Myth #1: Generational “Theory” Was Meant To Be Tested
The sheer number of empirical studies purporting to test generational “theory” would suggest that such theory was intended for testing. However, this is far from the case. The concept of generations as we know it stems from early functionalist sociological thought experiments, derived from foundational work by Mannheim (1927/1952) and others (e.g., Ortega y Gasset, 1933; see also Kertzer, 1983). Adopting the term in a largely historical, rather than familial or genealogical, sense, these authors offered “generations” as social units that account for broad societal and cultural change. Generations were suggested to emerge through “shared consciousness,” which developed across individuals (e.g., those at similar life stages) after common exposure to formative events (e.g., political shifts, war, disaster; see Ryder, 1965). This consciousness, in turn, was theorized to shape unique values, attitudes, and behaviors that characterize a given generation’s members, especially to distinguish one generation from its predecessor. These attributes subsequently impact how these individuals interact with and influence society.
Here, a tautology emerges: culture begets generations and generations beget culture. This is a potentially useful perspective for describing macro-scale interactions between social groups and the social environments in which they live—that is, it is useful as a functionalist sociological mechanism, as the concept of generations was intended. However, this perspective also implies that culture, and the generational groups it forms and is formed by, cannot be disentangled. Generational “theory” is not falsifiable, nor was it intended to be. Attempts to empirically study generations have extended these ideas into positivist and deterministic practices for which they were not intended. Even life course research (e.g., Elder, 1994), which centers on the impact of social change and forces on individuals’ lives as opposed to societal change, does not directly “test” for generational differences, per se. Instead, it uses generations conceptually in explicating the roles that historical, biological, and social time play in life trajectories.
In fact, Mannheim’s (1927/1952) work was partly a critique of the overemphasis on absolutist/biological perspectives in the study of social and historical development, including the objective treatment of time (Pilcher, 1994). This makes it all the more puzzling and problematic that generational “theory” has been applied to discrete quantitative increments (i.e., age and year ranges to define cohorts), and in a fashion that ignores the “non-contemporaneity of the contemporaneous” (i.e., the fact that being alive at the same time, or even being alive and of a similar age at the same time, does not mean history is experienced uniformly; Troll, 1970, p. 201). When considering the roots of “generations,” it is apparent that the concept has been re-characterized and misappropriated.
Myth #2: Generational Explanations Are Obvious
One appealing, if overstated, quality of generations is that there are unique characteristics that are (assumed to be) associated with various cohorts. Moreover, it is assumed that lines can be drawn between generations to distinguish them from one another on the basis of such characteristics. These characteristics, which are said to be influenced by the various events that supposedly give rise to generations in the first place, “make sense” in a way that give generations an air of face validity. For example, it seems very rational and indeed quite self-evident to many that living through the Great Depression made the Silent Generation more conservative and risk-avoidant and that helicopter parents and the rise of social media made Millennials narcissistic and cynical. These and other observed social phenomena such as job-hopping and materialism are frequently ascribed to generations. However, looking more deeply into the identification of these critical events, as well as the mechanisms by which generations supposedly emerge, reveals a far more complex, nuanced picture than a generational explanation would have us believe.
In order to understand why the events that created generations may, or may not, have been impactful, it is important to understand how the critical events purported to give rise to them are identified. As one example, in their popular book, Strauss and Howe (1991) offer a taxonomy of generations, developed by tracing historical records in search of what they called “age-determined participation in epochal events…” (p. 32). To Strauss and Howe, such events were deemed to be so critical that they contributed to the creation of a unique generation. This post hoc historical demographic approach benefits from the passage of time: it is far easier to identify critical events retrospectively, rather than when they are actually occurring. Although major events like economic depressions and wars likely qualify as epochal, dozens of other events have been proposed to be critical in the formation of generations, only to fade into historical oblivion within a matter of a few years.
For example, in defining supposedly seminal events in the development of the Millennial generation, Howe and Strauss (2000) cite the case of “Baby Jessica” (n.b. on October 14, 1987, 18-month-old Jessica McClure Morales fell into a well in her aunt’s backyard in Midland, Texas. After 56 h, rescue workers eventually freed her from the 8-in. well casing 22 ft below the ground; Helling, 2017). Why this event should help form a generation is uncertain, as is whether or not Millennials were or have been systematically impacted by her saga and subsequent rescue.
Rather than being obviously generational, explanations for many social phenomena are more likely to be associated with age or period effects, both of which are other time-based sources of variation that are often conflated with generational cohorts. Specifically, there are three sources of time-based variation that need to be accounted for to make claims about generations: age, period, and cohort effects (see Glenn, 1976, 2005). Age effects refer to variability due to time since birth, in that chronological age is simply an index of “life lived” (e.g., Wohlwill, 1970). Period effects refer to variability due to contemporaneous time and refer to the effects of a specific time and place (i.e., the year 2020). Finally, cohort effects are those that are typically taken as evidence for generations, referring to the year of one’s birth. To make claims about generations, therefore, it is necessary to rule out the effect of age (i.e., developmental influences) and period (i.e., contemporaneous contextual influences).
There are numerous examples of how these sources of variability are conflated and confused with one another. Consider that popular press accounts of Millennials have until recently painted them to be dedicated urban dwellers who favored ride-sharing services and eschewed traditional families (e.g., Barroso, Parker, & Bennet, 2020; Godfrey, 2016). However, adults in this age range have more recently been observed moving to the suburbs, buying houses and cars, and having children (e.g., Adamczyk, 2019). This is not a generational effect but rather a phenomenon attributable to the fact that Millennials are reaching the normative age where people get married, start families, and purchase houses. This is a product of age and context, not generation or period. The picture becomes even more complex given other contextual factors not necessarily bound to time, for example, when considering that the average age of first conception is higher in urban, compared to rural, areas (Bui & Miller, 2018).
Another example comes from data showing that high school and college students are less likely to hold summer jobs today than 20 years ago (Desilver, 2019). This is not a generational effect, but rather is attributable to contemporaneous economic conditions. As a final example, after 9/11, there was a modest increase in the number of people enlisting in the United States Army, which is an example of a period effect (Dao, 2011). However, this change has also been misattributed in various ways to a generational effect (e.g., Graff, 2019). Notably, in ~ 2019 (i.e., when those born in ~ 2001 turned ~ 18 and were eligible to join the army), there were historically low rates of enlistment (Goodkind, 2020). If this rate had been particularly high, one might conclude evidence for a generational effect, such that people born in 2001 grew up in a time and place that demanded enlistment. However, this is not the case—growing up in a post 9/11 world did not make this cohort more likely than others to join the army.
In summary, whereas certain historical events might be easily identifiable as epochal, the extent to which recent events are defined as such might not be known for some time. Moreover, this idea assumes that epochal events actually matter for the formation of distinct generations, a key argument in generations theory that is by-and-large untested, and indeed untestable. Moreover, consider that “global” events (i.e., those that affect all members of a population regardless of age, not just those born in a particular time and place, like a global pandemic) almost certainly manifest as period, not generational cohort effects (Rudolph & Zacher, 2020a, 2020b). Generations and the events that are purported to give rise to them are far from obvious and to attribute current individual characteristics to the occurrence of specific events is misguided. Furthermore, many of the “obvious” generational effects often attributed to such events are much more likely due to other factors associated with age and/or period.
Myth #3: Generational Labels and Associated Age Ranges Are Agreed Upon
Whereas generational labels are well-known and widely recognized, the specific birth year ranges that define each generational grouping and the consistency with which such groupings are applied across time, studies, and location, vary substantially. For example, Smola and Sutton (2002, p. 364) identified a great deal of variation in the start and end years that define different generational groups and the names used to describe various generations, noting “generations…labels and the years those labels represent are often inconsistent” (p. 364).
In their meta-analysis, Costanza et al. (2012) found similar discrepancies with variations in start and end dates ranging from 3 to 9 years depending on the study, the variables of interest, and the source of the generational year ranges being used. Similar conclusions were reached by Rudolph et al. (2018) in their review of generations in the leadership literature.
Beyond these definitional inconsistencies, there are notable differences in the way researchers address cross-cultural variability in generational research. The ubiquity of the labels and their pervasiveness in the literature has led researchers from countries other than the USA to use labels (e.g., “Baby Boomers”) when doing so does not make sense, as the events that supposedly influenced individuals and gave rise to these generations in the first place clearly differ from country to country. Moreover, consider that the term “Millennials” is not meaningful in countries that use Chinese, Islamic, Jewish, Buddhist, Sakka, or Kolla Varsham calendars (Deal, Altman, & Rogelberg, 2010) and that generations are often labeled based on political or cultural events and epochs. For instance, members of the Greek workforce have been categorized into the Divided Generation, the Metapolitefsi Generation, and the Europeanized Generation (Papavasileiou, 2017). In Israel, generations are identified by wars and thus have shorter ranges (Deal et al., 2010). The German media has variously labeled younger people as Generation C64, Generation Golf, or Generation Merkel. In China, generations are pragmatically called the Post-50s generation, Post-60s Generation, and so on, whereas in India, the three main generational groups are labeled Conservatives, Integrators, and Y2K (Srinivasan, 2012).
One approach researchers have adopted for dealing with the complexities of cross-cultural variation in generational labeling is to ignore the issue and simply use US-based generational labels and years when studying individuals in other countries. For example, Yigit and Aksay (2015) looked at Turkish Gen X and Gen Y health professionals, roughly using US date ranges for these groups. A second approach has been to use the date ranges associated with US generations but assign country-specific labels to those same periods. Utilizing this approach, Weiss and Zhang (2020) picked birth year ranges and adopted or developed generational labels in three different countries. For example, for the years 1946–1965, they labeled the generations as the “68er Generation” in Germany, “Baby Boomer” in the USA, and the “New China Generation” in China. A third approach has been to develop country-specific generational groups based on local events that impacted people in that county, a strategy used by To and Tam (2014) who identified four distinct post-WWII generations in China.
Inconsistencies in labeling have significant conceptual and computational implications for the study and understanding of generations and especially so if one wishes to conduct comparative cross-national and/or cross-cultural research. Importantly, we would argue that the validity of the generations concept and its utility for understanding individual, group, and organizational phenomena is very limited due to a number of factors, including (a) researchers’ inability to agree on the start and end dates for different generations; (b) inconsistencies in the classification and labeling systems that characterize them; (c) disagreement on the specific significant influencing events that supposedly gives rise to them, such as the extent to which the timing of events plays a role, including the length of time that is associated with their influence, and the lag required to observe such influences; and (d) the issue of cross-cultural equivalencies. As such, defining generations represents a moving target, which is a significant liability for science and evidence-based practice.
Myth #4: Generations Are Easy To Study
Although there have been numerous attempts to study generations and generational differences, it is clear that these phenomena have not been studied very well. Indeed, it is not only difficult to study generations as they have been framed in the literature but also impossible. As noted above, research on generations is typically based upon birth year ranges, which is to say that they are derived from information about birth cohorts. A common problem emerges when one tries to study cohort effects in cross-sectional (i.e., single time point) research designs, which are the most commonly applied designs used to make inferences about generations (see Costanza et al., 2012). Namely, age, period, and cohort effects are confounded with each other in such designs.
This confounding is best understood through an example. Let us assume that a hypothetical cross-sectional study is conducted in the year 2020 (i.e., the year constitutes the “period effect” in this case). If we reduce the logic of generations a bit and define a cohort effect in terms of a single birth year (e.g., those born in 1980), then the effect of age (i.e., time since birth; 40 years) is completely confounded with cohort. This is because:
$$ {\mathrm{Period}}_{(2020)}={\mathrm{Age}}_{(40)}+{\mathrm{Cohort}}_{(1980)} $$
(1)
In this example, any differences that researchers observe as a function of assumed cohort variability may instead be due to the age of the individuals when they were studied. This pattern would likewise be extrapolated to any age–cohort combinations studied in a single period. The linear dependency among the three effects means that unique effects of age cannot be separated from whatever cohort effect might exist and vice versa.
One common attempt to circumvent this confounding is to artificially group members of different cohorts together to form generational groups. However, this practice is likewise fraught with the same issues raised just above. Another hypothetical cross-sectional study helps to illustrate why: in this study, let us assume that we want to define two arbitrary groupings of birth cohorts, representing people born between 1981 and 1990 (“Generation A”) and those 1991–2000 (“Generation B”), to disentangle age and cohort effects from one another. The variability due to birth cohort in each generation is 10 years; however, as in our previous example, the age range within cohorts is likewise 10 years. Thus, this approach does little to solve the dependency other than shifting the scaling of age. As the rank order of cohort versus age has not changed (relatively older people are in “Generation A” and relatively younger people are in “Generation B”), there is still a correlation between age and generational groups in this study. Moreover, this approach has other limitations, including the loss of statistical power to detect age effects (see Rudolph, 2015) and a confusing logic of cohort versus age effect interpretations (e.g., the oldest members of “Generation A” are closer in age to the youngest members of “Generation B” than to the average age of their own generational group).
From a research design standpoint, this issue of confounding represents an unresolvable problem, which has long been known and lamented in the literature (e.g., Glenn, 1976, 2005). Other research designs are unfortunately no better geared than cross-sectional designs to address this issue, or they do not address variability in cohort effects at all. For example, in typical longitudinal designs, cohort effects are held constant (i.e., from the first time point, people’s birth year does not vary) and period is allowed to vary (i.e., as data are collected from the same people across multiple time points). However, in such designs, period effects are conflated with age (i.e., as people “get older” across time). Expanded longitudinal approaches, such as cohort sequential designs (e.g., sampling 20-year-olds at each time point, T1 − Tk, adding successive cohorts of 20-year-olds at each time point) may be able to separate age/aging from period and cohort effects, depending on how “cohort” is defined. However, such studies require immense resources and time (e.g., 20+ years or more of data collection, including long-term data management and subject retention efforts; see Baltes & Mayer, 2001). As such, and perhaps not surprisingly, we are unaware of any applications of such designs to the study of generations at work.
An alternative that has been employed by some researchers (e.g., Twenge, Konrath, Foster, Campbell, & Bushman, 2008) is a cross-temporal approach, often employing time-lagged panels or cross-temporal meta-analyses (discussed further below). Cross-temporal approaches use data collections from members of different cohort groups, collected during different periods, holding age constant (e.g., data from panels of 25-year-olds and 50-year-olds collected in 2000, 2010, and 2020 or research done on college students every year from 1990 to the present). The logic of cross-temporal methods is to compare groups of similarly aged individuals (i.e., to “control” for age by holding its value constant) across time and then argue that cohort effects are more likely the cause of any observed differences than period effects. Among other issues, cross-temporal approaches have been criticized for their reliance on ecological correlations (i.e., correlations among variables that represent group means) and design assumptions (see Trzesniewski & Donnellan, 2010; Trzesniewski, Donnellan, & Robins, 2008) raising significant concerns about them as a way to study generations. Specifically, ecological correlations can misrepresent relationships when contrasted with correlations among individual observations (see Robinson, 1950).
Overall, the methodological and design challenges associated with studying generations are substantial and the conceptualization of generations as the intersection of age and period makes them impossible to study. Thus, studying generations is only “easy” to the extent that one is willing to ignore the issues raised here. Given these concerns, we echo the recommendations of Rudolph and Zacher (2017), who suggest that “…both research and practice would benefit from a moratorium on time-based operationalizations of generations as units for understanding complex dynamics in organizational behavior” (p. 125).
Myth #5: Statistical Models Can Help Disentangle Generational Differences
Given the design challenges noted above, it is perhaps not surprising that researchers have tried a variety of statistical techniques to resolve the age, period, and cohort confounding problem. Unfortunately, the great majority of generational studies to date have employed the least useful approach to doing so, pairing cross-sectional designs with comparisons of generational cohort means (e.g., typically via linear models, such as t tests or other variants of ANOVA-type models). As noted, cross-sectional approaches control for period effects but confound cohort and age effects with one another and this confounding cannot be resolved statistically through any means. To be clear, this is not a function of a lack of innovation regarding statistical modeling techniques. On the contrary, as long as age, period, and cohort are defined in time-related terms, they will be inextricably confounded with one another in cross-sectional research designs.
With respect to cross-temporal approaches, some researchers have implemented a specific technique referred to as “cross-temporal meta-analysis” (CTMA). CTMA shares certain features with traditional meta-analysis (e.g., studies assumed to be representative of a population of all possible studies on a given phenomenon are taken from the literature and synthesized). In a typical CTMA, age is more or less held constant by narrowing the sampling frame of studies included (e.g., by only considering studies of college age students). By holding age constant and looking at the effects of time on outcomes (i.e., by considering the relationship between year of publication and mean levels of a given phenomenon derived from contributing studies), CTMA models change over time in a phenomenon. However, although age is to some extent held constant, recall that cross-temporal methods inherently confound period and cohort effects with one another. Thus, any identified cohort effect cannot be unambiguously separated from period effects in CTMA. Although research employing CTMA has argued that generations are more likely than period effects to explain observed differences, such work also recognizes that period effects are equally likely explanations for any results derived therefrom (e.g., Twenge & Campbell, 2010). Furthermore, a recent paper by Rudolph, Costanza, Wright, and Zacher (2019) used Monte Carlo simulations to test the underlying assumptions of CTMA, finding that it may misestimate cohort effects by a factor of three to eight times, raising questions about both the source and magnitude of any differences identified.
A final analytic technique that has been occasionally employed to disentangle age, period, and cohort effects is cross-classified hierarchical linear modeling (CCHLM; Yang & Land, 2006, 2013). Applying CCHLM to generational research, age is treated as a fixed effect and period and cohort are allowed to vary as random effects. Importantly, however, decisions about how such effects should be specified are somewhat arbitrary, because it is also possible that cohort and period could be fixed and age random in the population, resulting in different outcomes and conclusions from such models that are largely dependent on analytic decisions rather than reflecting “true” population effects. Thus, without generally unknowable insights into “what” to hold constant in estimating such models, CCHLM results in ambiguous parameter estimates for age, period, and cohort effects.
To this end, a series of simulation studies by Bell and colleagues (Bell & Jones, 2014; see also Bell & Jones, 2013, for further commentaries) has shown that the Yang and Land methodology for separating age, period, and cohort effects simply does not “work.” Even ignoring this issue, CCHLM does little to solve the problem of age, period, and cohort confounding, because the three variables are still linearly dependent upon each other and hence computationally inseparable. Something (typically age) has to be held constant in such models to separate these variables from one another, and even then, ambiguities in how to interpret confounded effects of period and cohort still abound. In short, none of the statistical techniques that have been used to study generations can fully separate age, period, and cohort effects (see Costanza, Darrow, Yost, & Severt, 2017, for a full discussion) and cannot solve the conceptual or design problems noted earlier. This known issue has befuddled social scientists for quite some time. For example, more than 40 years ago Glenn (1976) referred to this problem as “a futile quest.”
Myth #6: Generations Need To Be Managed at Work
Given the proliferation of research and popular press articles identifying generational differences, it is not surprising that practitioners and academics have suggested that people in different generations need to be managed differently at work (e.g., Baldonado, 2013; Lindquist, 2008). There are two main problems with these recommendations.
First, as has been noted, research generally does not and cannot support the existence of generational differences. Conceptual, theoretical, methodological, and statistical issues abound in this literature, and absent clear, convincing, and valid evidence for the existence of generational differences, there is no justification for managing individuals based on their supposed generational membership (NASEM, 2020a, 2020b; Rudolph & Zacher, 2020c). Eschewing the notion of generations does not mean that one must ignore that individuals change over the course of their lifespan or that their needs at different stages in their careers will vary. However, it is important to note that there is not a credible body of evidence to suggest that such changes are generational or that they should be managed as “generational differences” at work.
Indeed, as already noted, much of what lay people observe as “generational” at work is likely more accurately attributed to either age or career stage effects masquerading as generational differences. There is a broad and well-supported literature on best practices for HR, leadership, and management (e.g., Kulik, 2004) and customizing policies and practices based on those recommendations rather than generational stereotypes makes much more sense. Furthermore, there is a burgeoning literature on the positive influence that age-tailored policies (e.g., age-inclusive human resource practices that foster employees’ knowledge, skills, and abilities, motivation, effort, and opportunities to contribute, irrespective of age) for building positive climates for aging at work and supporting worker productivity and well-being (see Böhm, Kunze, & Bruch, 2014; Rudolph & Zacher, 2020d). For example, research suggests that workers of all ages benefit from flexible work policies that allow for autonomy in choosing the time and place where work is conducted (see Rudolph & Baltes, 2017).
Second, as alluded to earlier, management strategies that are based on generations have the potential to raise legal risks for organizations. For example, in the USA, provisions of The Civil Rights Act of 1964, the Age Discrimination in Employment Act of 1967, and the Americans with Disabilities Act of 1991 disallow the mistreatment of individuals from certain groups based on a variety of characteristics. Although generational membership is not directly covered by such legislation, under the ADEA, age is a protected class for workers aged 40+. Given the conflation of generational effects with age, life, and career stage, employment-related decisions tied to generations could be interpreted as prima facie evidence of age-related discrimination (e.g., Swinick, 2019). Indeed, organizations that market themselves to and build personnel practices around generations and generational differences have been implicated in age discrimination lawsuits (e.g., Rabin vs. PriceWaterhouseCoopers, 2017). Combined with the absence of valid studies supporting generationally based differences, organizations open themselves up to an unnecessary liability if they manage individuals based on generational membership (Costanza & Finkelstein, 2015; for a related discussion of various policy implications of managing generations, see Rudolph, Rauvola, Costanza, & Zacher, 2020).
Recently, Costanza, Finkelstein, Imose, and Ravid (2020) reviewed the applied psychology, HR, and management literatures looking for studies about how organizations should manage generations in the workplace. They identified a range of inappropriate inferences and unsupported practical recommendations and systematically refuted them based on legal, conceptual, practical, and theoretical grounds. We echo their conclusion here, regarding advice from managing based on generational membership (p. 27): “Instead of customizing HR policies and practices based on such [generational] differences, organizations could use information about their overall workforce and its characteristics to train recruiters, develop and refine policies, and offer customizable benefits packages that appeal to a broad range of employees, regardless of generation.”
That said, we do not think that the idea of generations should be ignored altogether in the development of management strategies. Instead, the focus should be shifted away from managing assumed differences between members of different generations and toward managing perceptions of generations and generational differences. Considering evidence that people’s beliefs and expectations about age and generations feed into the establishment of stereotypes that interfere with work-relevant processes (e.g., King et al., 2019; Perry, Hanvongse, & Casoinic, 2013; Raymer, Reed, Spiegel, & Purvanova, 2017; Van Rossem, 2019), this is a particularly important consideration and is, in and of itself, a topic worthy of further study.
Myth #7: Members of Younger Generations Are Disrupting Work
While it may feel “new” to blame members of younger generations for changes in the work environment, this is a form of uniqueness bias: we think our beliefs and experiences are new, when in reality similar complaints have been levied against relatively younger and older people for millennia. Indeed, generationalized beliefs about the inflexibility and “out of touch” nature of older generations, or the laziness, self-centeredness, and entitlement of younger generations, have repeated with remarkable consistency across recorded history (Rauvola, Rudolph, & Zacher, 2019). One of the more obvious examples is in referring to generations with self-referent terminology: New York Magazine wrote about youth in the so-called “Me” Decade (Wolfe, 1976) over 30 years prior to Twenge’s (2006) work on “Generation Me,” Time Magazine’s (Stein, 2013) publication on the “Me Me Me” generation, and even the British Army’s recent use of the phrase “Me Me Me Millennials…Your Army needs you and your self belief” in recruitment ads (Nicholls, 2019).
Lamentations about young people “killing things” are far from radical as well. Modern claims are made about youth ending an absurd number of facets of life, ranging from institutions such as marriage and patriotism to household products like napkins, bar soap, and “light” yogurt (Bryan, 2017). Moreover, similar concerns have been voiced throughout the years regarding the rise and fall of consumer preferences, including concerns about young people upending and revolutionizing romantic relationships and transportation (e.g., Thompson, 2016), or being corrupted by new forms of popular media like the radio in the 1930s (Schwartz, 2015).
A more realistic explanation exists for both shifts in consumer preferences as well as changes and disruptions in the nature of work: the contemporaneous environment, and innovations and unexpected changes therein. To take a recent example, the global COVID-19 pandemic has tremendously impacted and transformed how and where work is conducted (Kniffin et al., 2020; Rudolph et al., 2020). While “non-essential” workers are conducting more work virtually and with more flexible hours, other workers deemed “essential” are working in environments with new health and safety protocols and often with different demands and resources in place (e.g., with respect to physical equipment, coworker and customer contact). Even more workers have been furloughed or laid off altogether, with the need to turn to alternative forms of work to maintain income or, when feasible, resorting to early retirement (see Bui, Button, & Picciotti, 2020; Kanfer, Lyndgaard, & Tatel, 2020; van Dalen & Henkens, 2020).
These changes have led to a dramatic pivot for many organization, managers, and individual workers, far surpassing the speed and degree to which more gradual, “generational” workplace changes have supposedly occurred. Not only this, but such changes have had outcomes for workers and society that contradict what generational hypotheses would predict. For example, generational stereotypes suggest that relatively older workers would struggle with technological changes at work while relatively younger workers would thrive. However, the move to work-from-home arrangements has resulted in positive benefits for some, including helpful and flexible accommodations, or health and safety protections, as well as new challenges for others, such as the need to balance childcare or eldercare with work while at home, while still others face newfound isolation and lack of in-person social support coupled with great uncertainty (Alon, Doepke, Olmstead-Rumsey, & Tertilt, 2020; Douglas, Katikireddi, Taulbut, McKee, & McCartney, 2020). These changes create a diverse set of advantages and disadvantages for individuals of all ages. Rather than blaming those of younger generations for disrupting work and life more generally, societal trends and events are a more appropriate, fitting, and ultimately addressable explanation (i.e., through non-ageist interventions and policies).
Myth #8: Generations Explain the Changing Nature of Work (and Society)
Generations are an obvious and convenient explanation for the changing nature of work and societies. However, as discussed previously, convenience and breadth in applying generational explanations does not translate into validity. Because they can easily and generally be applied to explain age-related differences, generations give a convenient “wrapper” to the complexities of age and aging in dynamic environments (i.e., both within and outside of organizations). However, this wrapper restricts and obscures the complexities inherent to both individuals and the environments in which they operate. Generations are highly deterministic, suggesting that individuals “coming of age” at a particular time (i.e., members of the same cohort) all experience aging and development uniformly (i.e., cohort determinism; Walker, 1993). With so many other demonstrable age-related and person-specific factors (e.g., social identities, personality, socioeconomic status) that have bearing on individuals’ attitudes, values, and behaviors, as well as how these interact with contextual and environmental influences, the prospect of generations overriding all such explanations is implausible. Assuming otherwise wipes away a tremendous amount of potentially useful detail and heterogeneity.
Moreover, this perspective stipulates that events in a given time period impact younger people and not older people, such that historical context only influences individuals up to a certain (early) point in their development. This aligns with the idea that identity is “crystallized” or “ratified” at a certain age and development or change is more or less halted thereafter (Ryder, 1965). However, ample evidence suggests that this is far from the case, with age-graded dynamics in such areas as personality emerging across the breadth of the lifespan (e.g., Bianchi, 2014; Donnellan, Hill, & Roberts, 2015; Staudinger & Kunzmann, 2005) and alongside external forces (e.g., economic recessions). Our ability to dismiss crystallization claims is not merely empirical: although current methods and analyses used cannot fully disentangle age from cohort, lifespan development theory promotes the ideas of lifelong development, multiple intervening life influences, and individuals’ agency in shaping their identity and context (e.g., Baltes, 1987). Accordingly, it is more rational and defensible to suggest that individuals’ age, life stage, social context, and historical period intersect across the lifespan. These intersections, in turn, produce predictable as well as unique effects that translate into different attitudes, values, and behaviors, but not as a passive and predetermined function of an individual’s generation.
Myth # 9: Studying Age at Work Is the Antidote to the Problems with Studying Generations
Age and aging research are neither remedies for nor equivalent approaches to the study of generations. First, there are a broad range of phenomena encompassed in both research on “age at work” and “aging at work” (e.g., see discussion of “successful aging” research components in Zacher, 2015a). These two areas are related but distinct, spanning the study of age as a discrete or sample-relative sociodemographic (i.e., age as a descriptive device, especially between person), age as a compositional unit property (e.g., age diversity in a team, organization), and age as a proxy for continuous processes and development over time (i.e., age representing the passage of time, especially within-person in longitudinal research). Each of these forms has a multitude of potential contributions to our understanding of the workplace, and these contributions should not (and cannot) be reduced to generational cohort-based generalizations. Second, and as noted earlier, although aging research is confounded by cohort effects, it draws on sound theories, research designs, and statistical modeling approaches (Bohlmann, Rudolph & Zacher, 2018). The study of generations at work, however, relies upon theories unintended for formal testing and flawed data collection methods and analyses (Costanza et al., 2017).
Moreover, whereas both age and aging research treat time continuously, generational research groups people into cohort categories. This results in a loss of important nuance and information about individuals, with results prone to either over- or underestimated age effects. The practice of cohort grouping also creates a “levels” issue in generational research to which age and aging research are not subject: studying aging focuses on the individual level of analysis, whereas (sociological) generational research “groups” individuals into aggregates and then incorrectly draws inferences about individual outcomes. This mismatch of levels can produce ecological or atomistic fallacies (i.e., assumptions that group-level phenomena apply to the individual level and vice versa), depending on whether group- or individual-level data are used to draw conclusions (Rudolph & Zacher, 2017). Thus, although age and aging research present robust opportunities for understanding how to support the age-diverse workforce, generational research provides incomplete conclusions about, and unclear implications for, understanding trends in the workplace. Studying age alone is not a substitute for generational research; rather, it transcends generational approaches and engenders more useful and tenable conclusions for researchers and practitioners alike.
Myth #10: Talking About Generations Is Largely Benign
Talking about generations is far from benign: it promotes the spread of generationalism, which can be considered “modern ageism.” Just as “modern racism” is characterized by more subtle and implicit, yet no less discriminatory or troubling, racist beliefs about black, indigenous, and people of color (BIPOC; e.g., McConahay, Hardee, & Batts, 1981), generationalism is defined by sanctioned ambivalence and socially acceptable prejudice toward people of particular ages. These beliefs are normalized and pervasive, reiterated across various forms of popular media and culture to the point that they seem innocuous. However, generationalism leads to decisions at a variety of levels (e.g., individual, organizational, institutional) that are harmful, divisive, and potentially illegal.
Media outlets play a large role in societal tolerance and acceptance of generationalism (Rauvola et al., 2019). New “generations” are frequently proposed in light of current events, and age stereotyping becomes further trivialized with each iteration. Adding to this, an abundance of generational labels “stick” while others do not—“iGen,” “Generation Wii,” “Generation Z,” and “Zoomers” all vie to define the “post-Millennial” generation (Raphelson, 2014), and “Generation Alpha” (a name inspired in part by naming conventions during the 2005 hurricane season; McCrindle & Wolfinger, 2009) now faces competition from “Gen C” to define the next generation. “Gen C” (or “Generation Corona;” see Rudolph & Zacher, 2020a, 2020b) has gained traction in the media alongside the recent COVID-19 pandemic, with some suggesting that “coronavirus has the potential to create a generation of socially awkward, insecure, unemployed young people” (Patel, 2020). These labels differ markedly by country as well, as noted earlier, adding to the trivialization and confusion. More and more, these labels are also used to add levity, and/or to avoid blatant ageism, to deep-seated sociopolitical divides and conflicts portrayed in the media. Take, for example, the rise of “OK Boomer” alongside resentment toward conservatism (Romano, 2019), or the labeling of the “Karen Generation” to encapsulate white privilege and entitlement, especially among middle- to upper-class suburban women (Strapagiel, 2019).
Although often treated as harmless banter, this lexicon filters into influential research and policy-based organizations (e.g., “Gen C” in The Lancet, 2020), legitimizing the use of generational labels and associated age stereotypes in discourse and decision-making. As suggested above, in many countries, age is a protected class and the use of generations to inform differential practices and policies in organizations (e.g., hiring, development and training, benefits) poses great risk to the age inclusivity, and the legal standing, of workplaces (see also Costanza et al., 2020). Whether a generational label is new and catchy or accepted and seemingly mundane, it is built on the back of modern ageism, and generationalism—just like other “isms”—is far from benign.