A new article by Jean Twenge and colleagues is getting extraordinary attention from the news media and is being hotly debated in fields concerned with youth civic engagement.* The article argues that young people today–the so-called Millennial Generation–are relatively self-interested, unengaged, distrustful, and unconcerned about social justice and the environment compared to their predecessors. The main evidence is a set of trends like those shown below, which are taken from the vast and highly-respected federal survey of high school students known as Monitoring the Future. These particular questions are behavioral (e.g., Do you think about social problems? In your personal actions, do you make an effort to conserve energy and help the environment?).

Several of the declines are significant, and one could add other such trends. Participation in face-to-face meetings, interest in the news, working on community projects, joining unions, attending religious services, and joining clubs and groups have all declined for youth since the 1970s. Although youth turnout was higher in the presidential elections of 2004 and 2008 than it had been during the 1990s, turnout in off-year elections was consistently higher from 1974-1994 than in 1996-2010. (Off-year elections are arguably better measures of youth engagement than presidential elections, since the latter are more volatile and subject to choices by a few campaign managers.)

On the other hand, as Twenge et al. concede, the youth volunteering rate has climbed since the 1970s. They dismiss that increase as a result of external forces (mandates and rewards for community service), but most large-scale trends in youth behavior can be attributed to context rather than choice. Certainly, whether you vote or join a union has a lot to do with the behavior and resources of parties and unions, respectively. The long-term decline in trust for other people is deeply troubling, but few analysts would be satisfied with saying that young people refuse or fail to trust other citizens. The social context is surely responsible.

Twenge et al. move from the trends shown above (notice all those wavy, non-parallel lines) to two kinds of broad generalization. One is a thesis that civic engagement is down–overall. That conclusion requires combining and weighing many different variables. If one chose presidential-election turnout, Internet-use, and volunteering as the key measures of engagement, the trend would be up instead of down. So the narrative is sensitive to choices about what to include and emphasize.

Their second generalization involves generations. Almost any civic trend will show frequent ups and downs–as in the graph above. A generation is a way of simplifying that complex story. You are even simplifying if you lump together all the young Americans who were born in the same specific year. Distinguishing contemporaries by their educational attainment, race, gender, religion, and ideology would reveal dramatic gaps at any given point as well as divergent trends over time. In fact, differences in civic engagement by social class utterly dwarf changes in average youth civic engagement over time.

Below is an illustration from a CIRCLE report that is supposed to make the point that young people differ an awful lot in their civic engagement. Averages are basically uninformative. (Click the graphic for the report.)

Twenge et al. interpret and simplify complex data in a way typical of psychologists. They administer a battery of survey measures to a particular group of youth (182 San Diego State University undergraduates) to see how the measures cluster. They do not choose their measures randomly but use items that will help them explore specific psychological constructs, such as self-esteem, altruism, conformity, communitarianism, and narcissism. They conclude that rising narcissism really underlies the trends in the larger population.

There is nothing technically wrong with this method. The authors do not use their undergraduate sample to make generalizations about frequencies or trends in the national population, but only to build a conceptual model for interpreting national trends. For that purpose, their sample need not be nationally representative, although it would be interesting to see whether the same patterns emerge, for example, among high school dropouts.

My concern is not technical but philosophical. As I argued at more length recently, civic engagement is not primarily a psychological construct having to do with motives, personal values, and behaviors. It is rather (or also) a set of  actions and values that a good society needs from its citizens. Which behaviors and values are most important depends on one’s theory of a good society. Perhaps volunteering and voting in presidential elections are most valuable, in which case we are in good shape. Or perhaps trust is more important, in which case the decline has been profound and inexorable. The answer lies not in psychological research but in our conception of a good community. Further, the causes of civic engagement and disengagement are unlikely to reside inside young people’s heads. If, for example, being contacted by a political campaign is an indicator of civic engagement, then the reason for the decline from 1972-2o08 was a set of choices by campaigns, not by youth.

In short, I think Twenge et al. provide a helpful reminder that the civic engagement of our young people has changed in many ways, at least some of which are troubling. Their article could be misleading, however, if it led readers to assign too much importance to personality traits or to ignore enormous differences among young people, or if it seemed to prove that certain behaviors are most important for democracy.

*“Generational Differences in Young Adults’ Life Goals, Concern for Others, and Civic Orientation, 1966-2009,” Jean M. Twenge, PhD, and Elise C. Freeman, MA, San Diego State University; W. Keith Campbell, PhD, University of Georgia; Journal of Personality and Social Psychology, online. PDF currently available here.