Age- associations in British politics: Implications for the sociology of aging

Debates regarding the status of age in social analysis are foundational to the sociology of aging, with scholars con-tinually questioning the role of age as a social force. The contemporary politicization of age in British politics sheds useful light on this debate. During the past decade, age has emerged as a potent predictor of political preference in the United Kingdom, encompassing numerous intertwined po litical economic developments. At face value, the emergence of age as a key political variable substantiates the status of age in social analysis. However, I argue that it is articulations of age- stratified politics, as much as the associations themselves, that should be of

in later life and eventually inspired the Old Age Pension Act 1908 (Boyer & Schmidle, 2009). In the mid-20th century, Parsons (1942) outlined the sociological case for conceptualizing age itself as a source of structure, though his interest proved fleeting. The empirical importance of age was eventually cemented via the popular theories of disengagement (Cumming & Henry, 1961) and activity (Havighurst, 1961). Disengagement theory cast later life as a juncture at which aging people disengaged from society, relinquishing statuses and roles, and entering a period of inactivity that was mutually beneficial for them and society. In contrast, activity theory argued that older people should resist disengagement through the pursuit of activity, ensuring a positive older age. Importantly, both delineated age as a social classification that warranted sociological attention. Today, it is difficult to envisage a sociology of aging without appeals to the analytic status of age.
The emergence of a sociology of aging is itself symptomatic of late-19th and 20th century demographic change (as opposed to mid-20th century longevity increases), particularly in rapidly urbanizing and industrializing settings wherein older people made up growing proportions of those classified as paupers (e.g., Davenport, 2020). This rendered age a major societal concern as it became increasingly tied to poverty, especially amidst growing state commitments to populace health and security. This is evident in the institutionalization of the lifecourse through the creation of age-based education, work and retirement (Anderson, 1985). The formulation of age as an axis of social stratification predates specific uses in legislation. For instance, 19th century advances in record-keeping widened access to precise chronological ages in industrializing states. While taken for granted today, some historic populations did not have access to detailed information regarding chronological age (Bytheway, 2005). As populations became both old and modern, chronological age became a key organizing principle of both nation states and personal biography (Hacking, 1990). The 20th century was, therefore, characterized by the increasing importance of age, which began to receive greater sociological attention as a result, largely in terms of labor relations and welfare provision rather than the issue of political preferences discussed below. The assumed value of age, and particularly later life, has subsequently become commonplace in social analysis. The underlying supposition is that variables such as "60+," "65+," "older people," or "later life" denote a meaningful category, with important social ramifications.
As the sociology of aging has itself aged, the status of age has garnered skepticism. Some critics (the continuists) wonder if age ever really mattered, while others (the poststructuralists) suspect that it once did, but that its significance has recently diminished. The continuists predate sociological interest in aging. Indeed, the sociology of aging arose in opposition to continuist arguments, which historically encouraged a lack of sociological engagement with age. This line of thought flows throughout psychoanalysts' implicit stability models during the early-20th century (Daatland, 2002), continuity theory during the 1960s and 1970s (Atchley, 1971(Atchley, , 1989, and into some contemporary work on ageism (Bytheway, 2005). For continuists, the older person is essentially unchanged by age, besides some inconsequential greying. An influential-related area of thought is cumulative (dis)advantage (CAD). CAD posits that inequality across various measures increases over the lifecourse (Dannefer, 2003).
It brings age-based differentiation into question by revealing substantial inequality within older cohorts. There can be greater difference between two 65-year-olds than a 65-year-old and a 50-year-old (Dannefer, 1987). From a CAD perspective, the diversity within any age-group undermines the analytic use of age, and by extension its structural influence.
The poststructuralists differ from the continuists, supporting the once substantial, but now diminishing, social salience of age. Hepworth (1989, 1991) have suggested the emergence of a postmodern lifecourse, no longer demarcated by age-standardized transitions, but instead individually negotiated to maintain continuity throughout life. This argument is rooted in the cultural turn and broader critiques of structuralism within the sociology of aging (Powell & Longino, 2002;Twigg & Martin, 2014). Amidst societal trajectories toward diversity and fragmentation, age is considered to have lost its 20th century homogenizing qualities (Gilleard & Higgs, 2014). Gilleard and Higgs (2013, p. 375) claim that "a sociology of old age is increasingly difficult to demarcate [because] the social categorization of old age now contains more than any social category of this type can usefully hold." Similarly, Bytheway (2005) has criticized the illusory homogenizing effect of age-based categorization, reiterating Billig's (1987) argument that thinking of people in terms of categories leads us to overexaggerate their similarities. This is particularly applicable to the commonplace operationalization of the oldestold as a single open-ended category (Bytheway, 2005). The 76-year-old and 103-year-old, both designated "75+," are highly dissimilar in terms of that variable's focus (chronological age) before one even considers how dissimilar they may be in other respects.
The continuists and poststructuralists are minorities, for a sociology of aging would be unsustainable if age was roundly dismissed. Far more popular are arguments that age is socially constructed, having no essential meaning beyond that which is ascribed to it within specific moral and political economies. The constructedness of age is largely invisible in everyday life, so that these ascriptions are typically taken for granted as natural qualities of age (Powell & Hendriks, 2009). Most notably, a sizeable scholarship has charted the positioning of old age as a public fiscal concern in relation to shifting political economic conditions. First, pension-supported retirement from formal labor participation in later life was extended to many populations through the 20th century. Second, fiscal crises in the 1970s and 80s fueled public concern with government expenditure and political movements to limit spending. Consequently, older populations were constructed as a costly burden on imperilled public finances (Mortimer & Moen, 2016;Phillipson, 2013;Walker, 2000). Critically for this paper, the convergence of the late 2000s financial crisis and mid-20th century birth cohort entering retirement have created new conditions for the social construction of age (Phillipson, 2013). While much sociology of aging is dedicated to identifying and dissecting these constructions, the field has also been heavily involved in solidifying constructions of aging as a growing social problem for welfare states. The preponderance of social scientific analyses of future pension coverage or care availability perpetuates particular political ascriptions, for example, the fiscal harms of aging (Powell & Hendriks, 2009). Recognizing such ascription of meaning to age is a useful route through direct causal claims relating to age and other social phenomena, and is pursued as such in this paper.
Some argue that, while postmodern fragmentation has degraded the analytic validity of several social categories, global population aging is strengthening the social salience of age, so that "aging populations now exert a major influence on all aspects of social and economic life" (Phillipson, 2015, p. 80). Moreover, while age-groups contain, and hence, risk concealing, a diverse amalgamation of individuals, this is not dissimilar to the variety that is held within other common sociological categorizations. Ultimately, sociologists continue to engage with age as a meaningful variable, despite noteworthy critique. Here, I argue that contemporary socio-political phenomena can provide new insights into the sociological status of age, particularly as it is constructed. I explore recent trends toward the age-stratification of British political preferences as an example of the changing nature of age as a meaningful social category, and the ways in which that age-stratification is measured, interpreted and articulated. This paper is not principally concerned with establishing determinants of political preference per se (for such discussions see Liu et al., 2020;Neundorf & Smets, 2020;Richards et al., 2020;Rouse & Ross, 2018), but instead deals with the practices of measurement, depiction, interpretation and commentary that amass around those purported preferences and render age-associations socially meaningful. I will argue that beyond the associations themselves, it is value-laden institutional and public engagements with those associations that essentialize age as a social force. I will begin by briefly outlining the recent rise of age as a notable concern in British politics.

| IS AG E THE NE W CL A SS?
Traditionally, British political preferences have been heavily class-associated. Crudely conceived, the Conservatives have been the party of upper-and middle-class voters, while Labor has relied on a working-class voter base. This trend was evident throughout the postwar period and into the 2010 general election (Heath, 2015). By 2017, class differences-categorized as AB (highest class), C1, C2 and DE (lowest class)-were dramatically weakened (Curtis, 2017;Ipsos MORI, 2017), all but disappearing by 2019 (McDonnel & Curtis, 2019). In 1974, the combined AB/C1 vote was 56% Conservative and 19% Labor, while the DE vote was 22% Conservative and 57% Labor (Ipsos MORI, 2010). By 2019, not only did the Conservative party (47%) receive more DE votes than the Labor party (34%), but also the proportion of the DE vote that went to the Conservatives was larger than that of the AB (42%) or C1 (43%) groups (McDonnel & Curtis, 2019). Figures 1 and 2 show votes by class for the two main parties between 1974 and 2019.
As the association between class and preference has weakened, the association between age and preference has strengthened. The age-stratification of voting habits grew throughout the 2010s, until age became one of the strongest associations between conventional demographic variables and political preferences. In 2010, the proportion of the 18-24 vote and the 65+ vote going to Labor was identical (31%) (Ipsos Mori, 2010). In 2015, a 20-point difference had opened up (43% and 23%), and by 2017 this had grown to 37 points (62% and 25%) (Ipsos Mori, 2015. Furthermore, 2017 was notable for being the first election in which age was a better predictor of voting intention than class (Sloam & Henn, 2018). These trajectories are evident in party vote share, with younger and older populations voting in increasingly dissimilar directions across the decade. Figures 3 and 4 show Conservative and Labor vote shares by age-group from 1987 to 2019.
As the association between age and political preference has strengthened, alongside weakening class associations, political commentators have popularized the notion that age is now the dominant influence in British politics. Several have claimed that age is the new class (e.g., Bell, 2019;Curtice, 2019;Curtis, 2017;Sloam & Henn, 2018). Renowned politics professor Sir John Curtice (2019) writes: "It is clear that age, not social class, is the division that nowadays lies at the heart of British party politics." At face value, this claim is straightforward.
The effect size of one traditional demographic association has declined as another has increased, and the resulting association between age and political preference is unprecedented in modern political history (Ehsan & Sloam, 2018). This represents an important moment for the sociology of aging, with contemporary societal issues speaking directly to key disciplinary debates. However, the issue requires considered engagement, beginning with recognition of what age-associations contain and possibly obscure.

| WITHIN CONTEMP OR ARY TRENDS
Taking for granted the claim that age is the new class in British politics, we might immediately wonder why. I offer some partial answer here, but it is not my intention, nor do I think it possible, to thoroughly explain the trend. Instead, I highlight some contributing factors to exemplify the ways in which age-groups relate to other phenomena in social analyses, potentially concealing more than they reveal. I will consider three factors: university education, internal migration and home ownership. Each is a political economic constraint that has altered the characteristics of age-groups over recent decades, and in doing so has strengthened the association between age and political preference.
The first trend is the increasing proportion of younger people receiving a university education over the previous three decades. Between 1990 and 2010, the number of people attaining their first degree from a British higher education institution increased from below 80,000 to over 330,000 (Bolton, 2012). This growth was aided by the re-designation of polytechnics as universities in 1992, coupled with the New Labour government's promotion of widening university participation from 1997 (Robertson, 2010). In 2018, the Higher Education Initial F I G U R E 2 Labour vote by social class (AB/C1 and C2/DE) in British general elections between 1974 and 2019, based on data from Ipsos MORI (2010 F I G U R E 3 Conservative vote by age-group in British general elections between 1987 and 2019, based on data from Ipsos MORI (2010 Participation Rate (HEIPR) rose above 50% for the first time, representing the proportion of people aged 30 who will have entered higher education by 2030 (Brant, 2019). This indicates that growing participation rates over the previous three decades are being maintained and will likely continue.
The growing proportion of people accessing higher education is important for understanding age-group characteristics, political preferences and their relations. Receiving higher education is associated with more leftleaning political preferences, and being in full-time education is an even stronger predictor (Ehsan & Sloam, 2018).
Over recent years, this has become increasingly evident in British voting behavior (Surridge, 2019). In the 2019 general election, 29% of voters with a university degree opted for the Conservatives, compared with 58% of voters whose highest qualifications were below A-level (McDonnel & Curtis, 2019).
The reasons behind this association are subject to ongoing debate (Carl, 2017). Hypotheses include a mixture of proxy and absolute effects, for example: those from more liberal family backgrounds are more likely to enter higher education; children with authoritarian parents are less likely to pursue education and more likely to be illiberal, cognitive ability is positively associated with both liberalism and educational attainment; higher education increases personal security which facilitates empathetic values, experiential aspects of higher education beget liberal values-be that interaction with new people or indoctrination by socially liberal teaching staff. Analyzing data from the 1970 Birth Cohort Study, Surridge (2016) has tested each of these hypotheses to show that, while background proxy factors such as social location and ability partly contribute to the association, they do not fully explain it. Differences between specific study subjects suggest that experiential aspects play an important role, particularly through socialization. Ultimately, any causal relationships between education and political preference are likely far more complex than merely intelligence or indoctrination. Nonetheless, irrespective of causation, what is important here is that the association is strong, sustained and widespread.
The expansion of higher education means that younger cohorts are more likely to be highly educated than older cohorts, most university attendees being younger people. In 2017/18, 1,881,380 students at U.K. universities were aged below 30, compared with 461,495 aged above 30 (HESA, 2019). In the 2011 Census, over 40% of 25-34-year-olds had a university level qualification compared with less than 18% of those aged 65+ (ONS, 2011). Given the association between education level and political preferences, it follows that stratification of higher education by age-group will contribute to corresponding stratification of political preferences by age-groups. The recent political economy of education has thus skewed the relationship between political preferences and age.
F I G U R E 4 Labour vote by age-group in British general elections between 1987 and 2019, based on data from Ipsos MORI (2010 Another contributing factor is unequal rural-urban aging. Between 2006 and 2016, the median age of residents increased in villages and towns, but remained stable in cities (Baker, 2018). Furthermore, between 1981 and 2011, the ratio of people aged above 65 to those aged below 65 increased for villages and smaller towns, but decreased for larger towns and cities (Warren, 2017). The expansion of university education contributes to this trend because enrolling students account for a fifth of internal migration (Swinney & Williams, 2016).

Cities such as Bristol and Manchester experience a pronounced spike in inward migration among 18-22-year-olds,
while rural areas such as the Cotswolds and the Derbyshire Dales experience a similar spike in outward migration (ONS, 2016). Graduates typically remain in metropolitan areas following their degrees due to preferable economic circumstances (Fielding, 2012). Consequently, the United Kingdom is now one of the world's most geographically age-segregated nations (Sabater et al., 2018).
Unequal rural-urban aging is important because population density is associated with political preference across North America and Western Europe. Cities tend toward the political left and the countryside tends toward the right (Wilkinson, 2018). In the United Kingdom's 2017 general election, village residents were more likely to vote Conservative (51%) than Labor (29%), while city dwellers preferred Labor (54%) to the Conservatives (31%) (Warren, 2019). Various explanations for this association have been offered (see Rodden, 2010;Wilkinson, 2018), but for our purposes, what matters is that recent changes in the United Kingdom's political economy have altered the age associations of a characteristic that is also associated with political preference. Unequal urban-rural aging is hence another contributor to the age-stratification of British politics.
A third contributing factor is housing status, which is associated with political preference and has recently become strongly age-associated in the United Kingdom. In the early-1990s, several factors contributed to record private homeownership. These included large postwar housing stock, rising incomes, improved mortgage accessibility and a post-1950s politics of "property owning democracy" that incentivized and facilitated ownership.
However, the financial crises of the 1970s slowed house building, restricting supply and increasing prices, particularly after the late-1980s (Ronald, 2008;Yamada, 1999). Since the mid-1990s, house price increases (152%) have dramatically outpaced wage growth (22%), pricing later cohorts out of the market, and the proportion of young homebuyers has declined sharply. By age 27, 43% of those born in the late-1970s were homeowners, compared with 33% born in the early-1980s and 25% born in the late-1980s (Cribb et al., 2018). Today, 75% of people aged 65 + own their home, compared with 25% of 25-34-year-olds, a reversal of historic trends (Resolution Foundation, 2019). Figure 5 shows homeownership by age-group in the United Kingdom between 1967 and 2017.
The age-stratification of homeownership influences contemporary politics because housing tenure is associated with political preference. Homeowners tend to favor right-leaning politics while renters favor the left (Stubager, 2003). This is evident in modern British voting trends. In the 1997 general election, homeowners typically preferred the Conservatives (41%) over Labor (32%), while those living in rented accommodation were more likely to vote Labor (48%) than Conservative (26%) (Ipsos MORI, 1997). This division remained twenty years later in the 2017 election, with homeowners voting Conservative (55%) more than Labour (30%), and renters leaning in the opposite direction (Conservative 31%, Labour 54%) (Apostolova et al., 2019). Again, there are various explanations for the association (see Stubager, 2003), but irrespective of the reasons, recent political economic changes have differentiated a characteristic of age-groups that also correlates with political preference.
University education, geographic distribution and housing tenure have all recently transformed in ways that have exaggerated age differences. A recent report by United for All Ages, (2020) found that the United Kingdom is among the most age-segregated countries due to a range of social and economic factors, including homeownership and urban-rural distribution. Each phenomenon, traditionally associated with political preference, is now also strongly associated with age, bolstering statistical concordance between age and voting behavior. There are many other considerations. For example, ethnicity is associated with age and political preference, as well as housing tenure, education level and population density (Apostolova et al., 2019;GOV, 2018), and this discussion could be extended sevenfold to include further variables.

| G ENER ATI ON S A S SO CIAL LO C ATI ON S
I have presented education, population density and housing to emphasize how recent interrelated political economic changes contribute to contemporary observations of age-associated politics, in a manner resembling Mannheim's (1952) famous characterization of generations as age-groups forged in specific socio-historical contexts. Mannheim was concerned with positivist and romantic-historical theorizations of generations. Positivists treated generations as a function of biological lifespan, with the birth and death of successive generations driving steady political progress as each group transformed the circumstances created by their immediate forefathers. Thus, each generation initially sought political reform, and later sought to conserve their political creations.
Romantic-historicists approached generations as qualitatively experiential, disconnected from notions of fixed time and biological lifespan. The concept of generations was intuitively ascribed to transient cultural movements in specific historical moments, imposing abstract temporal coherence on diverse phenomena. For Mannheim, generations were somewhere between positivist and romantic depictions, representing vaguely bounded dynamic groups, loosely unified by a shared social location, even when unrecognized by the individual. He argued that generations were social locations because social contexts change over time, meaning that to be a human in one time is a different experience to being a human in another time. This means that individuals within any given generation have a certain range of potential experience which predisposes them to certain modes of thought, feeling and behavior, however, generations are not inherently coherent and distinct units.
A sizeable sociology of generations exists beyond Mannheim's influential work. Some scholars have pursued grandiose schemas of predictable generational cycles (e.g., Strauss & Howe, 1991, 1998, but such structuralfunctionalism offers little to this paper. Notably for both sociologies of generations and aging, C. Wright Mills  (2019) class consciousness weakened, some generation scholars posited that age might come to define politics and identity as class had (e.g., Abrams, 1970). Second, the sociology of generations has often centered on associating groups with radical political transformation, for example, the German youth of the interwar period (Mannheim, 1952) and the postwar Baby Boomers (Edmunds & Turner, 2005). Finally, efforts to label generations, define their unique qualities and identify inter-generational cultural conflicts have garnered greater academic attention in the 21st century, in line with cultural sociology more broadly (Bristow, 2016). These observations highlight two problems of knowledge contained within one question: what makes a generation? On the one hand, this is a question of what particular social circumstances create a corresponding social location. On the other hand, the question speaks to cultural practices of imagining generations as meaningful collectives and the socio-political implications of those imaginings. This paper speaks to both, but focuses on the latter as more immediately fruitful for operationalizations of "age" in sociology.
Related to conceptualizations of generations are longstanding debates regarding the relative influence of age-, cohort-and period-effects in determining political preferences. Those who argue that age determines political preferences often point to straightforward associations between age and preference, typically proposing that conservativism increases with age as one becomes more attached to the past and the status-quo. Cohort-effect advocates typically argue that particular age-groups have inherent political characteristics that are somewhat consistent across their lives, and therefore, irrespective of age. Finally, arguments in support of period-effects center on the importance of immediate context in temporarily determining political preferences across ages and cohorts, meaning that age-and cohort-effects oscillate through history (Schwadel & Garneau, 2014;Vallée-Dubois, Dassonneville & Godbout, 2020). Empirically disentangling these causal hypotheses is a particular vexing challenge because age, cohort and period are all proxy variables for time, meaning that they are linearly dependent, and therefore, generate multicollinearity in analyses (Fosse & Winship, 2019;Glen, 1976;Smets & Neundorf, 2014). Again, unresolved (and perhaps unresolvable) debates concerning the "causes" of associations are not of concern to this paper so much as the interpretative practices of analysis and articulation that amass around and lend symbolic vitality to those associations. Mannheim's conceptualization of generations as social locations, combined with recognition of the social construction of age, is useful here in charting a route through the age/cohort/period debate because each plays a role, and none is causally absolute.

Approaching generations as both history-dependent locations and value-laden social inventions, what is nota-
ble for the purposes of this paper is that the vast interplay of characteristics behind the headline age-associations in contemporary discussion of British political preferences is often unexplicated. Few political commentaries deal with the socio-political entanglements underpinning these trajectories, and in doing so perpetuate the staunch positivist approach to generations critiqued by Mannheim. In claiming that age is the new class, analysts risk attributing to age a degree of demographic determinism that is naïve, if not outright misleading. Furthermore, the claim begets a focus on age despite also representing an interesting class trajectory-the levelling out of a historically highly skewed association between class and political preference. I argue that, rather than the associations themselves, it is the uses of age as a political variable should invigorate the sociology of aging, particularly in relation to longstanding questions regarding the status of age in sociological analyses.

| AG E A S A VARIAB LE
Taken at face value, the growing potency of age as a predictor of political preference is among the most interesting recent developments in the sociology of aging. It seemingly offers evidence that age is a considerable structuring force in contemporary British society. This speaks directly to those longstanding questions regarding the sociological status of age, potentially inspiring a new response to those questions. Contrary to continuist and poststructuralist critiques of age as a vacant classification, the strengthening association between age and political preference suggests that age is gaining structural potency. Perhaps rather than progressing from an age-structured modern lifecourse into a de-structured postmodern lifecourse, we have moved into a new period of age-structuring, in which our political perspectives are increasingly age-structured. This observation alone might suffice to stimulate new sociologies of ageing.
However, we must be more critically reflective than the political commentariat, beginning with claims that position age as an analytic pauper that has suddenly risen from obscurity. The ascent of age-associated political preferences is simultaneously short-lived and protracted in some respects. It is only since the United Kingdom's 2016 referendum on membership of the European Union (EU) that age has received a semblance of the attention that it now garners in political commentary (Fox & Pearce, 2018). It may be that this rapid ascent has attained its zenith, ultimately proving to be a brief political moment rather than a sustained trend. On the contrary, it is not completely accurate to say that the age-association is a novel phenomenon. A small but sustained association has been evident across Western Europe throughout the late-20th century (Van der Brug, 2010), and the politicization of age divisions has some precedent in the 1980s United States (Shaw, 2018). Age is in part more salient today because the association has grown as that of class has declined.
The recognition of similar phenomena in the United States during the 1980s, as well as a smaller sustained trend across Western Europe over several decades, begs the question of international comparability. While the British example is stark, likely exacerbated by an especially divisive referendum on European Union membership (discussed below), various age-categorized political articulations are evident worldwide. Cannon and Kendig (2018) have shown that Australian Millennials believe that they are socio-economically disadvantaged compared to older generations. Rasmi and Daley, (2016) surveyed Arab adolescents and parents living in the United Arab Emirates and Canada to identify 35 socio-cultural and religious issues that were recurrently contested along generational lines. Following recent elections, Ismailbekova (2020) placed "intergenerational conflict at the core of Kyrgyzstan's turmoil" as young activist groups protested the lack of youth representation in parliament. In 2019, the death of a Chinese teenager who jumped from a bridge following an argument with his mother prompted heated debate regarding intergenerational conflict across the country (Yan, 2019). The same year, a 25-year-old New Zealand lawmaker stirred international controversy for using the phrase "Ok Boomer" in parliament (Mezzofiore, 2019(Mezzofiore, , 2020. These diverse international examples of the politicization of age reveal that, across the world, age is a candidate for political essentialization in reference to various debates. The extreme contemporary British situation could hence feasibly be echoed in many contexts if a suitably factious political issue emerges onto which age can be transposed. Returning to the status of age as a variable, there are also questions regarding what an age category contains and represents. Just as we might be skeptical of a category such as "Afro-Caribbean" given the diversity that it contains, so we might wonder at the considerable heterogeneity held within a group defined as "65+." This is especially true from a CAD perspective, discussed above. Moreover, inequality is growing across G20 nations, meaning that later life is a more diverse experience, and older people a more diverse group, than ever before (Grenier et al., 2017). The interpretation of age-groups as distinct political tribes that act collectively has scant empirical basis (Phillipson, 2015). For example, Portes (2014) has shown that both inequality and redistribution are greater within age-groups than between them. Furthermore, there is an unavoidable arbitrariness to age-group boundaries, whereby the 64-year-old and the 65-year-old are separated by the "50-64" and "65+" classification (Campbell, Twenge & Campbell, 2017). As such, it is important to remain alert to the artificial homogenization and separation of people in demographic analyses (Bilig, 1987;Bytheway, 2005).
Sociologists should also be critical of assumed causation and the relationships between the numerous characteristics that age is representing, whether consciously or unconsciously. As with many social variables, age encompasses, and hence, risks concealing, numerous pertinent phenomena. Consider, for example, university education and home ownership. In contemporary political analyses of age-based conflict, "young" becomes a proxy for university-educated urban renters, while "old" represents nonuniversity-educated rural homeowners.
The collinearity of these classifications undermines the identification of primary relationships (if they exist at all) (Johnston et al., 2018). A major problem here is that 38% of voters aged 18-24 and 36% of those aged over 65 did not behave as they should have in the 2019 election, or rather, these huge populations did not act in the way they are presented as having acted. The use of age as a route into a dramatic population politics risks concealing other politically salient social factors. For example, stratifying politics by age, rather than housing status, naturalizes inequalities by attributing differences to an individual characteristic rather than political economic forces, potentially justifying inaction. Age becomes a political precis that is too easily, and sometimes conveniently, read as the full story.
Finally, we might also question the choice to use age in these kinds of analysis. In the EU referendum, 63% of over 75s voted to leave, a widely reported observation at the time (Ipsos MORI, 2016). Less well reported was an analysis stratified by television viewing preferences, which found that 62% of people who liked Mrs Brown's Boys (a popular sitcom) voted to leave (Smith, 2018). What makes a "75+" category more legitimate than a "Mrs Brown's Boys fans" category for political analysis? This is an intentionally provocative question, but it does highlight assumptions regarding the hierarchies of legitimacy within which analytic classifications operate (Elrick et al., 2014;Wemrell et al., 2016). Researchers can be guilty of imposing their own notions of meaningful difference onto study populations and underappreciating the extent to which these meanings are colinear, and perhaps co-constitutive (Elrick et al., 2014;Shim, 2002).

| P OLITI CIZING AG E-A SSO CIATED P OLITI C S
The depiction of politics as being directly age-structured is problematic in several respects, but it does also reflect substantive social transformations that warrant attention. Crucially for the sociology of aging, a secondary process of politicizing age-associated politics is evident. Complex structural changes lead to observations of age-associated politics, the articulation of which perpetuates shifting normative appraisals of the lifecourse and the characteristics of specific age-groups (Varriale, 2019). These socio-political uses of age-associations, wherein complex political economic trajectories are made to constrain new types of age-group membership, offer considerable fodder for invigorating a sociology of aging and revealing the nature of age as a social category.
The politicization of age is evident throughout traditional and social media. An example of stark normative appraisals of age-based politics is the depiction of age-groups in the news media following the EU referendum, wherein hostile commentary regarding age became a popular means of interpreting British politics (Fox & Pearce, 2018). Referendums are typically binary, offering a positive and a negative answer to a question, in this case whether the United Kingdom should maintain its EU membership. The binarism of such a question is more conducive to the differentiation of two opposed groups than is the case in general elections, which typically entail a wider range of issues and options (Meredith & Richardson, 2019). This binarism, coupled with the intense emotion that characterized the referendum, likely contributed to the hyperbole surrounding age when it was revealed that 70% of voters aged 18-30 backed Remain, compared with 33% of those aged above 65 (Ehsan & Sloam, 2018).
The media response to this age-differentiation was intense. In the days following the result of the EU referendum, The Conversation (Založnik, 2016) wondered, "Wouldn't it be fairer if [younger people's] vote was worth more than the vote of someone with only a decade left to live?"; The Huffington Post (Ridley, 2016) noted that the young were "screwed by older generations. Thanks Granny"; and The Guardian (Cosslett, 2016) reassured readers that "If you're young and angry about the EU referendum, you're right to be." These conflictual articulations evidence the broader tendency for analyses of voting trends to focus on simplified inter-group conflicts (Jarness et al., 2019).
This genre of emotive media analysis of age-stratified politics is emulated in public interactions. One week after the 2019 general election, the top two tweets that appeared in a Twitter search for "#GE2019 age" were: The sentiments expressed here exemplify the ability for political age-stratification to provide evidence for one's own political preferences, whatever they are, enhanced through provocative and accusatory language. The old are either wise or selfish, the young indoctrinated or judicious. These invocations of age as a social identity resemble Meredith and Richardson's (2019) account of the emergence of "Brexiteer" and "Remainer" identities during the EU referendum. Their analysis of online newspaper comment sections revealed the important role of the internet as a medium through which these identities were discursively created and popularized, largely based on making derogatory comments about the out-group to imply the preferable nature of the in-group.
New political identities absorb the meanings of their forebears. Meredith and Richardson (2019) noted that old socially salient political identities such as "Tory" or "Leftie" were infused into new referendum identities, establishing some continuity of political divisions. In a similar manner, old stereotypes of privileged Conservative and idealistic Labor voters are injected into age-groups, as are more recent characterizations of intergenerational division. The post-2008 financial crash emergence of a "Boomer versus Millennial" narrative in British politics, whereby younger people (either subjugated or spoilt) are cast against older people (either selfish or sensible), is an important facet of contemporary age politics and the social identities that it supports (Shaw, 2018). Shaw (2018) has identified an almost 300% increase in the number of news articles referring to both "Boomer" and "Millennial" since 2015. She (2018, p. 2) writes: In line with Shaw's observations, Phillipson (2013Phillipson ( , 2015 argues that, building on the 20th century creation of older people as a social category (e.g., through industrialization, public health and welfare-see Achenbaum, 2010), an early 21st century confluence of factors has generated fertile conditions for new imaginings of age that essentialize age-based social categories, including through the ascription of political characteristics to those categories, and promote inter-category hostility. These factors include intensified population aging, the entry of the boomer generation into retirement, the 2008 financial crisis and greater emphasis on personal rather than state welfare responsibility. Phillipson notes that, within this milieux, the accuracy of arguments about age-groups' political characteristics and conflicts are less important than general perceptions of those characteristics. Depictions of novel integrational conflicts, reiterating older stereotypes, have saturated political discourses, through which they are amplified. The resulting symbolism seeps out into the eclectic political concerns of the day. Postelection examples include high-profile analyses and commentary regarding the age-stratification of attitudes toward the Duke and Duchess of Sussex (Landler, 2020;Owen, 2020) and even COVID-19, with competing claims that either the young (Connolly et al., 2020) or the old (Clarke, 2020) were behaving immorally. Again, events are interpreted to support one's personal judgments of age-groups and their relations.
The proliferation of engaging visualizations of election results facilitates this interpretive labor. Kennedy and Hill (2017) note that emotions play an important, and often ignored, role in people's engagements with data and their visualization. The datafication of social life is, therefore, not the rational experience that is often projected.
Instead, it is a process of evoking the feeling of the statistics involved. The use of neat single lines can simplify complex situations, reducing them toward a certitude that is imbued with neutrality by virtue of its statistical basis. As an example, Hill (2017) highlights the use of data visualizations by antiabortion campaigns to politically neutralize their arguments. Election Maps United Kingdom produced some stark visualizations of the United Kingdom's 2019 general election based on age-associations, widely circulated in support of contrasting political stances (see Grafton-Green, 2019). Graphs are hence political comments immaculately conceived. Ultimately, bland observations of curious age differentials are politically, and financially, insufficient. Observers are compelled to uncover the moral facets of these differentials and how they speak to conflicts between good and evil.
The moralization of age as a political entity creates a feedback loop whereby the potency of age as a social divide is rhetorically and institutionally enhanced. The public is increasingly exposed to analyses that advocate age as a vital consideration, and once we have discovered the age association, we cannot unsee it. Our perception is essentially changed. Further, the more cultural currency that age gains as a political problem, the more that problem is institutionalized. For example, the Resolution Foundation (2018)  British policy. The Institute for Fiscal Studies has begun to analyze intergenerational wealth (IFS, 2017(IFS, , 2019 and the polling organization Ipsos MORI (2013) has a workgroup dedicated to age-based analysis. The Economic and Social Research Council recently allocated £45million to establish a research center dedicated to "connecting generations" (UKRI, 2020). Taken together, a great deal of work now goes into deciphering society via age (Shaw, 2018). The United Kingdom has thus recently created the necessary infrastructures to produce sophisticated statistical and normative analyses of politics in reference to age, amplifying the power of age.
In sum, observations of Britain's new age-associated politics are not neutral depictions. Age becomes a window onto our existence, through which diverse populations and complex phenomena can be rendered comprehensible, especially in reference new identities and the relationships between them. Evoking the Thomas Theorem, what is important here, beyond the underlying associations themselves, is the general conviction that those associations exist and are meaningful, for it is this that leads to their measurement by dedicated institutions and artful depiction in ways that emphasize the purported meaningfulness. This is arguably a more noteworthy recent development than age associations, but has received far less critical attention. The newfound generative capacity of age, or rather, institutional and public interpretations of age, informs our lives as much as it informs political commentaries. King (2013) has charted the situated accomplishment of age identities in relation to the differing identities of other age-groups and notions of their relations. Thus, an important socio-political facet of being the "young" is not being the "old," and vice versa. Age-group membership is evaluated and attained within historical, social, cultural, political and economic contexts that differ over time (King, 2013). In the United Kingdom, age-appropriate political preferences are increasingly tied to age identities. It is important to note here that this is not so much a question of subjective self-identification, but is manifest in broader cultural imaginings of identity groups beyond our-selves. Indeed, many theorists argue that our late-modern self-identities have become unrooted from such structural moorings as class and age, and instead respond to multitudinous shifting social relations (Hyde & Jones, 2015). Hence, to be comprehensibly "old" or "young" in the public imaginary is to hold a fitting politics, which may be rather disconnected from one's self-perception. This is not to deny that historic political economic changes have altered the characteristics of age-groups and thereby enhanced the association between age and political preference. Instead, it is an attempt to acknowledge that the measurement, interpretation and articulation of this association is bound up with normative appraisals that lend age social power, warranting sociological attention.

| IMPLI C ATI ON S FOR THE SO CI OLOGY OF AG ING
The sudden politicization of age poses a challenge for the sociology of aging, partly because it catalyzes the aforementioned longstanding disciplinary tensions wherein social scientific work on aging perpetuates particular constructions of age (Powell & Hendriks, 2009). How can one be simultaneously committed to promoting the importance of age as a social variable and denouncing the othering effects of categorizing populations in reference to their age (often to derogatory ends)? How is one to promote age-based social analyses, while also berating those who engage with older individuals via population-level age-associations? The sociology of aging has long survived despite this core inconsistency-that age is simultaneously remarkable and inconsequential-by engaging with it only infrequently. As a result, four years after the EU referendum, The Gerontologist, The Journals of Gerontology and Aging and Society collectively contain one article that mentions "Brexit"-itself a single mention of the age-stratified vote as the "deviant behaviour" of older adults (Wanka, 2018:138).
Others have sought solace in the micro, discarding gerontology's traditional functionalism and instead championing often-ignored critiques of age-group analysis as abstract and simplistic. The protestations of CAD and poststructuralism-that age is a poor variable because age-groups are crammed full of diversity-have become central to sociological engagement with age-based politics. For example, commenting on an online post of Election Maps United Kingdom's aforementioned depictions of age-stratified voting behaviour, one scholar noted: "I really object to lumping one age group all together. Ageist in the extreme … We are all individuals." Another stated: "I think it's high time data was collected and presented as 65-74, 75-84, 85-95, etc. in this kind of analysis." This turn to diversity is intellectually sound to an extent. There are young homeowners in rural areas who have not attended university, just as there are older people living in rented accommodation in large cities while completing their first degree. It is an argument that must be regularly rearticulated in response to reactionary political accusations of intergenerational inequity and conflict (Phillipson, 2015). Yet, appeals to intragenerational diversity cannot abolish the data presented above in Figures 3 and 4, much less silence the cacophony of age-focused political commentary.
A better option might be to approach the phenomena of age-differentiated political preferences as a useful epistemological moment. The contemporary politicization of age-associated political preferences can tell us several things about age and its sociological implications. For example, the repeated presentation of age-associations as being somehow directly explanatory neglects entirely the complex amalgamation of social, political and economic phenomena that are contained within age-associations. Furthermore, it can conceal those phenomena.
Attributing outcomes to age can shield education or housing inequalities from scrutiny that is instead directed toward "Millennials," "Boomers," and their purported conflicts. The attribution of societal problems to age-conflict hence serves those who benefit from the continuation of said problems (Bristow, 2015(Bristow, , 2019. While empirically suspect, articulations of age-based political tribes and conflicts are themselves examples of the potent salience of age as a sociological concern. Interpretations and articulations of age are doing some serious socio-political work here, constructing particular socio-political imaginings of age.
The recent trend toward the stronger age-association of political preferences in the United Kingdom is a "real" observation. Anybody can access data, run analyses and verify this personally, and, more importantly, anybody can read about it across various media outlets. However, it is not this association itself that should be of sole interest to the sociology of aging. It is almost impossible to fully unpack the complex amalgamation of contributing factors that are bound up with the association between age and politics, and what we can readily know in terms of "causation" we already do, for example, education, urbanicity, housing, ethnicity, etc. Instead, it is the assumptions, practices and claims surrounding the creation and proliferation of age-associations to which we should turn our attention. What epistemic infrastructures facilitate the creation of provocative depictions of intergenerational politics? How are such depictions enacted in support of, or in opposition to, specific political ends? How do people's engagements with these depictions shape their relations of meaning with age, aging and age-groups?
Our most pressing concern should be how the politics of age is reflected back onto the world. By unpacking enactments of age-differentiated politics, we can gain a broad appreciation of how new types of age-group, made up of new types of people, enter political discourse (e.g., the emergence of the Boomer and the Millennial as key political actors, saturated with other political stereotypes of Brexiteer, Remainer, Tory and Leftie). Agecategorizing assumptions are evident across various forms of media and our everyday lives, but they also inhabit institutional politics. In 2010, government minister David Willetts (2010) published "The Pinch," in which he berated Boomers for their systemic mistreatment of Millennials. Ultimately, it is this making up of people, through the proliferation of systems of meaning that constrain what it is to be a given age, that speaks directly to longstanding sociological questions regarding the status of age as a force in human life. The notion that age-groups have essential characteristics is exaggerated in a manner that partly self-fulfills the exaggeration. Age may not be a stark socio-political divide per se, but if it is repeatedly imagined and articulated as such, then, it becomes that divide. Contemporary politicizations of age can thus be interpreted as novel iterations of the social construction of age that was widely charted in relation to political economy through the late 20th century, as discussed above (e.g., Mortimer & Moen, 2016;Phillipson, 2013;Walker, 2000).
In conclusion, the use of age-associations as explanatory is deeply suspect, as noted by continuists and poststructuralists. It conceals a range of important considerations, and in their place creates new forms of age-related phenomena that then self-justify the very same analytic efforts through which they are generated. This is a cautionary tale for a sociology of aging that too readily uses age as a mundane variable without sufficient qualification. However, it is also a call-to-arms for the sociology of aging to engage with interpretations and articulations of age as generative of new identities and relations of meaning. Irrespective of their empirical basis, commentaries berating the Boomers and graphs depicting Millennial inadequacies now enjoy a certain salience in the public sphere, lending to age a considerable power. Scholars must engage head-on with this power rather than ignoring it or turning to convenient micro-sociologies. In answer to those longstanding debates: age matters because we ask whether it does.