Self-esteem is, very simply, the set of feelings you have about yourself. Unlike self-knowledge, which refers to how much you know about yourself, the core of self-esteem is formed around whether you like yourself or not. Clinicians use low self-esteem as one possible symptom when they diagnose the psychiatric condition of major depressive disorder. They don’t necessarily care whether low self-esteem causes the depression or vice versa. However, personality researchers have long wondered about the chicken-and-egg problem of self-esteem and depression. Certainly, if you dislike yourself, you’ll be more likely to be depressed. Conversely, if you’re depressed, you’ll be more likely to feel bad about who you are as a person.
The only way to disentangle the highly related concepts of self-esteem and depression is through longitudinal research, in which people are followed up over time. Fortunately, a major meta-analytic study of 77 studies on depression, conducted by University of Basel researchers Julia Sowislo and Ulrich Orth, now can give us some answers.
A meta-analysis is a special type of investigation in which researchers measure what they call “effect size.” This is a statistical index that provides a number to attach to the magnitude of a finding. It goes beyond statistical significance (such as the frequently-used “p<.05”) to assign an absolute weight to the results. When psychologists have the results of meta-analyses, we can feel very confident that we’re not just basing our conclusions on a handful of studies or on rare probabilistic outcomes.
Before I get to the results of Sowislo and Orth’s meta-analysis, I’d like to take you through the logic of how they contrasted the competing directions of self-esteem to depression vs. depression to self-esteem. The vulnerability model predicts that people who have low self-esteem are at risk for depression. The lower your self-esteem, according to this view, the more likely it is that you will see the events in your life as reinforcing your negative sense of self. Someone cancels a date with you, and you conclude that they did so because they don’t want to be with you. Your low self-esteem causes you to twist your interpretation of this event in such a way as to confirm your already unfavorable sense of self. A non-depressive interpretation of this event would take into account the many possible reasons that people cancel dates with other people (i.e. they’re busy or something unexpected came up). A non-depressed person would accept the cancelation and go ahead to make another date. A depressed person, in contrast, would figure that the relationship is over and not even try to reschedule, or even refuse to talk to that individual again.
The scar model takes the opposite approach to understanding low self-esteem and depression. According to this view, depression is the cause of low self-esteem. The sad feelings that characterize your daily moods eventually wear away at your self-esteem, causing it to suffer “cuts” that never fully heal. Returning to our earlier example, if you’re already depressed, the fact that someone cancels a date with you causes you to question even further how likeable you are. The scar model says that your self-image suffers from rejection because you are already depressed.
Sowislo and Orth’s study devised two statistical models comparing the strength of the path leading from self-esteem at Time 1 to depression at Time 2 (the vulnerability model) vs. the path leading from depression at Time 1 to self-esteem at Time 2 (the scar model). They were also able to compare the combined effects of depression and self-esteem at Time 1 on both outcomes at Time 2. They found a total of 53 studies that met their criteria for examining self-esteem in terms of overall self-evaluations and depression as a continuous measure from low to high (as well as anxiety, which I’m not covering here). The 53 studies included 77 samples ranging in age from childhood to later life. All had to be longitudinal, meaning that the samples were studied over time. The studies spanned the globe, included males and females, and went beyond either the typical college student sample or a sample of people seen in clinical practice. The studies were published in the years between 1984 and 2010, with half after the year 2000. The largest sample in the study consisted of over 6800 individuals, and the longest time between tests was 13 years.
All in all, it’s a pretty impressive collection of data and, fortunately, the authors provide enough details in the paper so that it’s clear just exactly where the samples came from and how they were studied. Furthermore, the variety of studies within the collection strengthen the validity of the study’s conclusions.
The findings almost all overwhelmingly support the vulnerability model of self-esteem and depression. The lower people’s self-esteem at Time 1, the higher their depression at Time 2. There was only modest support for the scar model. These effects occurred across the board in all of the samples tested There were no effects of such factors as gender, meaning that the effects were similar for males and females nor did the effects vary by age of the participants. It didn’t even matter which exact measures they used to measure self-esteem and depression, whether the samples were clinical or non-clinical, or how long the time lag was between testings. Over time, low self-esteem is a risk factor for depression, regardless of who is tested and how.
We’ve already seen how the vulnerability effect may operate. You are a person with low self-esteem and you therefore take personally, and in a negative way, the events that occur in your life. The authors believe people with low self-esteem try to not to disprove but to verify their negative self-concept by seeking negative feedback from the people in their social network. They think about their inadequacies, focus on the negative feedback they receive from others, ponder that feedback, and as a result become more depressed. We don’t know for sure if this is the case, but one of the studies that the authors examined strongly suggested this line of argument. Some scarring may also take place among depressed people. Their negative mood leads them to be perceived more negatively by others, which leads them to feel hurt and rejected.
There were, of course, limitations to the study. The most obvious is that all of the investigations were correlational. It would hardly be feasible, or ethical, to induce low self-esteem to determine its experimental effect on a person's level of depression.
By cutting across generations, the authors were also able to address the question of whether self-esteem and depression are showing the effect of historical period. The findings they reviewed suggested that there were no changes across generations within the 20th century in self-esteem. They also did not find generational increases in such related concepts as self-enhancement and narcissism, as is sometimes claimed about today’s young adults. Even if those absolute levels were to show change, however, the relationships between self-esteem and depression were stable across the generations.
In summary, we can conclude that there is pretty convincing evidence to support the vulnerability effect of low self-esteem on depression. Using this knowledge, you can improve your own mood by trying these reasonably straightforward approaches:
It’s rare to find studies in psychology on topics such as self-esteem and depression that allow for any causal arguments to be made. However, from this large and comprehensive study, we can conclude that the best way to protect your positive mood is to find ways to boost your self-esteem.
Follow me on Twitter @swhitbo for daily updates on psychology, health, and aging. Feel free to join my Facebook group, "Fulfillment at Any Age," to discuss today's blog, or to ask further questions about this posting.
Copyright Susan Krauss Whitbourne, Ph.D. 2013
Sowislo, J., & Orth, U. (2013). Does low self-esteem predict depression and anxiety? A meta-analysis of longitudinal studies. Psychological Bulletin, 139(1), 213-240. doi:10.1037/a0028931