Multiple linear regression of DASS-42 data: Exploring potential predictors


Exploring the DASS-42 data and personality traits

A web-based Shiny application had been previously developed to explore the relationships between primary scales in the Depression Anxiety Stress Scales 42-item self-report instrument (DASS-42; Lovibond & Lovibond, 1996). The data set contained over 30,000 online responses to the DASS-42 representing a substantial sample of psychometric data to further analyse. However, the web application examined bivariate correlations only (e.g. relationship between depression and anxiety) and did not incorporate other important variables such as demographic variables and level of education.

We sought to examine how the DASS-42 depression and anxiety scores differ across available responses to the personality traits. Fortunately, Open Psychometrics administered the online DASS-42 with items from the Ten Item Personality Inventory (???) which is a brief assessment of the Big Five personality dimensions across. The ten items are listed below:

I see myself as:

  1. _____ Extraverted, enthusiastic.

  2. _____ Critical, quarrelsome.

  3. _____ Dependable, self-disciplined.

  4. _____ Anxious, easily upset.

  5. _____ Open to new experiences, complex.

  6. _____ Reserved, quiet.

  7. _____ Sympathetic, warm.

  8. _____ Disorganized, careless.

  9. _____ Calm, emotionally stable.

  10. _____ Conventional, uncreative.

These items were responded on a 1-7 Likert scale, i.e. 1=Disagree strongly and 7=Agree strongly.

A quick glance at the relationships between the scales are presented below. The exploratory results indicate that the depression, anxiety and stress scales are moderately-to-higly correlated with each other. Specifically, depression total scores and anxiety total scores are moderately correlated. Depression and anxiety symptomology has a common prevalence across the population and are highly comorbid with each other (???) and the basic results below appear to be supportive of that. However, these involve only two variables; there are countless hidden variables that could explain the relationships that have not been analysed.

We undertook then a quick investigation of how depression scores varied across the personality traits. This only examines the relationship between two variables only and is very likely to be confounded by other available variables in the data set that has not been taken into account. A number of key observations are reported below:

  • Respondents who rated themselves as high on TIPI9: Calm, emotionally stable (i.e. Strongly Agree with the personality trait) were associated with the lowest depression total scores compared to other respondents for the same item based on comparison of median total scores.

  • Respondents who rated themselves as low on TIPI4: Anxious, easily upset (i.e. Strongly Disagree with the personality trait) were associated with the lowest depression total scores compared to other respondents for the same item based on comparison of median total scores.

Anxiety scores were similarly examined against the same personality traits. The following observations were made:

  • Respondents who rated themselves as high on TIPI9: Calm, emotionally stable (i.e. Strongly Agree with the personality trait) were associated with the lowest anxiety total scores compared to other respondents for the same item based on comparison of median total scores.

  • Perhaps unsurpisingly, respondents who rated themselves as low on TIPI4: Anxious, easily upset (i.e. Strongly Disagree with the personality trait) were associated with the lowest anxiety total scores compared to other respondents for the same item based on comparison of median total scores.

The relationship between the depression and anxiety scores seem to suggest comorbidities between the two emotional states as measured by the DASS-42. . Furthermore, the relationship between the anxiety scores and stress scores conveyed the highest correlation between all three studied relationships. The DASS-42 is used widely in clinical settings and this analysis may provide some more evidence for continuing to assess for life events contributing to acute and chronic stress given the strong association identified. However, this analysis does not identify any mediators and moderators and caution is warranted.

Interestingly, analysis of the depression and anxiety scores across the personality items highlighted that those respondents who strongly endorsed themselves as “Calm, emotionally stable” reported the lowest depression and anxiety scores. Personality generally persists throughout our lives and these results do not suggest that personality causes certain levels of depression and anxiety symptoms. Psychological treatment planning could potentially take into account client personality traits to tailor the right treatment and potential enhance the patient-therapist relationship. Further analysis is needed to adjust for other relevant variables that could influence the relationship between the scores between the DASS-42 and TIPI though.

The source code for this analysis can be found in the below GitHub repository.

Lovibond, S. H., & Lovibond, P. F. (1996). Manual for the depression anxiety stress scales. Psychology Foundation of Australia.