Sleep has long been a fundamental concern in nursing. Nursing students during their clinical practice play a crucial role in helping patients to integrate various types of sleep interventions into their daily practice to improve their patients’ sleep quality. Because of the ease of dissemination and accessibility of sleep hygiene education, patients are taught about sleep hygiene to manage their sleep in clinical settings. All nursing students should be knowledgeable about sleep to promote patients’ sleep quality during hospitalization. Study revealed that nursing students who have more knowledge about sleep, are more positive attitudes toward sleep hygiene and are more likely to deliver sleep hygiene education to their patients. Knowledge cannot be directly observed, therefore the evidence needs to be of sufficient quality and quantity to make a sound judgement about the students’ level of sleep knowledge.
The Rasch model is a psychometric model for analyzing categorical data, such as answers to questions on a questionnaire responses. In the Rasch model, the probability of a specified response is modeled as a function of the respondent's abilities and the item difficulty. The mathematical theory underlying Rasch models is a special case of item response theory.
Therefore, the present study examined the psychometrics properties of the knowledge of sleep physiology via the Rasch models using a secondary survey dataset from Taiwan.
Methods:
The Sleep Knowledge response data that was taken from a previous survey study from Taiwan was used for analysis. This substudy was conducted in 2014 and included 151 participants aged 21-42 years, sampled from a part of previous study of nursing students’ sleep quality and sleep knowledge. Because there were currently no standardized questionnaires available to be used for measurement of student nurses’ knowledge of sleep physiology, a sleep knowledge questionnaire comprising 9 self-rated items on the Sleep Knowledge of Physiology subscale was extracted from the original 16-item of Sleep Knowledge Scale to examine the psychometric properties using item response theory. Each item offered right/wrong options. The psychometric properties of the Knowledge of Sleep Physiology scale were examined through Rasch analyses of the 151 students’ responses on the instrument.
We used two measures to determine the precision and accuracy of the Knowledge of Sleep Physiology scale. First, the standard error (SE) of measurement for each of the item difficulty was estimated, which provide an estimate of its precision. The SE of measurement for each item was calculated using the students’ responses to that particular item, by determining the difference between the true item difficulty and the estimated item difficulty. Second, the Rasch model was used to examine the data fit. This was useful for investigating how accurately the variable could be used to predict sleep knowledge. Rasch analysis was performed using Mplus8 software.
Results:
The mean age (SD) was 31.56 (5.4) years. Of the study sample (n = 151), 100% were females. The KSP score ranged from 0 to 9. The mean sum score standard deviation (SD) was 6.18 (2.57) for the entire study sample. The correct response rate for the 9-item KSP scale ranged from 0.58 to 0.82.
To examine the internal consistency of the Sleep Knowledge Scale, the responses of the 151students were analysed using Cronbach's alpha. The results showed an acceptable internal consistency (0.72) for Cronbach's alpha. The results from the Rasch analysis revealed that all items fitted the Rasch model. The level of difficulty for items were between -0.328~-1.429 logit units. From the item characteristic curve of all items, the most difficult item was about the pharmacology of sleep medicine, the easiest one was the item about sleep waves.
Conclusion:
Educational tests have been commonly used to measure nursing students' knowledge levels. To ensure measurement quality, it is fundamental to provide evidence of content validity that ensures. Whether the content of the test is congruent with testing purposes researchers heavily rely on the judgment of content validity. However, the drawback of the classic test theory is that it prevents the test content from satisfactory results due to the complication of testing situations, such as latent trait of the test-takers and the difficulty of the test items .
The major superiority of Rasch measurement is that it diagnoses noise in test data and converts ordinal/ dichotomous item response or test raw score into a linear measure such that subsequent parametric statistical analysis becomes feasible and inter-person difference can be quantified. Although Rasch model has not been commonly recognized in nurse education, recent decades have witnessed the blooming of Rasch measurement. In this paper, we concluded that the use of the Rasch model offers more confidence in measuring students’ knowledge of sleep physiology, but minor item modifications should be considered.