Non-iterative Conditional Pairwise Estimation for the Rating Scale Model

We investigate two non-iterative estimation procedures for Rasch models, the pair-wise estimation procedure (PAIR) and the Eigenvector method (EVM), and identify theoretical issues with EVM for rating scale model (RSM) threshold estimation. We develop a new procedure to resolve these issues—the conditional pairwise adjacent thresholds procedure (CPAT)—and test the methods using a large number of simulated datasets to compare the estimates against known generating parameters. We find support for our hypotheses, in particular that EVM threshold estimates suffer from theoretical issues which lead to biased estimates and that CPAT represents a means of resolving these issues. These findings are both statistically significant ( p < .001) and of a large effect size. We conclude that CPAT deserves serious consideration as a conditional, computationally efficient approach to Rasch parameter estimation for the RSM. CPAT has particular potential for use in contexts where computational load may be an issue, such as systems with multiple online algorithms and large test banks with sparse data designs.

Cross-cultural adaptation and psychometric properties of the Kessler Distress Scale (K10): an application of the rating scale model

Background The Kessler Distress Scale (K10) is a self-report scale for the assessment of non-specific psychological distress in the general and clinical population. Because of its ease of application and good psychometric properties, the K10 has been adapted to several cultures. The present study seeks to adapt the K10 to Brazilian Portuguese and estimate its validity evidence and reliability. Methods A total of 1914 individuals from the general population participated in the study (age = 34.88, SD = 13.61, 77.7% female). The adjustment indices were compared among three different measurement models proposed for the K10 through confirmatory factor analysis (CFA). The items’ properties were analyzed by Andrich’s Rating Scale Model (RSM). Furthermore, evidence based on relations to other variables (depression, stress, anxiety, positive and negative affects, and satisfaction with life) was estimated. Results CFA indicated the adequacy of the bifactor model (CFI= 0.985; TLI= 0.973; SMR= 0.019; RMSEA= 0.050), composed of two specific factors (depression and anxiety) and one general factor (psychological distress), corresponding to the theoretical hypothesis. Additionally, it was observed multiple-group invariance by gender and age range. The RSM provided an understanding of the organization of the continuum represented by the psychological distress construct (items difficulty), which varied from −0.89 to 1.00; good adjustment indexes; infit between 0.67 and 1.32; outfit between 0.68 and 1.34; and desirable reliability, α= 0.87. Lastly, theoretically coherent associations with the external variables were observed. Conclusions It is concluded that the Brazilian version of the K10 is a suitable measure of psychological distress for the Brazilian population.

Analisis Rasch model Indonesia Problematic Internet Use Scale (IPIUS)

Indonesia Problematic Internet Use Scale (IPIUS) consists of six dimensions, namely a preference for online social interaction, escaping, negative outcomes, compulsive internet use, cognitive preoccupation, and emotional reactivity. University students in the covid-19 pandemic are constantly learning to use the internet that causes compulsive internet use. This study explores IPIUS using analysis technique rating scale model, especially the compulsive internet use dimension which consists of ten items. The quantitative research method is survey research using an online questionnaire. Participants in the research were 395 university students who do lectures online. The sampling technique used to obtain participants was convenience sampling. The results showed nine-item compulsive internet use dimension is good quality, unidimensional, and item reliability was better than previous research using confirmatory factor analysis. IPIUS is a stable instrument to measure problematic internet use for university students in Indonesia. Keywords: compulsive internet use; covid-19; problematic internet use; rating scale model, university students Abstrak Indonesia Problematic Internet Use Scale (IPIUS) terdiri dari enam dimensi, yaitu: preference for online social interaction, escaping, negative outcome, compulsive internet use, cognitive preoccupation, dan emotional reactivity. Mahasiswa di era pandemic covid-19 secara terus-menerus melakukan pembelajaran daring yang mengakibatkan menjadi compulsive internet use. Penelitian ini bertujuan mengekplorasi IPIUS dengan menggunakan teknik analisis rating scale model, khususnya dimensi compulsive internet use yang terdiri dari sepuluh item. Metode penelitian kuantitatif yang digunakan berjenis survey research menggunakan kuesioner online. Partisipan penelitian sebanyak 395 mahasiswa yang melakukan kuliah secara daring. Teknik sampling yang digunakan untuk mendapatkan partisipan adalah teknik convenience sampling. Hasil penelitian menunjukan sembilan item pada dimensi compulsive internet use merupakan item yang berkualitas baik, bersifat unidimensi, dan nilai reliabilitas item yang dihasilkan lebih baik dibandingkan penelitian sebelumnya yang menggunakan confirmatory factor analysis. IPIUS merupakan instrumen yang cocok mengukur problematic internet use pada mahasiswa di Indonesia. Kata kunci: compulsive internet use; covid-19; problematic internet use; rating scale model mahasiswa

Rasch Analysis of an Indonesian version of the General Self-Efficacy Scale-12: A Comparison of Rating Scale Model (RSM) and Partial Credit Model (PCM)

Self-efficacy is one of the most researched topics in education and psychology, which grounded within social cognitive theory. Various instruments are available that can be used to measure general self-efficacy. As a foundational instrument in the measurement of general self-efficacy, the study investigated the psychometric properties of the General Self-Efficacy Scale-12 (GSES-12). Although the instrument has been adapted into Bahasa, the contribution of the Rasch measurement model in validating this instrument has never done before. By utilizing various advantages of the Rasch model, the polytomous Rasch model (RSM and PCM) were used. The data used in this study are secondary data from previous research. The sample consisted of 303 students (132 male, 171 female) with an age range 18 - 22 years old (mean age = 19.56, SD = 1.20). The results of Rasch's analysis show that the psychometric characteristics of GSES-12 are adequate. The data fit to the Partial Credit Model compared to Rating Scale Model. All assumptions in applying the Rasch model are fulfilled, and also the item fit to the model. Implications and suggestions for future research are also discussed.

The Development of Chinese Version of Transcultural Nursing Self-Efficiency Scale: Using Rasch Model Analysis

Introduction: With the huge advances of globalization and urbanization, the nursing services in China are meeting the challenge of varied culture-specific needs in practices. Tools to measure transcultural sensitivity thus has great value in promoting transcultural nursing practices. Method: The Chinese version of Transcultural Nursing Self-Efficacy Scale (TSET-CV) was translated and validated with a sample of 2,346 nurses in China. Utilizing the analytical framework of the rating scale model, each item of the three subscales, namely the cognitive, practical, and effect subscale, was analyzed in terms of item analysis, item fitting, difficulty, reliability, item function, and validity. Results: The analysis suggested that 5-point Likert-type scale was more suitable than the 10-point Likert-type scale originally used. Item 27 and Item 79 showed deficient properties, which was deleted in the TSET-CV. The most accurate ranges measured by the TSET-CV matched the range of the distribution of the subjects’ ability. Discussion: Unlike the classical test theory, the rating scale model based on the Rasch’s model can provide sample-independent psychometric properties of items. The revised TSET-CV has great potential in diagnosing deficiency among nursing student in transcultural self-efficiency, which can promote the development of corresponding education strategies.

Defining statistical measurement levels of Intuitive Eating with the Rasch Rating Scale Model

Intuitive eating is an intrinsic ability to moderate the amount and type of food, promoting a healthy diet and self-regulation of one's own weight. Different instruments (IES-H, IES-T, IES-2) have been developed to assess intuitive eating between different population groups from different countries. The construct validity and invariance of the 23-item Intuitive Eating Scale-2 (IES-2) has been widely validated by Confirmatory Factor Analysis and Exploratory Factor Analysis. However, these Classical Test Theory (CTT) methods have not always confirmed the same 4-factor structure.Rasch analysis, a Modern Test Theory method (MTT), has been used as an alternative approach to examine the psychometric properties of various health and medical instrument (HADS, KIDSCREEN-52, LANSS, MHRM). One significant difference between CTT and MT is the method of calculating a composite score. A CTT total sore is based on the summation of raw categorical scores, whereas these raw categorical scores are converted to interval-scaled measures into a Rasch composite score.Data was collected from 625 respondents was fitted to the Rasch Rating Scale Model. The data fitted the model adequately, as less than 5% and 1% of absolute standardised residuals were found to be ≥ 2 and ≥ 3, respectively. A Principle Component Analysis of Rasch residuals (PCAR) was used to determine the unidimensionality of the IES-2 and its subscales, after checking and adjusting for lack of item fit and proper rating scale functioning. PCAR indicated that all 23 items could not function as a single total unidimensional Rasch measure. However, the same item structure for the 4 subscales, originally proposed by Tylka and Kroon Van Diest, was confirmed by PCAR. The relationship between the respondents and the IES-2 items in each subscale could be explained using a Wright map, allowing both to be represented on the same logit scale.Statistical different levels of intuitive eating were determined for each subscale from a table representing the relationship between the lowest and highest possible raw scores and their Rasch measures. Wright maps showed the position of respondents between cut-off lines indicating different statistical levels along each unidimensional subscale of intuitive eating. The majority of respondents were classified into 2 out of 4 levels in the subscale Unconditional Permission to Eat, 3 out of 5 levels in Eating for Physical Rather Than Emotional Reasons, 3 out of 5 levels in Reliance on Internal Hunger and Satiety Cues and 2 out of 4 levels in Body-Food Choice Congruence.