We recently talked to Maurits Masselink about his paper, "The longitudinal association between self-esteem and depressive symptoms in adolescents: Separating between-person effects from within-person effects", which is scheduled for publication in an upcoming issue of the European Journal of Personality. Maurits is a PhD student at the University Medical Center Groningen.
Read more about the article below!
Q: Hi Maurits, can you tell us what your study is about?
In our study we investigated the associations between self-esteem and depressive symptoms among adolescents over 1-1.5 year time intervals. We tested two opposing theoretical models, the vulnerability model, which states that self-esteem makes one vulnerable to develop depressive symptoms, and the scar model, which states that experiencing depressive symptoms can have a long lasting negative impact on self-esteem. Although the vulnerability and scar models have been extensively investigated, previous research shares an important limitation. Associations between self-esteem and depressive symptoms can occur between persons and within persons. An example of a between-person association is that adolescents who have low self-esteem compared to others are also the ones having more depressive symptoms compared to others. The vulnerability and scar models describe within-person processes, that is, they assume that changes in self-esteem result in changes in depressive symptoms or vice versa within the same individual. The problem with most previous research is that the vulnerability and scar models were commonly investigated using a Cross-Lagged Panel Model (CLPM), which is not suitable to investigate within-person processes. Results from a CLPM reflect a blend of between- and within-person effects. Research has shown that results from a CLPM can dramatically differ from the actual within-person associations. These differences may be in the presence of effects, the predominance of one effect over the other, and even the sign of effects may differ. This suggests that results from earlier CLPM research investigating the association between self-esteem and depressive symptoms cannot be trusted to reflect actual within-person processes.
A major strength of our research is that we used a method of analysis that is capable of separating between-person and within-person effects, the Random Intercept Cross-lagged Panel Model (RI-CLPM). Our results showed that when between- and within-person effects are separated, there were strong between-person effects between self-esteem and depressive symptoms. Adolescents who reported lower self-esteem than other adolescents, also reported more depressive symptoms than other adolescents. Within persons, we found small effects indicating that low self-esteem was associated with the experience of more depressive symptoms 1-1.5 years later, but not the other way around. These results are thus in line with the vulnerability model, but the effect sizes were small.
Our findings suggest that, despite the strong between-person association, self-esteem may not be a major vulnerability factor for developing depressive symptoms among adolescents. If this is the case, interventions aimed at increasing self-esteem would be of little use to prevent the development of depressive symptoms or to improve depressive symptoms. However, before such substantial statements about implications can be made, replication research is needed. A suitable topic for future research is to investigate the often mentioned hypothesis when small temporal effects are found, namely that small temporal effects may cascade and eventually result in large effects over multiple time points. Another pressing issue is to investigate the associations on a smaller time scale than yearly intervals. During daily life adolescents experience many situations which potentially affect self-esteem and mood, for example failing an important test at school or being rejected by peers. For a more detailed picture about the associations between self-esteem and depressive symptoms, it will be necessary to investigate these associations from day to day, or from moment-to-moment during the day.
In our study we also explored whether results would have been different using a normal CLPM. The results happened to be quite similar, but we want to emphasize that the fact that we found similar effects across the RI-CLPM and CLPM, is in no way a guarantee that this would have been the case in previous research or will be the case in future research. An important message that we want to convey is that the method of analysis should always be in line with the research question. When the goal is to investigate a within-person research question, the method of analysis should be a within-person method.
Q: Where do you see yourself in the (near) future?
In the near future I hope to finish my PhD. After finishing my PhD I hope to continue doing research, but not necessarily related to the topic of self-esteem and depression. I have several research interests that relate to ESM research. For example, I take part in the No Fun No Glory project. In this project we used ESM data to give personalized lifestyle advice to young adults with anhedonia in order to improve their pleasure experience. We had promising results, but we have multiple ideas on how to further improve our intervention, which I would like to explore. Another research interest is investigating whether it actually is the personalized aspect that drives the effect, or whether it is sort of a placebo/horoscope effect and randomly generated feedback would work just as well. Regardless of the research topic in which I may get involved, I aim to advocate the aims of the Open Science Movement. For the social sciences to advance and make an actual impact on society, it is imperative that open science principles are integrated in every research step. I incorporate open science initiatives in my own research, with pre-registrations and sharing of codes and data files, and hold that as a personal requirement for any future research.
Q: Do you have any tips or advice for young researchers?
Based on my experience while working on this paper, I would advise young researchers to share their research ideas with the broader scientific community at an early stage and to be open to collaborations. Initially only Study 1 was included in my plans, and I planned to use the conventional CLPM for the analyses. However, I came into contact with Loes Keijsers who pointed out the limitations of the CLPM and introduced me to the RI-CLPM. After presenting my findings at a conference, it turned out that Janne Vanhalst and Ben Hankin had data to extend and replicate my findings. This resulted in the inclusion of Study 2 and Study 3. Adding additional studies and collaborating with others resulted in a greatly improved manuscript, valuable new connections and new research skills.