You start dressing differently. From Single to Married, finally Happily Married.

Does getting married make you happier? The proverbial wisdom of many societies has been that marriage is essential to a good, happy life. But, in modern societies, this is more debated. Gallup polls and the General Social Survey (GSS) show that fewer Americans believe marriage is crucial to happiness or that it’s essential that lifelong romantic partners marry.

What makes these trends especially galling is that they are, in some cases, empirically wrong. Indeed, married people are happier than unmarried people: across nearly five decades of surveys, data from the GSS shows that 36% of people who have ever been married (including divorced, separated, and widowed people) say they are “pleased” while just 11% are “not too happy,” compared to 22% and 15% for people who have never married. Despite changing public views, the truth is married people really are happier.

But there’s a more charitable way to interpret the views of the growing number of people skeptical of marriage's benefits. Maybe married people are happier, but it’s not because of marriage. Indeed, perhaps delighted people are just more likely to get married! There is some solid empirical evidence for this view. One very influential paper showed that in several decades of German longitudinal data, self-reported happiness began to rise just before getting married, peaked in the year of marriage, and then declined within a year, with more significant effects for women than men. Other papers have built on this, such as a prominent recent paper that compared cohabiting and married people and found that higher happiness in marriage was due to other contextual factors, not marriage itself. These findings and others like them have led many people to discount the happiness benefits of marriage as just a product of bias and selection: happy people get married; marriage itself doesn’t make people happy. This research, however, turns out to be even more pessimistic than it sounds: the same methods that suggest marriage doesn’t impact happiness also tend to indicate that many other life experiences like unemployment and widowhood also don’t impact happiness much. In other words, this research implies that the reason poor, unemployed, lonely, or disabled people are less happy isn’t that bad things happened to them: they’d be unhappy no matter what.

Other recent research, however, has challenged this pessimistic view. Data from an extensive British panel survey show that marriage increased long-run happiness, suggesting that “selection” isn’t the whole story. Perhaps marriage doesn’t make Germans happy, but it does seem to make the English happy. Likewise, a paper using panel data from Taiwan found somewhat more durable positive effects of marriage (though there was a lot of variety in happiness trajectories, and Christians might get more happiness from marriage than other people).

However, research on happiness is bedeviled by a problem: only a few datasets track people over time and measure their satisfaction. Many longitudinal surveys (like the National Longitudinal Surveys of Youth in the United States) don’t ask about joy. Indeed, the United States has proved to be a significant blind spot in happiness literature, as few longitudinal surveys in America ask about happiness.

Luckily, since 2006, the GSS has included a longitudinal follow-up component. Because the GSS asks about happiness and marital status, this data can be used to see how happiness changes around crucial family events, particularly marriage, divorce, and widowhood. With several rotating panels of respondents totaling about 9,600 two-year-change observations, the GSS provides a reasonably small sample size with a relatively short follow-up period, so it can’t provide an exhaustive test of the effect of marriage on happiness. Within that sample, there are just a few hundred marital status changes. But some respondents, about 4,300, were followed for four years rather than two, making longer-term analysis possible.

Despite the small sample size and short time window, the GSS is a longstanding survey with high data quality. As one of the few U.S. surveys with any longitudinal happiness data, it provides an invaluable window into whether marriage might boost happiness in the U.S. as it does in Taiwan and the United Kingdom and in the short run in Germany. Figure 2 shows the estimated effect on happiness (on a scale ranging from 1 to 3) of four changes in family and social life. It shows results from three models:

 first, a model looking at all two-year changes in happiness individually, so showing short-run changes in satisfaction after marriage; second, a model looking at all four-year changes in satisfaction separately, so showing slightly longer-run changes in satisfaction after marriage; and third, a more technical model using “fixed effects” which estimates effects of marital status on satisfaction within individuals across multiple periods. All models also show the impact of any person spending about 2 more evenings a month hanging out with friends, estimated among the same respondents. Because the frequency of spending time with friends changes a lot more than marital status, results for social time with friends are evaluated more precisely.