The value of longitudinal research
Longitudinal research studies are increasingly being used in the social and medical sciences to answer important questions about human growth and development. They are fundamentally different from one-off, cross-sectional studies. Whereas cross-sectional studies compare data from different groups, longitudinal studies track the same individuals over time. They therefore control for two problems: the lack of comparability across different groups, and the inability to answer questions regarding continuity in individual development.
Although longitudinal studies only evaluate one group over time, which means that any findings might simply reflect cultural or other conditions relevant at the time of data collection, they are nevertheless considered far superior to cross-sectional ones. But, they are also logistically more difficult to handle.
As with cross-sectional studies, the questions answered with longitudinal data sets are incredibly varied. In the social sciences, some studies have examined the long-term effects of adolescent peer group membership on later social and academic development, the precursors and consequences of personality change, childhood predictors of adult criminality, the effects of cognitive ability on mortality, the effects of adolescent personality on adult health status, and much more.
Tracking the same individuals over time and assessing their social, emotional, health, occupational and other outcomes is not only an enormous privilege because of the window provided into these developing lives, but also extremely exciting from a scholarship point of view.
There are a number of well-known large databases in the UK, US, New Zealand and elsewhere. For example, a fascinating series of papers have emanated from the “Scottish Mental Surveys” first begun in the 1930s and still tracking those individuals now in their 80s.
Two large-scale national longitudinal studies in Australia are The Household, Income and Labour Dynamics in Australia (HILDA) Survey which began in 2001 and also includes variables of interest to sociologists and psychologists. This sample comprises several thousand individuals. The other is the Longitudinal Study of Australian Children (LSAC) which has been tracking 10,000 children and families and which gave rise to the TV series Growing Up in Australia. Both HILDA and LSAC provide researchers with access to their data and thus huge potential for publications in the world’s leading journals. This is important as researchers are spared the trouble and expense of collecting large-scale data over many years, supervising research assistants, and monitoring retention and attrition rates. Instead, they can spend their time focusing exclusively on analysing variables of interest and producing good quality papers for high impact journals.
It turns out that in psychology, and I am sure in other disciplines too, journal editors love well designed and well executed longitudinal studies. For all researchers, but especially early career researchers still developing their track record, these data sets provide a very useful and convenient way of building a high quality publication record. With colleagues at the University of Wollongong, I have used the HILDA data set to answer important questions regarding the effects of personality change during adulthood on health outcomes some years later. It is now accepted that personality is not static. Is personality change related to health outcomes? It turns out that increases in the personality traits conscientiousness and sociability during adulthood are associated with better mental and physical health outcomes. This is not surprising; conscientious people are more persistent and dependable and thus more likely to follow medical advice or monitor their health. Increases in conscientiousness over time are even more beneficial for mental and physical health. On the other hand, increasing emotional instability and neuroticism over time is bad for your health.
There are also other smaller longitudinal studies. For instance, I have been involved in the Wollongong Youth Study since 2004 and the Australian Character Study since 2010. Both are funded by the Australian Research Council and have been providing important new data on the development of several thousand young people across two Australian states. For instance we found that, even though girls outperform their male counterparts academically in high school, their levels of self-esteem decline much faster than that of boys. Although these changes might be due to biological and hormonal changes or cultural effects, we have no data to test such predictions.
Interestingly, changes in some personality traits are linked to teenagers’ perceptions of parental behaviours. Those who recalled their parents as warm and loving, yet firm, had better wellbeing than those who recalled their parents as autocratic. Interestingly, warm and democratic parenting moderated the extent of decline in some personality traits over time. Although some of this might sound intuitive, this is the first longitudinal evidence to demonstrate these parental effects.
Longitudinal studies require good planning, persistence, and teamwork. If you can’t “grow your own”, then consider the many national and international data sets that are available to help build your research track record.