Mental Health Indicators for Under-Represented Population, "Adult Male Survivors of Childhood Sexual Abuse": Psychometric Properties for Short Forms of the Depression Anxiety Stress Scales (DASS)

Background/Rationale

The Depression Anxiety Stress Scales (DASS; Lovibond & Lovibond; 1995) is one of the most widely used self-reported measures that assess clinical symptoms of depression, anxiety, and stress. Since the introduction of the original 42-item DASS, a few short forms were further identified: DASS-21 (Lovibond & Lovibond, 1995), DASS-14 (Wise, Harris, & Oliver, 2017), DASS-12 (Yusoff, 2013), and DASS-9 (Kyriazos, 2018). Using shortened measures in practice has many advantages because of higher response rates, better data quality, and the opportunity to measure additional attributes without overtaxing participants (Allen, 2016). However, compared to DASS-21, the other three short forms—DASS-14, DASS-12, and DASS-9—are relatively new and have not been copiously tested regarding their reliability and factor structure. Particularly, the literature includes no study testing psychometric properties of the DASS short forms in a clinical sample, such as male survivors of sexual abuse who suffer mental health problems. Therefore, this study aims to test the reliability and factor structure of the DASS short forms by using exploratory/confirmatory factor analyses that provide health care professionals with practical and informed guidance for selecting valid and reliable alternatives in their practice.

Methods/Methodology

The current study employed secondary data analysis. The data were collected by the Sexual Assault Crisis Centre of Essex County (SACC), which leads 18 government-funded male survivor programs in 12 county regions located in western Ontario, Canada from January 5^th, 2011 to April 12^th, 2018. A total of 463 adult male survivors of childhood sexual abuse participated in the research. Using exploratory factor analysis (EFA) and confirmatory factor analysis (CFA), the authors tested the factor structure of the four short forms (DASS-21, DASS-14, DASS-12, and DASS-9) of the original DASS scale. SPSS-24 and AMOS-24 were used for data analyses.

Results

Communalities, internal consistency reliability, scree plots, and factor loadings for EFA with Oblimin (Delta = 0) show that the factor structure of the four short forms is relevant. However, the analyses also produced another short form of DASS (16 items). Therefore, CFA was conducted for the five short forms (DASS-21, 16, 14, 12, and 9). The model fit indexes after modifications show the five short forms are a good fit: RMSEA (0.027-0.044), RMR (0.027-0.039), GFI (0.941-0.986), CFI (0.974-0.995), and IFI (0.974-0.995). With respect to internal consistency reliability, a wide range of internal consistency reliability (questionable to excellent) were reported: DASS-21/16 (0.81-0.93), DASS-14 (0.73-0.92), DASS-12 (0.77-0.86), and DASS-9 (0.68-0.81).

Conclusions and Implications

CFA suggests that all five short forms have a good fit and can be used as an alternative to the original DASS-42 scale. The reliability analyses also indicate that all the short forms exceed the minimum acceptable level for internal consistency. Given that service providers often develop treatment strategies based on a limited understanding and a lack of evidence-based knowledge of valid and reliable measurements, the proposed short forms, due to their reduced number of items when compared to the original 42-item DASS, are potentially easier to administer and thus more efficient data collection instruments than the original full-length scale.