THE NATIONAL ACADEMIES
Advisers to the Nation on Science, Engineering, and Medicine
National Academy of Sciences
National Academy of Engineering
Institute of Medicine
National Research Council
Division of Behavioral and Social Sciences and Education Committee on the Youth Population and Military Recruitment
March 5, 2002
Lieutenant General John A.Van Alstyne
Deputy Assistant Secretary of Defense for Military Personnel Policy
Room 3E767, The Pentagon Washington, DC 20310–4000
Dear General Van Alstyne:
The Committee on Youth Population and Military Recruitment was established in 1999 at the request of the Office of Accession Policy in the Office of the Under Secretary of Defense to examine a broad range of questions concerning the characteristics of the 21st century youth population and subpopulations that are likely to influence recruitment efforts. The committee is now completing its work on a major report that examines such issues as the implications for recruiting and advertising strategies of trends in youth values, the appeal of options available to youth that compete with military service, and the changing nature of work in the military and civilian sectors. In addition to its major task, the committee has also been asked to write occasional, narrowly focused letter reports addressing specific issues of concern to the department that are within the committee’s overall mandate.
Last fall, the Office of Accession Policy requested a letter report that assesses the scientific quality of the popular literature characterizing various generations, with a particular focus on “millennials” —a term coined and described in detail by Neil Howe and William Strauss (2000) for those born in or after 1982. The impetus for this request was the extensive use of that literature, and most particularly of the concept of millennials, for developing and implementing recruiting strategies by the Department of Defense (DoD), although its scientific quality had not been examined. The most frequently cited references are the books by Howe and Strauss; the latest in the series and the one most frequently used by military recruiters and advertisers is entitled Millennials Rising (Howe and Strauss, 2000). Two recent examples of reliance on the popular literature for descriptions of today’s youth and their differences from earlier generations are the Army’s recruiting research program for designing recruit advertising (Parlier, 2001) and a DoD sponsored program for structuring survey research on youth attitudes for use by military recruiters and advertisers (Hoskins, 2001). In addition, the Strategic Studies Institute (Wong,
2101 Constitution Avenue, NW, Washington, DC 20418 Telephone (202) 334–3026 Fax (202) 334–3584 national-academies.org
2000), has recently completed an analysis of attrition behavior of junior officers based on popular characterizations of “boomers” and “Xers” as portrayed by Howe and Strauss (1993), Legree (1997), and Zemke et al. (2000). The use of this literature is not limited to the United States: a report of the Department of Defense Canada (Wait, 2001) also draws on work on generations by Adams (1997, 2000) to contrast “elders,” “boomers,” and “Gen Xers.”
In this letter we examine two key ideas presented in the popular literature: there are distinct generations with sharp differences among them, and there are large and dramatic differences among youth cohorts in different generations. For example, on “the next great generation” (Millennials Rising, p.4):
As a group millennials are unlike any youth generation in living memory…. Over the next decade the millennial generation will entirely recast the image of youth from downbeat and alienated to upbeat and engaged—with potentially seismic consequences for America.
We also consider the quality of the research and analysis used to support these ideas. The committee reviewed eight books that are heavily cited by the DoD in analyzing and characterizing generations and in designing recruiting and advertising strategies: Generations (Strauss and Howe, 1991), Generation X: Tales of an Accelerated Culture (Copland, 1991), 13th Turning (Howe and Strauss, 1993), The Official Guide to Generations (Mitchell, 1995), American Generations (Mitchell, 1998), The Fourth Turning (Strauss and Howe, 1998), Generations at Work: Managing the Clash of Veterans, Boomers, Xers and Nexers in Your Workplace (Zemke et al., 2000), and Millennials Rising (Howe and Strauss, 2000).
Concept of Generations
The popular generational literature (e.g., the “baby boomers,” “generation X,” the “millennials”) synthesizes information from demographic projections, surveys based on selective samples, magazine articles, and newspaper reports. This information is not a part of the peer-reviewed scientific literature.
The style of this literature is engaging and entertaining. Its message for millennials is often positive, even refreshingly optimistic. It communicates a buoyant enthusiasm for youth, for the future, and for the country. There is intuitive appeal to these ideas. Indeed, everyday language is sprinkled with associated labels. It is commonplace to speak of the “founding fathers,” the “Depression-era generation,” and “boomers.” Although such generalizations are the stuff of casual talk and common in the popular press, they are not the focused concepts and explanatory devices of social science research, and they rarely stand up under the careful scrutiny of that research. The seminal work by Ryder (1964; 1965) and the exhaustive review by Riley, Foner, and Waring (1988) explore the theoretical and methodological complexity and confusions in comparing the characteristics or life-course experiences of members of one age group with those of another without specifically defining explanatory variables but rather indexing them only by age and date. That is, one cannot use only people’s ages and fixed dates to compare cohorts; one must specify the events and experiences that are hypothesized to lead to cohort differences and systematically test those hypotheses.
The recent popular accounts of “generations” should not be confused with more scholarly efforts by demographers and economists to describe and interpret patterns of cohort change. That the cohorts born during the 20th century have varied radically in size is an established demographic fact. A number of theoretical and empirical investigations have examined the implications of variation in the sizes of birth cohorts for social and economic welfare. Some of this work investigates the effects of the absolute sizes of birth cohorts; other work focuses on the sizes of birth cohorts relative to the sizes of the birth cohorts of their parents (e.g., Easterlin, 1987). Although the empirical evidence for hypotheses generated from this research is mixed (see Macunovich  for a review), it constitutes a valuable effort to systematically assess the economic effects of demographic changes. Unlike the popular literature, the scholarly literature does not draw arbitrary and abrupt lines between generations, but rather attempts to quantify the ways that cohorts may vary on independent and dependent variables. It presents testable empirical claims that researchers can investigate using data drawn from multiple time periods and locales. The scholarly literature neither proposes nor provides evidence in support of the currently popular notions of generational differences regarding values, attitudes, and beliefs.
Historical and sociological scholarship has identified distinctive characteristics of individuals in particular birth cohorts who have been subject to important social changes, such as the Depression (see Elder, 1974). They do this retrospectively. Thus, Thomas and Znaniecki (1918) used diaries and letters; Lazarsfeld, Berelson, and Gauder (1944; 1960) used panel studies; and Elder (1974) used longitudinal analysis. If one studies multiple cohorts and identifies systematic relationships between cohort characteristics and subsequent patterns of behavior, one can use those analyses to make predictions about subsequent cohorts. But such relationships cannot be inferred without advance specification of explanatory variables and collection of systematic data on multiple cohorts. The popular literature does not base its predictions on such systematic analysis.
The popular literature postulates a recurring cycle of generational characteristics. The literature is unclear about the relationship between this cycle and specific events. However, there is no scientific or historical basis for postulating or predicting that key events and social changes occur in a predictable cycle or that any given event will prove influential in the long term. An event such as the terrorist attack on the World Trade Center may prove such an historical spark. But one cannot know this in advance. Thus, there is no scientific basis for the claim in some of the popular literature that parallels between characteristics of earlier generations and today’s youth is so striking that the destiny of millennials will unfold along predictable lines. Furthermore, the definition of a generation in the popular literature is inconsistent in bracketing generations, i.e., defining the beginning and end points of a generation. For example, the generational scheme defined and used by Howe and Strauss posits 1982 as the transition from generation X to the millennial generation, while the scheme used by New Strategist Publications, another popularizer of the generational concept, posits 1977 (Mitchell, 1995).
Differences Between Youth Cohorts
The claim that one “generation” differs from another is at times taken to imply that there are sharp changes in attitudes and behaviors from one generation to the next. An example is a recent DoD sponsored survey in which 15- to 19-year-olds were labeled millennials and compared on the variable “teamwork” with 20- to 21-year-olds who were labeled generation X. The focus of the survey was youth attitudes as portrayed in the popular literature and the implications for military recruiters and advertisers (Hoskins, 2001). Yet scientific research shows that most indicators of youth attitudes and values change slowly and smoothly over time, if they change at all (Sax et al., 1999). Narrow and specific behaviors certainly do change, such as those driven by the technologies of cell phones and the Internet. These certainly are relevant to such issues as the choice of media for advertising messages, but these changes do not imply fundamental changes in attitudes and values.
Long-term data on youth have been collected by several longitudinal surveys: Monitoring the Future Project, National Longitudinal Survey of Youth, the Youth Attitude Tracking Survey, the Alfred P.Sloan Study (University of Chicago), and the Youth Development Study (University of Minnesota). For youth attitudes, the most comprehensive of these is the Monitoring the Future Project, which has carefully examined representative samples of youth from the mid-1970s to the present (Bachman et al., 2001) and earlier volumes in the same series (e.g., Bachman et al., 2000). Monitoring the Future is a nationwide survey of youth attitudes and behaviors conducted by the Survey Research Center at the University of Michigan. A national stratified random sample of high school seniors, ranging in size from 14,000 to 19,000 annually, have been asked to answer the same questions over the past 25 years.
Several findings from this project are important for military recruiting. First, ratings of the importance of various life goals (e.g., finding purpose and meaning, making a lot of money) show a high degree of stability over two decades: the rank ordering of the goals is virtually unchanged. Second, ratings on the importance of a set of 24 job characteristics (e.g., difficult and challenging, high status and prestige, amount of vacation) are also very similar over two decades. Third, views about the importance of work in young people’s lives have been largely stable: there has been a slow and modest decline in the portion considering work as a central part of life, but only of roughly one-half of 1 percent per year. When changes occur gradually, an analysis that focuses on comparing generations can mistakenly create the impression of a sharp change. Collapsing the data into two artificial averages creates the illusion of a sharp change between generations when, in fact, no sharp change occurred.
There has also been a slight decline in the proportion of respondents who consider that work is important (1) as a chance to make friends and (2) as an opportunity to make contact with a lot of people. These latter two job characteristics were rated as very important by only 39 percent and 26 percent of the respondents, respectively, in the most recent surveys (a finding that may run contrary to the notion that millennials are more team oriented).
Fourth, one example of a gradual change of great importance is youth interest in postsecondary education. Monitoring the Future and other studies that track the value youth place on postsecondary education (e.g., Current Population Survey [U.S. Census Bureau, 1999]; Bachman et al., 2001; Youth Attitude Tracking Study [Wilson et al., 2000]) uniformly report a
continuing upward trend in how contemporary youth value postsecondary education. Indeed, longitudinal studies of U.S. youth indicate that value for postsecondary education is the single most compelling differentiating factor for contemporary youth. The policy implications of this finding are important for both military recruiting strategy and the design of programs that combine military service and higher education opportunities.
Quality of Research and Analysis
The kind of data and analysis used in the popular literature on youth generations also raises serious questions about the quality and usefulness of that work. Scientific approaches to the study of youth involve systematic collection and evaluation of data; the popular approaches tend to use selective data. The representativeness of a sample is crucial when collecting survey or interview data. Social scientists devote careful attention to ensuring that samples represent the population of interest. In contrast, Millennials Rising draws heavily on a 1999 survey of 655 students in the class of 2000 in four public high schools in Fairfax County, Virginia, one of the wealthiest counties in America. For the three high schools named in the survey (the fourth was not listed), the SAT math scores were well above the national average of 514. For one school the verbal score was slightly higher than the national average of 505, whereas the scores for the other two schools were 51 and 76 points higher than the national average (see www.fairfax.gov). In short, generalizing the picture of youth attitudes and behaviors emerging from four high schools in Fairfax County to the rest of the country is not scientifically sound.
There are numerous other interpretational problems with the data in the popular literature. One is the use of retrospective comparisons, in which people are asked to compare current young people with their recollection of earlier cohorts: to compare a 15-year-old today with a 15-year-old of 10–15 years ago. Scientific research has clearly demonstrated that such recall is prone to bias (see e.g., Dawes, 1988). Another frequent problem in the popular literature is the presentation of single point-in-time data, with the implicit suggestion that the data reflect some type of change, although there are no data reported for previous cohorts.
In sum, the committee concludes that two critical features of the popular writing on generations run counter to scientific findings. First, the notion of distinct generations with clear differences between them is not supported by social science research. Second, contrary to claims of large and dramatic differences among youth cohorts in different generations, high-quality longitudinal research documents a high degree of stability in youth attitudes and values. Change is limited, and when it does occur, it occurs gradually. In addition, the popular literature is often based on selective, not systematic, data and analysis and on nonrepresentative samples.
The committee does not believe that it is appropriate to give credence to popular portrayals of “generations” as a key explanatory concept for understanding youth attitudes and behaviors. More particularly, we question specific claims regarding characteristics of the current youth cohort, such as claims that this cohort differs dramatically on dimensions such as teamwork, achievement, modesty and good conduct. We do not assert that all of the conclusions drawn in the popular literature are incorrect; some of the insights and impressions may prove to
be accurate. We do advise against uncritical acceptance of claims for generational characteristics, and we encourage careful examination of the scientific bases for any such claims.
We hope this brief letter is helpful to you in evaluating DoD’s recruitment strategies.