More Polarized but More Independent

Abstract

In three nationally representative surveys of U.S. residents (N = 10 million) from 1970 to 2015, more Americans in the early 2010s (vs. previous decades) identified as Independent, including when age effects were controlled. More in the early 2010s (vs. previous decades) expressed polarized political views, including stronger political party affiliation or more extreme ideological self-categorization (liberal vs. conservative) with fewer identifying as moderate. The correlation between party affiliation and ideological views grew stronger over time. The overall trend since the 1970s was toward more Americans identifying as Republican or conservative. Older adults were more likely to identify as conservative and Republican. More Millennials (born 1980-1994) identify as conservative than either GenXers or Boomers did at the same age, and fewer are Democrats compared with Boomers. These trends are discussed in the context of social identification processes and their implications for the political dynamics in the United States.

Keywords

political views political party affiliation ideological self-categorization age generations

No sooner did the United States become an independent nation than it created two political parties that fought for the “hearts and minds” of its citizenry. Although the names and ideological profiles of the parties have changed over time, a two-party system continues to provide a basis for the political identification of Americans. Even when modern-day Americans identify themselves as Independents, they do so in reference to two parties: Republicans or Democrats.

Subjective identification with a political party has far-reaching consequences. It provides a framework for understanding of and orientation in the political world (G. Cohen, 2003) and shapes that world by informing political decision-making and voting behavior (Bartels, 2000). Political identification, perhaps more than any other social category, evokes open competition for positive distinctiveness (Tajfel & Turner, 2010), giving credence to Plato’s claim that humans are political animals.

Because political identification is consequential, it is important to understand its developmental trajectory and its shifts over historical time. If political identification changes over time, does it follow a recognizable developmental pattern as people age and/or do people move with the generational or cultural zeitgeist? If it is stable, is it equally strong over time? Knowing answers to these questions could inform predictions about voting behavior and by extension, sociopolitical climate.

Complicating projections about political outcomes is the issue of ideological self-categorization (e.g., liberal or conservative) and its temporal (in)stability. Party affiliation is a proxy but not a perfect indicator of ideological self-categorization. The latter stems from “beliefs about the proper order of society and how it can be achieved” (Erikson & Tedin, 2003, p. 64). It is typically classified along a single liberal (left)—conservative (right) dimension (Jost, Federico, & Napier, 2009), depending on preferences for social change and equality versus tradition and acceptance of inequality. Although individuals outside of political elites may not always position themselves consistently along the liberal–conservative dimensions, their self-categorization on this dimension is a potent predictor of their voting intentions (Jost, 2006).

In addition, models from cultural psychology suggest that cultural change may shift political views, creating change over time in these beliefs (Markus & Kitayama, 2010). Generational differences generally arise from cultural change (Twenge, 2014), with these differences demonstrating the effect of the larger culture on the individual. Generational differences in political party preferences and ideological self-categorization are currently uncertain. Several polls (e.g., Newport, 2014) indicate that Millennials are more likely to identify as Democrats and as liberals compared with Boomers (born 1946-1964) and Generation Xers (GenX; 1965-1979).¹ Even among Millennials who identify as Republicans, fewer describe themselves as conservatives compared with other generations (Kiley & Dimock, 2014). However, one-time polls cannot separate the effects of age and generation. Millennials may be more liberal and more likely to be Democrats because they are young, and not because their generation is more left-leaning. As the popular saying goes, “If you are not a liberal at 25, you have no heart. If you are not a conservative at 35 you have no brain.”

Generational trends are better understood by examining generational differences in ideological self-categorization using data collected over time, preferably with age held constant or statistically controlled. For example, a time-lag design compares respondents of the same age in different years, holding age constant while time period and generation vary (Schaie, 1965). Other studies examine people of many ages across time; this type of data can be analyzed using a special form of hierarchical linear modeling (HLM) known as age–period–cohort (APC) analysis (Yang, 2008; Yang & Land, 2013). APC analysis separates the effects of age from those of generation or cohort (defined as change that affects only the young) and time period (defined as change that affects everyone equally; Campbell, Campbell, Siedor, & Twenge, 2015).

Time Period and Generational Differences in Political Affiliation and Ideological Self-Categorization

Most observations about generational political tendencies in the United States rely on one-time polls or anecdotal evidence rather than surveys taken across time. For example, generational observers Strauss and Howe (1991) characterize GenX as more Republican-leaning, conservative, and patriotic than the “hippie” Boomers who proceeded them. Millennials are seen as more liberal and Democrat than GenXers (Kiley & Dimock, 2014). These perceptions suggest that Boomers and Millennials are the more liberal and Democratic generations, whereas GenXers are more Republican and conservative. However, to our knowledge, this claim has never been tested using data collected at multiple times that can separate the effects of age from those of time period and generation.

In addition to possible generational trends in affiliation with the two traditional parties, some recent evidence also suggests that more Americans are identifying as Independents. The previously observed trend toward party loyalties and partisanship (e.g., Bartels, 2000) could be ending, or it was initially constricted and limited by the frame of reference when the analysis was conducted. This appears particularly evident among Millennials (Jones, 2015; Pew Research Center, 2015). Once again, however, these one-time polls cannot separate the effects of age, period, and cohort, raising the possibility that young people might have always been more likely to identify as Independents compared with older people. Alternatively, perhaps disassociating from political parties and political institutions in general has increased over time (Dugan, 2015) in concert with the established reduction of confidence in nearly all major societal institutions (Riffkin, 2014; Twenge, Campbell, & Carter, 2014). Thus, the observed increase in the number of Americans declaring themselves Independent in the political arena may well reflect these broader cultural trends.

Age and Political Affiliation and Ideological Self-Categorization

Any attempt at a projection of generations’ political trajectory must necessarily address the role of age in political party identification and ideological self-categorization. Previous research has yielded inconsistent findings. Some research suggests that aging has an impact on party affiliation independent of generational factors. Using cross-sectional nationwide surveys from 1946 to 1958, Crittenden (1962) concluded that age seemed to account for a noticeable shift toward Republicanism. Subsequent reanalysis of the same data, however, suggested that the shift toward Republicanism was instead attributable to generation (Cutler, 1969). Further corroborating this interpretation, a broader data set spanning 24 years (1945-1969) determined that aging had no effect on political ideology (Glenn & Hefner, 1972). Subsequent research found inconclusive if not disconfirming evidence for prevailing stereotypes about a trend toward conservatism with age (Danigelis & Cutler, 1991; Danigelis, Hardy, & Cutler, 2007; Glenn, 1974). Adding to this complexity is the possibility that the association may vary for particular segments of the population (Saltzman Chafetz, & Ebaugh, 1983) and with economic cycles (Milojev, Greaves, Osborne, & Sibley, 2015).

Contemporary investigations on the role of age in party identification and ideological self-categorization found a mix of similarities and differences between the young and old. There is evidence that compared with the youth, older people have comparable percentages of liberals/Democrats but higher percentages of conservatives/Republicans and lower percentages of moderates and disengaged (DeSilver, 2014; Pew Research Center, 2014a). Other evidence indicates that generational imprinting is the strongest and most reliable predictor of ideological self-categorization and party affiliation over the life span—at least among White voters. Using Presidential polling surveys from 1952 to 2012, Ghitza and Gelman (2014) claim that political events happening over a person’s life span inform his or her voting, with the events during teenage and young adult years having the most impact.

Understanding age, time period, and generational differences in political affiliation and ideological self-categorization is important because each of these factors can have a separate but related role in shaping the political landscape of the United States. Unless these dynamics are parsed out, it is impossible to understand current, much less future, impacts of different groups in society on the political composition of the United States. For example, mistaking generational for age trends may lead to an erroneous (and costly) conclusion that “young people will grow out of their current political leanings.” Conversely, mistaking age for generational trends may lead to another erroneous (and equally costly) conclusion that “this generation will stay with their current political leanings.”

Time Period and Generational Differences in Political Polarization

Political polarization is often defined as diverging ideology between members of the public on specific issues (Fiorina, 2014), or as increasing separation between groups that are clearly defined (Hill, 2005; Hill & Tausanovitch, 2015), such as political parties (e.g., Republicans vs. Democrats) or ideological categories (e.g., liberals vs. conservatives). Several observers have speculated that political polarization has increased in recent decades, with Americans abandoning moderation in favor of more extreme beliefs (e.g., Hill, 2005). For example, a 2014 Pew Center poll found that Democrats have become more liberal and Republicans more conservative, with more individuals placing themselves at the polar opposite ends of the political spectrum than in 1994 (Pew Research Center, 2014b). However, this study did not examine data before 1994, nor did it attempt to separate the effects of age, time period, and generation on political polarization. In addition, some authors claim that political polarization has not increased, arguing that there are just as many moderates and more Independents in recent years than in the past (Fiorina, 2014; Fiorina & Abrams, 2008), and thus no evidence that the public has polarized in its policy views over time (Hill & Tausanovitch, 2015).

This seeming inconsistency may reflect lack of conceptual clarity. For example, when conceptualized as issue partisanship or the association between issue attitudes and party identification, polarization appears much more evident than when conceptualized as issue alignment or the association between pairs of issues (Baldassarri & Gelman, 2008). For the purposes of this study, we conceptualized polarization as extremity of party and ideological self-categorization, as these variables are important in their own right. (For example, identifying as “extremely conservative” is more polarized than identifying as simply “conservative.”) These factors, regardless of how they may be related to policy positions, reflect movement that pushes these groups apart. Arguably, the move apart may be integrally associated with an assertion of uniqueness by the groups as they seek to establish and assert distinct identities.

The Current Study

The current study aims to provide a comprehensive overview of age, time period, and generational (cohort) differences in political party affiliation (Democrat, Republican, Independent) and ideological self-categorization (liberal, conservative, moderate) and the relationship between the two. We also explore whether political polarization (i.e., the extremity of ideological self-categorization) has increased, examining those who identify as “strongly” affiliated with a party or “very” or “extremely” conservative or liberal, versus those identifying as moderate or only slightly liberal or conservative. To reach these goals, we draw from three large, nationally representative surveys conducted over several decades. We analyze time period and generational effects among 12th graders in the Monitoring the Future (MtF) study covering 1976-2014 (N = 563,102), among entering college students from the American Freshman (AF) survey covering 1970-2015 (N = 9.4 million). Both of these studies use time-lag designs with age held constant, so any variation must be due to generation or time period. Finally, we examine age, time period, and generational differences among adults 18 and older from the General Social Survey (GSS) covering 1972-2014 (N = 58,542) using new HLM techniques (APC models) that can separate the three effects (Yang, 2008).

Method

Samples

MtF (Johnston, Bachman, O’Malley, & Schulenberg, 2015) has surveyed a nationally representative sample of high school seniors (12th graders) every year since 1976. The most recent year available for the purpose of this analysis was 2014 (N 1976-2014 = 563,102). High schools and the students within those schools are chosen using a random sampling procedure to represent a cross-section of the population in terms of region, gender, race, and socioeconomic status (SES). Similar schools are replaced for those that decline to participate. Participation rate is between 66% and 80% for schools and between 79% and 83% for students (Johnston, Bachman, O’Malley, Schulenberg, & Miech, 2014).

The AF project, part of the Cooperative Institutional Research Program (CIRP) administered by the Higher Education Research Institute (HERI), has surveyed a nationwide sample of first-year students at 4-year colleges or universities every year since 1966 (e.g., Pryor, Hurtado, Saenz, Santos, & Korn, 2007), with the political questions first asked in 1970 and 2015 the most recent year available. Originally, some 2-year colleges participated, but the data—including that for past years—are now reported only for students at 4-year institutions (N 1970-2015 = 9,407,435). Most respondents are 18 years old. Data for campuses are included only if more than 75% of first-time full-time freshman students participated. The survey weights its results to be demographically representative of all first-year students at 4-year institutions in the United States; this population is more female, more White, and higher in SES than the population of those who do not attend college. We obtained the aggregated data from HERI’s publicly available research reports.

The GSS is a nationally representative sample of U.S. adults, collected in most years between 1972 and 2014 (N = 58,893; for the questions in the current survey, N ranges from 49,789 to 58,542). The GSS data and codebooks are available online (see Smith, Marsden, Hout, & Kim, 2015). As recommended by the survey administrators, we excluded the Black oversamples from 1982 and 1987 and weighted by the variable WTSSALL, which makes the sample representative of individuals instead of households and corrects for other sampling biases (Smith et al., 2015). However, differences between weighted and unweighted analyses were minor.

Measures

Twelfth graders (MtF)

MtF included two questions on political party affiliation and ideological self-categorization asked of all participants in each year. The first was as follows: “How would you describe your political preference?” Response choices were 1 = “strongly Republican,” 2 = “mildly Republican,” 3 = “mildly Democrat,” 4 = “strongly Democrat,” 5 = “American Independent Party,” 6 = “no preference, independent,” 7 = “other,” and 8 = “don’t know, haven’t decided.”

In 1996, “American Independent Party” was changed to “Independent” and “no preference, independent” was changed to “no preference.” We recoded party affiliation so Republican corresponds to higher numbers, with 1 = “strongly Democrat,” 2 = “mildly Democrat,” 3 = “mildly Republican,” and 4 = “strongly Republican.” Given the wording change, for the initial analyses of party affiliation, we excluded “American Independent Party,” “Independent,” “no preference, independent,” “no preference,” “other,” and “don’t know” responses. In secondary analyses, we examined “Independent” and “no preference” responses after 1996 when the wording became consistent, still excluding “other” and “don’t know” responses.

The second item assessing ideological self-categorization was as follows: “How would you describe your political beliefs?” Response choices were 1 = “very conservative,” 2 = “conservative,” 3 = “moderate,” 4 = “liberal,” 5 = “very liberal,” 6 = “radical,” and 8 = “none of the above, or don’t know.” We recoded ideological self-categorization so conservative corresponds to higher numbers, with 1 = “radical,” 2 = “very liberal,” 3 = “liberal,” 4 = “moderate,” 5 = “conservative,” and 6 = “very conservative.” We excluded “none of the above, or don’t know” from the primary analyses.

To examine political polarization, we recoded the variables to reflect the percentage of respondents who answered in the extreme categories. For party affiliation, that included those responding “strongly Democrat” and “strongly Republican” (coded 1) versus those choosing “mildly Democrat” or “mildly Republican” (coded 0). For ideological self-categorization, this included those responding “radical,” “very liberal,” or “very conservative” (coded 1) with “conservative,” “liberal,” or “moderate” (coded 0).

Entering college students (AF)

Starting in 1970 and every year through 2015, the survey asked, “How would you characterize your political views?” with response choices of “far left,” “liberal,” “middle of the road,” “conservative,” and “far right.” We combined “far left” and “liberal” into a single liberal category, and “far right” and “conservative” into a single conservative category. We also combined “far left” and “far right” into an “extreme views” category. The survey did not include a question on party affiliation.

Adults (GSS)

For party affiliation, the GSS asked since 1972, “Generally speaking, do you usually think of yourself as a Republican, Democrat, Independent, or what?” with response choices of 0 = strong Democrat; 1 = not strong Democrat; 2 = independent, near Democrat; 3 = independent; 4 = independent, near Republican; 5 = not strong Republican; 6 = strong Republican; 7 = other party; and 8 = don’t know. For the primary analyses, we excluded “other party” and “don’t know” responses. We recoded the 0 to 6 scale to a 1 to 7 scale.

For ideological self-categorization, the GSS asked beginning in 1974, “We hear a lot of talk these days about liberals and conservatives. I’m going to show you a seven-point scale on which the political views that people might hold are arranged from extremely liberal (point 1) to extremely conservative (point 7). Where would you place yourself on this scale?” with choices of 1 = extremely liberal, 2 = liberal, 3 = slightly liberal, 4 = moderate, 5 = slightly conservative, 6 = conservative, 7 = extremely conservative, and 8 = don’t know. We excluded “don’t know” responses. Political polarization was coded as in the other data sets, with “strong Democrat” and “strong Republican” combined and “extremely liberal” and “extremely conservative” combined.

Procedure

Data collected over time can be analyzed in many ways, including grouping by 20-year generation blocks, by decades, or by individual year. As a compromise between specificity and breadth, we separated the data into 5-year intervals in the tables. We report the effect sizes (d) comparing the earliest and latest year blocks and the maximum d using the lowest and highest numbers. We also include figures with all of the year-by-year data for most variables. We present some results by generation for the 12th-grade and entering college student samples.

We report both continuous variables (e.g., the 1-7 scale for liberal to conservative in the GSS) and dichotomous variables (e.g., the percentage who identify as Republicans). Given the very large Ns, we focus on effect sizes rather than statistical significance testing. With the large sample sizes included here, virtually all ds with absolute values of .03 or larger were statistically significant at p < .05 (and ds of .01 or larger in AF).

For the APC models, we estimated random coefficient models allowing intercepts to vary across time periods (years) and generations (cohorts). Thus, an intercept (mean political affiliation) score is calculated (using empirical Bayes estimation to correct for unreliability) for each survey year and cohort. These intercepts represent the effects of time period and cohort after controlling for each other and both linear and quadratic effects of age. This model has three variance components: one for variability in cohort effects (τ_u0), one for variability in period effects (τ_v0), and a residual term containing unmodeled variance within cohorts and periods. We used 10-year intervals for cohorts as they provided reliable cohort effect estimates at a much less cost to specificity than it would have been the case with 20-year cohorts. Variance in time period and cohort effects indicates meaningful period and cohort differences, respectively (Yang & Land, 2013). Thus, the technique allows for the statistical separation of the effects of generation/cohort, time period, and age. Age and age-squared were entered as fixed effects predictors. Weighting could not be used for the mixed-effects analyses because proper probability weighting for variance component estimation requires taking into account pairwise selection probabilities, which is not possible in current statistical software (i.e., R).

Results

Political Party Affiliation and Ideological Self-Categorization

As 12th graders, Millennials in the 2000s and early 2010s were slightly more likely to identify as Republican and conservative than Boomers were at the same age in the 1970s; a similar pattern appeared among entering college students and adults (see Tables 1 -3 and Figures 1 -4). Over the entire time period, Americans in the late 1970s favored Democrats the most of any era, followed by the early Obama years in the late 2000s. However, the Reagan era of the 1980s saw the largest percentage of Republicans. Comparing generation groups, 52% of Millennials were Democrats, compared with 60% of Boomers and 49% of GenXers. Thus, Millennials were less likely to be Democrats than Boomers (d = −.15) but more likely than GenXers (d = .07). The percentage of 12th graders identifying as Republican increased after 2008, rising from 42% in 2008 to 51% in 2012 and 49% in 2014 (see Figure 1; this includes only those identifying as either Democrat or Republican). Adults’ political party affiliation follows a similar pattern over time as 12th graders, with Democrats in the clear majority in the 1970s and Republicans winning out in the 1980s (see Figure 5A).

Table 1.

American 12th Graders’ Political Party Affiliation and Ideological Self-Categorization, Monitoring the Future Survey, 1976-2014.

Variable	n	76-79	80-84	85-89	90-94	95-99	00-04	05-09	10-14	d (maximum)	d (period)
Political party
Democrat vs. Republican (1-4; high = Republican)	270,259	2.31 (0.95)	2.42 (1.01)	2.58 (1.04)	2.52 (1.05)	2.46 (1.02)	2.43 (1.06)	2.43 (1.06)	2.50 (1.05)	.26	.19
Democrat	270,259	58% (0.49)	52% (0.50)	44% (0.50)	48% (0.50)	51% (0.50)	52% (0.50)	53% (0.50)	51% (0.50)	−.29	−.16
Strongly Democrat	270,259	22% (0.42)	22% (0.41)	20% (0.40)	22% (0.41)	21% (0.41)	24% (0.43)	24% (0.43)	21% (0.41)	.11	−.03
Mildly Democrat	270,259	36% (0.48)	30% (0.46)	24% (0.43)	26% (0.44)	30% (0.46)	28% (0.45)	28% (0.45)	29% (0.46)	−.27	−.14
Republican	270,259	42% (0.49)	48% (0.50)	56% (0.50)	52% (0.50)	49% (0.50)	48% (0.50)	47% (0.50)	49% (0.50)	.29	.16
Strongly Republican	270,259	12% (0.32)	16% (0.37)	21% (0.41)	21% (0.41)	19% (0.39)	20% (0.40)	20% (0.40)	22% (0.41)	.27	.27
Mildly Republican	270,259	30% (0.46)	31% (0.46)	35% (0.48)	31% (0.46)	30% (0.46)	28% (0.45)	27% (0.44)	28% (0.45)	−.17	−.05
Strong affiliation (Democrat or Republican)	270,259	34% (0.47)	38% (0.49)	41% (0.49)	43% (0.49)	40% (0.49)	44% (0.50)	44% (0.50)	43% (0.50)	.21	.18
Independent	184,749	—	—	—	—	14% (0.35)	13% (0.33)	13% (0.34)	16% (0.37)	.09	.06
No preference	184,749	—	—	—	—	27% (0.44)	22% (0.42)	18% (0.39)	21% (0.41)	−.22	−.14
Other party	431,505	2% (0.14)	2% (0.13)	2% (0.14)	3% (0.16)	2% (0.14)	2% (0.15)	2% (0.15)	3% (0.16)	.06	.05
Don’t know	561,454	25% (0.43)	27% (0.45)	23% (0.42)	20% (0.40)	23% (0.42)	19% (0.39)	22% (0.42)	25% (0.43)	−.19	−.02
Political beliefs
Liberal to moderate to conservative (1-6)	393,838	3.77 (1.04)	3.87 (1.06)	3.86 (1.08)	3.81 (1.14)	3.78 (1.15)	3.79 (1.14)	3.84 (1.14)	3.90 (1.15)	.12	.12
Liberal	393,838	35% (0.48)	30% (0.46)	31% (0.46)	34% (0.48)	36% (0.48)	36% (0.48)	36% (0.48)	34% (0.47)	.13	−.02
Moderate	393,838	44% (0.50)	46% (0.50)	43% (0.50)	40% (0.49)	40% (0.49)	40% (0.49)	38% (0.48)	37% (0.48)	−.17	−.14
Conservative	393,838	21% (0.41)	24% (0.43)	25% (0.44)	25% (0.43)	24% (0.43)	24% (0.43)	27% (0.44)	29% (0.45)	.18	.18
Very liberal or radical	393,838	9% (0.29)	9% (0.28)	9% (0.29)	11% (0.31)	12% (0.32)	12% (0.32)	11% (0.32)	10% (0.30)	.04	.11
Very conservative	393,838	4% (0.19)	5% (0.21)	5% (0.22)	6% (0.24)	6% (0.24)	6% (0.24)	7% (0.26)	8% (0.27)	.19	.19
Very liberal or conservative	393,838	13% (0.34)	13% (0.34)	14% (0.35)	17% (0.38)	18% (0.38)	18% (0.38)	18% (0.39)	19% (0.39)	.15	.15
None of the above or don’t know	563,102	26% (0.44)	30% (0.46)	29% (0.45)	26% (0.44)	31% (0.46)	26% (0.44)	33% (0.47)	39% (0.49)	.27	.27

Note. d (maximum) = difference in standard deviations from the lowest group of years to the highest; d (period) = difference in standard deviations from the early 1970s to the 2010s. SDs given in parentheses.

Table 2.

American Entering College Students’ Political Views, American Freshman Survey, 1970-2015.

Political views	n	70-74	75-79	80-84	85-89	90-94	95-99	00-04	05-09	10-15	d (maximum)	d (period)
Far left or liberal	9,407,436	37% (0.48)	28% (0.45)	22% (0.42)	25% (0.43)	28% (0.45)	26% (0.44)	28% (0.45)	32% (0.46)	31% (0.46)	−.33	−.13
Middle of the road	9,407,436	47% (0.50)	54% (0.50)	57% (0.49)	52% (0.50)	50% (0.50)	52% (0.50)	50% (0.50)	44% (0.50)	47% (0.50)	−.26	.00
Far right or conservative	9,407,436	16% (0.37)	18% (0.38)	21% (0.41)	22% (0.42)	22% (0.42)	22% (0.42)	22% (0.42)	24% (0.43)	22% (0.42)	.20	.15
Far left	9,407,436	2.5% (0.16)	1.8% (0.13)	1.7% (0.13)	1.9% (0.14)	2.1% (0.14)	2.4% (0.16)	2.9% (0.17)	3.0% (0.17)	2.8% (0.16)	.09	.02
Far right	9,407,436	.7% (0.07)	.8% (0.08)	1.1% (0.10)	1.4% (0.12)	1.4% (0.12)	1.5% (0.12)	1.6% (0.12)	1.8% (0.13)	1.7% (0.13)	.11	.11
Strong/extreme views (far left or far right)	9,407,436	3.2% (0.17)	2.6% (0.15)	2.9% (0.17)	3.3% (0.17)	3.5% (0.18)	3.9% (0.19)	4.5% (0.20)	4.7% (0.21)	4.7% (0.21)	.12	.08

Table 3.

American Adults’ Political Party Affiliation and Ideological Self-Categorization, General Social Survey, 1972-2014.

Variable	n	72-74	75-79	80-84	85-89	90-94	95-99	00-04	05-09	10-14	d (maximum)	d (period)
Political party
Democrat vs. Republican (1-7; High = Republican)	57,633	3.39 (1.98)	3.46 (1.90)	3.60 (1.95)	3.76 (2.05)	3.88 (2.04)	3.82 (1.94)	3.86 (2.00)	3.79 (1.97)	3.70 (1.93)	.24	.16
Democrat	57.633	46% (0.50)	42% (0.49)	39% (0.49)	38% (0.49)	36% (0.48)	33% (0.47)	33% (0.47)	33% (0.47)	33% (0.47)	−.27	−.27
Republican	57,633	23% (0.42)	22% (0.42)	24% (0.43)	29% (0.46)	31% (0.46)	28% (0.45)	28% (0.45)	27% (0.44)	23% (0.42)	.20	.00
Independent (Independent Near Democrat, Independent Near Republican, Independent)	57,633	31% (0.46)	36% (0.48)	37% (0.48)	32% (0.47)	33% (0.47)	39% (0.49)	38% (0.49)	41% (0.49)	44% (0.50)	.27	.27
Strongly Democrat	57,633	19% (0.39)	16% (0.37)	16% (0.36)	16% (0.37)	14% (0.35)	13% (0.34)	15% (0.35)	16% (0.37)	16% (0.37)	−.17	−.08
Strongly Republican	57,633	8% (0.27)	7% (0.25)	8% (0.28)	11% (0.31)	12% (0.32)	10% (0.30)	12% (0.33)	11% (0.32)	10% (0.30)	.17	.07
Strong/extreme views (Strongly Democratic or Republican)	57,633	26% (0.44)	23% (0.42)	24% (0.43)	27% (0.44)	26% (0.44)	23% (0.42)	27% (0.44)	27% (0.45)	26% (0.44)	.07	.00
Weak views (not strong Democrat or not strong Republican)	57,633	42% (0.49)	42% (0.49)	39% (0.49)	41% (0.49)	41% (0.49)	38% (0.49)	35% (0.48)	32% (0.47)	30% (0.46)	.26	.26
Other party	58,542	4.3% (0.20)	.05% (0.07)	1.0% (0.10)	.07% (0.08)	1.1% (0.10)	1.8% (0.13)	1.5% (0.12)	1.6% (0.12)	2.5% (0.16)	−.30	−.10
Ideological self-categorization
Liberal vs. conservative (1-7; High = Conservative)	49,789	3.98 (1.29)	4.04 (1.31)	4.16 (1.28)	4.12 (1.33)	4.15 (1.37)	4.17 (1.37)	4.16 (1.40)	4.15 (1.40)	4.12 (1.43)	.14	.10
Liberal	49,789	31% (0.46)	29% (0.45)	25% (0.43)	26% (0.44)	27% (0.44)	26% (0.44)	26% (0.44)	26% (0.44)	27% (0.44)	−.14	−.09
Moderate	49,789	40% (0.49)	39% (0.49)	41% (0.49)	39% (0.49)	37% (0.48)	38% (0.48)	39% (0.49)	39% (0.49)	39% (0.49)	−.04	−.02
Conservative	49,789	30% (0.46)	32% (0.47)	34% (0.48)	35% (0.48)	36% (0.48)	35% (0.48)	35% (0.48)	35% (0.48)	34% (0.47)	.13	.09
Extremely liberal	49,789	1.6% (0.12)	2.4% (0.15)	2.3% (0.15)	2.3% (0.15)	2.5% (0.16)	2.3% (0.15)	3.5% (0.18)	2.9% (0.17)	3.7% (0.19)	.12	.12
Extremely conservative	49,789	2.4% (0.15)	2.2% (0.15)	2.9% (0.17)	2.5% (0.16)	3.3% (0.18)	3.3% (0.18)	3.6% (0.19)	3.8% (0.19)	4.2% (0.20)	.11	.09
Strong/extreme views (extremely liberal or extremely conservative)	49,789	4.0% (0.20)	4.6% (0.21)	5.2% (0.22)	4.8% (0.21)	5.8% (0.23)	5.6% (0.23)	7.1% (0.26)	6.7% (0.25)	7.9% (0.27)	.15	.15
Moderate views (slightly liberal or slightly conservative, or moderate)	49,789	71% (0.45)	71% (0.45)	73% (0.45)	69% (0.46)	68% (0.47)	67% (0.47)	66% (0.47)	65% (0.47)	64% (0.48)	−.19	−.15
Don’t know	51,800	4.5% (0.21)	4.8% (0.21)	3.0% (0.17)	4.3% (0.20)	3.3% (0.18)	5.1% (0.22)	4.0% (0.20)	3.3% (0.18)	3.2% (0.18)	−.08	−.07

Figure 1.

American 12th graders’ political party affiliation, 1976-2014; high = Republican.

Figure 2.

Percentage of American adults identifying as Democrats, Republicans, and Independents, 1972-2014.

Figure 3.

American 12th graders’ ideological self-categorization, 1976-2014.

Figure 4.

American entering college students’ ideological self-categorization, 1970-2015.

Figure 5.

(A) Time period and (B) cohort (generational) effects in political party affiliation (Republican = higher), controlling for age, U.S. adults, 1972-2014.

Between 2002 and 2014, 12th graders’ ideological self-categorization also became more conservative, d = .16 (see Figure 3). In 2002, 23% identified as conservative, which rose to 28% in 2014. Among entering college students, only 15% identified as conservative or far right in 1971, which rose to 26% in 2006, declining to 22% in 2015. The largest change was a substantial conservative shift between the early 1970s and the early 1980s (see Figure 4). Nevertheless, more 12th graders, and just as many college students, identified as conservative in the early 2010s as did in the early 1980s, a time often characterized as conservative.

By generation groups, 20% of Boomers, 25% of GenXers, and 27% of Millennials identified as conservative as 12th graders. As entering college students, 17% of Boomers, 22% of GenXers, and 23% of Millennials identified as far right or conservative. Thus, Millennials were more conservative than both Boomers (d = .14 for 12th graders, d = .15 for college students) and GenXers (d = .04, d = .03). Among adults, the number identifying as liberal declined slightly between the 1970s and the 2010s, and the number identifying as conservative rose slightly (see Table 3).

The APC model for political party on the GSS adults showed relatively weak effects of age and age-squared (b = .002, .0001; t = 1.55, 3.73, respectively). However, there was substantial variance in intercepts for both time period (SD = .16 [95% CI .11, .20]) and cohort (SD = .13 [.07, .20]). Examining party as a continuous variable (with positive numbers = Republican, and negative = Democrat), the largest time period shift appeared between 1972 (Z = −0.15) and 1990 (Z = 0.10), with Republicans gaining d = .25 (see Figure 5A). Between 1990 and 2012 (Z = −0.07), the trend reversed toward more Democrats, d = −.17.

The cohort born in the 1920s had the most Democrats (Z = −0.12), whereas GenXers born in the 1960s had the most Republicans (Z = 0.07), d = .19 (see Figure 5B). Millennials born in the 1980s (Z = −0.04) were more likely to identify as Democrats than GenXers born in the 1960s, d = −.11. Two-decade averages approximating the generations reveal a similar pattern: Silent generation (1920s-1930s, Z = −0.09), Boomers (1940s-1950s, Z = −0.02), GenX (1960s-1970s, Z = 0.05), and Millennials (1980s-1990s, Z = −0.03). Thus, Millennials do not appear to be unusually Democratic, falling near the average across all generations when age and time period are controlled.

The largest change over time in adults’ political affiliation is the increase in Independents. Only 30% of adults identified as Independents in 1989, compared with 46% in 2014, an all-time high (see Figure 2). The logistic APC model predicting Independent party identification showed sizable effects of age (odds ratio = 0.98, Z = −18.40) and age-squared (odds ratio = 1.0001, Z = 4.18) indicating that younger people—controlling for time period and cohort—are more likely to identify as independent (49% at age 18, 40% at age 32, 30% at age 55, and 22% at age 89). In addition, although there was variation in intercepts due to both time period (SD = .19 [.14, .24]) and cohort (SD = .08 [.03, .11]), most of the variation was due to time period. In terms of time period effects, only 26% of adults were Independents in 1972, but their numbers rose to 44% in 2014, also an all-time high. Cohort/generation explains very little variance in the increase in Independents, with Silents born 1920s-1930s (34%), Boomers born 1940s-1950s (36%), GenX 1960s-1970s (33%), and Millennials 1980s-1990s (33%).

Given the constraints in the data set, we were only able to examine 12th graders identifying themselves as “Independent” or expressing “No preference” since 1996. Twelfth graders identifying as Independents rose from 11% to 18% between 2005 and 2014 (see Figure 6).

Figure 6.

Percentage of American 12th graders identifying as Independent or expressing no party preference, 1996-2014.

The APC model predicting ideological self-categorization showed large effects of age (b = .01, t = 16.22) and age-squared (b = −.001, t = −6.19), indicating that—controlling for time period and cohort—older people are more politically conservative on average (see Figure 7). Specifically, Americans, on average, identified as more politically liberal (Z = −0.24) at age 18, increasingly more conservative from their 20s to their 60s (Z₂₀ = −0.22, Z₃₀ = −0.10, Z₄₀ = −0.01, Z₅₀ = 0.07, Z₆₀ = 0.12, Z₇₀ = 0.16) and largely stabilized after 70 (Z₈₀ = 0.17, Z₈₉ = 0.17). There was some rather minimal variation in intercepts due to both time period (SD = .05 [.03, .07]) and cohort (SD = .05 [.02, .08]). Time period effects for ideological self-categorization (conservative = higher) shifted the most between 1975 (Z = −0.02) and 1996 (Z = 0.09), d = .11, becoming more liberal again in 2012 (Z = 0.00) and 2014 (Z = 0.01). Generational effects were also small, with those born in the 1950s the most liberal (Z = −0.01) and those born in the 1960s the most conservative (Z = 0.09). Millennials born in the 1980s were close to the average (Z = 0.02) and those born in the 1990s slightly more conservative (Z = 0.06). Two-decade averages approximating the generations reveal a similar pattern: Silent (1920s-1930s, Z = 0.06), Boomers (1940s-1950s, Z = −0.01), GenX (1960s-1970s Z = 0.06), and Millennials (1980s-1990s, Z = 0.04). This suggests a generational shift from liberal to conservative between Boomers born in the 1950s and GenXers born in the 1960s, a move back toward liberalism for those born in the 1970s and 1980s, and a slight return to conservatism for those born in the 1990s.

Figure 7.

Age effects on (A) political party (Republican = higher) and (B) ideological self-categorization (conservative = higher), U.S. adults, aged 18 to 96.

Political Polarization

Millennial 12th graders and entering college students in the early 2010s were more politically polarized than Boomers and GenXers were at the same age (see Tables 1 and 2, and Figures 8 and 9). More 12th graders in the 2010s identified as strong Democrats or strong Republicans than in earlier decades. Fewer expressed no preference for a political party, declining from 29% in 1998 to 23% in 2014 (see Figure 2). In addition, fewer 12th graders and entering college students identified as moderates in the 2010s compared with the late 1970s and 1980s (see Figures 3 and 4) and more identified as radical, very liberal, or very conservative (see Tables 1 and 2, and Figure 9). By generation groups, 50% of college student Boomers identified as middle-of-the-road, compared with 53% of GenXers and 47% of Millennials.

Figure 8.

Percentage of American 12th graders identifying as strongly Democrat or strongly Republican, 1976-2014.

Figure 9.

Percentage of American 12th graders identifying as “very” liberal/radical or “very” conservative, 1976-2014.

The same was true for adults, with twice as many identifying as “extremely” liberal or conservative between the 1970s and the 2010s, and slightly fewer identifying as moderate or only “slightly” liberal or conservative (see Table 3 and Figure 8). Although the number of moderates among adults has not changed, the number identifying as only slightly liberal or conservative has decreased, suggesting a move away from the center similar to that found among 12th graders. In the APC analyses, moderates (combining slightly liberal, slightly conservative, and moderate) decreased from 73% in 1978 to 65% in 2010 (it then rose slightly to 67% in 2014).

For the most part, this polarization was primarily driven by those on the right (Republicans and conservatives), even though those on the left (Democrats and liberals) still have a greater proportion identifying in the extreme than those on the right. More 12th graders and college students in recent years (vs. previous eras) identified as strongly Republican, very conservative, or far right, while the percentage identifying as strongly Democrat, very liberal, or far left has stayed about the same, decreased, or increased at a slower pace compared with those on the right. More adults identify as strongly Republican but fewer as strongly Democratic. However, identifying as extremely liberal or extremely conservative has increased at about the same rate among adults. Thus, overall, the increase in political polarization is mostly due to more Americans identifying as strongly Republican or very conservative.

We examined this another way in the 12th-grade sample by calculating the mean strength of party affiliation or identification within groups. These analyses show similar results, with greater polarization primarily driven by those on the right (see Table 4). That is, Republicans have become more likely to identify as strongly Republican, and conservatives as very conservative, while Democrats’ and liberals’ affiliation strength has remained about the same.

Table 4.

American 12th Graders’ Strength of Party Affiliation and Political Views, 1976-2014.

	76-79	80-84	85-89	90-94	95-99	00-04	05-09	10-14	d (maximum)	d (period)
Mean Republican	3.29 (0.45)	3.34 (0.48)	3.38 (0.49)	3.40 (0.49)	3.38 (0.49)	3.41 (0.49)	3.42 (0.49)	3.44 (0.50)	.33	.33
Mean Democrat	1.62 (0.49)	1.58 (0.49)	1.54 (0.50)	1.55 (0.50)	1.59 (0.49)	1.53 (0.50)	1.54 (0.50)	1.58 (0.49)	.17	−.07
Mean Conservative	5.19 (0.39)	5.19 (0.39)	5.20 (0.40)	5.24 (0.43)	5.25 (0.44)	5.26 (0.44)	5.26 (0.44)	5.29 (0.45)	.24	.24
Mean Liberal	2.63 (0.67)	2.58 (0.72)	2.59 (0.70)	2.55 (0.71)	2.55 (0.71)	2.56 (0.70)	2.60 (0.64)	2.61 (0.64)	.12	−.03

Note. Decreases in values for Democrats and Liberals reflect increasing extremity, whereas increases in values for Republicans and Conservatives reflect increasing extremity. d (max) = difference in standard deviations from the lowest group of years to the highest; d (period) = difference in standard deviations from the early 1970s to the 2010s. SDs given in parentheses.

The logistic APC model predicting party extremity showed substantial linear (odds ratio = 1.026, Z = 24.87) and quadratic (odds ratio = 0.9997, Z = −7.88) effects such that older people, compared with younger people, were more likely to identify as “strong” in their affiliation as a Democrat or Republican (see Figure 10A). On average, and controlling for time period and cohort effects, Americans were least likely to endorse any “strong” affiliation at age 18 (13%), but were increasingly more likely to endorse a “strong” affiliation as they aged (Age₂₀ = 14%, Age₃₀ = 19%, Age₄₀ = 24%, Age₅₀ = 30%, Age₆₀ = 34%; Z₇₀ = 37%, Z₈₀ = 40%, Z₈₉ = 40%). Although there was some variation in intercepts due to time period (SD = .12 [.09, .16]) and cohort (SD = .08 [.00, .10]), a visual inspection of these effects indicated no clear pattern, suggesting that this variation is due inconsistent fluctuations (see supplemental R code). The logistic APC model predicting political view extremity showed only a weak linear trend (odds ratio = 1.014, Z = 2.94) with no curvilinear effect (odds ratio = 1.00, Z = −0.52), such that older people, compared with younger people, were slightly more likely to identify as “extreme” in their ideological self-categorization (see Figure 10B). On average, Americans were least likely to identify as either “extremely liberal” or “extremely conservative” at age 18 (3.9%), but were increasingly more likely to endorse extreme views as they aged (Age₂₀ = 4.0%, Age₃₀ = 4.6%, Age₄₀ = 5.3%, Age₅₀ = 6.0%, Age₆₀ = 6.8%; Z₇₀ = 7.6%, Z₈₀ = 8.4%, Z₈₉ = 9.2%). There was also variation in intercepts due to time period (SD = .12 [.04, .18]) and cohort (SD = .25 [.08, .37]). There was a general increase in extreme views by cohort, with the fewest with extreme views in the 1900s-born cohort and the most in the 1980s-born (Millennial) cohort. The time period variation was weaker and showed little discernable pattern of change.

Figure 10.

Age effects on (A) strong political party identification and (B) extreme ideological self-categorization, U.S. adults, aged 18 to 96.

Finally, we examined ideological polarization between the parties, examining the correlations between political party affiliation (Democrats and Republicans only) and ideological self-identification as liberal vs. conservative over time. The goal was to establish whether the relationship between party and ideology changed over time with Democrats more likely to be liberal and Republicans more likely to be conservative. We found that they were, with the correlation between party and ideology highly positively related to year of data collection, r(27) = .95, p < .001, for adults, and r(38) = .96, p < .001, for 12th graders (see Figure 11). In other words, the tie between political party and ideology became tighter in more recent years.

Figure 11.

Trends in the correlation between political party identification (Democrat vs. Republican only) and ideological self-categorization (liberal vs. conservative) among (A) adults, and (B) 12th graders.

Discussion

In three nationally representative samples, more Americans in the 2010s (vs. previous decades) identify as Independents and more, especially on the right, express polarized ideological self-categorization (e.g., identifying as “very” or “extremely” conservative). In addition, the tie between ideological identification and political party has grown steadily stronger. Thus, more Americans view themselves at political extremes, but more are disassociating from the two major political parties and are striking out as Independents.

These two trends are only seemingly contradictory. Although our data cannot speak to the possible reasons, it is plausible that the observed increase in the percentage of Independents could reflect a growing dissatisfaction with the entrenchment of the two major U.S. parties (Jones, 2015). This conjecture is supported by the findings that Independents can be quite partisan in their judgments (Hawkins & Nosek, 2012), and that Independents tend to vote consistently for one party (Newport, 2014; Pew Research Center, 2015). This suggests that the Independent identification does not reflect real ideological self-categorization. Rather, it is more a reluctance to declare affiliation with a major party than a genuine independence, which would be evident in more cross-party voting. Although it has been argued that “conventional wisdom regarding the ‘decline of parties’ is both exaggerated and outdated” (Bartels, 2000, p. 35), it appears that trend toward self-identifying as Independent has become more robust.

The APC analyses revealed that time period differences in party identification were curvilinear, with more Americans identifying as Democrats in the 1970s, the tide shifting toward Republicans in the 1980s and 1990s, and Democrats winning out again in the late 2000s and early 2010s. Generational differences showed that G.I.s and Silents included the most Democrats, followed by Boomers and Millennials; GenX had the most Republicans. However, Millennial 12th graders and college students were more conservative and more likely to identify as Republican than Boomers were in the 1970s and were slightly more conservative than GenXers overall. The number of 12th graders identifying as conservative increased between 2000 and 2014, and Millennials in the 2010s were similar to GenXers in the 1990s in their party identification. Millennial 12th graders were more likely to identify as Democrats during a brief period around Obama’s first election to the Presidency, but during the 2010s returned to a more even split between those identifying as Democrats versus Republicans. If Millennials become more Republican and conservative as they age as previous generations have done, they will not be the highly Democratic and liberal generation many had anticipated. Although the effect sizes here are small by most standards (e.g., J. Cohen, 1988), they are large enough to determine the outcomes of closely contested presidential elections (e.g., Kennedy vs. Nixon, Bush v. Gore). For example, a change from 49% to 50% is only d = .02, but Bush v. Gore was settled by a margin considerably smaller than 1%.

As for political polarization, Millennials were more politically polarized than Boomers and GenXers were at the same age, likely due to a time period effect. We also found that people become more politically polarized as they age. Given this, Millennials could go on to be an extremely politically polarized generation as they grow into mature adulthood, as they are beginning adulthood already more polarized.

The APC models further showed that older people are more conservative and Republican than younger people, even after the effects of generation and time period are accounted for. The present research fails to confirm previous assertions that generational imprinting completely explains the age differences in ideological self-categorization (Ghitza & Gelman, 2014). Whereas generational imprinting could be a factor at play (though the present research does not directly investigate generational imprinting), the present findings independently demonstrate a positive relationship between age and conservatism and Republicanism. Our analyses do lend support for the adage about liberalism of the young and conservatism of the mature. This conclusion is quite novel and is inconsistent with research over the past four decades that has vehemently claimed no association between age and conservatism (e.g., Danigelis et al., 2007). This inconsistency may arise from assessment differences, as past research predominantly assessed changes over time on specific issues (e.g., “support for civil liberties”), as opposed to actual changes over time in ideological self-categorization. A critical feature of past methods may be an erroneous assumption of strong correlations between ideological self-categorization and positions on specific issues, thereby resulting in conclusions that might not actually address the initial inquiry into the association between age and conservatism. Overall, the conclusion that older age leads to more conservatism is consistent with research showing that Big Five openness (which is connected to political liberalism) declines at older ages (Roberts, Walton, & Viechtbauer, 2006).

Some may find it surprising that more 12th graders, college students, and adults now identify as conservative, given the increasing acceptance of liberal views around issues such as same-sex marriage, tolerance for religious nonbelievers, and gender equality (e.g., Donnelly et al., 2016; McCarthy, 2015; Twenge, Carter, & Campbell, 2015). This is only a seeming contradiction, predicated on the assumption of a strong association between political self-categorization and ideological self-categorization on specific (ideological) issues. This assumption, however, has received little empirical support as only a few and shrinking number of issues clearly delineate conservative from liberal orientation (Baldassarri & Bearman, 2007). These issues (e.g., legalized abortion) are emphasized in pundits’ and political candidates’ debates likely because any venture to the outside territory may be counterproductive, revealing that their targeted constituents share some significant views with their opponents. This point is best illustrated by absence of correlation between ideological self-categorization (conservative vs. liberal) and views on tolerance and equality (Brandt, Reyna, Chambers, Crawford, & Wetherell, 2014; Wetherell, Brandt, & Reyna, 2013). It is possible that some issues that used to be ideological have ceased to be so, or at least lost their ideological fervor, as illustrated by increasing support for same-sex marriage among Republicans (McCarthy, 2015). The dynamic nature of ideological attitudes suggests that although more high school seniors, college students, and adults may have identified as conservative in 2014 than in 1970, they are not necessarily subscribing to the same brand of conservatism present in 1970. Finally, it is worth mentioning that the unidimensional scale of assessment used in these surveys may not have fully captured the complexities of ideological attitudes (Weber & Federico, 2013). Hence, identical scores may reflect different underlying processes, including distinctly different underlying moral profiles (Federico, Weber, Ergun, & Hunt, 2013).

Further underlying the dynamic nature of our findings is the possibility that they may reflect conformity to group norms as posited by social identity theory (Tajfel & Turner, 2010) and self-categorization theory (Turner, 1985). Liberals and conservatives may have been conforming to what they perceive to be the normative position of their group. In this process, the conditions under which individuals have conformed to this norm has been a function of the frame of reference (Turner, 1985), informed by their group membership as they seek to enhance the status of the group to boost their own self-esteem (Tajfel & Turner, 2010). Polarized political elites dominating the frame of reference aid and abet polarization of the populace. In tandem, individuals then use elites as a reference and role model for “party line” positions to adopt and defend, given their initial ideological self-categorization as liberal or conservative (Zaller, 1992). This then contributes to a new norm for each group that is more extreme than the former norm (Hogg, Turner, & Davidson, 1990) as conservatives and liberals, Democrats and Republicans seek to maximize differences between themselves. Consequently, following the precedent set by political elites and seeking to maximize differences between each other will produce a trend of increasing extremity in party identification and ideological self-categorization. This is evident in the decline in individuals identifying as moderate, combined with an increase in individuals identifying in the extremes of their group (i.e., very liberal/conservative; strongly Democrat/Republican), a trend indicative of the social categorization theory processes essentially being stuck on repeat.

Much past work purporting to demonstrate the presence of political polarization (e.g., Abramowitz & Saunders, 2008) has come under criticism, primarily because of the data from which conclusions about political polarization were drawn. Although, for instance, election returns, vote reports, or even approval ratings may be significant political variables to analyze, they are not sufficient sources of evidence for or against polarization (Fiorina, Abrams, & Pope, 2008). Unlike past investigations of political polarization, the present study does not rely on these forms of data, and instead tracks reported degrees of party affiliation and ideological self-categorization over time in nationally representative samples. Importantly, however, our findings about polarization of ideological self-categorization do not map onto the recent finding that divergence in public ideology, measured as dispersion of views on ideological policy issues, has hardly changed between 1956 and 2012 (Hill & Tausanovitch, 2015). How do we reconcile polarized ideological self-categorization with the status quo in divergence of positions on ideological policy issues? Should not the former derive from or at least reflect the latter? It could be argued that rather than following the laws of formal logic, ideological self-categorization subscribes to psychologic whereby consistently small differences between social categories trigger a race toward their respective extremes. This may reflect “narcissism of small differences” (Blok, 1998). Narcissism of small (or minor) differences reflects a tendency for minor differences between people who are otherwise highly similar to form the basis of hostility and conflict between them (Blok, 1998). As noted by Bourdieu (1984), “social identity lies in difference, and difference is asserted against what is closest, which represents the greatest threat” (p. 481). Liberals and conservatives, Democrats and Republicans are not moving further apart on particular issues, and in many instances, they actually have great overlap and many shared views. It may be the case that they are not aware of this overlap and instead use symbolic issues and emphasize symbolic differences to shape their identity, and in turn these small and overemphasized differences form the basis for more conflictual rather than conciliatory relationships.

This process could be reinforced by selective attention to those few issues on which social categories do differ and continue to widen their differences (e.g., abortion; Baldassarri & Bearman, 2007). Public myopia for the lack of polarization on most issues and selective focus on the polarizing issues is likely helped by the genuinely polarized political elites (Theriault, 2006) who are hard at work limiting public discourse to the latter. Importantly, actual polarization on the select few partisan issues is associated with an exaggeration of polarization on other issues, which, in turn, is associated with greater political activism (Westfall, Van Boven, Judd, & Chambers, 2015). Because voting, working toward getting “the right candidates” elected, and other forms of political activism shape political reality, it matters little that these very real outcomes may originate from not-so-real polarization on the vast majority of political issues. What matters is the select few polarized issues that feed polarized ideological self-categorization, as we suspect is the case.

The present research demonstrates a growing ideological self-categorization entrenchment in Americans identifying as very or extremely liberal or conservative. Furthermore, the present findings demonstrate that over time, age cohorts move toward holding stronger party identification and more extreme ideological self-categorization. Combine this with past research demonstrating how people perceive Republican to be the exact opposite of Democrat, and liberal to be the exact opposite of conservative (Heit & Nicholson, 2010), it becomes apparent that governance may continue to become more difficult due to party polarization, though not necessarily because of issue polarization. Although only a few individual issues may currently be highly polarized (e.g., abortion, illegal immigration), because of party polarization just about any issue could become partisan if it serves a party purpose (e.g., appointment and confirmation of judges). As the present research shows, the general population is also growing more extreme. Even though political elites may be contributing to polarization of the American public (Zaller, 1992), given that political elites have already been polarized for some time (Theriault, 2006), the growing polarization of the American public may in turn also drive political elite polarization to even greater heights, further inhibiting efficient governance.

Although political polarization may yield some benefits, such as easily identifiable parties characterized by highly divergent policy platforms from which citizens can choose (Lavine, Jost, & Lodge, 2015) or assurance that controversial issues will be thoroughly scrutinized by individuals having opposing motivated preconceptions (Zaller, 1992), increased political polarization has several downsides. Political polarization underlies significant repercussions for interest representation, political integration, and social stability (Baldassarri & Gelman, 2008). Furthermore, individuals who hold extreme positions perceive greater polarization than those holding less extreme positions (Van Boven, Judd, & Sherman, 2012). In the process of “naïve realism,” these individuals perceive polarization not necessarily because they believe their own attitudes are extreme, but because they perceive the out-group to be more biased and self-interested than themselves. Thus, as ideological self-categorization and political party identification become increasingly polarized, we should expect to see more in-group versus out-group information processing and the resultant conflict. Whereas society as a whole may indeed see an increase in tolerance, party and ideological dynamics are likely to grow in antagonism. Given growing dissatisfaction with elected representatives (Dugan, 2015), political parties may be digging their own graves by continuing to push polarization.

If more people continue to identify with the extremes of their wings—very liberal, very conservative—the observed political stalemate and dysfunction will become even more pronounced. Correspondingly, it could be expected that the professed value of tolerance for out-group views will diminish. Regardless of likely negative implications of political polarization, though, there may be some hope in a silver lining contingent upon the mobilization of Independents. Whereas polarization may continue to intensify, it may intensify among smaller and smaller groups. This may be especially obvious given the growth of individuals who now identify as Independent and the shrinking numbers of those identifying as Republican or Democrat. If Independents as a new “silent majority” become more vocal, they may change the frustratingly confrontational and paralyzing dynamics created by the polarized political parties. In recent general elections, we have already been seeing this to an extent, as candidates from both major parties engage in a “race to the middle” seeking to brand themselves as more moderate than they were in the primary to appeal to this “silent majority.” Independents may have a critical and unique opportunity to de-escalate party polarization should they seek to take advantage of their growing numbers.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Supplemental Material

The online supplemental material is available at .

Notes

References

Abramowitz

A. I.

Saunders

K. L.

(2008). Is polarization a myth? The Journal of Politics, 70, 542-555.

Baldassarri

Bearman

(2007). Dynamics of political polarization. American Sociological Review, 72, 784-811.

Baldassarri

Gelman

(2008). Partisans without constraint: Political polarization and trends in American public opinion. American Journal of Sociology, 114, 408-446.

Bartels

L. M.

(2000). Partisanship and voting behavior, 1952-1996. American Journal of Political Science, 44, 35-50.

Blok

(1998). The narcissism of minor differences. European Journal of Social Theory, 1, 33-56. doi:10.1177/136843198001001004

Bourdieu

(1984). Distinction: A social critique of the judgement of taste. Cambridge, MA: Harvard University Press.

Brandt

M. J.

Reyna

Chambers

J. R.

Crawford

J. T.

Wetherell

(2014). The ideological-conflict hypothesis: Intolerance among both liberals and conservatives. Current Directions in Psychological Science, 23, 27-34.

Campbell

W. K.

Campbell

S. M.

Siedor

Twenge

J. M.

(2015). Generational differences are real and useful. Industrial and Organizational Psychology: Perspectives on Science and Practice, 8, 324-331.

Cohen

(2003). Party over policy: The dominating impact of group influence on political beliefs. Journal of Personality and Social Psychology, 85, 808-822.

10.

Cohen

(1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.

11.

Crittenden

(1962). Aging and party affiliation. Public Opinion Quarterly, 26, 648-657.

12.

Cutler

N. E.

(1969). Generation, maturation, and party affiliation: A cohort analysis. Public Opinion Quarterly, 33, 583-588.

13.

Danigelis

N. L.

Cutler

S. J.

(1991). Cohort trends in attitudes about law and order: Who’s leading the conservative wave? The Public Opinion Quarterly, 55, 24-49.

14.

Danigelis

N. L.

Hardy

Cutler

S. J.

(2007). Population aging, intracohort aging, and sociopolitical attitudes. American Sociological Review, 72, 812-830. doi:10.1177/000312240707200508

15.

DeSilver

(2014). The politics of American generations: How age affects attitudes and voting behavior. Retrieved from http://www.pewresearch.org/fact-tank/2014/07/09/the-politics-of-american-generations-how-age-affects-attitudes-and-voting-behavior/

16.

Donnelly

Twenge

J. M.

Clark

M. A.

Shaikh

Beiler-May

Carter

N. T.

(2016). Attitudes toward women’s work and family roles in the United States, 1976-2013. Psychology of Women Quarterly, 40, 41-54.

17.

Dugan

(2015, August 12). U.S. Congress and its leaders suffer public discontent. Gallup Organization. Retrieved from http://www.gallup.com/poll/184556/congress-leaders-suffer-public-discontent.aspx

18.

Erikson

R. S.

Tedin

K. L.

(2003). American public opinion (6th ed.). New York, NY: Longman.

19.

Federico

C. M.

Weber

C. R.

Ergun

Hunt

(2013). Mapping the connections between politics and morality: The multiple sociopolitical orientations involved in moral intuition. Political Psychology, 34, 589-610. doi:10.1111/pops.12006

20.

Fiorina

(2014, June 23). Americans have not become more politically polarized. Washington Post. www.washingtonpost.com/blogs/monkey-cage/wp/2014/06/23/americans-have-not-become-more-politically-polarized/

21.

Fiorina

M. P.

Abrams

S. A.

Pope

J. C.

(2008). Polarization in the American public: Misconceptions and misreadings. The Journal of Politics, 70, 556-560.

22.

Fiorina

M. P.

Abrams

S. J.

(2008). Political polarization in the American public. Annual Review of Political Science, 11, 563-588.

23.

Ghitza

Gelman

(2014). The great society, Reagan’s revolution, and generations of presidential voting (Unpublished working paper). New York, NY: Columbia University.

24.

Glenn

N. D.

(1974). Aging and conservatism. The ANNALS of the American Academy of Political and Social Science, 415, 176-186.

25.

Glenn

N. D.

Hefner

(1972). Further evidence on aging and party identification. Public Opinion Quarterly, 36, 31-47.

26.

Hawkins

C. B.

Nosek

B. A.

(2012). Motivated independence? Implicit party identity predicts political judgments among self-proclaimed independents. Personality and Social Psychology Bulletin, 38, 1437-1452.

27.

Heit

Nicholson

S. P.

(2010). The opposite of Republican: Polarization and political categorization. Cognitive Science, 34, 1503-1516.

28.

Hill

(2005). Divided we stand: The polarizing of American politics. National Civic Review, 94(4), 3-14.

29.

Hill

S. J.

Tausanovitch

(2015). A disconnect in representation? Comparison of trends in congressional and public polarization. The Journal of Politics, 77, 1058-1075. doi:10.1086/682398

30.

Hogg

M. A.

Turner

J. C.

Davidson

(1990). Polarized norms and social frames of reference: A test of the self-categorization theory of group polarization. Basic and Applied Social Psychology, 11, 77-100. doi:10.1207/s15324834basp1101_6

31.

Johnston

L. D.

Bachman

J. G.

O’Malley

P. M.

Schulenberg

J. E.

(2015). Monitoring the future: A continuing study of American youth, 2013 (ICPSR35218-v2). Ann Arbor, MI: Inter-University Consortium for Political and Social Research [distributor], 2015-03-26.

32.

Johnston

L. D.

Bachman

J. G.

O’Malley

P. M.

Schulenberg

J. E.

Miech

R. A.

(2014). Monitoring the Future: A Continuing Study of American Youth (12th-Grade Survey), 2014 (ICPSR36263-v2). Ann Arbor, MI: Inter-University Consortium for Political and Social Research [distributor], 2016-03-14. http://doi.org/10.3886/ICPSR36263.v2

33.

Jones

J. M.

(2015, January 7). In U.S., new record 43% are political Independents. Gallup Organization. Retrieved from http://www.gallup.com/poll/180440/new-record-political-independents.aspx

34.

Jost

J. T.

(2006). The end of the end of ideology. American Psychologists, 61, 651-670.

35.

Jost

J. T.

Federico

C. M.

Napier

J. L.

(2009). Political ideology: Its structure, functions, and elective affinities. Annual Review of Psychology, 60, 307-337.

36.

Kiley

Dimock

(2014, September 25). The GOP’s millennial problem runs deep. Retrieved from http://www.pewresearch.org/fact-tank/2014/09/25/the-gops-millennial-problem-runs-deep/

37.

Lavine

Jost

J. T.

Lodge

(2015). Political cognition and its normative implications for the “democratic experiment”: Theory, evidence, and controversy. In Mikulincer

Shaver

P. R.

Borgida

Bargh

J. A.

(Eds.), APA handbook of personality and social psychology, Attitudes and social cognition (Vol. 1, pp. 721-752). Washington, DC: American Psychological Association.

38.

Markus

H. R.

Kitayama

(2010). Cultures and selves: A cycle of mutual constitution. Perspectives on Psychological Science, 5, 420-430.

39.

McCarthy

(2015, May 19). Record-high 60% of Americans support same-sex marriage. Gallup Organization. Retrieved from http://www.gallup.com/poll/183272/record-high-americans-support-sex-marriage.aspx

40.

Milojev

Greaves

Osborne

Sibley

C. G.

(2015). Stability and change in political conservatism following the global financial crisis. Personality and Social Psychology Bulletin, 41, 127-139.

41.

Newport

(2014, July 10). Party identification varies widely across the age spectrum: Young Americans lean Democratic, independent as well. Gallup Organization. Retrieved from http://www.gallup.com/poll/172439/party-identification-varies-widely-across-age-spectrum.aspx

42.

Pew Research Center. (2014a). Beyond Red vs. Blue: The political typology. Retrieved from http://www.people-press.org/2014/06/26/the-political-typology-beyond-red-vs-blue/

43.

Pew Research Center. (2014b). Political polarization in the American public: How increasing ideological uniformity and partisan antipathy affect politics, compromise, and everyday life. Retrieved from http://www.people-press.org/2014/06/12/political-polarization-in-the-american-public/

44.

Pew Research Center. (2015, April 7). A deep dive into party affiliation. Retrieved from http://www.people-press.org/2015/04/07/a-deep-dive-into-party-affiliation/

45.

Pryor

J. H.

Hurtado

Saenz

V. B.

Santos

J. L.

Korn

W. S.

(2007). The American Freshman: Forty-Year Trends, 1966-2006. Los Angeles, CA: Higher Education Research Institute.

46.

Riffkin

(2014, June 19). Public faith in Congress falls again, hits historic low. Gallup Organization. Retrieved from http://www.gallup.com/poll/171710/public-faith-congress-falls-again-hits-historic-low.aspx

47.

Roberts

B. W.

Walton

K. E.

Viechtbauer

(2006). Patterns of mean-level change in personality traits across the life course: A meta-analysis of longitudinal studies. Psychological Bulletin, 132, 1-25.

48.

Saltzman Chafetz

Fuchs Ebaugh

H. R.

(1983). Growing conservatism in the United States: An examination of trends in political opinion between 1972 and 1980. Sociological Perspectives, 26, 275-298.

49.

Schaie

K. W.

(1965). A general model for the study of developmental problems. Psychological Bulletin, 64, 92-107.

50.

Smith

T. W.

Marsden

Hout

Kim

(2015). General Social Surveys, 1972-2014 [Machine-readable data file]. Chicago, IL: National Opinion Research Center [producer]; Storrs, CT: The Roper Center for Public Opinion Research, University of Connecticut [distributor]. Retrieved from http://www3.norc.org/GSS+Website/Download/

51.

Strauss

Howe

(1991). Generations. New York, NY: Random House.

52.

Tajfel

Turner

J. C.

(2010). An integrative theory of intergroup conflict. In Postmes

Branscombe

N. R.

Postmes

Branscombe

N. R.

(Eds.), Rediscovering social identity (pp. 173-190). New York, NY: Psychology Press.

53.

Theriault

S. M.

(2006). Party polarization in the US Congress: Member replacement and member adaptation. Party Politics, 12, 483-503. doi:10.1177/1354068806064730

54.

Turner

J. C.

(1985). Social categorization and the self-concept: A social cognitive theory of group behavior. Advances in Group Processes: Theory and Research, 2, 77-122.

55.

Twenge

J. M.

(2014). Generation me: Why today’s young Americans are more confident, assertive, entitled—and more miserable—than ever before (2nd ed.). New York, NY: Atria Books.

56.

Twenge

J. M.

Campbell

W. K.

Carter

N. T.

(2014). Declines in trust in others and confidence in institutions among American adults and late adolescents, 1972-2012. Psychological Science, 25, 1914-1923.

57.

Twenge

J. M.

Carter

N. T.

Campbell

W. K.

(2015). Time period, generational, and age differences in tolerance for controversial beliefs and lifestyles in the United States, 1972–2012. Social Forces, 94, 379-399.

58.

Van Boven

Judd

C. M.

Sherman

D. K.

(2012). Political polarization projection: Social projection of partisan attitude extremity and attitudinal processes. Journal of Personality and Social Psychology, 103, 84-100.

59.

Weber

C. R.

Federico

C. M.

(2013). Moral foundations and heterogeneity in ideological preferences. Political Psychology, 34, 107-126. doi:10.1111/j.1467-9221.2012.00922.x

60.

Westfall

Van Boven

Judd

C. M.

Chambers

J. R.

(2015). Perceiving political polarization in the United States: Party identity strength and attitude extremity exacerbate the perceived partisan divide. Perspectives on Psychological Science, 10, 145-158.

61.

Wetherell

G. A.

Brandt

M. J.

Reyna

(2013). Discrimination across the ideological divide: The role of value violations and abstract values in discrimination by liberals and conservatives. Social Psychological and Personality Science, 4, 658-667.

62.

Yang

(2008). Social inequalities in happiness in the United States, 1972 to 2004: An age-period-cohort analysis. American Sociological Review, 73, 204-226.

63.

Yang

Land

K. C.

(2013). Age-period-cohort analysis: New models, methods, and empirical applications. Boca Raton, FL: CRC Press.

64.

Zaller

(1992). The nature and origins of mass opinion. Cambridge, UK: Cambridge University Press.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB