The Myth of the Spoiled Child

by Alfie Kohn

  43. Penn. This passage and the quotations that follow appear on pages 113–17.

  44. One could raise questions about whether the poll respondents are truly representative of the general population and also about the accuracy of people’s self-reports, but these questions also apply to many studies published in scholarly journals. What distinguishes the latter is a reliance on measurement instruments whose validity and reliability have been established. A fair amount of thought would go into constructing a definition of permissiveness, which might be based on whether respondents strongly agree with, agree with, are undecided about, disagree with, or strongly disagree with a series of statements (overall summaries of their parenting practices as well as examples of how they tend to deal with specific challenges that arise). The relationships among those individual responses would be carefully measured. Ideally, the responses would be compared to those provided by the children in some of the same households or even to samples of actual parent-child interaction as observed and coded by the experimenters. Now compare that methodical validation process to a poll that includes a single question: “Would you describe your parenting as strict or permissive?”

  45. The single exception is his reference to a 1968 poll finding about attitudes toward spanking. See below.

  46. For example, a classic study, which has been confirmed by more recent research, found that the babies most likely to fuss and cry were those whose parents hadn’t responded promptly to their cries earlier. Parents who held back for fear of spoiling often set a vicious circle into motion: Ignoring the baby’s cries led to more crying as the baby grew, which further discouraged the parent from responding, which made the baby even more irritable, and so on (Bell and Ainsworth). More generally, research has confirmed that mothers who were most worried about spoiling their babies were least likely to provide a warm, caring, emotionally supportive environment (Luster et al.).

  47. For a more detailed examination of time-out, a technique sometimes classified in the psychological literature as “love withdrawal,” see Kohn 2005a.

  48. Another national poll, taken in 2005, reported that 72 percent of American adults think it’s “OK to spank a kid.” See http://www.surveyusa.com/50StateDisciplineChild0805SortedbySpank.htm.

  49. “Even as the percentage of Americans who approve of spanking has fallen dramatically in the last half century, the actual incidence of it has barely budged” (Burnett, p. 18). For example, “data from a nationally representative longitudinal study of over 21,000 children . . . found that by the time these children reached the fifth grade in 2003, 80% had been corporally punished by their parents” (Gershoff and Bitensky, p. 232).

  50. For examples, see Kohn 1999b.

  51. E. Shapiro.

  52. For example, Gentile et al.; Twenge 2013a; Twenge and Campbell 2008; Twenge and Foster.

  53. Twenge 2006; Twenge and Campbell 2010.

  54. One might be able to devise a psychological scenario in which these two characteristics are theoretically compatible, but empirically there is a significant negative correlation between them (Trzesniewski and Donnellan 2010, p. 59). Also, Twenge is extremely critical of how young people today are ostensibly too concerned with being happy, yet she then criticizes them (or their priorities) on the grounds that they’re less happy than earlier generations—a claim that other researchers have failed to replicate, incidentally (see below).

  55. For more on the hazards of generalizing from such “cross-temporal meta-analytic” studies, see Trzesniewski and Donnellan 2010.

  56. “Adaptive narcissism is characterized by healthy ambitions, energy, creativity, and empathy, supported by an underlying sense of self that is firm and cohesive” (Cramer, p. 19). This isn’t just a theoretical interpretation; some of the dimensions listed on the Narcissistic Personality Inventory are empirically correlated with healthy adjustment (Watson et al.), particularly for younger adults (Hill and Roberts). Twenge’s assertion that “narcissism is one of the few personality traits that psychologists agree is almost completely negative” (2006, p. 68) is patently false.

  57. Ramsey et al., p. 232, found a negative relationship between them. For a list of several studies that found them to be positively correlated—albeit modestly—see Foster et al., p. 471. But again, a positive correlation between self-esteem and overall narcissism scores may be misleading because the strength and direction of that relationships depend on which subscales—that is, which components of narcissism—one chooses to look at. See Ackerman et al.

  58. Whereas the prevalence of narcissism apparently decreases with age (see note 66 below), “Self-esteem increases during young and middle adulthood, reaches a peak at about age 60 years, and declines in old age” (Orth et al. 2010, p. 652; also see Erol and Orth—and Wagner et al. for data showing that any decline in old age is minimal). On pages 125–26, I discuss the conceptual differences between self-esteem and narcissism or grandiosity.

  59. “It is not altogether clear what this would mean because the NPI measures a multifaceted construct and the individual facets are not always strongly intercorrelated” (Trzesniewski et al. 2008b, p. 910).

  60. Donnellan and Trzesniewski, pp. 7–8.

  61. Brent Roberts, personal communication, January 2013.

  62. Donnellan et al.

  63. Trzesniewski et al. 2008a, p. 184; Trzesniewski and Donnellan 2010; Donnellan and Trzesniewski, pp. 4–6.

  64. Erol and Orth; Orth et al. 2010. Quotation from the former study, p. 614.

  65. Liberman 2013.

  66. Roberts et al. Quotations appear on pp. 100–1. Even Twenge and her collaborators, in a different study that collected self-reports from thousands of people at once on the Internet, found that older people were less narcissistic than younger people (Foster et al.).

  67. Konrath et al. The change “is admittedly minor (i.e., about one third of a scale point) . . . and still leaves today’s college student around the midpoint in these traits” (p. 187). Even assuming this finding is replicated by other studies, which remains to be seen, it’s not clear what significance to attach to the timing of this decline, which doesn’t correspond to Twenge’s, or anyone else’s, claims about when young people were supposed to have begun exhibiting more narcissistic tendencies. In any case, a lower score on these measures may not have much significance. As Dan Batson, a leading authority on altruism and empathy, commented about this study, “The idea that [students are] less capable of caring than they were 20 years ago—that just seems unlikely. I don’t think we change like that. But our situation may have changed. One may feel pressure to pull back on the scope of one’s concern . . . and say, ‘I’ve got to deal with the needs that are pressing right here’” (quoted in O’Brien). One further concern: Are self-reported levels of empathy meaningful, or is it possible that “what we’re measuring is not changes in what they really feel, but rather changes in what they think that they should want us to think that they feel?” (Liberman 2012).

  68. Pryor et al., p. 26; Astin et al., p. 45. Of course this surge may reflect students’ perception that volunteer work will enhance their chances of being admitted to a selective college.

  69. This was the finding of a 2012 Pew Research survey. See www.pewsocialtrends.org/2012/02/09/young-underemployed-and-optimistic/5/.

  70. This conclusion is offered by researcher Keith Macky and his colleagues, quoted in Donnellan and Trzesniewski, p. 8. Emphasis added.

  71. On these last two points, see Arnett 2007 and 2010.

  72. Jencks, p. 89.

  73. All societies require a certain degree of conformity, he said, but when “skepticism and resistance to established authority ceases, democracy becomes a mere façade for preserving the status quo” (Jencks, p. 76).

  74. Agnew is quoted in a book by journalist Jules Witcover, which is cited in Ehrenreich, p. 71. Presumably these rhetorical flourishes were created by one of Agnew’s two speechwriters, Patrick Buchanan and William Safire.

  75. Ehrenreich, pp. 71–72.


  76. Flacks, pp. 142–43. Like Jencks, Flacks points out that the characterization of these parents as permissive is mistaken to begin with: “The humanist values of activist parents leads . . . toward an emphasis on encouraging the autonomy of the child and a skepticism about conventional moral standards. But such child-rearing practices do not seem to be characterized by indulgence, or by a failure by the parents to assert standards. As a matter of fact, activists’ parents typically had high expectations and standards—for instance . . . they strongly and directly influenced their children to maintain interest in intellectual and creative activity, in academic work, and in socially useful activity.” Flacks also found that student activism “is strongly related to the political interests and attitudes of their parents” (p. 142). But that doesn’t mean that parenting style is completely unrelated to children’s activism, as I’ll explain in chapter 8.

  77. This assertion was made by Kevin Ryan, founder of the Center for Character and Social Responsibility at Boston University. It’s contained in an undated list entitled “Ten Tips for Raising Children of Character,” which appears on a number of websites.

  78. For an argument that, “without corroborating evidence, we cannot trust people’s memories about early childhood,” see Halverson. Quotation appears on p. 436.

  79. Givertz and Segrin; Otway and Vignoles; Ramsey et al.; Watson et al. 1992. Another researcher looked at “pampering” (which he divided into four subcategories) and compared them to overall scores on the NPI as well as to seven of its scales. Of forty comparisons, only eight reached conventional levels of statistical significance and those were all fairly small. There was no relation between overall pampering in childhood and the summary narcissism score (Capron).

  80. Cramer, p. 26. This was one of the rare studies that followed a sample from early childhood to early adulthood. It included assessments by the children’s parents.

  81. Taylor et al. As the authors succinctly explained the result, “The child learns to be aggressive by being treated directly with aggression” (p. 1063).

  82. I cite a series of studies in Kohn 1992, esp. pp. 138–43, 238. Ironically, some of the same people who criticize young people for being self-centered are also enthusiastic proponents of setting children against one another in contests and are critical of parents and teachers who look for alternatives to competition (e.g., Twenge 2006). For more about competition, see pp. 34, 81–84, 107–10.

  CHAPTER 2

  1. Unconditional Parenting (Kohn 2005a).

  2. For a list of examples, see Kohn 2005a, p. 4.

  3. Gosman, p. 180; Shaw, p. 115; Mogel, pp. 72, 124.

  4. Hart et al.

  5. Undoubtedly the relationship is reciprocal, but developmental experts have established conclusively that parenting styles, which reflect stable patterns and attitudes, affect children’s behavior above and beyond the extent to which they’re affected by that behavior.

  6. Statistics Canada.

  7. Rhee et al.

  8. For a comprehensive review of the research, see Gershoff. For examples of studies conducted since that review, see Taylor et al.; Gromoske and Maguire-Jack; and MacKenzie et al.

  9. Both styles were related to what are called internalizing symptoms (anxiety, depression), but even here the association to authoritarian parenting was stronger than the relation to permissive parenting (Cohen et al.)

  10. Givertz and Segrin.

  11. Baumrind is strikingly conservative in her beliefs about children, parenting, and human nature. She strongly supports the use of extrinsic motivators and “contingent reinforcement,” which she assumes are required in the service of family “structure.” She approves of spanking, dismisses criticisms of punishment as “utopian,” and declares that parents who don’t use power to compel obedience will be seen as “indecisive” (Baumrind 1996). Elsewhere, she has written, “The parent who expresses love unconditionally is encouraging the child to be selfish and demanding” (Baumrind 1972, p. 278). These values are baked into the notion of “authoritative” parenting, which, she acknowledges, is “no less power-assertive than the disciplinary style of authoritarian parents” (Baumrind 2012, p. 41), despite the fact that it is positioned as—and derives its appeal from appearing to be—a moderate alternative to the two extremes.

  12. This begins with Baumrind’s original investigations. When a researcher named Catherine Lewis looked carefully at the data Baumrind had provided, she found that positive outcomes for authoritative parenting weren’t at all a function of the use of firm enforcement. Children whose parents were high in warmth but low in “demandingness” did just as well as children whose parents were high in both—probably, Lewis suggested, because control in the traditional sense isn’t actually required to create structure and predictability as Baumrind and many others assumed (Lewis 1981).

  By the same token, Baumrind originally blurred the important differences between “permissive” parents who were neglectful and those who were deliberately democratic. Children in the latter group were not at risk, and that suggests, as another psychologist put it, that “a close look at Baumrind’s actual data may reveal significant support for child-centered parenting” (Crain, p. 18) even though she personally opposes that style. (Eventually Baumrind did introduce these and other distinctions by adding several new categories, including “democratic” parenting, to her original three.)

  Subsequent research supports the view that child-centered parenting, rather than the kind Baumrind prefers, has positive effects. A huge study of teenagers (Lamborn et al. 1991) found benefits from what was described as “authoritative” parenting, but that term had been defined to mean that parents were aware of, and involved with, their children’s lives, not that they were even the least bit punitive or controlling. Another study (Strage and Brandt) similarly cited Baumrind by way of suggesting that parents need to be both supportive and demanding, but it turned out that being demanding when their children were young was unrelated, or even negatively related, to various desirable outcomes. By contrast, the extent to which the parents had been supportive, and also the extent to which they had encouraged their children’s independence, had a strong positive relationship to those same outcomes. A third study (Weiss and Schwarz) used college students as subjects and classified them according to six parenting styles. Those raised by permissive (“nondirective”) parents fared quite well overall, their outcomes closely resembling those from “authoritative” families. That fact was clear only because the investigators distinguished permissiveness from a more detrimental “unengaged” approach in which parents weren’t just neglecting but rejecting. In all, the “distinguishing factor in lateadolescent outcome” wasn’t control or firmness but “parental supportiveness” (quotation on p. 2111).

  13. All of these ideas are described at some length in Kohn 2005a.

  14. The three studies: Kochanska et al.; Joussemet et al. 2005; and Garber et al. Also, a review of research over more than four decades and across cultures finds that the extent to which people experienced their parents as accepting rather than rejecting is strongly associated with the quality of their psychological adjustment (Rohner). For an overview of research on the benefits of autonomy-supportive parenting, see Joussemet et al. 2008.

  15. Many studies have found this to be true. For three recent examples, see Zhou et al., Knafo and Plomin; Mikulincer et al.

  16. See Hoffman and Saltzstein; and Krevans and Gibbs. Strayer and Roberts found that “fathers who were authoritarian and mothers who used anxiety and guilt control” had children who were more likely to be angry. That anger, in turn, was negatively related to empathy (p. 246).

  17. I made this point in a book called The Brighter Side of Human Nature (Kohn 1990), drawing from the work of Martin Hoffman, Ervin Staub, and others. It was empirically confirmed by a group of researchers in 2005, who explained that people who felt securely attached as children are more likely to “perceive others not only as sources of security and support, b
ut also as suffering human beings who have important needs and therefore deserve support. . . .

  Thus, if we wish to help children and adults develop their natural potential for compassion and altruism, one way to do so would be to help them achieve attachment security” (Mikulincer et al., pp. 818, 837).

  18. Trzesniewski and Donnellan (2010, p. 70) invoke this bias to explain conclusions about “Generation Me.” They credit identification of the bias itself to Amos Tversky and Daniel Kahneman.

  19. Eibach et al.; and Eibach and Libby. The first quotation is from the former, p. 918; the second is from the latter, p. 419. Such beliefs may be reversible. These researchers discovered that, while many people think adolescents today are less moral than adolescents in earlier times, adults who were specifically asked (by the experimenters) to reflect on how their own moral perspectives had changed became less likely to make that assumption.

  20. Mogel, p. 62.

  21. Ibid., p. 79.

  22. Okay, our position.

  23. Shaw, p. 131.

  24. For example, see Bird and Bird; Gosman.

  25. Zuk.

  26. Psychologist: Arnett 2008, p. 677. Essayist: Hoagland, p. 34.

  27. Luster et al.

  28. I offer a number of examples pertaining to classroom management and character education in Kohn 1996, chapter 1; and Kohn 1997a, respectively.

  29. Mamen, pp. 24, 29. Mogel’s book is replete with similar examples.

  30. Brooks 2010.

  CHAPTER 3

  1. Orenstein and Goldstein, respectively.

  2. Atlantic blog: Lahey. Australian study: Locke et al.

  3. Blog post: Belkin 2010. Column: Belkin 2009. Law review article: Bernstein and Triger.

  4. This may have been the import of a New Yorker cartoon in which a mother standing next to her son confides to his teacher, “It’s a lot of pressure on me not to pressure him.” Indeed, one small study—featuring in-depth interviews of fifty-six elementary school students and their parents—found that middle-class parents “did not—as commonly thought—swoop in to ‘save’ the student from problems in school” (Sparks, p. 11). In fact, these parents seemed “very conscious of the helicopter parent stereotype, saying things like ‘I’m not one of those helicopter moms.’ . . . They recognized that they could not (and possibly should not) always be there to do for their children, and understood that schools expect children to take responsibility for their own learning needs.” (The latter quotation is from a personal communication from the study’s author, Jessica McCrory Calarco, in August 2012.)