BACKGROUND: Old age respondents may differ systemically in their responses to measures of psychological distress over and above their actual latent distress levels when compared to younger respondents. The current study aimed to investigate the potential for age-related bias(es) in the Kessler 6 Psychological Distress Scale (K6) items. METHODS: Data from the 2007 Australian National Survey of Mental Health and Wellbeing were analyzed using Item Response Theory to detect the presence of item bias in each of the K6 items. The potential for item bias was assessed by systematically comparing respondents classed as young (16-34 years), middle aged (35-64 years), and old aged (65-85 years). The significance and magnitude of the item bias between the age groups was assessed using the log-likelihood ratio method of differential item functioning. RESULTS: After statistical adjustment, there were no biases of significant magnitude influencing the endorsement of K6 items between young and middle-aged respondents or between middle-aged and old age respondents. There was a bias of significant magnitude present in the endorsement of the K6 item addressing levels of fatigue between young and old age respondents. CONCLUSIONS: Despite the identification of significant item bias in the endorsement of K6 items between the age groups, the magnitude and influence of the bias on total K6 scores is likely to have little influence on the overall interpretation of group data when comparing psychological distress across the lifespan. Researchers should be cautious, however, when examining individual levels of fatigue related to psychological distress in older individuals.
BACKGROUND: Old age respondents may differ systemically in their responses to measures of psychological distress over and above their actual latent distress levels when compared to younger respondents. The current study aimed to investigate the potential for age-related bias(es) in the Kessler 6 Psychological Distress Scale (K6) items. METHODS: Data from the 2007 Australian National Survey of Mental Health and Wellbeing were analyzed using Item Response Theory to detect the presence of item bias in each of the K6 items. The potential for item bias was assessed by systematically comparing respondents classed as young (16-34 years), middle aged (35-64 years), and old aged (65-85 years). The significance and magnitude of the item bias between the age groups was assessed using the log-likelihood ratio method of differential item functioning. RESULTS: After statistical adjustment, there were no biases of significant magnitude influencing the endorsement of K6 items between young and middle-aged respondents or between middle-aged and old age respondents. There was a bias of significant magnitude present in the endorsement of the K6 item addressing levels of fatigue between young and old age respondents. CONCLUSIONS: Despite the identification of significant item bias in the endorsement of K6 items between the age groups, the magnitude and influence of the bias on total K6 scores is likely to have little influence on the overall interpretation of group data when comparing psychological distress across the lifespan. Researchers should be cautious, however, when examining individual levels of fatigue related to psychological distress in older individuals.
Authors: Zephanie Tyack; Kerrie-Anne Frakes; Petrea Cornwell; Suzanne S Kuys; Adrian G Barnett; Steven M McPhail Journal: BMC Health Serv Res Date: 2013-10-11 Impact factor: 2.655
Authors: Carla Morgan; Elizabeth Mason; Jill M Newby; Alison E J Mahoney; Megan J Hobbs; John McAloon; Gavin Andrews Journal: Internet Interv Date: 2017-10-24