Reading focus
Syllabus: How Does Depression Arise and Recur?
Identify the alternatives removed by comparing identical twins and those left by retrospective event ratings and unshared circumstances.
Why you’re reading it. Apply the earlier causal questions to a specific observational strategy.
A question to consider. Does finding an association within identical-twin pairs settle the direction of the feedback proposed by Hammen? Explain which uncertainty remains.
Why this reading
An association between stressful events and depression can reflect an environmental effect, shared vulnerability, or several processes together. Kendler and colleagues compare population and co-twin estimates to probe these possibilities. The paper provides a compact example of causal reasoning with an observational design—and of how much its interpretation depends on measurement and assumptions.
Before you read
A co-twin comparison examines twins who differ in an exposure. Identical twins share inherited genetic variation and much family background, but can differ in other circumstances. Event dependence concerns whether the person’s behavior may have contributed to an event. An odds ratio compares odds, not the probability that an event caused a particular episode.
Read for
- Work through what the identical-twin comparison controls and what unshared influences it leaves unresolved. Why is it still different from randomly assigning exposure?
- Examine how events, contextual threat, and depression onset were reconstructed over the previous year. Pay attention to the assumed order of events and onset within the same month.
- Compare the estimates across analyses and the uncertainty around them. What extra assumptions enter the authors’ attempt to partition causal and noncausal components?
Consider afterwards
- Two identical twins differ in job loss and subsequent depression. Which alternative explanations survive the fact that they share genes and family background?
- What would longer prospective observation add to the event and onset measures used here?
Editorial context
The study concerns White female twins from Virginia and relies on retrospective reports of the preceding year. Recall, event classification, timing, and unshared confounding limit causal interpretation and generalization. Maciejewski and Mazure’s contemporary comment questions the comparison of estimates and the classification of interpersonal events. The paper’s approximate causal fraction should therefore be read as an assumption-dependent interpretation, not an established proportion of depression caused by stress. The comment is not a retraction.