In an analysis last year, Volunteerism in America, the Social Capital Project found that rates of volunteerism have either held steady or risen over the past forty years—a rare indicator of the health of our associational life that has not worsened over the period. Our initial report, What We Do Together, also highlighted the increase in hours of volunteering per person over time.

Harvard political scientist Robert Putnam came to a more mixed conclusion in his social capital opus, Bowling Alone: The Collapse and Revival of American Community. In analyses of the DDB Needham Life Style survey, Putnam found a rise in the number of hours of volunteering per person from 1975 to 1998. He also cited the Gallup trend we analyzed, which at the time showed a rise in volunteering rates from 1977 to 1991.

But he noted as well that data from the Independent Sector showed a decline in hours volunteered between 1987 and 1999. And the DDB Needham data indicated that the share of adults working on community projects (and the hours spent working on them) had fallen between 1975 and 1999, even as the same dataset indicated a rise in hours volunteered generally. Finally, Putnam speculated that the evidence he summarized might indicate that volunteerism was becoming unmoored from religious and other organizations, and he noted that “the increase in volunteering in recent decades is concentrated in the one generation most resistant to civic disengagement”—Americans older than age sixty.

These caveats are worth emphasizing, and we have added a link to this post to Volunteerism in America. To further muddy the picture, we offer two additional thoughts on the evidence. First, in another instance of the same data source yielding conflicting results, Figure 1 in Volunteerism in America includes Independent Sector data showing a rise in volunteerism rates during the period in which Putnam notes Independent Sector found a decline in hours spent volunteering.

Second, the Current Population Survey (CPS) evidence we analyzed reinforced Putnam’s DDB Needham results showing the rise in volunteerism much larger among older Americans. In fact, volunteerism rates rose meaningfully only among the oldest of the four age groups we considered.

However, if the evidence from Gallup that we cited in Volunteerism in America is correct, volunteerism rose among the population as a whole, in which case it’s entirely possible that it rose for younger groups as well as older ones. In new analyses using 1988 Gallup microdata and comparing results to a Gallup post using 2013 data, we confirmed that between these two years volunteerism increased notably for adults ages 18-29 (59 percent rise), 30-49 (35 percent), 50-64 (39 percent), and 65+ (61 percent). (Overall, volunteerism rose by 44 percent.)

And to add another complication, Putnam also drew our attention to the fact that Gallup has also found differing results from most other surveys on the trend in church attendance. In the General Social Survey (GSS), a widely respected academically rooted survey, the share of adults attending church almost every week or more fell from 41 percent in 1972 to 30 percent in 2000. Gallup figures show the share of adults attending in the last seven days stable at around 41 percent in 1972, but no lower in 2000.* The gap narrows after 2000, as the GSS series stabilizes (29 percent in 2016) and the Gallup series declines (36 percent). If Gallup data produces overly rosy trends in civic engagement generally—be it church attendance or volunteering—then we might privilege the volunteering results from the CPS and DDB Needham.

Unfortunately, it is unlikely that researchers will be able to get to the bottom of this discrepancy. The existing data are what they are, and there is little reason to think that a new data archive will be discovered.

Meanwhile Gallup and the Census Bureau will continue to measure volunteerism. Last September, the CPS survey included a supplement called the Volunteer and Civic Life Assessment. The new supplement consolidates questions on volunteerism and civic engagement from past CPS surveys. Perhaps eventually the Census Bureau, the Corporation for National and Community Service, and Gallup could collaborate to figure out what accounts for the disparate results from the two surveys.

^{*Thanks to Amy Lakeman, a doctoral candidate in Harvard’s Government department for assistance in understanding the debates around church attendance data, and to Robert Putnam for putting us in touch with her.}