Narcissism in Gen Y: Is it Increasing or Not? Two opposing perspectives

A group of strangers sits in front of you in the courtroom. Who are these people? What are their individual experiences? How will those experiences influence their decisions in your case? Some might start to answer these questions by putting jurors into categories – how many are male, how many are female, how many of each race or ethnicity. Most would also notice age, and would wonder how that will affect jurors’ perceptions. But age is actually two variables – age itself, or how experiencing more years of life affects one’s perspective – and generation, or how one’s attitudes and personality have been shaped by the culture of the times.

Until recently, much of the material on generational differences was based on conjecture. Authors such as William Strauss and Neil Howe (Generations, Millennials Rising) argued that generations came in cycles of four (for example, they predicted that those born after 1982 would resemble the “Greatest Generation” who fought WWII and would thus be civically-oriented rule followers). However, outside of some broad behavioral data from the U.S. Census, they had no real data to support their theories – nothing that would confirm or disconfirm the psychological differences captured in their ideas.

I have spent my research career trying to change that, conducting almost 20 studies that explore how generations differ from each other psychologically. Initially, I had no agenda and was a fan of Strauss and Howe’s books. In the end, however, much of what my research uncovered was inconsistent with Strauss and Howe’s theories. At least in terms of psychological differences, generations do not occur in cycles; instead, the changes are primarily linear, with each generation taking the previous generations’ traits to the next level. There is no sudden shift in personality for someone born before or after 1982 (Strauss and Howe’s cutoff for what they call the “Millennial” generation). Thus generational labels such as Boomers, Xers, and GenY are of limited use. What’s more important is the number of birth years separating two people – e.g, 20 years or 40 years. Although I occasionally use generational labels (such as Generation Me or GenMe to describe today’s young people), I primarily rely on labels such as “older” and “younger” generations; those in the middle in terms of age (today, the GenXers in their 30s and 40s) will typically fall in the middle in terms of traits and attitudes.

The Data Don’t Disagree:  Younger Jurors Are Likely To Be More Narcissistic  

There is another major difference between my findings and Strauss and Howe’s theories: the most consistent finding across all of my studies was an increase in individualistic traits. Many other authors (e.g., David Myers in The American Paradox, Francis Fukuyama in The Great Disruption) have documented the shift toward more individualistic behaviors and attitudes, and the trend appears in psychological data as well. I’ve found that younger generations score significantly higher in individualistic traits such as assertiveness, agency (including acting as a leader and being dominant), self-esteem, and even extremes of individualism such as narcissism (Twenge, 1997, 2001; Twenge & Campbell, 2001; Twenge et al., 2008). This, of course, contradicted Howe and Strauss’ idea that those born after 1982 would be concerned more with society than with themselves.

The finding regarding narcissism resulted in a research paper that purported to show different effects (Trzesniewski et al., 2008). However, it is somewhat strange that these allegedly opposing findings have been framed as a “debate,” mostly because our datasets and theirs actually show very similar results. Our study of narcissistic traits, drawn from the responses of college students across the country, showed an increase in narcissistic traits between 1982 and 2006 using 85 samples, with the change especially strong since 2000. Trzesniewski and colleagues’ (2008) data from UC Davis also show an increase in narcissism between 2002 and 2007, with the year-by-year rate of change the same or even larger than the increase in our analysis (Twenge & Foster, 2008). Thus both datasets show an increase in narcissism during the most recent decade.

The only place we differ is in what happened before 2002, and the divergence stems from a substantial difference in the data used in each evaluation. Their analysis includes just two datapoints prior to 2002 (one from 1982 and one from 1996). Our analysis includes 58 datapoints collected between 1982 and 2001 that also include their two. In addition, these two datapoints are from a different campus (UC Berkeley) than their later data (from UC Davis). Thus it is very possible that differences by campus are causing or contributing to the variation – with time and campus perfectly confounded, drawing conclusions is difficult. After 2002, when the data are all from one campus, they show the same rise in narcissism we found. In short, everyone agrees narcissism has been rising the last five years.

A completely independent study has also found evidence for a generational difference in narcissism. A group sponsored by the National Institutes of Health interviewed a nationally representative sample of almost 35,000 Americans in 2005 and asked them if they had ever experienced certain symptoms; they then determined if these symptoms met the criteria for Narcissistic Personality Disorder (NPD, the clinical form of the trait). Because the researchers asked about lifetime experience, one would expect that older people, who have lived many more years, would be more likely to have experienced the disorder. However, the data showed the opposite: Only 3% of people over age 60 had experienced NPD, compared to 9% of people in their twenties – even though a 25-year-old has had only 7 years to experience the disorder (which cannot be diagnosed until one is 18), compared to 40 years for someone who is 65 (Stinson et al., 2008). Even if older people failed to recall some earlier instances of symptoms, this suggests a strong generational increase in narcissism.

There is also cultural evidence supporting the conclusion: Plastic surgery and procedures have increased fivefold in just 10 years; even invasive surgeries like breast augmentations quadrupled between 1997 and 2007 (American Society for Aesthetic Plastic Surgery, 2008). The square footage of U.S. homes nearly doubled in 30 years (National Association of Homebuilders, 2006), and levels of debt rose from 16% of disposible income in 1978 to 19% in 2007  (United States Federal Reserve, 2008). The circulation of celebrity gossip magazines has surged even as other types of magazines have faltered. High school students are markedly overconfident about their future educational and career prospects (Reynolds et al., 2006). Other changes are less quantifiable but more stunning: In 2006, it became possible to hire fake paparazzi to follow you around when you go out at night. Most packages include a copy of one of the shots on a fake celebrity magazine cover. These changes are so large and pervasive that they suggest the change in individual-level narcissism is a pale reflection of the much bigger sea change in the culture. (Twenge & Campbell, 2009; we discuss these changes at length in The Narcissism Epidemic, also addressing where they came from).

There Are Meaningful Generational Differences

Trzesniewski and Donnellan (in press) also present a critique based on their analysis of data from a nationwide sample of high school students called Monitoring the Future. This paper concludes that there are few generational differences. However, the data show many statistically significant generational differences, most of which are consistent with my findings. Since the 1970s, high school students have become more likely to agree that “people should do their own thing,” that having lots of money is important, and that it’s okay to buy things you don’t need. There are also small increases in self-satisfaction and believing one is more intelligent than peers. Consistent with a cultural boosting of self-feelings, GenMe students are twice as likely to report earning an “A” average even though fewer do many hours of homework (Twenge, in press). More recent generations are also more likely to anticipate that they will be “very good” at important adult roles such as spouse, parent, and worker (Twenge & Campbell, 2008). As before, this is not a “debate” with competing data, as both datasets show similar results.

However, not all of the results are consistent with our data. For example, Monitoring the Future does not show an overall increase in self-esteem, as our data does across middle school, high school, and college samples. It does show a small increase in self-liking, however (Twenge & Campbell, 2008). In science as in law, the best approach is to rely on the preponderance of the evidence, and most of the evidence, from most sources, shows an increase in positive self-feelings over the generations.

There are other meaningful generational differences as well. Other data showed increases in anxiety and poor mental health, perhaps the price of focusing on the self rather than relationships (Twenge, 2000). Younger generations have a more external locus of control, meaning they are more likely to believe that their fate (and that of the country and world) is determined by luck and powerful others rather than their own actions (Twenge et al., 2004). The upside of individualism is evident as well: Less prejudice, more tolerance, and more opportunities regardless of a person’s background (Twenge, 2006). Younger generations take it more or less for granted that women and minorities will participate fully in society and in the workforce.

Generational Differences Are Clearly Relevant

Trzesniewski and Donnellan (in press) also argue that generational changes are too small to be relevant. However, many generational changes are in fact quite large. For example, 82% of 1990s college students scored higher on anxiety than the average college student in the 1950s (Twenge, 2000); 65% of college students in 2006 scored higher in narcissism than the average 1982 student (Twenge et al., 2008).

Yet even the smaller generational differences are worth noting. Whether a given difference is large enough to matter is an extremely subjective judgment. The effect of secondhand smoke on lung cancer, which has spawned restrictive laws across the country, is less than one-fourth the size of the change in narcissism over 24 years in the nationwide data. It is less than a tenth of the size of the change in anxiety over 50 years. It is half the size of the change in agreeing it is good to “do your own thing,” which the other authors considered too small to deign to discuss in their paper. It is also smaller than the increase in thinking one is more intelligent than one’s peers, which the other authors also dismissed as negligible.

There are several reasons to take relatively small effects seriously. First, small changes at the average multiply into larger ones at the extremes. In a bell curve, a small shift at the middle pushes the tail of the distribution higher so that two or three times as many people now fall into the extreme category. Narcissism is a good illustration – the small to moderate change at the average translates into three times as many people in their twenties having experience with NPD as compared to people in their sixties. So if one generation scores higher in assertiveness on average, there will certainly be those who fit this generalization and those who are notable exceptions. An average change does not mean that everyone fits the profile. But there will be two or three times as many who are highly assertive in the younger generation compared to the older one.

Many generational differences are the same size (or in many cases larger) than gender differences. Yet few would question the wisdom of paying attention to the gender composition of a jury. Gender is also an illustration of small average changes multiplying into larger ones at the extreme: On average, the gender difference in aggression is just a little larger than the change over time in narcissism in 24 years. Yet our courtrooms and prisons are filled with many, many more men than women who have committed violent crimes.

Trzesniewski & Donnellan (in press) have also argued that we believe there are generational differences because older people’s perceptions have changed over time. However, all of the data presented in this article – and by Trzesniewski and Donnellan themselves — come from reports from young people themselves. Thus this explanation does not fit the data. In fact, a recent Harris poll suggests that members of the younger generations actually have more negative views of their generation that those in older generations do (Harper, 2008).

It is also very important to note that these changes are not just generational: They are cultural (this is covered in more depth in our book, The Narcissism Epidemic). As authors like Myers and Fukuyama pointed out, Western cultures shifted decisively away from duty and social rules throughout the 1960s, 1970s, and into the 1980s and beyond, actively promoting the notion that the individual should come first. As I document in Generation Me (Twenge, 2006), parents began to tell children they could be anything they wanted to be. Popular songs declared that loving yourself was the greatest love of all. Schools began programs to raise children’s self-esteem. Pop psychology taught that you have to love yourself first before you can love someone else. When a culture changes, the younger members experience and reflect the changes most strongly; however, the young do not raise themselves.

Generational Differences Are Important To Your Trial Practice

In practical terms, what do these generational differences mean for the pool of potential jurors sitting in front of you? Younger jurors are more likely to be blind to differences in gender, race, and sexual orientation. A case of gay domestic violence, for example, would not faze many of them, whereas many older jurors would be confused or disturbed by such a case. Due to their more external locus of control, younger jurors will be more willing to believe arguments that defendants acted due to outside circumstances (e.g., childhood abuse, coming from a disadvantaged background). The older the juror, the more likely he or she respects authority and the rules of society. They may be more willing to automatically believe the testimony of law enforcement. In addition, the older generation is more likely to do what they are told. In contrast, younger generations need to be told why they are doing something. Younger generations may feel more entitled to perks and time off. Especially among those high in narcissistic traits (especially among the 9% with experience with NPD), younger jurors may feel less empathy for people (whether victims or perpetrators). A more engaging style may be necessary to engage younger jurors, who are used to multimedia and multitasking; many schools, for example, have abandoned the straight lecture because it puts this generation to sleep. 

Overall, there are meaningful and important generational differences in personality traits and attitudes. Many other factors, such as gender, race, and class, will also influence jurors’ viewpoints, but generation should certainly be considered. We might ask, “Why not consider generation?” Even if a generational effect is small, it could mean the difference between a favorable trial result and an unfavorable verdict for your client. 

Jean M. Twenge is Associate Professor of Psychology at San Diego State University, and is also the author of Generation Me: Why Today's Young Americans Are More Confident, Assertive, Entitled — and More Miserable Than Ever Before. Her research interests include generational differences, the effects of modern culture, and social rejection. When not researching she can be found reading or swimming, though usually not doing both at once. More about her research and speaking activities can be found at www.jeantwenge.com and http://www.psychology.sdsu.edu/new-web/FacultyLabs/twenge/TwengeResearchProjects.htm.

W. Keith Campbell is Associate Professor of Psychology at the University of Georgia and the author of When You Love A Man Who Loves Himself. His research interests include narcissism and romantic relationships, the relationship between narcissism and self-esteem, and the intersection between narcissistic personality traits and narcissistic personality disorder. He also enjoys surfing and fishing. More about his research can be found at http://wkeithcampbell.googlepages.com/home.

 

References

American Society for Aesthetic Plastic Surgery (2008). Cosmetic Surgery National Data Bank Statistics, 1997-2007. http://www.surgery.org/press/statistics-2007.php (viewed online 9/18/08).

Fukuyama, F. (1999). The great disruption: Human nature and the reconstitution of social order. New York: Free Press.

Harper, J. (2008). Boomer narcissism, ageism debunked: Poll exposes generation myths. Washington Times, August 26, 2008.

Myers, D. G. (2000). The American Paradox: Spiritual hunger in an age of plenty. New Haven: Yale University Press.

National Association of Homebuilders (2006). Housing facts, figures, and trends. http://www.soflo.org/report/NAHBhousingfactsMarch2006.pdf (viewed online 6/8/2008)

Reynolds, J., Stewart, M., MacDonald, R., & Sischo, L. (2006). Have adolescents become too ambitious? High school seniors’ educational and occupational plans, 1976 to 2000. Social Problems, 53, 186-206.

Stinson, F. S., Dawson, D. A., Goldstein, R. B., Chou, S. P., Huang, B., Smith, S. M., Ruan, W. J., Pulay, A. J., Saha, T. D., Pickering, R. P., & Grant, B. F. (2008). Prevalence, correlates, diability, and comorbidity of DSM-IV Narcissistic Personality Disorder: Results from the Wave 2 National Epidemiologic Survey on Alcohol and Related Conditions. Journal of Clinical Psychiatry, 69, 1033-1045.

Trzesniewski, K. H., Donnellan, M. B. (in press). Rethinking “Generation Me:” A study of cohort effects 1976-2006. Perspectives in Psychological Science.

Trzesniewski, K. H., Donnellan, M. B., & Robins, R. W. (2008). Do today’s young people really think they are so extraordinary? An examination of secular trends in narcissism and self-enhancement, Psychological Science 19, 181–188.

Twenge, J. M. (1997). Changes in masculine and feminine traits over time: A meta-analysis. Sex Roles, 36, 305-325.

Twenge, J. M. (2000). The age of anxiety? Birth cohort change in anxiety and neuroticism, 1952-1993. Journal of Personality and Social Psychology, 79, 1007-1021.

Twenge, J. M. (2001a). Birth cohort changes in extraversion: A cross-temporal meta-analysis, 1966-1993. Personality and Individual Differences, 30, 735-748.

Twenge, J. M. (2001b). Changes in women’s assertiveness in response to status and roles: A cross-temporal meta-analysis, 1931-1993. Journal of Personality and Social Psychology, 81, 133-145.

Twenge, J. M. (2006). Generation Me: Why Today’s Young Americans Are More Confident, Assertive, Entitled–and More Miserable Than Ever Before. New York: Free Press.

Twenge, J. M. (in press). Birth cohort differences in the Monitoring the Future dataset: Further evidence for Generation Me. Perspectives in Psychological Science.

Twenge, J. M., & Campbell, W. K. (2008). Increases in positive self-views among high school students: Birth cohort changes in anticipated performance, self-satisfaction, self-liking, and self-competence. Psychological Science, 19, 1082-1086.

Twenge, J.M., & Campbell, W. K. (2009). The Narcissism Epidemic: Living in the Age of Entitlement. New York: Free Press.

Twenge, J. M., & Foster, J. D. (2008). Mapping the scale of the narcissism epidemic: Increases in narcissism 2002-2007 within ethnic groups. Journal of Research in Personality, 42, 1619-1622.

Twenge, J. M., Konrath, S., Foster, J. D., Campbell, W. K., & Bushman, B. J. (2008). Egos inflating over time: A cross-temporal meta-analysis of the Narcissistic Personality Inventory. Journal of Personality, 76, 875-901.

Twenge, J. M., Zhang, L., & Im, C. (2004). It’s beyond my control: A cross-temporal meta-analysis of increasing externality in locus of control, 1960-2002. Personality and Social Psychology Review, 8, 308-319.

United States Federal Reserve (2008). Household Debt Service and Financial Obligations Ratios. http://www.federalreserve.gov/releases/housedebt/default.htm (viewed online 6/8/2008)

 

 

---

 

Are Today’s Young People Really That Different from Previous Generations?

 A Skeptical Perspective on “Generation Me”

by Kali H. Trzesniewski and M. Brent Donnellan

Commentators suggest that socio-cultural changes in recent decades, such as the rise of the culture of self-worth, declines in social connectedness, changes in parenting practices, and increases in perceptions of threat, have coalesced to create a relatively unique generation of young people (e.g., Bellah, Madsen, Sullivan, Swidler, & Tipton, 1985; Lasch, 1979; Putnam, 2000; Twenge, 2006).  Many of these accounts portray recent generations in a negative light.  For example, Americans born in the 1970s, 1980s, and 1990s have been grouped into a single cohort, and labeled “Generation Me” because they seem to have a heightened sense of egotism, increased self-esteem, and unrealistically high expectations for their future (Twenge, 2006).  This generation is also alleged to exhibit elevated levels of misery and other symptoms of psychological distress (Twenge, 2006; but see Arnett, 2007).  In 2008, 60 Minutes ran a story about members of the so-called “Millennial” generation (individuals born between 1982 and 2002) in the workplace and proclaimed that a “new breed of American worker is about to attack everything you hold sacred” (Safer, 2008).  The program described Millennials as cynical, unaccustomed to hard work, and having fragile egos because their “…childhoods filled with trophies and adulation didn't prepare them for the cold realities of work” (Safer, 2008).  All in all, there seems to be considerable concern over the apparently unique characteristics of today’s young people.

However, concerns about generational decline are not new.  Past concerns are most clear when examining the generational gaps of the 1960s and the 1920s.  For example, a 1967 Time Magazine article about the “hippies” stated, “To their deeply worried parents throughout the country, they seem more like dangerously deluded dropouts, candidates for a very sound spanking and a cram course in civics—if only they would return home to receive either” (The Hippies, 1967).  An unnamed businessman offered a similar perspective about the “flappers” of the 1920s in a 1926 issue of the Dallas Morning News: “The worst part is that they don't care what people – their mothers and fathers and uncles and aunts – think of them. They haven't any sense of shame, honor or duty…Daily doses of good, old-fashioned discipline is what restless youth need” (Richardson, 1926).  The psychologist G. Stanley Hall wrote in 1904 that “Modern life is hard, and in many respects increasingly so, on youth.  Home, school, church, fail to recognize its nature and needs and, perhaps most of all, its perils” (p. xiv).  It wouldn’t surprise us if we heard older individuals make any of the preceding remarks about the young people of 2009.  

As social scientists we approach questions about generational change from an empirical perspective and with a healthy dose of skepticism.  This skepticism is informed by our knowledge of the complexities of birth-cohort analyses as well as our conviction that scientific research should not be used to fuel unfounded stereotypes of young people (see also Arnett, 2007).  Indeed, an appreciation for the long history of sentiment regarding generational decline suggests that recent pronouncements about generational excesses should be taken with a grain of salt.  In this essay we describe our perspectives on the methodology of generational studies and reiterate some of our concerns with the “Generation Me” perspective of today’s young people. 

What Are Cohort Effects and How Should They Be Studied?  

Social scientists who study generational changes are interested in the study of birth cohort effects. A birth cohort is simply a group of individuals who share similar birthdays and are therefore thought to have experienced a common set of developmental experiences. To date, there is no universal standard for defining birth cohorts and different authors lump and split birthdates into different cohorts in fairly idiosyncratic ways. Nonetheless, the basic idea of a birth cohort effect is that individuals born in the 1960s, for example, are somehow different than individuals born in the 1970s because they had different life experiences, experiences which exerted a relatively consistent effect across all members of the birth cohort (e.g., perhaps all people exposed to the Watergate hearings as young adolescents became particularly cynical).

Unfortunately, the study of cohort effects is exceedingly difficult.  At first blush, the most difficult challenge is the need for historical data.  If a researcher wants to know if adolescents in 2009 have higher self-esteem than adolescents did in 1959, then such a researcher needs data on the same self-esteem measure from both groups.  This hypothetical researcher would also ideally want to document that the properties of the measurement instrument have not changed over time such that the measure was more or less valid for making trait inferences in 1959 versus 2009. 

One solution to this challenge of needing historical data to identify cohort effects is to use a method known as the cross-temporal meta-analysis (see Twenge, 2006).  This approach takes advantage of the extensive questionnaire data about attitudes and personality traits that different researchers have collected over the last decades.  The strategy is straightforward: Collect all of the studies that included a particular measure on an age-restricted sample (e.g., college students, adolescents) and compute the association between the year of data collection and average scores on that measure.  If average scores from college students on a given measure have changed from say the 1950s to the 2000s, then researchers might infer that there has been a cohort-related change in that attribute.  A fundamental question is whether such an inference is valid.

We have argued that such inferences are usually not valid (e.g., Trzesniewski, 2008a).  Simply put, we believe that researchers need data that are suitable for making population-based inferences and the data used in most relevant cross-temporal meta-analyses are not appropriate (see Costello, Erkanli, & Angold, 2006 for an appropriate use of the cross-temporal meta-analytic method for making population inferences).  To be sure, researchers interested in generational changes are interested in generalizing to a very broad class of people (e.g., all Americans born within a certain time frame).  This means that data used for generational analyses should be collected using a scientific sampling plan, a plan which permits valid inferences from a sample to a defined population of interest (e.g., all young people born from 1972 to 1980).  This principle is taught in virtually all methods courses in the social sciences.  As noted by Pedhazur and Schmelkin (1991), “… the incontrovertible fact is that, in nonprobability sampling, it is not possible to estimate sampling errors.  Therefore, validity of inferences to a population cannot be ascertained” (p. 321).  This concern with valid population-based inferences is the reason that there are professional pollsters and companies that specialize in assessing public opinion. A good resource for understanding some of the more technical issues behind scientific polling is available in a report entitled “20 Questions a Journalist Should Ask about Poll Results” produced by the National Council on Public Polls (http://www.ncpp.org/?q=node/4).

In short, our objection is not with the cross-temporal meta-analytic method per se, but rather our objection is that the method is usually applied to studies that are inappropriate for population-based inferences.  The reality is that the college student samples that are incorporated into many of these cross-temporal meta-analytic studies were not generated using probability-based sampling methods.  They are convenience samples, and as noted by Schwarz, Groves, and Schuman (1998), “[s]uch samples, be they college students who voluntarily sign up for a study or readers who respond to a questionnaire printed in a magazine, do not allow inferences to any population because their representativeness is unknown” (p. 145).  In other words, the studies analyzed in the typical cross-temporal meta-analysis are not appropriate for making inferences to a defined population, such as all young people of a particular generation.  Aggregating the data cannot solve this inferential dilemma. 

We should also point out that college students from conventional 4-year institutions, the primary participants in most psychological research, are a fairly select group of young people.  They represent approximately 20% of all American youth aged 18 to 24 (U.S. Census Bureau, 2005a, Table 9; U.S. Census Bureau, 2005b, Table 2).  Thus, even if these college student samples were designed to make inferences to that population of Americans (and remember they are not), it would still leave a clear majority of 18 to 24 years olds unrepresented in the samples analyzed by researchers using college student samples.    

Given our perspective on valid population inferences, we have been relatively critical of the evidentiary basis for many of the “Generation Me” effects Twenge and her colleagues identify in their work.  Twenge has responded that “it is odd they [Trzesniewski, Donnellan, and our colleague, Richard Robins] would argue for the analysis of only randomly sampled, nationally stratified data. Very little such data exists” (Twenge, 2008, p. 1450).  Here we are reminded of the valuable insights of the statistician John Tukey: “The data may not contain the answer. The combination of some data and an aching desire for an answer does not ensure that a reasonable answer can be extracted from a given body of data” (p. 74 – 75).     

Importantly, there are actually existing resources that can be used to study cohort effects in a way that is valid for making population-based inferences. One such resource is the Monitoring the Future (MTF) project, an ongoing, nationally representative study of high school seniors that began in the mid 1970s (see Bachman, Johnston, & O’Malley, 1996; Johnston, O’Malley, Schulenberg, & Bachman, 1998 for a more detailed description).  Across the 30 years of the study, over 450,000 high school seniors have participated in this project.  Schools and participants are selected using carefully designed sampling techniques to provide a final sample each year that is generally representative of the population of high school seniors living in the 48 coterminous states who agreed to participate.  Over 95% of schools approached agreed to participate and over 80% of students within those schools participated.  Of those who did not participate, the most common reason was being absent from school.  Although the study is voluntary (as are all survey studies conducted by ethical social scientists), only a very small portion of the students actively refused to participate (1%).  Accordingly it is reasonable to assume the final sample of participants represents U.S. high school students quite well.  Data were collected following standardized procedures via closed-ended questionnaires administered in classrooms by University of Michigan representatives and their assistants.  This study is close to ideal for testing cohort effects because a large, representative sample of individuals of the same age have been asked the same questions for 30 years, and a full set of raw data from 1976 to 2006 is readily available in electronic format for interested researchers.

What Can Be Learned from The Monitoring the Future Survey (1976-2006)?   

We conducted an analysis of change in 31 constructs that were measured in the MTF to try to address many of the "Generation Me" claims made by Twenge and her colleagues. We quantified generational changes by simply correlating year of administration with observed scores. Overall, we found little support for generational differences, typically finding small correlations, if any, between year and levels of each construct. In fact, we found no meaningful change in 22 constructs (Trzesniewski & Donnellan, in press). In particular, we found correlations that were close to zero (defined here as any correlation smaller than |.10|; Cohen, 1988) for the constructs summarized in Table 1.  

Table 1:  Constructs Not Showing Generational Change (i.e., r  with year of data collection less than |.10|)

Egotism	Self-Enhancement
Individualism	Self-Esteem
Locus of Control	Hopelessness
Happiness	Life Satisfaction
Loneliness	Antisocial Behavior
Time Spent Working	Political Activity
Time Spent Watching Television	The Importance of Social Status
The Importance of Religion

Only nine effects were substantial enough to be considered a small effect (i.e., a correlation somewhere between |.10| and |.30|), and none were large enough to be considered a large effect (i.e., a correlation above |.50|).  Most notably, members of more recent cohorts expressed fewer worries and concerns about social issues (r = -.23), they were less trusting of others (r = -.12), and were more cynical of institutions (e.g., r = .11).  Members of more recent cohorts expressed less interest in keeping up with materialistic trends (r = -.14); however, at the same time they were more tolerant of blatant consumerism and the marketing of unnecessary material goods (r = .17).  In line, with “Generation Me” ideas, members of more recent cohorts tended to have higher expectations for their future (e.g., r = .23 for expecting to graduate college) and they seemed less convinced that working hard will lead to desired jobs (r = .11). 

Our primary conclusion from these analyses was that the bulk of the results pointed to generational consistency rather than generational differences.  In fact, the direction of some correlations ran counter to the “Generation Me” profile.  For example, the correlation for self-esteem (r = -.02) was actually negative in sign (i.e., if anything, more recent cohorts had lower scores than older cohorts), whereas the correlations for happiness (r = .02) and life satisfaction (r = .02) were positive in sign (i.e., if anything, more recent cohorts had higher scores on these variables than older cohorts, counter to a prediction of increased misery).  However, given how extremely close these effects are to zero, we did not consider these constructs as showing any meaningful change over time.

All in all, we found little reason to conclude that the average member of “Generation Me” is dramatically different from members of previous generations.  Today’s youth seem to have the same self-esteem scores as previous generations and they appear to be just as happy and satisfied as previous generations.  In fact, based on the variables we examined, today’s youth seem to have psychological profiles that are generally similar to youth from the past 30 years.  The potential qualifications to this generalization are that we found that more recent generations have higher expectations for their educational careers and they are a bit more cynical and distrusting than previous generations (although these trends seem to predate “Generation Me”; see Robinson & Jackson, 2001).  Nonetheless, using the Monitoring the Future datasets, we found little evidence to support the conclusion that the youth of 2006 were considerably different from the youth of the late 1970s. 

What About Narcissism?

There has been considerable attention to whether today’s youth are more narcissistic than previous generations.  Indeed, Twenge and her colleagues have recently raised concerns about an emerging “epidemic” of narcissism (e.g., Twenge & Foster, 2008).  One limitation of our MTF analyses is that there was no direct measure of narcissism.  Narcissism is a complex construct and it is important to understand what it is and how it is typically assessed before getting into issues of generational changes.  Narcissism has a long history in psychoanalytic writings as it was perhaps first discussed Havelock Ellis (1898) who wrote of a “Narcissus-like tendency” for sexual urges to be directed toward the self in the form of autoeroticism.  That said, it is difficult to precisely define narcissism as it is understood by contemporary psychologists because researchers in psychiatry, clinical psychology, and personality/social psychology do not always agree about the best ways to conceptualize this construct (e.g., Cain, Pincus, & Ansell, 2008). 

The vast majority of social and personality psychology research uses a measure called the Narcissistic Personality Inventory (NPI; Raskin & Terry, 1988) to study narcissism.  Cain et al. (2008) estimate that over 75% of social/personality research on narcissism since the mid-1980s has used this measure.  In many ways, the NPI’s content has come to define the construct of narcissism for social and personality researchers.  However, Cain et al. (2008) also note that the “content of the NPI total score may reflect a confusing mix of adaptive and maladaptive content” (p. 643).  Likewise, Emmons (1987) noted that “Narcissism [as assessed by the NPI], rather than being a unidimensional construct, consists of four moderately correlated factors tapping the domains of leadership, self-admiration, superiority, and interpersonal exploitiveness [sic].” (p. 15).  There is debate over the underlying structure of the measure and different groups argue that there are two, three, four, or seven reasonably separable constructs embedded within the NPI.  The reality is that the NPI total score seems to capture some amalgam of confidence, leadership, and social potency along with more socially harmful elements of personality such as a sense of entitlement and a willingness to exploit others.  This fact makes it difficult to give a clear and unambiguous interpretation of what is represented by a relatively high score on this measure, as it might reflect heightened levels of socially toxic traits (e.g., entitlement), somewhat obnoxious traits (e.g., vanity), socially adaptive traits (e.g., leadership), or some indeterminate mixture of these components.

Despite concern about interpretation of the total NPI score, this measure has been used to examine whether narcissism is increasing in today’s youth (e.g., Trzesniewski et al., 2008b; Twenge et al., 2008).  Here we should emphasize that none of the results from these studies should be used to generalize to all American young people because none were collected with scientific sampling plans.  At most, these studies should spark curiosity and fuel debate about an important topic.  The earliest NPI data point is from Raskin and Terry (1988) who collected data from students at U.C. Santa Cruz and U.C. Berkeley from 1979 to 1985.  The average response at that time was to endorse 39% of the items in a narcissistic fashion.  We reported that the average score for a large sample of University of California at Davis undergraduates was about 38% in 2006 (Trzesniewski et al., 2008b).  This figure was quite similar to the mean reported by Raskin and Terry (1988) and thus we were not terribly impressed by the evidence for secular changes at U.C. schools from Northern California, the epicenter of the self-esteem movement.  That said, we also issued this caveat about our own work: “…consistent with our sampling concerns, these results should be interpreted cautiously because they are based on convenience samples” (Trzesniewski et al., 2008a, p. 910).  Twenge et al. (2008) estimated that the average NPI score in 2006 was 43% versus 38% in 1982.  The absolute difference between these estimates amounts to around 2 items on the NPI.  In all, the recent discussion about an “epidemic” of narcissism boils down to a small number of items and a fairly small range of average scores (38% to 43%) on a measure that can be controversial to interpret.  Viewed from such a perspective, it is not clear that any of these NPI findings should be used to make the case that there is an emerging epidemic of narcissism. 

Furthermore, the actual meta-analytic results in Twenge et al. (2008) paint a more nuanced picture of secular increases in narcissism than is often portrayed in summaries and media accounts.  Simply put, increases may be due to increased confidence or self-sufficiency in young women attending college today.  Twenge et al. (2008) reported that the trend for increasing narcissism was apparently more pronounced for women than men.  The trend for an increase in men was not even statistically detectable based on their sub-analysis of the 44 studies that had information reported by gender (out of 85 total studies).  This gender effect may have profound implications for how the purported secular change in narcissism is best interpreted, especially in light of the perspective that the NPI measures multiple traits, including things like leadership, social potency, and confidence.  The Twenge et al. analysis might simply indicate that today’s generation of women who attend college are more confident and assertive than previous generations of women who attended college.  This seems to be an equally plausible but far less nefarious interpretation for any observed changes in the NPI (e.g., as an indicator of an epidemic).  In short, we are very reluctant to conclude that any observed changes in the NPI are an indicator an epidemic of pathological traits.

Finally, we should add that full blown Narcissistic Personality Disorder (NPD) or the personality disorder that is diagnosed by clinicians and psychiatrists is very rare (Lenzenweger, Loranger, Korfine, & Neff, 1997; Torgersen, 2005; see Cain et al., 2008; p. 648 for a review).  In fact, NPD appears to be one of the least commonly observed “official” personality disorders (Torgersen, 2005).  These considerations have made us very skeptical of the idea that there is good evidence for an epidemic of narcissism.  To be clear, we are not saying that the claim has been definitely falsified rather we are saying that there are reasons to be skeptical of such a claim.

Summing Up

Questions about generational differences are inherently difficult to address and this is bound to create discussion, dissent, and even outright controversy.  As it stands, we believe that the best evidence thus far points to consistency over the past 30 years, whereas Twenge and her colleagues believe the evidence points to widespread generational changes.  We imagine that readers might be puzzled by the seemingly inconsistent portraits of today’s youth described in this increasingly contentious debate.  Moreover, to many non-social scientists it might seem painfully obvious that today’s young people are ruder, cruder, and more self-important than previous generations, based on their own observations and experiences. 

In response to any of these anecdotal impressions of youth we point to the considerable amount of psychological research suggesting that great care should be exercised when forming generalizations about entire groups of people (e.g., all individuals born in a particular decade) based on limited perceptions that might be unduly influenced by extremely memorable exemplars.  Consider how easy it is to think of a particular young person who is arrogant, egotistical, miserable, extremely lazy, or any of the other less flattering adjectives used in reference to that segment of the population.  This ease may simply reflect an availability bias, reflecting the fact that more “memorable individuals” stick out in our memories (Tversky & Kahneman, 1974).  It is also the case that once a stereotyped generalization is established, a confirmation bias may set in whereby individuals selectively attend and recall pieces of information that are consistent with that generalization.  It appears that individuals are less likely to pay attention and remember pieces of information that are inconsistent with their preconceptions.  Thus, there may be psychological reasons for the impressions held by adults about kids “these days” that have little to do with actual generational differences (for a review of social cognitive mechanisms involved in stereotype formation and maintenance see Hamilton & Sherman, 1994). 

We are also convinced that there is reasonably good evidence for normative personality development with age (see e.g., Caspi, Roberts, & Shiner, 2005) such that individuals, on average, appear to become more conscientious, more emotionally stable, and more agreeable as they age.  It might simply be that older adults in 2009 overlook the personality changes that they themselves experienced as they matured from young people into full-fledged adults with the “typical” responsibilities of adulthood such as earning a living, participating in a committed romantic partnership, and taking care of children.  Thus, the somewhat negative impressions of the characteristics of today’s young people might just reflect contrasts between the personalities and preoccupations of those in their 40s and 50s as compared with the personalities and preoccupations of those in their late teens and early 20s.

The more difficult question then remains – is there any way to reconcile our position with that of Twenge and her colleagues?  Overall, we believe that the differences in our respective positions boil down to two main issues: (1) differences in the willingness to generalize to entire populations of young people from convenience samples that were not collected using scientific sampling methods and; (2) differences in the interpretation of effect sizes associated with available generational trends.  The first issue seems to us to be fairly cut-and-dry as we believe that relatively few social scientists would disagree with our position that inferences from convenience samples are more like “faith-based” appeals rather than scientifically valid inferences.

The second issue regarding effect sizes is more of a matter of style and preference as the interpretation of trends requires the application of subjective judgment.  For example, let’s assume that the rise in narcissism truly amounts to an increase of two scale points on the NPI from 1982 to 2006 (the estimate from the Twenge et al. meta-analysis).  Is a change of two scale points practically or theoretically important and noteworthy?  Different researchers will attach different degrees of significance to a change of two scale points on a self-report measure that is based on an arbitrary metric (see Blanton & Jaccard, 2006).  We admit that we are unsure as to the real-world meaning of such a change and we strongly question whether a two scale point increase has any bearing on who should be selected for a jury trial.  We submit that no one knows when a change of a few scale points on the NPI becomes problematic or whether two scale points on this kind of measure should provide an impetus for societal concern.  The bottom line is that we are uncertain as to what kind of applied significance should be attached to any apparent changes on the NPI.

In light of these issues, we argue that a conservative approach to generational change is warranted given the quality of the existing evidence and the real possibility that negative stereotypes of today’s young people might be reinforced by potentially flawed studies (Arnett, 2007).  To be frank, we are concerned that the findings reported by Twenge and her colleagues may help justify negative impressions of youth given that apparently long-held stereotypes (see the opening paragraph) are now somehow “validated” by scientific research.  At this point, we are simply not convinced that the evidence warrants any strong conclusions about generational changes that have applied significance.  In line with our conservative approach, the editors of a special issue of the Journal of Managerial Psychology summarized the literature with respect to the evidence and implications of generational differences at work: “many of the empirical findings are less strong and consistent than popular sentiment suggests.  Indeed, there may be more variation among members within a generation than there is between generations” (Macky, Gardner, Forsyth, 2008, p. 860).  At this point, we strongly concur.

Kali H. Trzesniewski ([email protected]) is an Assistant Professor of developmental psychology at the University of Western Ontario.  Dr. Trzesniewski's research interests are at the intersections of developmental psychology, personality psychology, and social psychology.  Her current research is focused on biological and social-cultural influences on the development of the self and self-related processes.    

M. Brent Donnellan ([email protected]) is an Assistant Professor of social and personality psychology at Michigan State University.  Dr. Donnellan is currently an Associate Editor at the Journal of Research in Personality and he almost always investigates topics related to personality traits, personality development, or interpersonal relationships. He and Dr. Trzesniewski have worked on projects since they were both in graduate school at the University of California, Davis. 

*Both authors contributed equally to this article. 

 

References

Arnett, J. J. (2007). Suffering, selfish, slackers? Myths and reality about emerging adults. Journal of Youth and Adolescence, 36, 23-29.

Bachman, J. G., Johnston, L. D., & O’Malley, P. M. (1996). The Monitoring the Future project after twenty-two years: Design and procedures. (Monitoring the Future Occasional Paper No. 38). Ann Arbor, Michigan: Institute for Social Research, The University of Michigan.

Bellah, R.N., Madsen, R., Sullivan, W.M., Swidler, A., & Tipton, S.M. (1985). Habits of the heart: Individualism and commitment in American life. Berkeley: University of California Press.

Blanton, H., & Jaccard, J. (2006). Arbitrary metrics in psychology. American Psychologist, 61, 27-41.

Cain, N. M., Pincus, A. L., & Ansell, E. B. (2007). Narcissism at the crossroads: Phenotypic description of pathological narcissism across clinical theory, social/personality psychology, and psychiatric diagnosis. Clinical Psychology Review, 28, 638-656.

Caspi, A., Roberts, B. W., & Shiner, R. L. (2005). Personality development: Stability and change. Annual Review of Psychology, 56, 453-484.

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

Costello, E. J., Erkanli, A., & Anglod, A. (2006). Is there an epidemic of child or adolescent depression? Journal of Child Psychology & Psychiatry, 47, 1263-1271.

Ellis, H. (1898). Auto-eroticism: A psychological study. Alienist and Neurologist, 19, 260-299.

Emmons, R. A. (1987). Narcissism: Theory and measurement. Journal of Personality and Social Psychology, 52, 11-17.

Hall, G. S. (1904). Adolescence: Its psychology and its relation to physiology, anthropology, sociology, sex, crime, religion, and education (Vol. 1). New York: Appelton.

Hamilton, D. L., & Sherman, J. W. (1994). Stereotypes. In T. K. Srull & R. S. Wyer (Eds.), Handbook of Social Cognition (2nd ed., Vol. 2, pp. 1-68). Hillsdale, NJ: Erlbaum.

Johnston, L. D., O’Malley, P. M., Schulenberg, J. E., & Bachman, J. G. (1998). The aims and objectives of the Monitoring the Future study and progress toward fulfilling them. (2nd ed.) (Monitoring the Future Occasional Paper No. 34). Ann Arbor, Michigan: Institute for Social Research, The University of Michigan.

Lasch, C. (1979). The culture of narcissism: American life in an age of diminishing expectations. New York: Norton.

Lenzenweger, M. F., Loranger, A. W., Korfine, L., & Neff, C. (1997). Detecting personality disorders in a nonclinical population: Application of a 2-stage procedure for case identification. Archives of General Psychiatry, 54, 345-351.

Macky, K., Gardner, D., & Forsyth, S. (2008). Generational differences at work: Introduction and overview. Journal of Managerial Psychology, 23, 857-861.

Pedhazur, E. J., & Schmelkin, L. P. (1991). Measurement, design, and analysis: An integrated approach. Hillsdale, New Jersey: Lawrence Erlbaum.

Putnam, R.D. (2000). Bowling alone: The collapse and revival of American community. New York: Simon & Schuster.

Raskin, R., & Terry, H. (1988). A principal-components analysis of the Narcissistic Personality Inventory. Journal of Personality and Social Psychology, 54, 890–902.

Richardson, V. (1926, July 11). What about this flapper generation. Dallas Morning News, Section VII, 5-6. Retrieved April 10, 2009, from http://freepages.history.rootsweb.ancestry.com/~jwheat/flappers.html.

Robinson, R. V., & Jackson, E. F. (2001). Is trust in others declining in America? An age-period-cohort analysis. Social Science Research, 30, 117-145.

Safer, M. (2008, May 23). The “Millennials” are coming. 60 Minutes. Retrieved June 29, 2008 from http://www.cbsnews.com/stories/2007/11/08/60minutes/main3475200.shtml.

Schwarz, N., Groves, R. M., &Schuman, H. (1998). Survey Methods. In D. T. Gilbert, S. T. Fiske , & G Lindzey (Eds). Handbook of Social Psychology Vol 1 (4^th Ed). (pp. 143-179). New York: McGraw-Hill.

 

Torgersen, S. (2005). Epidemiology. In J. M. Oldham, A. E. Skodol, & D. S. Bender (Eds.), Textbook of personality disorders (pp. 129–141). Washington, DC: American Psychiatric Publishing, Inc.

Tukey, J. W. (1986). Sunset salvo.  The American Statistician, 40, 72-76.

Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185, 1124-1131.

Trzesniewski, K. H., & Donnellan, M. B. (in press). Rethinking “Generation Me”: A study of cohort effects from 1976-2006.  Perspectives in Psychological Science.

Trzesniewski, K. H., Donnellan, M. B., & Robins, R. W (2008a). Is “Generation Me” really more narcissistic than previous generations? Journal of Personality, 76, 903-917.

Trzesniewski, K. H., Donnellan, M. B., & Robins, R. W. (2008b). Do today's young people really think they are so extraordinary? An examination of secular changes in narcissism and self-enhancement. Psychological Science, 19, 181-188.

Twenge, J. M. (2006). Generation Me: Why today’s young Americans are more confident, assertive, entitled – and more miserable than ever before. New York: Free Press.

Twenge, J.M. (2008). Generation Me, the origins of birth cohort differences in personality traits, and cross-temporal meta-analysis. Social and Personality Psychology Compass, 2/3, 1440–1454.

Twenge, J. M., & Foster, J. D. (2008). Mapping the scale of the narcissism epidemic: Increases in narcissism 2002-2007 within ethnic groups. Journal of Research in Personality, 42, 1619–1622.

Twenge, J. M., Konrath, S., Foster, J. D., Campbell, W. K., Bushman, B. J. (2008). Egos inflating over time: a cross-temporal meta-analysis of the Narcissistic Personality Inventory. Journal of Personality,76, 875-902.

U.S Census Bureau. (2005a). School enrollment. Retrieved August 26, 2007, from http://www.census.gov/population/www/socdemo/school/cps2005.html, Table 9.

U.S Census Bureau. (2005b). Source of the data and accuracy of the estimates for the October 2005 CPS Microdata File on School Enrollment. Retrieved August 26, 2007, from http://www.census.gov/population/www/socdemo/school/2005SchoolS&A.pdf, Table 2.

 

The Limited Value of “Generation Me” in the Jury Box*

by M. Brent Donnellan and Kali H. Trzesniewski

We appreciate the comments and thought-provoking responses to our original paper.  All in all, we have gained greater insight into whether any of the current discussions of generational change (or lack thereof) have practical relevance for trial consultants.  In this response we briefly comment on this issue as well as mention a few of the other salient issues that have emerged in this exchange.

Debates Over “Generation Me” Have Limited Value for Trial Consultants

            A major role of the trial consultant is to provide guidance on the selection of potential jurors for specific trials.  Many of the responses from the trial consultants hinted that our exchanges with Twenge and Campbell provided little in the way of guidance for this basic task.  We agree.  It is interesting and even fun to debate research methodology, speculate about cultural changes, and engage in pop sociology; however, those are ultimately tangential issues for the day to day tasks of the jury consultant. Attorneys facing a pool of potential jurors are presented with a cross-section of people of different ages and backgrounds. The most basic message of personality psychology is that there are substantial individual differences in the traits that are likely to be relevant for a particular trial (e.g., prejudice, authoritarianism, empathy, intelligence).  A trial consultant might like to know whether any of these relevant individual differences are linked with age, a variable that is completely redundant with birth cohort in this context. Knowing for certain whether the average 22 year old of 2009 is different from the average 22 year old of 1979 is of little value. The relevant issue is how much predictive value the age of a potential juror at this moment in time has for understanding how that individual will behave in the context of a specific trial.  Today’s trial consultants must deal with age differences as they manifest themselves in 2009. 

             Let’s assume that a legal team determines that the attributes assessed by the Narcissistic Personality Inventory (NPI; Raskin & Terry, 1988) are relevant for a particular trial so that the team is interested in potential age differences on this measure.  Foster, Twenge, and Campbell (2003) reported one of the most well known studies on this topic.  They conducted a cross-sectional study in which they collected responses to the NPI from a sample of participants who responded to an advertisement on the internet.  Note that with this design, age and birth year are perfectly correlated just as they are in any real jury selection context.  Foster et al. found that the correlation between age and NPI scores was -.17 (p. 476); participants aged 20 to 24 scored about 15.50 on this measure, whereas participants aged 40 to 44 scored about 12.50.  The overall standard deviation was 6.7.  Let’s ignore for the sake of this argument that Foster et al. was a convenience sample of self-selected participants and stipulate that their pattern of results generalizes to the population of potential jurors. 

Just how diagnostic is age in this context?  We assert that age (and thus generation) is relatively uninformative in the context of actual jury selection when considering NPI scores.  We will simply quote Foster et al. (2003) directly on this point “[a]ge explains approximately 4% of the variance in reported [NPI scores], not a large amount by any measure” (p. 481).  This conclusion seems to stand in stark contrast to the more optimistic conclusions about the importance of generation made by Twenge and Campbell.

To be sure, attorneys for both sides are permitted to ask questions of potential jurors both through direct questioning and through questionnaires as part of voir dire in hopes of winnowing away unqualified or unsuitable jurors.  In other words, all parties involved in the legal dispute have a chance to get to know something about the individual qualities of potential jurors. Against that background, we think that those engaged in jury selection would be far better off if they focused on direct indicators of case-relevant traits rather than dwell on distal demographic variables like age or generational status.  Drs. Keene and Broda-Bahm and Ms. Foley seem to agree on this point and all three provided helpful applied guidance on this topic.  Wasting a preemptory challenge on a potential juror simply because she or he was part of Twenge’s so-called “Generation Me” does not strike us as good practice.  We would push our recommendation a bit farther by suggesting that the possible stereotypes inspired by “Generation Me” type-accounts might get in the way of removing the most “unfit” potential jurors for a given trial.  In sum, there is nothing that we read from Twenge and Campbell or from the professional trial consultants that made us rethink our tentative conclusion that current academic discussions over potential generational changes in narcissism or any other “Generation Me” trait have little applied value in this context.

Responding to Twenge and Campbell

            As it stands, we have other replies to Twenge and her associates that are either published or in press in academic journals (Donnellan & Trzesniewski, in press; Donnellan, Trzesniewski, & Robins, 2009).  We refer readers to those original papers for more detailed responses.  Here we want to simply make three points.

            First, we have never publicly speculated as to why Twenge and Campbell believe what they write.  Unfortunately, a reader might get that impression based on their description of one of our papers: “Trzesniewski & Donnellan (in press) have also argued that we believe there are generational differences because older people’s perceptions have changed over time.”  Our argument was actually based on an empirical article that is very much worth considering in this context (Eibach, Libby, & Gilovich, 2003).  Eibach et al. noted that there is a long history whereby members of an older generation criticize and pass moralistic judgments on the younger generation that begs for explanation.  Borrowing on their work, we suggested that one explanation for negative perceptions of “Generation Me” might be cognitive biases that prompt older adults to mistake changes that have occurred within themselves for changes in the outside world.  We refer readers to the original Eibach et al. paper for further insight and empirical evidence about this proposed process. 

            Second, Twenge and Campbell assert that the U.C. Davis data regarding NPI scores support their conclusions about an epidemic of narcissism.  This is not entirely accurate (see also Donnellan et al., 2009), and it is important to emphasize that this same assessment has also been reached by an outside research team.  Specifically, Roberts, Edmonds, and Taylor (in press) added a new NPI mean from a recent college student study conducted at the University of Illinois (Mean = 15.69) and the U.C. Davis means into the Twenge et al. (2008) meta-analytic database.  This aggregation had the effect of largely eliminating the original secular increase in narcissism reported by Twenge et al.  This hardly seems like the outcome that one would expect if all of the NPI data are in agreement. 

            Finally, Twenge and Campbell attempt to place their effect sizes for generational change in a meaningful context by comparing their work with “benchmark” effects in medical science.  However, comparisons between psychological effect sizes and medical effect sizes can be very misleading (Ferguson, in press).  Ferguson even warns that “Comparisons between medical research and psychological research should simply be avoided.” One of the major reasons for his recommendation is that there are potential errors in the ways that psychologists typically calculate medical effect sizes that generally depress the relevant medical coefficients.  Another reason for his recommendation, particularly relevant for our perspective, is that effect sizes in medicine are based on dependent variables that have a clear meaning that take no special training to appreciate (e.g., death or morbidity).  The majority of generational effect sizes referenced by Twenge and Campbell are not intuitive because they refer to the connections between year of data collection and variables that are measured on an arbitrary metric.  Most people have a good sense as to what it means to die from lung cancer whereas few people, including us, have a good grasp of the applied significance of a two point difference in NPI scores from 1982 to 2006. 

We can press this point about effect size interpretation even further.  The following statement is consistent with existing research: The gender difference in tendermindedness (r = .41; see Hyde, 2005, p. 585) is relatively close in magnitude to the newly calculated effect size for the connection between aspirin use and reduced risk of heart attacks (r = .52; Ferguson, in press).  Although the sentence may sound impressive, the issue is whether it helps to clarify the relative importance of gender or aspirin to these very different outcomes.  We doubt it.  Moreover, we can add yet another benchmark effect size that further illustrates our concerns with effect size comparisons made by Twenge and Campbell.  Ferguson (in press) calculated that the correlation between ESP and accuracy was about .16 (see Bem & Honorton, 1994).  This is the size of the Twenge et al. (2008) individual effect size for increases in the NPI from 1982 to 2006 (see Trzesniewski, Donnellan, & Robins, 2008).  Thus, it might be fair to point out that the generational change in the NPI evident in Twenge et al. (2008) was just as strong as the effects of ESP on accuracy in one experimental paradigm.  The bottom line is that we are not convinced that any of these comparisons ultimately helps to clarify the practical significance of the original generational findings. 

Conclusion

            At this point there is probably nothing more that can be done to reconcile our views with those of Twenge and Campbell.  In closing, we would simply point out that our reading of the existing literature on generational change is more or less consistent with an earlier report issued by the National Research Council in which a panel of prominent scholars considered the implications of potential generational changes in youth attitudes and behaviors for military recruiting (Sackett & Mavor, 2003).  We believe that their conclusions hold equally well today:

“[C]ontrary 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” (p. 305-306).

*Both authors contributed equally.

We thank Martin Carr, Christopher Ferguson, Christopher Hopwood, Barbara O’Brien, Aaron Pincus, and Donna Shestowsky for helpful insights on this response.

 

References

Bem, D., & Honorton, C. (1994). Does psi exist? Replicable evidence for an anomalous process  of information transfer. Psychological Bulletin, 115, 4-18.

Donnellan, M. B. & Trzesniewski, K. H. (in press). How should we study generational change – or should we? A critical examination of the evidence for “Generation Me.” Social and  Personality Compass

Donnellan, M. B. &Trzesniewski, K. H., & Robins, R. W. (2009). An emerging epidemic of narcissism or much ado about nothing? Journal of Research in Personality, 43, 498-501.

Eibach, R. P., Libby, L. K., & Gilovich, T. D. (2003). When change in the self is mistaken for change in the world. Journal of Personality and Social Psychology, 84, 917-931.

Ferguson, C. J. (in press). Is psychological research really as good as medical research? Effect size comparisons between psychology and medicine. Review of General Psychology.

Foster, J. D., Campbell, W. K., & Twenge, J. M. (2003). Individual differences in narcissism: Inflated self-views across the life span and around the world. Journal of Research in Personality, 37, 469-486.

Hyde, J. S. (2005). The gender similarities hypothesis. American Psychologist, 60, 581-592.

Raskin, R., & Terry, H. (1988). A principal-components analysis of the Narcissistic Personality Inventory. Journal of Personality and Social Psychology, 54, 890–902.

Robert, B. R., Edmonds, G., & Taylor, E. (in press). Every generation is Generation Me: Developmental changes are more important than generational changes in narcissism. Perspectives in Psychological Science.

Sackett, P. S & Mavor, A. S. (Eds.) (2003). Attitudes, aptitudes, and aspirations of American youth: Implications for military recruitment. Washington, DC: National Academies Press.

Trzesniewski, K. H., Donnellan, M. B., & Robins, R. W (2008). Is “Generation Me” really more narcissistic than previous generations? Journal of Personality, 76, 903-917.

Twenge, J. M., Konrath, S., Foster, J. D., Campbell, W. K., Bushman, B. J. (2008). Egos inflating over time: a cross-temporal meta-analysis of the Narcissistic Personality Inventory. Journal of Personality,76, 875-902.

---

A reply: Why generational differences have an impact

By Jean M. Twenge and W. Keith Campbell

 

Although we appreciate the cordial tone of this debate, there were several misconceptions that need to be remedied.

• The most serious misconception, raised by several authors, is that the differences we find are due to age and not generation. For example, Foley states, “normative personality development with age is a better explanation for any trend detected in data presented by Twenge and Campbell.” This is not possible, as our data rely on people of the same age, comparing them across historical time. For example, the study on narcissism compares college students who completed the measure at some point between 1982 and 2006. Thus it compares different generations when they were the same age, so age cannot be causing any of the differences. That is, in fact, the most significant strength of the research method we use.

• Foley states, “Just because some people reported they experienced symptoms on the NPI measurement does not mean they are narcissistic or behave in narcissistic ways.” Yes, it does. The NPI is the definitive measure of narcissism among normal populations. Thus when we report research about the negative outcomes associated with narcissism, these are the outcomes of someone who scores high on the NPI, because narcissism is defined as a high NPI score. This is also why the statement, “Maybe someone who scores high on the NPI is just a confident leader, not a narcissist” is also incorrect — unless one believes that confident leaders regularly lack empathy, shirk responsibility for failure, take more for themselves and leave less for other people, and act aggressively toward those who insult them, all outcomes associated with high NPI scores.

• Keene states, “A truly narcissistic person wouldn’t feel a need to undergo modification—they are already perfect. Cosmetic surgery is for the ‘worried well’—those who think that they aren’t quite good enough, young enough, appealing enough.  Anxious or depressed? Maybe. Narcissistic?  I don’t think so.” According to several medical journal articles, this statement is wrong. Narcissistic personality disorder (NPD), along with histrionic personality disorder and body dysmorphic disorder, are the three most prevalent personality issues in those seeking cosmetic surgery  (Malick, Howard, & Koo, 2008; Napoleon, 1993).

About 25% of cosmetic surgery patients have been diagnosed with NPD – a very high figure, considering that 6% of the U.S. population, on average, has ever suffered from NPD (Napoleon, 1993). 

• Keene states, “Yes, [NPD] does exist as a separate syndrome, but as Trzesniewski and Donnellan’s article points out, it is very unusual.” The National Institutes of Health study  we cited in our original article – the most definitive on the topic to date – finds that almost 10% of Americans in their 20s have already experienced NPD; the prevalence across all age groups was 6%. Thus, NPD is far from unusual. And if almost 1 out of 10 people in their 20s experiencing NPD isn’t an epidemic, we don’t know what is.  

• Trzesniewski and Donnellan cannot conclude from their MTF analyses that there are few generational differences, as they analyzed only 135 items of the more than 1,000 items in the Monitoring the Future dataset. As these 135 variables were not randomly chosen, and the MTF dataset does not contain all variables sensitive to cohort change, counting how many variables show change is not informative.  

• The “convenience samples” argument made by Trzesniewski and Donnellan and picked up by several of the commentators is problematic. There is a fair amount of debate about how to define a convenience sample. A collection of one’s friends would certainly qualify, as they would likely differ in substantial ways from a random sample of the general population. But college students? As long as they weren’t selected on a particular attribute, the worst that can be said is perhaps the results only apply to college students. This is not as selected a group as some have argued: 67% of high school graduates enroll in college. At the worst, one could conclude that the generational differences apply only to those who have had some college, which is still a good proportion of a jury pool. The most important point: This is not a challenge to the validity of the studies (as, for example, Foley states), but to their generalizability – a very different question. The studies are still valid (meaning they show a true effect) even if they apply to only part of the population.

In addition, many of the cohort changes do apply to broader populations, as the Monitoring the Future dataset shows many results consistent with the college student data. In addition, my meta-analyses have replicated many of the college student results among samples of elementary, middle school, and high school students. Interestingly, by the definition of a convenience sample used by Trzesniewski and Donnellan, the MTF dataset is also a convenience sample, as only about three out of four of students agree to participate, and many of those don't answer all of the questions. It also does not include anyone who dropped out before spring of their senior year of high school. The MTF study originators note that "nonresponse in the MTF is more common among boys, nonwhites, students in lower academic tracks, and students with lower grade point averages." So it’s not a perfect sample either, and the people who don't participate are systematically different from the people who do. 

• Trzesniewski and Donnellan state that there are more similarities among generations than differences. However, this is true of virtually every study finding average differences among groups. For example, there are many more similarities among men and women than differences, but few argue that sex differences don’t exist or are not meaningful, even though they are the same size (and often even smaller) than generational differences. Broda-Bahm makes a similar argument in saying, “one cannot presume that an individual carries the same attitudes as the majority of her gender, race, or age-cohort.” We agree – these are differences on average, and will not apply to every individual.

•Foley states, “This all indicates to me that the baby boomers are probably driving trends Twenge and Campbell describe, such as an increase in the square footage of homes and a fivefold increase in plastic surgery.” Yes, they did drive those trends, and we address this in the article – the rise in narcissism goes beyond generations, and beyond individual personality traits. And since the rise in narcissism and all of the other traits is linear, it is not just possible but likely that this trend started with the Baby Boomers. But the data clearly show it going up from there.

• Foley seems to misunderstand our argument about not using generational labels. The changes are linear, so the narcissism epidemic has not affected just one generation; narcissism has slowly continued to rise. The later one was born, the higher one is likely to score on narcissism. This obviates the need for debate over whether the cutoff between GenX and GenY should be 1970, 1977, 1982, or some other date. The GenMe label is not meant to replace these other generational labels; it is simply descriptive of personality traits that are now more common. 

References

Malick, F., Howard, J., & Koo, J. (2008). Understanding the psychology of the cosmetic patients. Dermatologic Therapy, 21, 47-53.

Napoleon,  A.  (1993). The presentation of personalities in plastic surgery. Annals of Plastic Surgery,  31, 193–208. 

 

Citation for this article: The Jury Expert, 2009, 21(3), 1-31.