Violent acts on university campuses are becoming more frequent. Enrollment rates of Latinos at universities is increasing. Research has indicated that youths are more susceptible to trauma, particularly Latinos. Thus, it is imperative to evaluate the validity of commonly used posttraumatic stress measures among Latino college students. The Impact of Event Scale-Revised (IES-R) is one of the most commonly used metrics of posttraumatic stress disorder symptomatology. However, it is largely unknown if the IES-R is measuring the same construct across different sub-samples (e.g. Latino versus non-Latino). The current study aimed to assess measurement invariance for the IES-R between Latino and non-Latino participants. A total of 545 participants completed the IES-R. One- and three-factor scoring solutions were compared using confirmatory factor analyses. Measurement invariance was then evaluated by estimating several multiple-group confirmatory factor analytic models. Four models with an increasing degree of invariance across groups were compared. A significant χ2 difference test was used to indicate a significant change in model fit between nested models within the measurement invariance testing process. The three-factor scoring solution could not be used for the measurement invariance process because the subscale correlations were too high for estimation (rs 0.92–1.00). Therefore, the one-factor model was used for the invariance testing process. Invariance was met for each level of invariance: configural, metric, scalar, and strict. All measurement invariance testing results indicated that the one-factor solution for the IES-R was equivalent for the Latino and non-Latino participants.

Objective: The stress response is a dynamic process that can be characterized by predictable biochemical and psychological changes. Biomarkers of the stress response are typically measured over time and require statistical methods that can model change over time. One flexible method of evaluating change over time is the latent growth curve model (LGCM). However, stress researchers seldom use the LGCM when studying biomarkers, despite their benefits. Stress researchers may be unaware of how these methods can be useful. Therefore, the purpose of this paper is to provide an overview of LGCMs in the context of stress research. We specifically highlight the unique benefits of using these approaches.Methods: Hypothetical examples are used to describe four forms of the LGCM.Results: The following four specifications of the LGCM are described: basic LGCM, latent growth mixture model, piecewise LGCM, and LGCM for two parallel processes. The specifications of the LGCM are discussed in the context of the Trier Social Stress Test. Beyond the discussion of the four models, we present issues of modeling nonlinear patterns of change, assessing model fit, and linking specific research questions regarding biomarker research using different statistical models.Conclusions: The final sections of the paper discuss statistical software packages and more advanced modeling capabilities of LGCMs. The online Appendix contains example code with annotation from two statistical programs for the LCGM.

The multidimensional graded response model, an item response theory (IRT) model, can be used to improve the assessment of surveys, even when sample sizes are restricted. Typically, health-based survey development utilizes classical statistical techniques (e.g. reliability and factor analysis). In a review of four prominent journals within the field of Health Psychology, we found that IRT-based models were used in less than 10% of the studies examining scale development or assessment. However, implementing IRT-based methods can provide more details about individual survey items, which is useful when determining the final item content of surveys. An example using a quality of life survey for Cushing's syndrome (CushingQoL) highlights the main components for implementing the multidimensional graded response model. Patients with Cushing's syndrome (n = 397) completed the CushingQoL. Results from the multidimensional graded response model supported a 2-subscale scoring process for the survey. All items were deemed as worthy contributors to the survey. The graded response model can accommodate unidimensional or multidimensional scales, be used with relatively lower sample sizes, and is implemented in free software (example code provided in online Appendix). Use of this model can help to improve the quality of health-based scales being developed within the Health Sciences.

Context: Patients in long-term remission of Cushing's syndrome (CS) commonly report impaired quality of life (QoL). The CushingQoL questionnaire is a disease-specific QoL questionnaire for patients diagnosed with CS. The developers of the CushingQoL recommend using a global (total) score to assess QoL. However, the global score does not capture all aspects of QoL as outlined by the World Health Organization (WHO). Objective: The aim of the study was to compare the performance of different scoring options to determine the optimal method for the CushingQoL. Design and patients: Patients in remission from CS (nZ341) were recruited from the Cushing's Syndrome Research Foundation's email listserv and Facebook page, and asked to complete the CushingQoL and a short demographics survey. Results: Using an exploratory analysis, adequate model fit was obtained for the global score, as well as a 2-subscale (psychosocial issues and physical problems) scoring solution. Confirmatory methods were performed to identify the optimal scoring solution. Both the global score and the 2-subscale scoring solution showed adequate model fit. However, a c 2 difference test indicated that the 2-subscale scoring solution was a significantly better fit than the global score (P!0.05). Conclusion: If doctors or researchers would like to tease apart physical and psychosocial issues, the 2-subscale scoring solution would be recommended, since this solution showed to be optimal in scoring the CushingQoL. Regardless of the scoring solution used, the CushingQoL has proven to be a valuable resource for assessing health-related QoL in patients with CS.

Context: When working with health-related questionnaires, outlier detection is important. However, traditional methods of outlier detection (e.g., boxplots) can miss participants with “atypical” responses to the questions that otherwise have similar total (subscale) scores. In addition to detecting outliers, it can be of clinical importance to determine the reason for the outlier status or “atypical” response.Objective: The aim of the current study was to illustrate how to derive person fit statistics for outlier detection through a statistical method examining person fit with a health-based questionnaire.Design and Participants: Patients treated for Cushing's syndrome (n = 394) were recruited from the Cushing's Support and Research Foundation's (CSRF) listserv and Facebook page.Main Outcome Measure: Patients were directed to an online survey containing the CushingQoL (English version). A two-dimensional graded response model was estimated, and person fit statistics were generated using the Zh statistic.Results: Conventional outlier detections methods revealed no outliers reflecting extreme scores on the subscales of the CushingQoL. However, person fit statistics identified 18 patients with “atypical” response patterns, which would have been otherwise missed (Zh > |±2.00|).Conclusion: While the conventional methods of outlier detection indicated no outliers, person fit statistics identified several patients with “atypical” response patterns who otherwise appeared average. Person fit statistics allow researchers to delve further into the underlying problems experienced by these “atypical” patients treated for Cushing's syndrome. Annotated code is provided to aid other researchers in using this method.