Published: 9 July 2021| Version 1 | DOI: 10.17632/822kj4vvyr.1


The core hypothesis was that the insomnia severity spectrum was unidimensional and potentially separable from fatigue, depression, anxiety, excessive dozing, but implicated with disability and poor quality of life. The main instrument used for modeling the insomnia spectrum was the Pittsburgh Insomnia Rating Scale(PIRS65), having 65 items, attached. The study was of research-grade, rather pure insomnia patients, compared to healthy controls. Attached main dataset presents concurrently obtained measures. Other datasets are of processed data for the comparison between sleep logs and the corresponding item's values on the PIRS65. A separate file contains the test:retest data. The descriptive file contains lists of questionnaires and variables. Other pertinent utility files are attached. A Powerpoint file presents output from the Multilog Item Response Theory (IRT) Software as models of all the PIRS65 IRT item characteristics.


Steps to reproduce

Largely specific insomnia patients were contrasted on a similar-aged sample of controls free of insomnia or mental illness. Before concurrent measures were taken, some subjects were asked to take the PIRS65 and to take it later without any intervening therapeutic interventions. For the week before the time that the concurrent measures were taken, the subjects completed the Pittsburgh Sleep Log nighty. The instruments used for concurrent comparison to the PIRS65 were the Spielman Insomnia Symptom Questionnaire, the Pittsburgh Sleep Quality Index, the Multidimensional Fatigue Scale, the Anxiety and Depression scales of the Brief Symptom Inventory, the Epworth Sleepiness Scale, and the SF-36. Classical and Item Response Theory (IRT) methods of study were deployed with this dataset. Receiver Operator Characteristic analysis was conducted, in light of the case:control nature of the data. Regression analyses were conducted in an exploratory manner to determine if any age or sex biases may have been present. Examination of IRT item characteristics was conducted to determine which items were superior for measuring the insomnia severity spectrum efficiently.


University of Pittsburgh


Psychometric Models