4

I've just come across the following PubMed article (referenced here and here) which claims that taking paternity leave causes one to live longer:

It has also been suggested that paternity leave [...] could benefit men's lifetime health.

[...]

The main finding was that [...] taking paternity leave between 30 and 135 days decreased the risk of all-cause mortality

[...]

the studied phenomena influence male mortality independently of each other.

I have no access to the full article but here's a summary courtesy of the Atlantic:

The researchers speculate that the men who opted for paternity leave in the late 1970s may have been more health-aware than their peers. But they conclude that there could also be something about increased parental involvement that moderates some of the detrimental effects of traditional masculinity on men’s health behaviors.

Here's according to Harvard Business review:

The study says increased involvement in parenting may reduce some of the detrimental effects of traditional masculinity on men’s health behaviors.

Is this conclusion scientifically sound? Are there other articles or reviews that find different results?

Sklivvz
  • 78,578
  • 29
  • 321
  • 428
  • I don't see the claim of causation. I see correlation and the wiggle word "could" being used. Of course if they're correlated one _could_ cause the other. – William Grobman Jan 09 '14 at 18:13
  • @WilliamGrobman there's as strong claim of it, and that's exactly what I am doubting. I posted the best excerpt I could find, but do read the whole abstract, it makes it quite clear. – Sklivvz Jan 09 '14 at 18:38

1 Answers1

2

Månsdotter (2010)

Månsdotter, Anna, and Andreas Lundin. "How do masculinity, paternity leave, and mortality associate?–A study of fathers in the Swedish parental & child cohort of 1988/89." Social Science & Medicine 71, no. 3 (2010): 576-583.

This study was not designed to assess causation.

Taken literally, this sentence in the abstract is an inaccurate representation of the study's results:

The main finding was that low 'masculinity' ranking increased the risk of all-cause mortality, and mortality from alcohol and violent causes, while taking paternity leave between 30 and 135 days decreased the risk of all-cause mortality.

Also, part of their conclusion makes an unproven assertion:

Overall, strategies aimed at less gender-stereotypical expectations on what a man “should do” are on the whole likely to benefit male health

However, the authors are consistent throughout the rest of the text to only talk about "association". Such as:

How do masculinity, paternity leave, and mortality associate? (The title)

the objective was to examine associations (also in the abstract)

paternity leave of more than 30 days but less than 136 days was associated with a decreased risk of all-cause mortality among Swedish men who had a child in 1988/89 (results section)

taking paternity leave is associated with a lower risk of mortality among middle-aged fathers in Sweden (in the conclusion)

They do mention causation when positing that there are "separate causal pathways" to explain the association between masculinity and longevity and the association between paternal leave and longevity. They don't claim to have determined what the nature of those causal pathways are (paternal leave -> longevity vs. confounding factor -> paternal leave and longevity vs. longevity -> paternal leave, etc.).

How this work has been received

This work has been cited uncritically in the following publications:

  • Sörlin, Ann, Lars Lindholm, Nawi Ng, and Ann Öhman. "Gender equality in couples and self-rated health-A survey study evaluating measurements of gender equality and its impact on health." Int J Equity Health 10 (2011): 37.

  • Sörlin, Ann, Ann Öhman, Nawi Ng, and Lars Lindholm. "Can the impact of gender equality on health be measured? a cross-sectional study comparing measures based on register data with individual survey-based data." BMC public health 12, no. 1 (2012): 795.

  • Brussoni, Mariana, Lise L. Olsen, Genevieve Creighton, and John L. Oliffe. "Heterosexual gender relations in and around childhood risk and safety." Qualitative health research 23, no. 10 (2013): 1388-1398.

  • Sörlin, Ann. "Health and the elusive gender equality: Can the impact of gender equality on health be measured?." PhD diss., Umeå University, 2011.

  • Elwér, Sofia, Lisa Harryson, Malin Bolin, and Anne Hammarström. "Patterns of gender equality at workplaces and psychological distress." PloS one 8, no. 1 (2013): e53246.

  • Maroto‐Navarro, Gracia, Guadalupe Pastor‐Moreno, Ricardo Ocaña‐Riola, Vivian Benítez‐Hidalgo, María del Mar García‐Calvente, María del Pilar Gutiérrez‐Cuadra, María T. Gijón‐Sánchez, María del Río‐Lozano, and Jorge Marcos‐Marcos. "Male and female involvement in the birth and child‐rearing process." Journal of clinical nursing 22, no. 21-22 (2013): 3071-3083.

I could find no other work citing Månsdotter 2010.

Any other studies on this question?

As for other work looking at paternal care and health, I could find none assessing the effect or association of paternity leave.

But, I found one that examined the association between "parental equality" and health (by the same authors as the study in the question):

  • Månsdotter, Anna, Lars Lindholm, Michael Lundberg, Anna Winkvist, and Ann Öhman. "Parental share in public and domestic spheres: a population study on gender equality, death, and sickness." Journal of epidemiology and community health 60, no. 7 (2006): 616-620.

    From the domestic sphere, it is significantly shown that traditional men (temporary care) have higher risks compared with equal men.

  • Clearly, causation implies association, thus the sentence you cite does not prove anything. – Sklivvz Jan 09 '14 at 20:07
  • I'm saying the authors aren't claiming causation and that the study was not designed to be able to detect causation. It was an epidemiological study. I'm saying there's no basis in their published paper to make the conclusion of causation. –  Jan 09 '14 at 20:09
  • They either claim causation or they don't. They do in the abstract, which is part of the article, right? So how can you say that they aren't? – Sklivvz Jan 09 '14 at 20:13
  • By reading the entire paper, it's clear to me that the wording in the abstract is not an completely accurate representation of their conclusions in the paper. That's how I can say that they aren't. Sometimes, people write sloppy sentences. There is no way they could have concluded causation based on their methodology, and they don't *anywhere else in the paper*. –  Jan 09 '14 at 20:15
  • The abstract clearly implies causation has been found in three sentences, how can you call it a "slightly inaccurate representation"? It's much, much harder to prove causation, thus if they don't, then the abstract is claiming something completely wrong. – Sklivvz Jan 09 '14 at 20:19
  • @GlenTheUdderboat but that's exactly what I am asking, except restricting it to one of the claims in the abstract. – Sklivvz Jan 09 '14 at 20:24
  • @Sklivvz Sure, it's wrong then. I was being generous to them, as I didn't take it to mean causation, given the title of the paper, and even the rest of the wording in the abstract: "the objective was to examine associations". I took it as a sloppy way to word their findings of association. –  Jan 09 '14 at 20:24
  • @Articuno did you find a free copy of the article anywhere? – Sklivvz Jan 09 '14 at 20:29
  • 1
    @Sklivvz Unfortunately not.. Do you know a way to share a pdf? –  Jan 09 '14 at 20:29
  • @Articuno dropbox? google drive? – Sklivvz Jan 09 '14 at 20:30
  • 1
    @Sklivvz https://www.dropbox.com/s/wqqfykcc069pfeg/1-s2.0-S0277953610003904-main.pdf –  Jan 09 '14 at 20:36
  • In the conclusion: `Overall, strategies aimed at less gender-stereotypical expectations on what a man “should do” are on the whole likely to benefit male health, and potentially reduce the gender gap in longevity` can only mean causation. – Sklivvz Jan 09 '14 at 21:05
  • Ya, I guess you can take that to mean they speculate that causation is the likely explanation, but they don't conclude that it *is*. If they did, that is not supported by the rest of the paper. –  Jan 09 '14 at 21:16
  • Actually, I don't see how any of their data supports their speculation that causation is even the likely hypothesis. –  Jan 09 '14 at 21:25
  • "cited uncritically" is ambiguous. It could mean they cited it approvingly (i.e. without any negative criticism) or it could mean they cited it without reading it and evaluating it (i.e. without any critical analysis). – Oddthinking Jan 11 '14 at 00:45
  • I meant the former, but I agree, it is ambiguous, and I shouldn't assume which of those is the case. –  Jan 11 '14 at 00:51