The Swedish SCOPA-SLEEP for assessment of sleep disorders in Parkinson s disease and healthy controls

Qual Life Res (2016) 25:2571–2577 DOI 10.1007/s11136-016-1318-2 BRIEF COMMUNICATION The Swedish SCOPA-SLEEP for assessment of sleep disorders in Par...
2 downloads 0 Views 370KB Size
Qual Life Res (2016) 25:2571–2577 DOI 10.1007/s11136-016-1318-2


The Swedish SCOPA-SLEEP for assessment of sleep disorders in Parkinson’s disease and healthy controls Peter Hagell1 • Albert Westergren1 • Shorena Janelidze2 • Oskar Hansson2,3

Accepted: 10 May 2016 / Published online: 19 May 2016 Ó Springer International Publishing Switzerland 2016

Abstract Purpose SCOPA-SLEEP is a rating scale for night-time sleep and daytime sleepiness (DS) proposed for use among people with Parkinson’s disease (PD) as well as others. We translated it into Swedish and assessed its psychometric properties in PD and age-matched healthy controls. Methods Following translation according to the dual-panel approach, the Swedish SCOPA-SLEEP was field-tested regarding comprehensibility, relevance and respondent burden (n = 20). It was then psychometrically tested according to classical test theory (data completeness, scaling assumptions, targeting, reliability and construct validity) using data from 149 people with PD and 53 agematched healthy controls from the prospective Swedish BioFINDER study. Results SCOPA-SLEEP took a mean of 3.5 min to complete and was considered easy to use and relevant. Missing item responses were \8 %, corrected item–total correlations were C0.47 (except for one DS item among controls), factor analyses suggested one dimension per scale, floor/ceiling effects were B17 %, reliability was C0.85 except for the DS scale among controls (0.65) and construct validity was supported. Conclusions Observations concur with previous evaluations, thus providing initial support for the Swedish SCOPA-SLEEP among people with PD. Further studies are & Peter Hagell [email protected] 1

The PRO-CARE Group, School of Health and Society, Kristianstad University, 291 88 Kristianstad, Sweden


Department of Clinical Sciences, Malmo¨, Lund University, Lund, Sweden


Memory Clinic, Ska˚ne University Hospital, Malmo¨, Sweden

needed to establish its generic properties and to understand its measurement properties in better detail. Keywords Parkinson’s disease  Rating scale  Reliability  Sleep  Translation  Validity

Introduction Sleep disturbances are common in Parkinson’s disease (PD) and associated with compromised functioning and well-being, but can be difficult to detect [1, 2]. Two main types of potentially treatable sleep disturbances in PD are compromised night-time sleep (NS) and daytime sleepiness (DS) [1]. One means of facilitating detection, assessment and treatment of sleep disorders is the use of rating scales [1, 3]. To this end, the SCOPA-SLEEP was developed as a brief and practical instrument for clinical and scientific assessment of NS and DS in PD [4]. The SCOPA-SLEEP [4] assesses NS (5 items), DS (6 items) and overall sleep quality (1 item). NS and DS items have 4 ordered response categories (see ‘‘Appendix’’), yielding summed total scores that range between 0–15 and 0–18, respectively (higher scores = worse); the overall sleep quality item has 7 response categories. In addition (but not included in the scoring), the SCOPA-SLEEP inquires about the use of sleeping tablets during the past month. Despite being recommended for use in PD [3], documented psychometric evidence and alternative language versions of the SCOPA-SLEEP appear sparse [4–6]. However, available studies have reported acceptable psychometric properties with, e.g. corrected item–total correlations C0.38, coefficient alpha C0.83, floor/ceiling effects \18 % and supported unidimensionality of both the NS



and DS scales. While SCOPA-SLEEP has been considered generically applicable [4], its psychometric performance in such circumstances appears undocumented. We translated the SCOPA-SLEEP into Swedish, assessed the translation and tested its psychometric properties according to classical test theory (CTT) among people with PD. Additionally, its psychometric properties were explored among age-matched healthy controls.

Qual Life Res (2016) 25:2571–2577

mentation, ADL, parkinsonian motor symptoms and complications of therapy, respectively) [9] during the ‘‘on’’phase (i.e. periods of good drug response and minimal PDrelated disability). Part II of the UPDRS (ADL) was also assessed for the ‘‘off’’-phase (i.e. periods of poor drug response and increased PD-related disability) from medical history [9]. Sample characteristics are reported in Table 1. One person with PD was on wake-promoting medication. Analyses

Methods The study was conducted in accordance with the Declaration of Helsinki and was approved by the local research ethics committee. All participants provided written informed consent. Translation and field testing of the Swedish SCOPASLEEP SCOPA-SLEEP was translated into Swedish using the dual-panel method [7]. It was first translated from English into Swedish by a bilingual panel of five individuals fluent in both languages, who together produced a consensus translation. A second panel (six lay people) reviewed the translation to ensure it was expressed in natural, everyday language. The Swedish SCOPA-SLEEP was then field-tested regarding comprehensibility, relevance and respondent burden using a convenience sample of 20 individuals (15 men) with a mean (standard deviation [SD]) age of 67.5 (6.4) years, diagnosed with PD since 9 (5.1) years. A passive observer noted the time taken to complete the questionnaire and then interviewed participants regarding their perceptions of the SCOPA-SLEEP by asking whether they found instructions and items easy to understand, and items easy to answer and relevant. Psychometric properties of the Swedish SCOPASLEEP The Swedish SCOPA-SLEEP was psychometrically tested using baseline data from 149 people with PD [8] and 53 age-matched healthy controls from the prospective Swedish BioFINDER study ( Data were collected by means of self-report and clinical examination by a movement disorders specialized neurologist and included medical history, the Schwab & England (SE) activities of daily living (ADL) scale [9], the Hospital Anxiety and Depression Scale (HADS) [10] and the Swedish SCOPASLEEP. People with PD were also assessed according to the Hoehn and Yahr (HY) staging of PD [11] and the Unified PD Rating Scale (UPDRS; parts I-IV, representing


Data were analysed using R version 3.2.2 (‘‘psych’’ package version 1.5.8; and FACTOR version 10.3.01 ( [12]. Field-test questions were analysed descriptively, and the time taken to complete the SCOPA-SLEEP was used as an indicator of respondent burden. Psychometric analyses were conducted according to CTT [13–16]. Data completeness was assessed by the percentages of missing item responses and computable total scores. Scaling assumptions (the legitimacy of summing item scores into a total score) were tested by corrected item–total correlations, which should support that items within each scale contribute substantially to the total score and represent a common variable (unidimensionality). Unidimensionality was also tested by exploratory factor analyses (EFA) for each scale based on polychoric correlations and parallel analysis to determine the number of factors [17]. EFA was not conducted on controls due to the small sample size. Targeting, i.e. how well respondents’ scores accord with scale coverage, was assessed by the distributional properties of scale scores. A well-targeted scale should have an average total score close to the scale midpoint and span most of its range, without excess skewness or floor/ceiling effects, i.e. percentages of participants with the lowest (floor) and highest (ceiling) possible scores. Reliability was assessed by coefficient alpha, and the standard error of measurement (SEM) was calculated (SD 9 H1 - alpha) to estimate score precision. SEM was expressed in the units of the respective scales, and as a percentage of the maximum possible scale scores (to facilitate comparison between scales). Construct validity was assessed by Spearman correlations between NS and DS scores and PD duration, HY, UPDRS (parts II–IV), HADS, SE and overall sleep quality. Based on previous experiences [4, 5], weak correlations (up to about 0.2) were expected with PD duration, HY, SE, UPDRS and HADS scores, but somewhat stronger correlations were expected with HADS. NS was also expected to correlate strongly ([0.60) and DS weakly with overall sleep quality. Furthermore, people with PD were expected to exhibit higher (worse) scores than controls [4].

Qual Life Res (2016) 25:2571–2577


Table 1 Sample characteristics

Age (years), mean (SD)

People with PD (n = 149)

Controls (n = 53)

64.9 (10.8)

65.3 (8.4)

Men/women, n (%)

90 (60)/59 (40)

p value 0.786



23 (43)/30 (57)


Married or living as married/single, n (%)

114 (77)/34 (23)

51 (96)/2 (4)


Years of education, mean (SD)

11.5 (4.9)

13 (3.3)


23 (43)/30 (57)


48 (92)


Working/not working, n (%)

44 (30)/104 (70)

Use of sleeping tablets last month, n (%)


Not at all

112 (78)

Less than once per week

5 (3.5)

2 (4)

Once or twice per week

4 (2.5)

1 (2)

More than three times per week Experienced sleep disturbance, n (%)f Overall sleep quality, median (q1–q3)e HADS anxiety, median (q1–q3)g HADS depression, median (q1–q3)


22 (15)

1 (2)

67 (46) 2 (1–3)

15 (28) 1 (0–2)

0.028b 0.002d

4 (1.5–7)

1 (0–5)


3 (1–6)

1 (0–2)


100 (100–100)


Schwab & England (ADL dependency), median (q1–q3)h

90 (90–95)

Duration of PD (years), mean (SD)

3.7 (5.0)


Hoehn and Yahr stage of PD, median (q1–q3)i

2 (1–2.5)


UPDRS I (mentation), median (q1–q3)j

1 (0–2)


UPDRS II (ADL) in ‘‘on’’, median (q1–q3)k

6 (3.5–10)


UPDRS II (ADL) in ‘‘off’’, median (q1–q3)


UPDRS III (motor score), median (q1–q3)l UPDRS IV (complications of therapy), median (q1–q3)m a b c

10 (5–14)


14 (8–22)


4 (1–5)


Independent samples t test Pearson’s Chi-square test Fisher’s exact test


Mann–Whitney U test


Single item included in the SCOPA-SLEEP; possible score range (overall sleep quality), 0–6 (higher scores = worse)


Single dichotomous item (yes/no) from the UPDRS IV


Possible score range, 0–21 (higher scores = worse)


Possible score range, 0–100 (higher scores = better)


Possible range, I–V (I, mild unilateral disease; II, bilateral disease without postural impairment; III, bilateral disease with postural impairment, moderate disability; IV, severe disability, still able to walk and stand unassisted; V, confined to bed or wheelchair unless aided)


Possible score range, 0–16 (higher scores = worse)


Possible score range, 0–52 (higher scores = worse)


Possible score range, 0–108 (higher scores = worse) Possible score range, 0–23 (higher scores = worse)


PD Parkinson’s disease, SD standard deviation, q1–q3 1st–3rd quartile (25th–75th percentile), HADS Hospital Anxiety and Depression Scale, UPDRS Unified Parkinson’s Disease Rating Scale, ADL Activities of Daily Living, NA not applicable

Results Field testing of the Swedish SCOPA-SLEEP Respondents completed the SCOPA-SLEEP in a mean of 3.5 (SD, 1; min–max, 2–5.5) minutes. The SCOPA-SLEEP was considered easy to understand (n = 19), answer (n = 16) and to be relevant (n = 16). Perceived lack of relevance was due to not experiencing item contents. Some participants commented that response categories were too

few (n = 2; DS scale) or too many (n = 1; overall sleep quality). Since these comments did not concern the translation, no modifications were made (see ‘‘Appendix’’ for the Swedish SCOPA-SLEEP). Psychometric properties of the Swedish SCOPASLEEP Table 2 summarizes the psychometric properties of the SCOPA-SLEEP. Data completeness was good, and scaling



Qual Life Res (2016) 25:2571–2577

Table 2 Data completeness, scaling assumptions, targeting, reliability and construct validity of the SCOPA-SLEEP among people with PD and healthy control subjectsa SCOPA-SLEEP NS scale People with PD

SCOPA-SLEEP DS scale Controls

People with PD


Data completeness Missing item responses (min–max %)b





Computable scale scores (%)









F1 loadings (min–max)d





F1/F2 % common variance explained









Scaling assumptions Corrected item–total correlation (min–max)c Item EFA (MRFA):


F1/F2 % common variance explained from PA Targeting Possible score range (midpoint)

0–15 (7.5)

0–15 (7.5)

0–18 (9.5)

0–18 (9.5)

Mean (SD) scoref

4.7 (3.3)

3.2 (2.5)g

3.8 (3.2)

2.2 (1.6)g

Median (q1–q3) scoref

4 (2–7)

3 (1–5)h

3 (2–5)

2 (1–3)h

Min–max scorei Floor/ceiling effects (%)j

0–15 3/1

0–10 17/0

0–18 13/1

0–8 15/0






Coefficient al (95 % CI)m

0.85 (0.80–0.90)

0.86 (0.74–0.91)

0.87 (0.83–0.91)

0.65 (0.43–0.77)

Coefficient a when item deleted (min–max)n





SEM (% of total score)p

1.28 (8.5)

0.94 (6.2)

1.15 (6.4)

0.95 (5.3)

Overall sleep quality





HADS anxiety





HADS depression





Schwab & England (ADL dependency)





Duration of PD (years)





Hoehn and Yahr stage of PD





UPDRS II (ADL) in ‘‘on’’





UPDRS II (ADL) in ‘‘off’’






Construct validityq


Qual Life Res (2016) 25:2571–2577


Table 2 continued SCOPA-SLEEP NS scale


People with PD


UPDRS III (motor score)





UPDRS IV (complications of therapy)






For interpretation criteria, please see references in the Methods section


Should be \10 %

People with PD



Should be [0.3 to support summation of raw item scores (i.e. each item contribute substantially to the total score), and [0.4 to support that items represent one common variable (i.e. unidimensionality) d

Kaiser–Meyer–Olkin measure of sampling adequacy, 0.81 and 0.85 for the NS and DS scales, respectively; Bartlett’s test, p \ 0.001 for both the NS and DS scales


Parallel analysis based on the 95th percentile of each factor’s percentage of common variance explained from permutation of 500 random correlation matrices


Should be close to scale midpoint


p \ 0.001 (independent samples t test), PD versus controls


p = 0.002 (Mann–Whitney U test), PD versus controls


Should span most of the scale’s score range


Should be \15–20 %


Should be between -1 and ?1


Should be C0.70 and preferably C0.80 Based on bootstrapping with 1000 iterations

m n

Should not increase


Increased when deleting item D6 (range for other items, 0.52–0.64)


SEM = SD 9 H1 - alpha


Spearman correlations

NS night-time sleep, DS daytime sleepiness, SD standard deviation, EFA exploratory factor analysis, MRFA minimum rank factor analysis, F factor, PA parallel analysis, q1–q3 1st–3rd quartile (25th–75th percentile), CI confidence interval, SEM standard error of measurement, PD Parkinson’s disease, HADS Hospital Anxiety and Depression Scale, UPDRS Unified Parkinson’s Disease Rating Scale, ADL Activities of Daily Living, NA not applicable

assumptions were supported except for DS item D6 among controls (corrected item–total correlation, 0.08; range among other DS items, 0.47–0.74). Targeting was generally satisfactory, although people tended to score at the lower quartile of the DS scale. Reliability was C0.85 except for the DS scale among controls (0.65). Construct validity was supported by correlations in general accordance with expectations and expected differences in scores between people with PD and controls.

Discussion The Swedish SCOPA-SLEEP was generally well perceived by people with PD and not associated with notable respondent burden. Furthermore, our psychometric observations largely concur with previous evaluations of the Dutch, Spanish and Thai SCOPA-SLEEP in PD [4–6], thus providing support for the Swedish SCOPA-SLEEP. However, we found less convincing evidence among healthy

controls, particularly for the DS scale. One source of this appeared to be item D6, which inquires whether respondents experience falling asleep during the day as a problem. Daytime sleepiness is less common and severe among healthy controls than in PD and therefore probably less of a problem for this group [1, 2]. Furthermore, item D6 is the only DS item that does not concern factual recall but perceived problems. This could suggest that it may not be conceptually coherent with the scale as a whole. However, we found no D6 anomalies in the PD sample. Therefore, and given the limited control sample, we do not take these observations as evidence against the appropriateness of SCOPA-SLEEP beyond people with PD but rather as a prompt for further studies. The study design did not allow for comparisons between SCOPA-SLEEP scales and other scales such as the Pittsburgh Sleep Quality Index and Epworth Sleepiness Scale, or for assessing test–retest reliability. Furthermore, our PD sample represented people in relatively early stages of PD, also compared to participants in previous SCOPA-SLEEP



evaluations [4–6]. It is thus not representative of the PD population at large or strictly comparable to previous studies. While this calls for some caution and probably explains the somewhat skewed score distributions, our observations are nevertheless in agreement with previous data. This suggests that the SCOPA-SLEEP appears appropriate for use in early PD, which is important since sleep disorders occur throughout the disease trajectory and useful scales need to be applicable in all stages of PD [3]. However, future studies of the Swedish SCOPA-SLEEP should consider more representative samples. Furthermore, to better understand the SCOPA-SLEEP it should be subjected not only to CTT but also to modern test theory analyses such as Rasch Measurement Theory, which provides detailed insights into the functioning of rating scales [13]. In conclusion, this study provides initial CTT-based evidence in support for the psychometric appropriateness of the Swedish SCOPA-SLEEP. This is the only documented PD targeted sleep assessment scale in Swedish so far. Further studies are needed to understand its

Qual Life Res (2016) 25:2571–2577

measurement properties in better detail, including its generic properties. Acknowledgments The authors wish to thank the participants, translators and lay people for their cooperation, and S. Hall, Y. Surova, K. Johansson, J. Reimer and A. Ho¨glund for data collection. The study was supported by the European Research Council, the Swedish Research Council, the Swedish Parkinson Foundation, the Swedish Parkinson Academy, the Basal Ganglia Disorders Linnaeus Consortium (BAGADILICO) at Lund University, the Swedish federal government under the ALF agreement, and Kristianstad University, Kristianstad, Sweden. The funding sources had no role in the design and conduct of the study; in the collection, analysis, interpretation of the data; or in the preparation, review, or approval of the manuscript. Compliance with ethical standards Conflict of interest None.

Appendix See Table 3.

Table 3 Item and response category wording of the SCOPA-SLEEP, English and Swedish versions English version

Swedish version

A. Use of sleeping tablets Response categories not at all – less than once a week – once or twice a week – more than 3 times a week A1. How often did you use sleeping tablets in the last months? (prescribed by a physician or not) A2. Which sleeping tablets did you use in the last month? [name; amount per month; dose per tablet]

Svarskategorier inte alls – mindre a¨n en ga˚ng i veckan – en eller tva˚ ga˚nger i veckan – mer a¨n tre ga˚nger i veckan A1. Hur ofta har du tagit so¨mntabletter under den senaste ma˚naden? (ba˚de receptbelagda och icke-receptbelagda) A2. Vilka so¨mntabletter har du tagit under den senaste ma˚naden? [namn; antal per ma˚nad; dos per tablett]

B. NS: Night-time sleep problems Response categories not at all – a little – quite a bit – a lot B1. In the past month, have you had trouble falling asleep when you went to bed at night? B2. In the past month, to what extent do you feel that you have woken too often? B3. In the past month, to what extent do you feel that you have been lying awake for too long at night? B4. In the past month, to what extent do you feel that you have woken up too early in the morning? B5. In the past month, to what extent do you feel you have had too little sleep at night?

Svarskategorier inte alls – i liten utstra¨ckning – i ganska stor utstra¨ckning – i stor utstra¨ckning B1. I vilken utstra¨ckning har du haft sva˚rt att somna na¨r du ga˚tt och lagt dig under den senaste ma˚naden? B2. I vilken utstra¨ckning har du vaknat ofta nattetid under den senaste ma˚naden? B3. I vilken utstra¨ckning har du legat vaken fo¨r la¨nge om natten under den senaste ma˚naden? B4. I vilken utstra¨ckning har du vaknat fo¨r tidigt under den senaste ma˚naden? B5. I vilken utstra¨ckning har du ka¨nt att du fa˚tt fo¨r lite nattso¨mn under den senaste ma˚naden?

C. Overall sleep quality Response categories very well – well – rather well – not well but not badly – rather badly – badly – very badly C1. Overall, how well have you slept at night during the past month?

Svarskategorier mycket bra – bra – ganska bra – varken bra eller da˚lig – ganska da˚lig – da˚lig – mycket da˚lig C1. Hur har din nattso¨mn varit under den senaste ma˚naden?

D. DS: Daytime sleepiness Response categories never – sometimes – regularly – often D1. How often in the past month have you fallen asleep unexpectedly either during the day or in the evening?


Svarskategorier aldrig – ibland – ganska ofta – ofta D1. Hur ofta under den senaste ma˚naden har du plo¨tsligt somnat under dag- eller kva¨llstid?

Qual Life Res (2016) 25:2571–2577


Table 3 continued English version

Swedish version

D2. How often in the past month have you fallen asleep while sitting peacefully?

D2. Hur ofta under den senaste ma˚naden har du somnat na¨r du suttit i lugn och ro? D3. Hur ofta under den senaste ma˚naden har du somnat na¨r du tittat pa˚ TV eller la¨st? D4. Hur ofta under den senaste ma˚naden har du somnat na¨r du pratat med na˚gon? D5. Hur ofta under den senaste ma˚naden har du haft problem att ha˚lla dig vaken under dag- eller kva¨llstid? D6. Har du under den senaste ma˚naden upplevt att det varit ett problem att du har somnat under dagen?

D3. How often in the past month have you fallen asleep while watching TV or reading? D4. How often in the past month have you fallen asleep while talking to someone? D5. In the past month, have you had trouble staying awake during the day or in the evening? D6. In the past month, have you experienced falling asleep during the day as a problem?

The full, formatted scale (including respondent instructions) is available from the corresponding author

References 1. Schrempf, W., Brandt, M. D., Storch, A., & Reichmann, H. (2014). Sleep disorders in Parkinson’s disease. Journal of Parkinson’s Disease, 4(2), 211–221. 2. Videnovic, A., & Golombek, D. (2013). Circadian and sleep disorders in Parkinson’s disease. Experimental Neurology, 243, 45–56. 3. Hogl, B., Arnulf, I., Comella, C., Ferreira, J., Iranzo, A., Tilley, B., et al. (2010). Scales to assess sleep impairment in Parkinson’s disease: Critique and recommendations. Movement Disorders, 25(16), 2704–2716. 4. Marinus, J., Visser, M., van Hilten, J. J., Lammers, G. J., & Stiggelbout, A. M. (2003). Assessment of sleep and sleepiness in Parkinson disease. Sleep, 26(8), 1049–1054. 5. Martinez-Martin, P., Visser, M., Rodriguez-Blazquez, C., Marinus, J., Chaudhuri, K. R., van Hilten, J. J., et al. (2008). SCOPAsleep and PDSS: Two scales for assessment of sleep disorder in Parkinson’s disease. Movement Disorders, 23(12), 1681–1688. 6. Setthawatcharawanich, S., Limapichat, K., Sathirapanya, P., & Phabphal, K. (2011). Validation of the Thai SCOPA-sleep scale for assessment of sleep and sleepiness in patients with Parkinson’s disease. Journal of the Medical Association of Thailand, 94(2), 179–184. 7. Swaine-Verdier, A., Doward, L. C., Hagell, P., Thorsen, H., & McKenna, S. P. (2004). Adapting quality of life instruments. Value Health, 7(Suppl 1), S27–S30. 8. Gelb, D. J., Oliver, E., & Gilman, S. (1999). Diagnostic criteria for Parkinson disease. Archives of Neurology, 56(1), 33–39.

9. Fahn, S., Elton, R. L., & Members of the UPDRS Development Committee. (1987). Unified Parkinson’s Disease Rating Scale. In S. Fahn, C. D. Marsden, D. B. Calne, & M. Goldstein (Eds.), Recent developments in Parkinson’s Disease (Vol. 2, pp. 153–163). Florham Park: MacMillan Healthcare Information. 10. Zigmond, A. S., & Snaith, R. P. (1983). The hospital anxiety and depression scale. Acta Psychiatrica Scandinavica, 67(6), 361–370. 11. Hoehn, M. M., & Yahr, M. D. (1967). Parkinsonism: Onset, progression and mortality. Neurology, 17(5), 427–442. 12. Lorenzo-Seva, U., & Ferrando, P. J. (2006). FACTOR: A computer program to fit the exploratory factor analysis model. Behavior Research Methods, 38(1), 88–91. 13. Hobart, J., & Cano, S. (2009). Improving the evaluation of therapeutic interventions in multiple sclerosis: The role of new psychometric methods. Health Technol Assess, 13(12), iii, ix-x, 1–177. 14. Ware, J. E, Jr, & Gandek, B. (1998). Methods for testing data quality, scaling assumptions, and reliability: The IQOLA Project approach. International Quality of Life Assessment. Journal of Clinical Epidemiology, 51(11), 945–952. 15. Saris-Baglama, R. N., Dewey, C. J., Chisholm, G. B., Kosinski, M., Bjorner, J. B., & Ware, J. E, Jr. (2004). SF health outcomesTM scoring software user’s guide. Lincoln: QualityMetric Inc. 16. Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw-Hill Inc. 17. Timmerman, M. E., & Lorenzo-Seva, U. (2011). Dimensionality assessment of ordered polytomous items with parallel analysis. Psychological Methods, 16(2), 209–220.


Suggest Documents