Age-related differences in internalizing psychopathology amongst the Australian general population

Two methodological criticisms have limited the reliability and validity of findings from previous studies that seek to examine change across the life span in levels of internalizing psychopathology using general population surveys. The first criticism involves the potential influence of cohort effects that confound true age-related changes whereas the second criticism involves the use of a single form of assessment to measure and compare levels of internalizing psychopathology. This study seeks to address these criticisms by modeling age-related change using multiple measures and multiple surveys. Data from 2 epidemiological surveys conducted 10 years apart in the Australian general population were combined and used for the current study. The latent construct of internalizing psychopathology was modeled using a combination of Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association, 1994) depression and anxiety diagnoses as well as items from the Kessler Psychological Distress scale (K10; Kessler et al., 2002). Confirmatory factor analysis (CFA) indicated that a single internalizing dimension provided good model fit to the data. Multigroup CFA indicated that strict measurement invariance of the model can be assumed across survey administrations and age bands, justifying comparisons of mean differences in latent trait levels. Significant changes in mean levels of latent internalizing psychopathology were evident between respondents aged 30-39 years old in 1997 and respondents aged 40-49 years old in 2007, suggesting a minor but significant increase in psychopathology across middle age. By contrast, a minor but significant decrease in psychopathology was noted when transitioning from late middle age (50-59 years old) to old age (60-69 years old). The majority of individuals in the general population will experience constant levels of internalizing psychopathology as they age, suggesting that the construct is relatively stable.