Can differences in personality and support from friends help explain depressive symptoms? A long-running Swedish twin study examined how these experiences related over time.
Within-twin comparisons changed parts of the picture. Extensive missing data and model differences also mean the findings cannot establish a simple causal chain.
Quick answer
Does the study prove that more friend support prevents depression?
No. The study identified associations between certain personality-related measures, friend support and depressive symptoms, including within identical-twin pairs. It was not a treatment trial, and missing data and measurement limitations are substantial.
Key takeaways
- SATSA included 2,034 Swedish twins and up to 23 years of observations.
- Social support referred to friends, not family or partners; the outcome was depressive symptoms.
- Some associations changed when identical twins were compared within pairs.
- Only 487 pairs had baseline depression data for both twins; extensive imputation limits certainty.
- The study did not test whether any intervention prevents or treats depression.
Why compare twins when studying depression?
Personality, relationships and depressive symptoms often move together over time. Someone who struggles to express disagreement may also experience hostility, strained friendships or low mood. But a correlation between those experiences does not show which came first, which is causal, or whether family background helps explain them.
A new analysis of the Swedish Adoption/Twin Study of Aging (SATSA) used repeated observations and twin comparisons to investigate a possible chain involving lack of assertiveness, hostility, support from friends and later depressive symptoms. The value of the design is not that it gives a simple answer, but that it can challenge one.
What did the researchers actually measure?
The researchers studied 2,034 individual twins from 1,017 pairs. The analysis covered observations across as much as 23 years, from 1984 to 2007, in a predominantly middle-aged and older Swedish cohort. The authors report an average age of 58.7 years; these findings should not automatically be generalized to teenagers or young adults.
Lack of assertiveness was assessed using a particular inhibition-of-aggression scale, not a global personality diagnosis. Hostility was a separate construct. Social support was a structural index based on how many friends participants met or could contact, rather than a direct measure of emotional closeness or help received. Family and partner support were not assessed. The outcome was depressive symptoms measured in the study, not a confirmed clinical diagnosis of major depressive disorder.
What was different about the twin analysis?
Ordinary analyses compare people who differ in personality and support. Those people may also differ in genetic predispositions, upbringing and earlier circumstances. Comparing twins within the same family reduces some of these shared differences. The researchers examined within-pair differences among 317 identical-twin pairs and 645 fraternal-twin pairs.
Identical twins share almost all inherited DNA, so differences within a pair can be especially informative. But twins still have different experiences, measurement errors and relationships. This remains observational research, not a randomized test of friendship, personality or depression treatment.
Finding one: the apparent pathway changed
In the identical-twin difference analysis, lower assertiveness did not clearly predict subsequent hostility. That contrasts with a positive association reported in the fraternal-twin analysis. The contrast is compatible with shared familial or genetic factors contributing to the overall association, although the design cannot identify a single explanation.
This matters because a story that looks straightforward when comparing unrelated people can look different when shared family background is partly held constant. It does not mean that learning to express one's needs is ineffective or that personality is fixed by genes.
Finding two: friendship support was part of the picture
In the authors' identical-twin serial model, higher hostility was associated with lower subsequent friend-network support (β≈−0.152), which was associated with higher subsequent depressive symptoms (β≈−0.238). The hostility-to-depression indirect path through friend support was statistically significant in this serial specification (b=0.020, 95% CI 0.005–0.038). The separate parallel model showed corresponding standardized paths of approximately −0.086 and −0.206 and retained direct associations for lack of assertiveness and hostility.
These estimates describe fitted relationships in a particular study and model. They do not show that hostility inevitably damages friendships, that social support is the only relevant mechanism, or that a particular person will become depressed.
Did the study prove that support mediates depression?
No. The researchers examined several serial and parallel mediation specifications. Some results were consistent with an indirect pathway through friend support, but estimates and statistical significance differed between models and inference methods. The serial model's hostility-to-depression indirect path was significant, whereas the corresponding indirect effect in the parallel model was not significant using a delta-method test (b=0.007, p=.120), despite a marginally positive Monte Carlo interval. The study was not preregistered.
An indirect statistical path is not the same as an established causal mechanism. To show that changing a particular behavior or support network prevents depression would require additional evidence, including carefully designed intervention studies.
The missing-data limitation is unusually important
Although the headline analytic sample contained 1,017 twin pairs, only 487 pairs had valid baseline depression data for both members. Complete observed values for all model constructs were available for just 10 identical and 18 fraternal pairs. The authors used imputation to handle missing information, reporting item-level imputation ranging from 24.4% to 83.0% across constructs and waves.
Imputation is a legitimate statistical method, but it cannot recover observations that were never collected with certainty. Attrition, changing measures over time and measurement error are particularly consequential for within-pair analyses. The results should therefore be read as suggestive rather than definitive.
How does this differ from research on loneliness?
Loneliness describes a subjective experience of insufficient connection. This SATSA study examined a different question: whether specific personality-related measures and reported support from friends were linked to depressive symptoms over time, and how the associations changed in twin comparisons.
The two lines of research can inform one another without being interchangeable. Neither establishes that simply adding more social contacts will treat depression, and neither justifies blaming someone for their mood or relationships.
What can readers take away?
Social relationships may matter for mental wellbeing, but this study does not provide a universal recipe. Its strongest contribution is methodological: some apparent pathways change when twins are compared within pairs, while others remain compatible with an environmental component. Major missing-data and model limitations prevent firm causal conclusions.
People experiencing persistent low mood deserve appropriate support and clinical assessment when needed. TwinPare Research presents the evidence for understanding, not as a diagnosis, personal risk score or substitute for mental-health care.
Source notes
The source has been verified and editorially reviewed for this article. The limitations below show which level of conclusion the sources support.
Editorial source review
This section shows how the article's key factual claims are linked to the source.
Phrasings that require caution
- Distinguish depressive symptoms from diagnosed major depression.
- Do not frame observational mediation as proven causation.
- Make missingness and imputation visible.
- Do not generalize friend support to all social relationships.
- No individualized diagnosis or treatment guidance.
| ID | Claim | Source support | Caution |
|---|