Originally posted on December 1, 2015.
“Happiness doesn’t bring good health,” headlines a December 9 New York Times article. “Go ahead and sulk,” explain its opening sentences. “Unhappiness won’t kill you.”
Should we forget all that we have read and taught about the effects of negative emotions (depression, anger, stress) on health? Yes, this is “good news for the grumpy,” one of the study authors is quoted as saying. In this Lancet study, which followed a half million British women over time, “unhappiness and stress were not associated with an increased risk of death,” reported the Times.
A closer look at the study tells a somewhat different story, however. Its title—“Does Happiness Itself Directly Affect Mortality?”—hints at an explanation for the surprising result. Contrary to what the media report suggests, the researchers found that “Compared with those reporting being happy most of the time, women who had reported being unhappy had excess all-cause mortality when adjusting only for age.” Said simply, the unhappy women were 36 percent more likely to die during the study period.
But the happy women also exercised more, smoked less, and were more likely to live with a partner and to participate in religious and other group activities. Controlling for those variables “completely eliminated” the happiness-longevity association, and that explains the headline.
In much the same way, one can reduce or eliminate the religiosity-health association by controlling for the factors that mediate the religiosity effect (social support, healthier lifestyle, greater positive emotion). Ditto, one can eliminate the seeming effect of a hurricane by “controlling for” the confounding effect of the wind, rain, and storm surge. A hurricane “by itself,” after eliminating such mediating factors, has little or no “direct effect.”
Likewise, happiness “by itself” has little or no direct effect on health—a finding that few researchers are likely to contest.
P.S. For more critique of the happiness-health study, see here.