approximately the same probability of choosing response options 2 to 9 among these
items, suggesting that the items do not adequately discriminate among students with varying
degree of unhelpful beliefs about sleep. This is also consistent with the item information curves
of Items 6, 8, and 11 which indicated that the items provided little information with low preci-
sion and low score accuracy to the overall scale. These findings may be characteristic of our
university student sample. For instance, a survey found that although 43% of students experi-
enced sleep difficulties, only 10% of their sample reported using other aids (e.g., medication)
to improve their sleep [6]. On the other hand, other research has shown more frequent use of
medication among students to alleviate sleep problems [5]. The lack of defined endorsement
for certain response options may be due to the varied use of sleeping aids among university
students. For Item 8, variable sleep schedules and patterns have been documented among uni-
versity students [5,50]; as such, students may not hold strong beliefs about the effects of sleep
variability on the quality of their sleep. Nonetheless, future investigations could explore ways
of improving the instrument by refining these items so that they can better discriminate
between people who hold unhelpful attitudes about sleep and people who do not. Item 11 had
the lowest discriminatory value among all the items indicating it does not appear to adequately
differentiate among students with varying levels of worry about sleep. Moreover, further
research is necessary to ascertain a 4-factor construct and to conduct a multi-dimensional IRT
for factor structure assessment.
Expectations and medication subscales
Of the four subscales, low test information (e.g., estimated total information values less than 3)
were observed in the expectations subscale for the full range of the latent trait and the medica-
tion subscale in most of the negative range of the latent trait (e.g., individuals who do not attri-
bute insomnia to a chemical imbalance or medication). Overall, our statistical findings are
consistent with those presented in Morin et al. [19] where the 16 DBAS items were subjected
to a CFA. Morin et al. [19] reported notably lower CFA coefficient (standard estimate) for the
expectations (B = 0.50) and medication (B = 0.43) subscales. Despite identifying low item-total
correlations for a few items within the medication and expectations subscales in the develop-
ment of the DBAS-16, the items were retained due to their content validity and clinical rele-
vance among a subgroup of individuals with insomnia [19]. As such, a possible explanation for
the low test information estimates for the expectations and medication subscales could be that
our nonclinical sample comprised of individuals experiencing varying levels of sleep difficul-
ties, with a small subset potentially meeting criteria for insomnia. This finding also under-
scores the notion that the expectations and medication subscales may be more appropriate in
identifying the degree of specific unhelpful beliefs among individuals with known insomnia to
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 10 / 15

aid in treatment planning (e.g., cognitive therapy for insomnia) rather than identifying
unhelpful beliefs about sleep among individuals who may be vulnerable to developing sleep
related issues. Since the DBAS-16 was developed to stimulate greater use of the DBAS among
the sleep community [18], understanding its utility in various settings and with non-clinical
samples is of importance.
Examining the items in both subscales also offers opportunities for item refinement. In
our analysis, Item 13 (“I believe insomnia is essentially the result of a chemical imbalance”)
was removed from the medication subscale IRT analysis due to low factor loading (B = 0.36),
which seems to have contributed to the failed convergence of the original IRT fit. This is con-
sistent with the CFA performed by Morin et al. [19] where the authors also reported low fac-
tor loading for Item 13 (B = 0.06). As discussed above, Item 6 (“To be alert and function well
during the day, I believe I would be better off taking a sleeping pill rather than having a poor
night’s sleep”), in the medication subscale also led characteristically flat item characteristic
curve in comparison to the rest of the items, suggesting that the item lacks predictive value
and discriminatory power. Item 15 (“Medication is probably the only solution to sleepless-
ness”) had the highest discriminatory value for the underlying construct indicating that this
item adequately discriminates among individuals who believe that medication is the only
solution to sleeplessness from those who do not. This finding is consistent with the item dif-
ficulty estimate; in order to have an equal probability (i.e., 0.50) of rating Item 15 as either 1
or higher (i.e., Strongly disagree or higher), an individual would have to hold some unhelpful
attributions of causes of insomnia (b1 = 0.20). This suggests that this item is “difficult”, and
any degree of agreement reflects unhelpful causal attributions of insomnia (i.e., believing
that medication is the solution to sleeplessness). However, our data also led to a high stan-
dard error for the b
9
estimate of Item 15, which is likely due to low frequencies of high
response scores for this item. As such, caution must be taken when interpreting the discrimi-
natory ability of Item 15. The high standard error for Item 15 and low frequencies for high
response options also corroborates the idea that our sample may have consisted of individu-
als experiencing varying levels of sleep difficulties, with a subset potentially meeting criteria
for insomnia.
The items within the expectations subscale (Item1 [“I need 8 hours of sleep to feel refreshed
and function well during the day] and Item 2 [“When I don’t get the proper amount of sleep
on a given night, I need to catch up the next day by napping or the next night by sleeping lon-
ger”]) provided greater information about individuals without unrealistic expectations about
sleep. The items within this subscale have been retained because they represent maladaptive
cognitions expressed by subgroups of people with insomnia and could inform treatment [19].
Our results are consistent with previous research [19,25,36] demonstrating low factor load-
ings and extend previous findings by highlighting item specific issues within this subscale
using an IRT analysis.
A possible explanation for these item specific issues may be the semantics or the content of
each item. Modifying the presentation of each item or creating more stringent respondent cri-
teria could aid in improving its discriminatory ability to assess broader groups of people
experiencing sleep difficulties. For example, the idea of needing to sleep for eight hours may be
a common expectation and as such, a strong endorsement of Item 1 may not indicate unhelp-
ful attitudes. It should be noted that in this sample, over 68% of the participants rated this item
as 7/10 or greater. Modifying the item to “I should always get 8 hours of sleep” may be useful.
Similarly, the idea of needing to compensate as a result of not attaining the “proper amount of
sleep” (Item 2) may be unclear. Perhaps considering changing the item to “I need to get the
proper amount of sleep on a given night or else I will need to catch up the next day by napping
or the next night by sleeping longer” could be helpful.
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 11 / 15

Consequences and worry/helplessness subscales
Our statistical findings are consistent with those presented by Morin and colleagues [18]
who reported high CFA coefficient for the consequences (0.94) and worry/helplessness
(0.88) subscales. Furthermore, total item information curves demonstrated that the items
within the worry/helplessness subscale provided the most information and highest level of
precision about individuals that excessively worry about sleep, followed by the conse-
quences subscale. This suggest that, perhaps, the items within the worry/helplessness and
consequences subscales best represent the items needed to measure unhelpful beliefs and
attitudes about sleep among university students. This finding will need to be re-evaluated in
a clinical sample. However, we also identified characteristically flat curves for Items 8 and
11 in the worry/helplessness subscale which suggest opportunities for item and response
category refinement. The items in the consequences and worry/helplessness subscales could
be useful for screening for extreme unhelpful beliefs about sleep in individuals in the gen-
eral population (e.g., student population) who may be vulnerable to developing sleep diffi-
culties. Our results increase confidence in the effectiveness of the items within these
subscales in discriminating between individuals who hold unhelpful beliefs about sleep at
varying degrees.
The item specific issues identified in our analyses are consistent with problems raised in
studies examining clinical and nonclinical samples [19,25,36]. That said, increased sleep-
related distress has been documented among university students and this population has
been identified as being vulnerable in the development of sleep problems [51,52]. Given the
high prevalence of sleep problems among university students, sleep-related measures have
been developed for and tested in this population. In particular, catastrophic thoughts impact-
ing sleep and the overall sleep quality of university students have been the subject of previous
investigations [15,16]. Consistent with previous research [28,29] that identified similarly
elevated levels of dysfunctional beliefs about sleep in student samples, our findings demon-
strate that students are particularly vulnerable to the presence of dysfunctional beliefs about
sleep given the mean scores obtained. Although the presence of unhelpful beliefs and atti-
tudes about sleep does not directly result in insomnia, consistent with the cognitive model of
insomnia [17], it may increase a student’s vulnerability to developing sleep-related disorders.
This is concerning given that reduced quality of sleep among university students has been
shown to negatively impact daily activities, academic performance, and mental health out-
comes [11–13].
Limitations
We recognize that our focus on university student research participants may limit the gener-
alizability of our findings. At the same time, psychometric investigation of the DBAS-16 in a
non-clinical sample with vulnerability to develop insomnia is important because it can be used
as a screening tool to identify those at risk of developing sleep problems later on. Nonetheless,
studies similar to ours, using clinical samples, are warranted.
The absence of information on certain sociodemographic characteristics (e.g., ethnicity,
race) poses some limitations on our ability to generalize these results. As such, it would be
important for future research to examine the impact of these characteristics on the psychomet-
ric properties of the DBAS-16. Small sample sizes may limit the assessment of a four-factor
structure in an IRT analysis. Although we tried to mitigate this with a larger sample size of 759
participants, further research is needed to conduct a multi-dimensional IRT analysis for factor
structure assessment and ascertain a four-factor construct.
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 12 / 15

Conclusion
To the best of our knowledge, this study represents the first IRT analysis of the DBAS-16 scale.
Overall, the DBAS-16 is a valuable instrument for assessing unhelpful beliefs and attitudes
about sleep. In our university student sample, certain items within the DBAS-16 were identi-
fied to best discriminate between those with low and high levels of unhelpful beliefs and atti-
tudes. However, various items were also shown to have low discriminatory value for
identifying students with maladaptive beliefs. Items 6, 8, and 11 were identified to not be per-
forming adequately, thus suggesting opportunities for item refinement and improvements to
the overall instrument. In addition, we identified low factor loadings for Items 10, 13, and 16.
Future research should examine the impact of omitting these items from instrument. Items
within the expectations and medication subscales may be more effective at identifying unhelp-
ful beliefs about sleep in clinical samples while the items in the consequences and worry/help-
lessness subscales are effective in general settings. Psychometric investigation of the DBAS-16
in non-clinical samples is important and can help identify people who may be at risk for devel-
oping sleep problems, future investigations should examine whether these results are repli-
cated in samples of people who present for treatment for insomnia.
Author Contributions
Conceptualization: Louise I. R. Castillo, Thomas Hadjistavropoulos.
Data curation: L. Odell Tan.
Formal analysis: Louise I. R. Castillo.
Investigation: Louise I. R. Castillo.
Methodology: Thomas Hadjistavropoulos, Ying C. MacNab.
Project administration: L. Odell Tan.
Resources: Thomas Hadjistavropoulos.
Supervision: Thomas Hadjistavropoulos, Ying C. MacNab.
Writing – original draft: Louise I. R. Castillo.
Writing – review & editing: Thomas Hadjistavropoulos, L. Odell Tan, Ying C. MacNab.
References
1. Chaput J, Yau J, Rao DP, Morin CM. Prevalence of insomnia for Canadians aged 6 to 79. Health Rep.
2018; 29: 16–20. PMID: 30566205
2. Morin CM, LeBlanc M, Daley M, Gregoire JP, Merette C. Epidemiology of insomnia: prevalence, self-
help treatments, consultations, and determinants of help-seeking behaviors. Sleep Med. 2006; 7: 123–
130. https://doi.org/10.1016/j.sleep.2005.08.008 PMID: 16459140
3. Ohayon MM. Epidemiology of insomnia: what we know and what we still need to learn. Sleep medicine
reviews. 2002; 6: 97–111. https://doi.org/10.1053/smrv.2002.0186 PMID: 12531146
4. Roth T, Roehrs T. Insomnia: epidemiology, characteristics, and consequences. Clin Cornerstone.
2003; 5: 5–15. https://doi.org/10.1016/s1098-3597(03)90031-7 PMID: 14626537
5. Lund HG, Reider BD, Whiting AB, Prichard JR. Sleep patterns and predictors of disturbed sleep in a
large population of college students. J Adolesc Health. 2010; 46: 124–132. https://doi.org/10.1016/j.
jadohealth.2009.06.016 PMID: 20113918
6. Forquer LM, Camden AE, Gabriau KM, Johnson CM. Sleep patterns of college students at a public univer-
sity. J Am Coll Health. 2008; 56: 563–565. https://doi.org/10.3200/JACH.56.5.563-565 PMID: 18400669
7. Pensuksan WC, Lertmaharit S, Lohsoonthorn V, Rattananupong T, Sonkprasert T, Gelaye B, et al.
Relationship between poor sleep quality and psychological problems among undergraduate students in
the Southern Thailand. Walailak Journal of Science and Technology. 2016; 13: 235. PMID: 27152114
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 13 / 15

8. Pilcher JJ, Ginter DR, Sadowsky B. Sleep quality versus sleep quantity: relationships between sleep
and measures of health, well-being and sleepiness in college students. J Psychosom Res. 1997; 42:
583–596. https://doi.org/10.1016/s0022-3999(97)00004-4 PMID: 9226606
9. Wilson KT, Bohnert AE, Ambrose A, Davis DY, Jones DM, Magee MJ. Social, behavioral, and sleep
characteristics associated with depression symptoms among undergraduate students at a women’s col-
lege: a cross-sectional depression survey, 2012. BMC Womens Health. 2014; 14: 1–9. https://doi.org/
10.1186/1472-6874-14-8 PMID: 24410897
10. Seun-Fadipe CT, Mosaku KS. Sleep quality and psychological distress among undergraduate students
of a Nigerian university. Sleep Health. 2017; 3: 190–194. https://doi.org/10.1016/j.sleh.2017.02.004
PMID: 28526257
11. Kelly WE, Kelly KE, Clanton RC. The relationship between sleep length and grade-point average
among college students. Coll Stud J. 2001; 35: 84–86.
12. Wong ML, Lau EYY, Wan JHY, Cheung SF, Hui CH, Mok DSY. The interplay between sleep and mood
in predicting academic functioning, physical health and psychological health: A longitudinal study. J Psy-
chosom Res. 2013; 74: 271–277. https://doi.org/10.1016/j.jpsychores.2012.08.014 PMID: 23497826
13. Trockel MT, Barnes MD, Egget DL. Health-related variables and academic performance among first-
year college students: Implications for sleep and other behaviors. J Am Coll Health. 2000; 49: 125–131.
https://doi.org/10.1080/07448480009596294 PMID: 11125640
14. Milojevich HM, Lukowski AF. Sleep and mental health in undergraduate students with generally healthy
sleep habits. PloS one. 2016; 11: e0156372. https://doi.org/10.1371/journal.pone.0156372 PMID:
27280714
15. Tan LO, Hadjistavropoulos T, MacNab YC. The catastrophic thoughts about insomnia scale (CTIS):
development and validation. Cognit Ther and Res. 2017; 41: 143–154. https://doi.org/10.1007/s10608-
016-9804-y
16. Aloba OO, Adewuya AO, Ola BA, Mapayi BM. Validity of the Pittsburgh sleep quality index (PSQI)
among Nigerian university students. Sleep Med. 2007; 8: 266–270. https://doi.org/10.1016/j.sleep.
2006.08.003 PMID: 17368977
17. Harvey AG. A cognitive model of insomnia. Behav Res Ther. 2002; 40: 869–893. https://doi.org/10.
1016/s0005-7967(01)00061-4 PMID: 12186352
18. Morin CM. Dysfunctional beliefs and attitudes about sleep: preliminary scale development and descrip-
tion. Behav Therapist. 1994; 17: 163.
19. Morin CM, Vallières A, Ivers H. Dysfunctional beliefs and attitudes about sleep (DBAS): validation of a
brief version (DBAS-16). Sleep. 2007; 30: 1547–1554. https://doi.org/10.1093/sleep/30.11.1547 PMID:
18041487
20. Calkins AW, Hearon BA, Capozzoli MC, Otto MW. Psychosocial predictors of sleep dysfunction: the
role of anxiety sensitivity, dysfunctional beliefs, and neuroticism. Behavioral Sleep Med. 2013; 11: 133–
143. https://doi.org/10.1080/15402002.2011.643968 PMID: 23136825
21. Carney CE, Edinger JD, Morin CM, Manber R, Rybarczyk B, Stepanski EJ, et al. Examining maladap-
tive beliefs about sleep across insomnia patient groups. J Psychosom Res. 2010; 68: 57–65. https://doi.
org/10.1016/j.jpsychores.2009.08.007 PMID: 20004301
22. Carney CE, Harris AL, Friedman J, Segal ZV. Residual sleep beliefs and sleep disturbance following
cognitive behavioral therapy for major depression. Depress Anxiety. 2011; 28: 464–470. https://doi.org/
10.1002/da.20811 PMID: 21400641
23. Choi SJ, Joo EY, Lee YJ, Hong SB. Suicidal ideation and insomnia symptoms in subjects with obstruc-
tive sleep apnea syndrome. Sleep Med. 2015; 16: 1146–1150. https://doi.org/10.1016/j.sleep.2015.04.
026 PMID: 26298792
24. Dhyani M, Rajput R, Gupta R. Hindi translation and validation of dysfunctional beliefs and attitudes
about sleep (DBAS-16). Ind Psychiatry J. 2013; 22: 80. https://doi.org/10.4103/0972-6748.123639
PMID: 24459381
25. Lang C, Brand S, Holsboer-Trachsler E, Pu¨hse U, Colledge F, Gerber M. Validation of the German ver-
sion of the short form of the dysfunctional beliefs and attitudes about sleep scale (DBAS-16). Neurologi-
cal Sciences. 2017; 38: 1047–1058. https://doi.org/10.1007/s10072-017-2921-x PMID: 28321516
26. Munezawa T, Morin CM, Inoue Y, Nedate K. Development of the Japanese version of Dysfunctional
Beliefs and Attitudes about Sleep Scale (DBAS-J). Jpn J Sleep Med. 2009; 3: 396–403.
27. Boysan M, Merey Z, Kalafat T, Kağan M. Validation of a brief version of the dysfunctional beliefs and
attitudes about sleep scale in Turkish sample. Procedia-Social and Behavioral Sciences. 2010; 5: 314–
317. https://doi.org/10.1016/j.sbspro.2010.07.095
28. Joshi K, Mishra D, Dubey H, Gupta R. Sleep pattern and insomnia among medical students. Somnolo-
gie-Schlafforschung und Schlafmedizin. 2015; 19: 205–211. https://doi.org/10.1007/s11818-015-0012-x
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 14 / 15

29. Ain NU, Ahmad H, Rehman S, Emman S, Sethi MR, Irfan M. Dysfunctional beliefs and attitudes about
sleep in undergraduate medical and dental students. J Postgrad Med Inst 2020; 34(1): 49–53.
30. Neilson BD, Shepherd MH, Dickerson C, Chaconas EJ, Young JL, Rhon DI. Relationship Between Atti-
tudes and Beliefs About Sleep, Sleep Disturbance, and Pain Interference in Patients With Spinal Pain.
Clin J Pain. 2022; 38: 541–549. https://doi.org/10.1097/AJP.0000000000001051 PMID: 35642567
31. Cro
¨nlein T, Wagner S, Langguth B, Geisler P, Eichhammer P, Wetter TC. Are dysfunctional attitudes
and beliefs about sleep unique to primary insomnia? Sleep Med. 2014; 15: 1463–1467. https://doi.org/
10.1016/j.sleep.2014.06.018 PMID: 25440733
32. Edinger JD, Wohlgemuth WK, Radtke RA, Marsh GR, Quillian RE. Does cognitive-behavioral insomnia
therapy alter dysfunctional beliefs about sleep? Sleep. 2001; 24: 591–599. https://doi.org/10.1093/
sleep/24.5.591 PMID: 11480656
33. Morin CM, Blais F, Savard J. Are changes in beliefs and attitudes about sleep related to sleep improve-
ments in the treatment of insomnia? Behav Res Ther. 2002; 40: 741–752. https://doi.org/10.1016/
s0005-7967(01)00055-9 PMID: 12074370
34. Roane BM, Dolan DC, Bramoweth AD, Rosenthal L, Taylor DJ. Altering unhelpful beliefs about sleep
with behavioral and cognitive therapies. Cognit Ther and Res. 2012; 36: 129–133.
35. Espie CA, Inglis SJ, Harvey L, Tessier S. Insomniacs’ attributions: psychometric properties of the dys-
functional beliefs and attitudes about sleep scale and the sleep disturbance questionnaire. J Psycho-
som Res. 2000; 48: 141–148.
36. Vand HDA, Vargahan FA, Birashk B, Habibi M, Haghighi KS, Fereidooni F. Validity and reliability of the
dysfunctional beliefs and attitudes about sleep scale-10 in iranian clinical population. Iran J Psychiatry
Behav Sci. 2018; 12.
37. Carney CE, Edinger JD. Identifying critical beliefs about sleep in primary insomnia. Sleep. 2006; 29:
342–350. https://doi.org/10.1093/sleep/29.3.342 PMID: 16553020
38. Hambleton RK, Jones RW. Comparison of classical test theory and item response theory and their
applications to test development. Educational measurement: issues and practice. 1993; 12: 38–47.
39. Kallestad H, Hansen B, Langsrud K, Hjemdal O, Stiles TC. Psychometric properties and the predictive
validity of the insomnia daytime worry scale: a pilot study. Cogn Behav Ther. 2010; 39: 150–157.
https://doi.org/10.1080/16506070903115317 PMID: 19802752
40. John OP, Donahue EM, Kentle RL. Big five inventory. J Pers Soc Psychol. 1991.
41. Andresen EM, Malmgren JA, Carter WB, Patrick DL. Screening for depression in well older adults: Eval-
uation of a short form of the CES-D. Am J Prev Med. 1994; 10: 77–84.
42. Buysse DJ, Reynolds CF III, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a
new instrument for psychiatric practice and research. Psychiatry Res. 1989; 28: 193–213. https://doi.
org/10.1016/0165-1781(89)90047-4 PMID: 2748771
43. 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. https://doi.org/10.1016/s1389-9457(00)00065-4
PMID: 11438246
44. Langkamp DL, Lehman A, Lemeshow S. Techniques for handling missing data in secondary analyses
of large surveys. Academic Pediatr. 2010; 10: 205–210. https://doi.org/10.1016/j.acap.2010.01.005
PMID: 20338836
45. Jakobsen JC, Gluud C, Wetterslev J, Winkel P. When and how should multiple imputation be used for
handling missing data in randomised clinical trials–a practical guide with flowcharts. BMC Med Res
Methodol. 2017; 17: 1–10. https://doi.org/10.1186/s12874-017-0442-1 PMID: 29207961
46. Embretson SE, Reise SP. Item response theory: Psychology Press; 2013.
47. Jose A, Olino TM, O’Leary KD. Item response theory analysis of intimate-partner violence in a commu-
nity sample. J Fam Psychol. 2012; 26: 198. https://doi.org/10.1037/a0027100 PMID: 22309817
48. Samejima F. Estimation of latent ability using a response pattern of graded scores. Psychometrika
monograph supplement. 1969.
49. Toland MD. Practical guide to conducting an item response theory analysis. J Early Adolesc. 2014; 34:
120–151. https://doi.org/10.1177/0272431613511332
50. Orzech KM, Salafsky DB, Hamilton LA. The state of sleep among college students at a large public uni-
versity. J Am Coll Health. 2011; 59: 612–619. https://doi.org/10.1080/07448481.2010.520051 PMID:
21823956
51. Jensen DR. Understanding sleep disorders in a college student population. J Coll Couns. 2003; 6: 25–34.
52. Jiang X, Zheng X, Yang J, Ye C, Chen Y, Zhang Z, et al. A systematic review of studies on the preva-
lence of insomnia in university students. Public Health. 2015; 129: 1579–1584. https://doi.org/10.1016/j.
puhe.2015.07.030 PMID: 26298588
PLOS ONE
IRT analysis of the DBAS-16 scale
PLOS ONE | https://doi.org/10.1371/journal.pone.0281364 February 2, 2023 15 / 15