A longitudinal examination of the psychometric properties of the English perceived stress scale- four (PSS-4) in mental health counsellors using item response theory

Abstract

We examined the longitudinal psychometric properties of the Perceived Stress Scale – 4 items version (PSS-4) using item response theory with a sample of 361 mental health counsellors. Participants completed the PSS-4 at three timepoints at six-month intervals in a one-year period. There were 290 participants who (80.3%) identified as female, 51 (14.1%) identified as male, eight (2.2%) identified as gender variant/non-conforming, seven (1.9%) wrote in their own gender identity (e.g., genderqueer, gender expansive), three (0.8%) identified as Transgender male, and two (0.6%) did not respond to the item. The racial and ethnic backgrounds were as follows: White (87.3%), Multiracial (5.5%), Latino or Hispanic or Spanish (2.8%), Black or African American (1.4%), Asian (0.8%), Middle Eastern (0.8%), and five did not respond to the item (1.4%). We found unidimensionality evidence of the PSS-4 across all three timepoints and response categories were monotonically ordered. We also found that across all timepoints, the average person location was lower than the average item location, suggesting that the PSS-4 may not be well-targeted for this sample of mental health counsellors. We observed no significant interactions between timepoints, hours worked per week, and length of employment. Implications of the findings, including a discussion of the utility of the PSS-4 as a global measure of stress and with mental health counsellors.