In two recent blog posts in reaction to PEW and Harvard Public Opinion Project reports on Millennials, John Sides warned against equating the millennial generation’s more liberal and Democratic-leaning preferences with a bright future for the Democratic Party. However, he closes his report acknowledging that the “political formation of younger Millennials isn’t over”. The problem with any claim regarding the future attitudes and behaviors of today’s youth is that they are impossible to predict. A series of articles in a recent issue of the journal Electoral Studies (edited by Anja Neundorf and Richard Niemi) draws attention to the importance of distinguishing between the effects of ageing and the effects of belonging to a certain cohort or generation when predicting political attitudes and behaviors over the life span. Sides’ earlier contribution to this blog raises the question of whether the Millennials (those born after 1980) are indeed a distinct generation or whether they are unique in comparison to the rest of the electorate because they are still young?
The problem in researching intergenerational differences in political behavior is that age and cohort differences are interrelated. Once we know someone’s age (at a given time) we can easily work out which (birth) cohort this person belongs to. To illustrate this problem, let’s take the example of turnout. If we were to research voter turnout three questions would be central. Do we decide to participate in elections because we are e.g. middle-aged (age effect)? Do we vote because it is a pretty exciting election (period effect)? Or do citizens decide to stay at home because they came of age during the Watergate Affaire and ever since think that politics is not to be trusted and it is best to stay as far away as possible from it (cohort effect)? The truth of the matter is that all these effects impact our vote decision. The challenge is to try to disentangle the three effects methodologically. This post is an attempt to illustrate these interrelated effects in light of the Millennial generation, which has been at the center of attention in the last two months.
We will first look at the age effect. It is common to blame young people for being apathetic, lazy, and losing moral (and civic) values. In politics, for example, young people are often blamed for having very low turnout levels or trust in institutions and political processes. But what happens if these youngsters grow into middle-aged voters? Do they remain disengaged or do they become more active? Figure 1 plots the average turnout by age in presidential elections between 1972-2008. The figure clearly shows that young people are indeed less likely to take part in presidential races than older voters. This is the so-called age effect according to which levels of political participation change with the life experience that comes with getting older.
Figure 1: Average turnout in presidential elections 1972-2008 by age (Graph by the authors).
According to figure 1, turnout rates especially increase between the age of 20 and 40, rising from below 40% to 70% participation. Interestingly, the low turnout rate among young people was just as much of a problem in the 1970s (among the baby-boomers) as it is today (among the Millennials). In 1976, 50% of the 18-28 year olds turned out to vote in the Carter election. In 2008, 51% of the Millennials aged 18-28 voted in the Obama election. Based on these numbers, we can conclude that young voters are indeed considerably less active in electoral politics.
Figure 1, however, does not tell us anything about the Millennial generation, which includes only those born after 1980, in comparison with other generations. We hence replicated the average turnout by birth years rather than by age. As Figure 2 clearly shows, all generations born after about 1930 are less and less likely to participate in presidential elections, with a particularly sharp decrease for those born after 1980. The youngsters born in the late 1980s show a very low and worrying level of turnout of only around 20%. These are the Obama kids, as the 2008 election was the first presidential election in which they were eligible to vote. Figure 2 confirms the negative picture about the Millennials painted in research reports and in the media.
Figure 2: Average turnout in presidential elections 1972-2008 by birth year (Graph by the authors).
However, this picture of the Millennials needs to be put into perspective by taking into account the trends that we observed in figure 1. After all, the Millennials are still young. The key to understanding turnout among different age groups (figure 1) and voters born at different point in time (figure 2) is to combine the two insights. The difference between age and cohort/generation effects is best illustrated with an example. Figure 3 tracks the average turnout for people who first voted under different US Presidents. Respondents who voted for the first time in the 1948 election of President Truman show the highest average turnout of up to 85% across the entire time period for which we have survey data (1972-2010). The Kennedy generation – who first voted in the 1960 election – slowly seems to catch up with this cohort and turnout levels across these two cohorts appear equal today. The Nixon, Reagan and Clinton cohorts are tracked through time from the election in which they were first allowed to vote. Their turnout levels are on the increase.
Figure 3: Turnout patterns of cohorts coming of electoral age under different US Presidents. Graph by the authors.
Two observations can be made from figure 3. First, for each group of voters, turnout increases as people grow older (analog to figure 1). While turnout levels increase as all groups of voters age, figure 3 also shows that the starting levels of voter turnout are different for groups of voters coming of age under different Presidents. For example, turnout among first-time voters was much lower in Reagan’s 1980 election (36%) compared to Clinton’s 1992 election (42%). This points to a cohort or generation effect. The idea behind the cohort/generation approach is that it is not so much the ageing as such that make one grow in political life, but rather that social, cultural, political and historical events shape the political participation patterns of people born in a particular era. These events affect younger generations disproportionally because they have not yet developed political habits. Events, moreover, have a long-lasting (positive or negative) effect on later patterns of political behavior.
Newly eligible voters do not all face the same political context. In a recent article, we find that the character of the first elections an incoming group of voters faces is crucial to this cohort’s future turnout levels. For example, the 2008 election was a high-stake election characterized by generally high turnout levels, which in turn mobilized voters as everybody around them voted. Moreover, Obama’s victory was marginal and the two parties had quite polarizing views on many political issues, which again in general boosts turnout. In contrast, Clinton’s reelection in 1996 can be considered a low-stake election. Clinton’s landslide victory of 379 electoral votes and the general low turnout of 49%, should have a negative effect on the generation that got to know electoral politics in this relatively “boring” election.
This leaves us with the question of whether the Millennial generation is expected to pick up political engagement or not. While we do not have a crystal ball, we believe it is not too late for this generation just yet. First, the Millennials are (still) growing up in a politicized environment, which should make them more likely to develop a habit of voting. Second they still need to make the transition to adulthood (i.e. the aging effect still needs to take place). At the moment, it is “too close to call” how this generation will turn out.
For people interested in disentangling the various effects of age – which political scientists refer to as “Age, Period, Cohort” effects – there is a host of new academic articles available on the topic in the special issue of Electoral Studies already mentioned above. For example, Larry Bartels and Simon Jackman employ a mathematical model of political learning to partisanship to estimate the critical years of heightened impact of period-specific effects or historical shocks on a citizen’s party attachments. In another example, Grigore Pop-Eleches and Joshua Tucker categorize historical socialization periods that identify aging and generational effects of experiences during Communism on people’s democratic and economic attitudes.
About the authors:
Anja Neundorf is Lecturer in Politics and Research Methods at the School of Politics and International Relations of The University of Nottingham.
Kaat Smets is Lecturer in Politics (Quantitative Methods) at the Department of Politics and International Relations of Royal Holloway, University of London.
Note: This originally appeared on