Do married people live longer than unmarried people, and how does the size of that difference compare between men and women?

This paper claims that married people in 1960s America died at lower rates than unmarried people — especially men — because marital roles provide psychological protection. The descriptive mortality gap is real and well-documented from national vital statistics, but the causal explanation rests entirely on the pattern of ratios across causes of death, with no measurement of psychological states, health behaviors, or any other proposed mechanism. The central claim is an interpretation the data cannot test.

Across 16 developed countries, unmarried adults died at higher rates than married adults, with divorced men worst off and men's gap larger than women's — these descriptive patterns are well supported by national death counts. The paper's more ambitious claims, that the gap has grown over time and that a 'selection' process makes small unmarried groups sicker, rest on a measure that can move in the opposite direction from what it seems to show and on an association the authors' own rival explanation fits equally well. Trust the rankings, but don't trust the explanations.

Using a large US health survey linked to death records, the paper finds that adults who were unmarried in 1989 — especially those who had never married — were more likely to have died by 1997. That descriptive finding is well supported. The paper's larger conclusions — that social isolation kills, hits hardest in young adulthood, and works through many diseases — are not supported. Isolation was never measured, the comparisons behind those claims were never tested, and the authors themselves attribute the most dramatic result to the HIV epidemic rather than to isolation.

This study uses 28 years of U.S. panel data to argue that marriage does not causally improve health and that even the modest benefit for women has disappeared in younger generations. The broad finding that healthier people select into marriage is well-supported, but the headline claim about generational decline rests on roughly 185 early-marrying women and is contaminated by an unacknowledged measurement problem: in this survey, the person rating a woman's health switches from herself to her husband at the exact moment she marries.

The paper claims that about half the health advantage of married people comes from healthier individuals marrying in the first place, while the other half reflects a genuine protective effect of marriage at older ages. The selection finding is well-supported, but the causal protection claim rests on a residual statistical association that cannot be separated from changing life circumstances, and every piece of mechanism evidence uses the very methods the paper itself demonstrated are biased for this question.

Using a survey of 165,000 older Americans in Medicare Advantage plans, the paper projects that married 65-year-olds will live about two years longer and stay independent about two years longer than unmarried ones, with the gap closing by the mid-80s. Those descriptive numbers are careful and believable. The headline conclusions built on them — that marriage protects health, and that roughly 60% of its apparent benefit is really healthier people being more likely to marry — are not something this design can establish.

Pooling 53 comparisons from prospective studies of a quarter-million older adults, this meta-analysis finds that married people over 65 die at roughly 10–12% lower rates than widowed, divorced, or never-married people, and the direction of that finding is well supported. The exact size is less certain than the paper's tight intervals suggest, and nothing here shows that marriage itself causes the difference rather than healthier people being more likely to be and stay married. The pattern is real, but be cautious about the precise number and the causal reading.

Pooling 641 estimates from 95 studies, this meta-analysis finds that never-married adults die at roughly 24–30% higher rates than married adults, and that figure is probably about right. But the paper's headline claims — that the gap is rising over time, that women have caught up with men, and that it is largest in East Asia — come from a regression model whose results the raw decade-by-decade and regional numbers do not show, and they rest on very thin data at the points that matter. Trust the modest overall gap; do not update on the trends.

This meta-analysis of 21 studies claims that unmarried men face a larger mortality penalty than unmarried women, concluding that men 'benefit more from marriage.' The descriptive pattern in the data is real, but the causal interpretation cannot be distinguished from the possibility that sicker men are simply less likely to marry — a competing explanation the paper barely engages. The discussion section cites a sparrow immunology paper and three other irrelevant studies as evidence for human biological mechanisms, and omits the most directly comparable prior meta-analysis's finding of no sex difference.

This meta-analysis claims unmarried people die at substantially higher rates than married people — 14% to 83% higher depending on marital category and cause of death. The numerical results cannot be trusted: the paper reports contradictory numbers for most of its own findings (three different estimates for widowed mortality appear in different sections), contains a confirmed copy-paste error that invalidates a central claim about sex differences, and pools cancer-survival studies with general-population cohorts as though they measure the same thing.

Using a large U.S. survey linked to death records, the paper finds that unmarried adults die at higher rates than married adults with similar age, education, income, and disability status — a result that is well estimated and consistent with prior work. Its distinctive claims go further: that this gap is marriage protecting people, that men get an extra benefit, and that never-married, divorced, and widowed people don't differ from one another. None of those three is supported by the data as analyzed. Trust the basic gap; don't update on the rest.

This letter makes a correct statistical point — that comparing results from separate models for men and women is not the same as testing whether those results actually differ — but directs it at a claim the original authors say they never made. The supporting evidence comes from studies measuring different health outcomes than the paper being criticized, and the letter's own prediction rests on personal belief rather than analysis.