What Should Children Learn? Americans’ Changing Socialization Values, 1986–2018

Abstract

Assessing changes in socialization values for children provides a unique window into how Americans perceive the landscape of their society. We examine whether, since the mid-1980s, Americans (b) emphasized survival values, like hard work, for children as economic precarity rose, or (a) prioritized self-expression values, like autonomy and compassion, as expected in postindustrial society. Analysis of 1986–2018 General Social Surveys (N = 23,109) supports the precarity thesis: preferences for hard work increased steadily whereas preferences for autonomy, the top-ranked quality throughout the period, declined. There was some indication of enduring self-expression values, as support for compassion increased and its relative importance to hard work stayed stable. Decomposition analyses show valuing hard work would have been even greater without demographic changes like an increase in college graduates. Aligning with earlier research, valuing obedience in children continued to decline. Our results extend theoretical work on complexities of socioeconomic links to parenting values.

What qualities matter most for children’s futures? Values for raising children are reflective of how Americans see the landscape of their society, and allow for a unique window into the economic and cultural climates of the country. Research on trends in Americans’ socialization values for children shows that from the early 20^th century to the early 1980s, the most preferred trait for children shifted from obedience to autonomy (Alwin 1984, 1988, 1989, 2001; Lenski 1961; Lynd and Lynd 1929). As Alwin (2001) noted, a key explanation for this shift is drawn from modernization theories, which contend that a nation’s economic development brings fundamental cultural changes (Inglehart 1997). According to Inglehart and colleagues, such changes can be observed across two dimensions (Inglehart 1997; Inglehart and Baker 2000; Inglehart and Welzel 2010). The first shift, from traditional to secular-rational values, occurs during the transition from agrarian to industrial societies. The second shift, from survival to self-expression values, occurs with the rise of a postindustrial society where the increase in economic security allows people to seek “higher order” needs like self-expression, freedom, and individual well-being. As such, the decline in emphasis on obedience for children from the early 20^th century to the early 1980s represented a shift away from traditional, survival values, whereas the increase in emphasis on autonomy for children reflected a move towards secular-rational, self-expression values (Alwin 2001; Inglehart and Welzel 2010).

In this article, we examine how Americans’ values for children shifted since the mid-1980s, focusing on the survival versus self-expression dimension. Have Americans’ values moved further to prioritize self-expression values, while shifting away from survival values, as expected in a postindustrial society? Drawing on scholarship on the nation’s economic climate, intensive parenting trends, and attitudes towards marriage and family life, we generate two contrasting predictions. First, much research underscores what we call the precarity thesis. Since the mid-1980s, changes in the labor market have increased a sense of economic insecurity among Americans (Fullerton and Wallace 2007; Hollister 2011; Kalleberg and Marsden 2013). We thus expect that Americans’ beliefs in desirable traits to instill in children have shifted away from traits designating self-expression values, like autonomy and compassion, to traits indicating survival values, like hard work, for achieving economic security. Second, in contrast to the precarity thesis, other trends suggest enduring expressivity such as attitudes moving away from traditional norms of marriage and childbearing (Cherlin 2004) and the increase in supportive attitudes toward LGBTQ partnerships and parenting (Baunach, 2012; Loftus 2001). We thus expect that the emphasis on self-expression values for children’s future, such as preferences for autonomy and compassion, may be continuing, despite the increase in economic insecurity.

We use data from the 1986 to 2018 General Social Survey (N = 23,109) to examine these predictions. Since 1986, the GSS, using a measure developed by Lenski (1961), has asked respondents to rank five traits in order from the most to the least important for children to learn to prepare themselves for life: obedience (“to obey”), autonomy (“to think for oneself”), diligence (“to work hard”), compassion (“to help others”), and likability (“to be popular”). Unlike prior research, which typically compared preferences for autonomy relative to obedience exclusively (e.g., Alwin 2001; Starks and Robinson, 2005), we focus on three traits that are well-suited to examine change in the survival versus the self-expression dimension of values: diligence or hard work, which signifies survival values, versus autonomy and compassion, which indicate self-expression values (Doepke and Zilibotti 2019; Inglehart and Welzel 2010). To account for the possibility that the aggregate change in Americans’ values for children merely reflects population turnover (Firebaugh 1997), we conduct decomposition analyses to examine the extent to which the change in Americans’ values for children is accounted by changes in other variables that have been known to be related to Americans’ values for children, such as education level, high occupational status and race/ethnicity (Alwin 1984, 1989; Suizzo 2007; Xiao 2000). In additional analyses, to extend the line of prior research that focused heavily on the value of obedience, we examine the traditional versus secular-rational dimension of value changes to assess whether there is a further move away from traditional values.

AMERICANS’ CHANGING SOCIALIZATION VALUES FOR CHILDREN

Sociologists have long been interested in understanding the link between social changes and changes in adults’ socialization values for children, or desirable traits that adults hope to instill in children (Alwin 2001). During the first three-quarters of the 20^th century, the major shift observed was from obedience (e.g., to obey parents, to respect authority) to autonomy (e.g., independence, to think for themselves, imagination) in children. Reviewing work by the Lynds (1929), Alwin (1988) found a move away from an emphasis on obedience from the 1920s to the 1970s in Middletown, Indiana. Using data collected in Detroit, Michigan, Alwin (1984) also found that from 1958 to 1983, preferences for instilling obedience in children declined while preferences for autonomy increased. Using the data from the GSS, Alwin (1989) found the same shift from obedience to autonomy from 1964 to 1984.

A major explanation for the shift in important qualities for children from obedience to autonomy relies on theories on the effects of economic development on cultural changes. Inglehart and colleagues have theorized that two stages of economic development lead to two sets of value changes (Inglehart 1997; Inglehart and Baker 2000; Inglehart and Welzel 2010). The first shift is from traditional to secular-rational values, which occurs during the transition from agrarian to industrial societies when various social changes occur, including urbanization, a spread of wage labor, rising education levels, rising standards of living, and a declining influence of traditional religions. These social changes lead to a shift away from traditional values, which prioritize obedience, respect for authority, traditional religions, traditional family values, and rejections of divorce, abortion, euthanasia, and suicide. The second type of value transition is from survival to self-expression values, which occurs in a post-industrialized society where mass prosperity and advancement of social security programs in welfare states allow a large part of the population to experience a sense of long-term economic security. Economic security allows people to take survival for granted and instead strive for higher-order needs such as self-expression, free choice, tolerance, and individuals’ quality of life. The core concept of the postindustrial higher-order needs thesis is echoed by Lesthaeghe’s (2010) second demographic transition theory which contends that the advanced economy has satisfied people’s material preoccupations and allowed the rise of existential or expression needs, which emphasize self-actualization and autonomy. All these changes in broader values in society, in turn, are reflected in changes in socialization values for children from ones reflecting traditional and survivalist values, such as obedience, conformity, and respect for authority, to ones reflecting secular-rational and self-expression values, such as autonomy in decision making, trust, and quality of life concerns (Alwin 2001; Inglehart and Welzel 2010; Lesthaeghe 2010).

Since the mid-1980s, has the shift in the emphasis from survival to self-expression values continued in American’s socialization values for children as expected in a post-industrial society? We examine this question focusing on three specific traits asked for respondents in the GSS that manifest the survival versus self-expression values dimension of values—autonomy, compassion, and hard work. As discussed above, prior research has shown that preferences for autonomy reflect secular-rational, self-expression values (Inglehart 1997; Inglehart and Baker 2000). Compassion—to help others—also indicates secular-rational, self-expression values, as are similar values tapping concern for others such as trusting people, importance of friends, and tolerance (Inglehart 1997; Inglehart and Baker 2000). In contrast, hard work signifies survival values, which has been shown using data from the World Values Survey (WVS) (Doepke and Zilibotti 2019; Inglehart 1997). Below we discuss the two contrasting predictions for change in Americans’ values for children since the mid-1980s: (a) the precarity thesis—the rise of the importance of hard work, a survivalist value, for children to gain economic security; and (b) the enduring expressivity thesis—a continuous shift towards emphasizing autonomy and compassion, self-expression values.

Precarity: The Rise of the Importance of Hard Work for Children since the Mid-1980s

The precarity thesis emphasizes that since the mid-1980s, as uncertainties about economic security have risen, Americans’ values for children may have shifted away from self-expression values, like as autonomy and compassion, to survival values, such as hard work. Much evidence suggest an increase in economic insecurity among Americans during the past several decades. Since the mid-1970s, many changes have occurred to work and the labor market in the United States. Several broader changes, including, but not limited to, an increase in global competition, technological changes, and a shift in power from managers to shareholders, put pressure on employers to adopt new strategies to reduce production costs, make prices more competitive, and generate profits at a short-term basis (Hollister 2011). To achieve this goal, employers have adopted new employment practices, such as downsizing, outsourcing, and increased use of contractors and temporary workers. These practices have resulted in increases in job instability and a sense of job insecurity among American workers (Hollister 2011; Kalleberg and Marsden 2013). Fullerton and Wallace (2007) showed that controlling for annual unemployment rates, U.S. workers’ perceptions of job insecurity increased during the 25-year period from 1977 to 2002. The increase in job insecurity may have led a large segment of Americans to feel as if economic well-being can no longer be taken for granted.

Rising economic inequality is a different, though related, change in the U.S. that may have influenced Americans’ views of the economic climate in their children’s future. Some of the changes discussed above, especially globalization and technological changes, have led to a polarization of types of jobs, one with low skills and low pay, and the other with high skills and high pay (Autor, Katz, and Kearney 2006). To secure a job with high skills and high pay, a college degree has become more important than before; and income inequality between those with and without college degrees has increased (Castex and Dechter 2014). Doepke and Zilibotti (2019) argue that the increases in economic inequality and higher returns to education have led Americans to emphasize that children must work hard and succeed in school. Their empirical study focused on cross-national comparisons with no data for trends in U.S. adults’ values for children since the 1980s, however. U.S. research has shown that, since the 1990s, arguably as a response to the rise of a college premium, parents have increasingly adopted a parenting style often called “intensive parenting,” pushing their children to work hard in school and extracurricular activities to build their talents, skills, and resumes from a young age to ensure children’s positioning in a competitive global marketplace (Lareau 2003; Milkie and Warner 2014; Nelson 2010; Putnam 2015). These studies indicate that the value of hard work for children among Americans may have increased in the past three decades.

The potential rise of emphasis on hard work as a desired quality in children is suggested in educational psychology research as well. In the early 1980s, U.S. mothers tended to believe in the importance of children’s innate ability, rather than children’s effort and hard work, for their academic success (Stevenson and Stigler 1994). More recent publications in the U.S. on this topic, both in scholarly work and in bestselling books, have emphasized the role of perseverance, motivation, and hard work in children’s success (Duckworth 2016; Valiente et al. 2008).

In all, drawing on these research findings, the precarity thesis leads us to expect that preferences for hard work have increased relative to both the preferences for self-expression values such as autonomy and compassion since the mid-1980s.

Enduring Expressivity: The Continuous Importance of Autonomy and Compassion for Children

Alternatively, we expect that growing economic precarity and inequalities may not necessarily stop Americans from prioritizing self-expression values for children. The enduring expressivity thesis is underscored by research in work and family realms that indicate durable cultural goals of expressivity. Bell (1973) emphasizes types of job experience over economic insecurity in facilitating value changes. As the majority of the U.S. labor force consists of service-providing jobs (U.S. Bureau of Labor Statistics. 2019), more people are engaged in occupations that require them to deal with the well-being of other people on daily basis, which leads to values emphasizing quality of life, self-expression, and tolerance toward diversity (Inglehart 1997). The decline in lifetime employment with the same employer, another change in the nature of the labor market, has shifted the employment narrative to emphasize the importance of individuals’ crafting their own career paths (Hollister 2011), which suggest needs for skills to think and act independently.

Americans’ continuing valuing of self-expressivity is also emphasized by some family scholars. Changes in the institution of marriage since the 1980s have been discussed as exemplars of the shift toward self-expression values (Cherlin 2004). Evidence is clear in changing attitudes toward marriage and family issues. Using data from GSS, Gubernskaya (2010) found that more Americans in the early 2000s than in the late 1980s disagree with the idea that people who want children ought to get married; and that people who had never have children lead empty lives, suggesting more acceptance of diversity in family life. Although Americans are more likely than people in other advanced industrial nations to be attached to the institution of marriage, the ideal marriage involves egalitarianism, mutual self-growth, and self-fulfillment, which reflect self-expression values (Cherlin 2004). Another area that indicates an enduring shift toward the value of self-expression is the steady increase in support for gay marriage since the early 1990s (Loftus 2001). Using GSS data, Baunach (2012) found that Americans who strongly agreed or agreed that same-sex couples should have the right to marry one another increased from 12.6 percent in 1988 to 47.4 percent in 2010; and this shift was not due to demographic turnover but to changes in attitudes.

Studies that examined trends in the value of helping others are scarce. Yet, the increasing percentage of middle and high schools which encourage or demand service to the community as a part of fulfilling requirements for graduation may relate to an increase in Americans’ valuing trust, giving, and compassion (Helms 2013). The school policies, requirements and practices might reflect a perceived need to encourage children to help others because it creates bonds, trust, and community strengths. Some students may work to help those less fortunate (e.g., serving meals to the homeless or building with groups like Habitat for Humanity) in an era with weak social safety nets, whereas other students may express their interests through choosing to volunteer with children’s camps or sports, or community events.

In all, drawing on this alternative lens based in research on the durable values of expressivity in the key realms of work and family, the enduring expressivity thesis predicts increases in preferences for self-expression values, such as autonomy and compassion, while decreases in preferences for survival values like hard work.

The Role of Demographics Shifts

When examining a shift in values at the aggregated level over time, it is necessary to account for demographic shifts in the population (Firebaugh 1997). We discuss three changes since the mid-1980s that may be important in understanding changing Americans’ values for children: (a) the increase in college graduates, (b) the increase in the proportion of the population in professional occupations, and (c) the increase in immigrants from Asian and Central or Latin American countries. We discuss how demographic changes may have led to changes toward expression values like autonomy and compassion relative to survivalist values like hard work.

Education.

A key change in the past three decades is the further increase in education levels, with more people graduating from college—from about 19 percent of adults 25 and older holding a Bachelor’s degree in 1986 to 31 percent in 2016 (U.S. Census Bureau, 2018a, 2018b). A number of studies have shown that higher levels of education are linked to adults’ preferences for autonomy for children (Alwin 1989; Kohn 1977; Park and Lau 2016; Suizzo 2007; Xiao 2000). Thus, the increase in college-educated Americans, who have gone through at least four years of advanced intellectual work, may have led to an increase in preferences for autonomy.

Occupation.

The changes in U.S. labor market discussed earlier involved changes in occupational structures. Blue-collar jobs have decreased, whereas jobs that require specialized skills—professional jobs—have increased (Autor, Katz, and Kearney 2006). An influential series of studies on parenting values found that occupations that involve freedom from close supervision, complex thought processes, and non-routine tasks—features related to professional occupations—were related to adults’ preferences for autonomy as values for children, whereas those occupations requiring close supervision and repetitive work were related to adults’ preferences for obedience in children (Kohn 1977; Lareau 2003). Thus, the increase in share of professional occupations in the U.S. population over the past three decades may have led to a further increase in preferences for autonomy relative to other traits, including hard work.

Race.

Racial-ethnic compositions have changed in the past three decades due to the increases in Latinx and Asian immigrants (Grieco 2010). Parenting values vary by race-ethnicity and nativity, with Asian and Latinx immigrants more likely than native-born counterparts to use an authoritarian parenting style which emphasizes obedience relative to autonomy in children (Nomaguchi and House 2013). Suizzo (2007) found that ethnic-minority parents—Chinese Americans, African Americans, and Mexican Americans—are more likely than White Americans to emphasize the importance of the traditional value of obedience over autonomy. Research has shown that Asian parents are more likely than White parents to emphasize children’s efforts as a key to academic success (Liu and Xie 2016; Stevenson and Stigler 1994). Thus we expect the increase in share of “other race” in the population relates to an increase in preferences for survivalist values like hard work relative to autonomy, since the mid-1980s.

In sum, major demographic changes in the past three decades suggest countervailing predictions as to how U.S. adults’ values for children might have changed during the period based on demographic shifts alone. The increase in immigrants from Asian and Central or Latin American countries suggest a possible increase in preferences for hard work relative to self-expression values such as autonomy and compassion. In contrast, the increases in college graduates and professional occupations over the past three decades may indicate a further increase in U.S. adults’ preferences for autonomy or compassion relative to other traits including hard work. Using decomposition analyses, we separate the effects of these demographic changes from changes in values for children, either toward emphasizing hard work as the economic precarity thesis posits or a continued prioritizing of autonomy and compassion as the enduring expressivity thesis contends.

The Current Study

This study empirically documents changes in Americans’ socialization values for children since the mid-1980s, using data from 17 waves of the GSS, spanning 32 years. We examine two competing theses on the basis of key ideas in modernization theories (Inglehart 1997; Inglehart and Baker 2000; Inglehart and Welzel 2010). First, we expect that as a sense of economic insecurity has increased, Americans’ beliefs that children need to work hard to prepare for life have increased over the past three decades (the precarity thesis). Second, alternatively, we expect that despite the increase in economic insecurity, value changes in postindustrial societies toward higher-order needs have continued to grow, which may be reflected in Americans’ increasing desires for children to inculcate self-expression values such as autonomy and compassion more than survival values such as hard work (the enduring expressivity thesis).

Any aggregate changes in values observed must take into account the effects of major demographic shifts during the same period (Firebaugh 1997). As discussed, we examine the role of the increases in college graduates and professional occupations, both of which are related to preferences for autonomy (Kohn 1977), and the increase in immigrants from Asian and Central or Latin American countries, which is related to preferences for hard work (Liu and Xie 2016) in contributing to the aggregate change in Americans’ preferred traits for children. In addition, the decreases in the married and in parents in the population and the aging of the population over the past three decades (Cherlin 2010) should be taken into account. We did not discuss any predictions on contributions of these trends to potential changes in adults’ values, however, as their associations with Americans’ preferred traits for children are less known in prior research. We also include other characteristics that have been known as key factors related to adults’ values for children, including gender, employment status, family income, political views, and religiosity (Alwin 1984, 1988, 1989; Starks and Robinson 2005; Xiao 2000) in the analysis.

METHOD

Data

Data for this paper are drawn from 17 years of the GSS conducted in 1986, 1988, 1990, 1991, 1994, 1996, 1998, 2000, 2002, 2004, 2006, 2008, 2010, 2012, 2014, 2016, and 2018. The GSS was collected annually until 1994 (except for 1992) and then biannually. To make the intervals consistent throughout the three decades (except for two intervals between 1990 and 1991 and between 1991 and 1994, because there was no survey in 1992), we do not include 1987, 1989 and 1993 surveys. Patterns of findings do not change whether we include these three years of surveys. The questions regarding qualities respondents prefer in children were asked regardless of the respondents’ parental status. The sample size for the present study is N = 23,109 for the descriptive analysis. This is smaller than the total sample of the 17 waves (n = 41,266), because the GSS used split samples and respondents who were in a split sample were asked these questions (Smith et al. 2019). For the decomposition analysis, we focused on two years, 1986 (n = 732) and 2018 (n = 1,542). There are a small percentage of missing cases for some variables in each year. We include these missing cases by using multiple imputation techniques suggested by Allison (2001). We perform PROC MI in SAS with five iterations for each year of survey and then combine the 17 imputed data sets into one pooled data set. Data are weighted to adjust for the sub-sampling of non-respondents and the number of adults in the household (Smith et al. 2019).

Measures

Adults’ values for children are measured using the set of questions developed by Lenski (1961): “If you had to choose, which thing on this list would you pick as the most important for a child to learn to prepare him or her for life? To obey; to be well-liked or popular; to think for himself or herself; to work hard; to help others when they need help? Which comes next in importance? Which comes third? Which comes fourth?” Thus, respondents were asked to provide relative rankings of the five traits of: obedience, likability, autonomy, hard work, and compassion. Each trait was measured as an ordered variable where higher values mean that the trait was ranked higher in importance among the five traits provided (i.e., 5 = most important; 4 = 2nd most important; 3 = 3rd; 2 = 4th; and 1 = least important).

Analysis of ranking measures is challenging because, unlike a rating measure, these measures are not independent as respondents are forced to rank them in comparison to the other four provided choices; thus, neither the mean ranking score of hard work, for example, nor the dichotomous variable indicating whether hard work is chosen to be the most important across the five traits, are appropriate for regression analyses (Allison and Christakis 1994; Alwin and Jackson 1982; Jackson and Alwin 1980). Of different analytical approaches used in prior research, we first use the percentage of respondents selecting a given trait, out of the five traits, as the most important trait for children to describe its trends, following Kalleberg and Marsden (2013). Second, following Alwin (1984) and Starks and Robinson (2005), we use ratio scales that represent the survival versus self-expression values, including autonomy relative to hard work and compassion relative to hard work, respectively, as our focus is on potential shifts in the survival versus self-expression dimension of Inglehart’s modernization theory. The measure of the mean ranking for autonomy (or compassion) relative to hard work is created by taking the rank given to autonomy (or compassion) minus the rank given to hard work. These relative measures range from −4 to 4. The positive values indicate a further shift toward self-expression values, whereas the negative values indicate a return to emphasizing survivalist values. Although the traditional versus secular-rational values theme in Inglehart’s theory is not a focus of this paper, in order to understand a possible move away from traditional values, we also create three measures that represent the traditional versus secular-rational values, including autonomy relative to obedience, compassion relative to obedience, and hard work relative to obedience. Note that we also conducted supplemental analyses using the procedure recommended by Allison and Christakis 1994—see the Analytical Plan section below.

The primary independent variable is year of the survey. We examine a total of 11 characteristics in decomposition analyses, including three focal characteristics: education, occupation, and race. College degree is a dichotomous variable where those with a four-year college degree or higher are assigned 1 and others are assigned 0. Occupation is measured in three categories based on the 1980 Census Occupation Code, including professionals (managerial and professional specialty occupations), non-professional white collar occupations (administrative, sales, and service occupations), and non-professional blue collar occupations (farming, forest, fishing, precision, production, craft, repair, operations, fabrications, and laborers; reference). Those who were not employed at the time of the survey reported the occupation that they had been typically engaged in. Race is measured as three dummy variables, including white (reference), black, and other race. The GSS began to ask respondents whether they are Hispanic origin since the 2000 survey but not 1998 or earlier.

Other demographic and socioeconomic characteristics include age, gender, marital status, parental status, employment status, and family income: Age is measured as three dummies, including aged 18 to 34, aged 35 to 54, and aged 55 or older. Gender is a dichotomous variable where women are assigned 1 and men are assigned 0. Marital status includes four dummy variables, married (reference), widowed, divorced/separated, and never married. Parental status is measured as three dummy variables, including childless (reference), those living with at least one child under age 18, and those who have adult children only. Employment status is measured as three dummy variables including nonemployed, employed part-time, and employed full-time (reference). Family annual income is a constructed variable by the GSS that is adjusted for inflation. We include two other variables that are related to values, political views and religiosity: Political view is measured by a seven-point scale in the GSS. We create three dummy variables, including conservative (including “slightly conservative”, “conservative”, “extremely conservative”), moderate (“moderate-middle of the road”, reference), and liberal (including “slightly liberal”, “liberal”, and “extremely liberal”). Religious service attendance in the past 12 months is an ordered variable ranging from 0 = never to 8 = several times a week.

Analytic Plan

To examine the survival versus expressivity axis of Inglehart’s modernization theory of value change, we examine the precarity versus the enduring expressivity thesis. To do so, first, we describe trends in the percentage of adults selecting each of the five trait as most important from 1986 to 2018. Second, we describe trends in three measures of preferring self-expression values to survival values from 1986 to 2018. Third, using the general model for decomposing aggregate change suggested by Firebaugh (1997), we decompose differences in each of these relative values between 1986 (Time 1) and 2018 (Time 2), which can be expressed as follows:

$$\begin{gathered} {\overline{Y}}_2-{\overline{Y}}_1=\left(a_2+b_2{\overline{x}}_2\right)-\left(a_1+b_1{\overline{x}}_1\right) \\ =\left(a_2-a_1\right)+\left(b_2-b_1\right){\overline{x}}_1+b_1\left({\overline{x}}_2-{\overline{x}}_1\right)+\left(b_2-b_1\right)\left({\overline{x}}_2-{\overline{x}}_1\right) \\ =\mathrm{\Delta }a+\mathrm{\Delta }b{\overline{x}}_1+b_1\mathrm{\Delta }\overline{x}+\mathrm{\Delta }b\Delta \overline{x} \end{gathered}$$

In this equation, change in the mean ranking for autonomy (or compassion) relative to hard work $\left({\overline{Y}}_2-{\overline{Y}}_1\right)$ is partitioned into four components, including changes in the intercept (Δa), changes in the effects of explanatory variables $\left(\mathrm{\Delta }b{\overline{x}}_1\right)$, changes in the levels (i.e., means) of explanatory variables $\left(b_1\mathrm{\Delta }\overline{x}\right)$, and interactions between changes in the effects and changes in the levels of explanatory variables $(\mathrm{\Delta }b\mathrm{\Delta }\overline{x})$. The sum of the second, third, and fourth components for explanatory variables indicates the overall contribution of explanatory variables to change in the mean ranking for autonomy (or compassion) relative to hard work. For the decomposition analysis, all explanatory variables are standardized in the ordinary-least-squares (OLS) regression models to solve the problem of arbitrary intercepts. For significance testing, we estimate standard errors using the bootstrap procedure in the Stata survey module with 500 replications (Amato et al., 2003).

In supplemental analyses, using the same set of analyses, we examine the other dimension of modernization—the traditional versus secular-rational values—to extend the rich literature heavily focused on analyzing the value of obedience in children. Following Allison and Christakis (1994), we also conduct another set of supplemental analyses using exploded logit models. For this procedure, we create a separate record for each trait for each respondent, where each record included the rank given to the trait, dummy variables for the five traits, an identification number for the respondent, and the independent variables. We estimate a model where the rank is the dependent variable, the independent variables, the dummy variables for four traits (excluding the reference trait), and the interaction terms between each trait dummy variable and year are included as explanatory variables, and the identification number is used to specify stratification.

RESULTS

Figure 1 presents trends in the percentage of adults selecting each of the five traits asked in the GSS as most important for children from 1986 to 2018. The percentage of adults who reported autonomy (“to think for oneself”) as the most important quality for children declined from 51.2 percent in 1986 to 41.6 percent in 2018. In contrast, the percentage of adults who reported diligence or “to work hard” as the most important quality for children increased markedly from 11.2 percent to 27.6 percent. The percentage of adults who reported compassion (“to help others”) as most important to prepare children for their future also increased from 13.7 percent to 20.7 percent. The percentage of adults who reported obedience (“to obey”) as most important declined steadily from 23.4 percent to 9.9 percent. A very small percentage of Americans reported likability (“to be popular”) as the most important quality for children in all years. In sum, although autonomy, a trait reflecting self-expression values, remains the top quality that U.S. adults selected as most important for children throughout the 32 years, the percentage declined. Indicating a shift back to emphasizing survival values, the percentage of U.S. adults selecting hard work as the most important quality more than doubled over the three decades. Preferences for compassion, the other trait indicating self-expression values, increased, though not as dramatically. Finally, preferences for obedience declined by half, suggesting a further move away from traditional values.

Figure 1.

Trends in % Americans Reporting a GivenTrait Among the Five Traits as the Most Important Quality for Children: 1986–2018

Now we turn our focus to the mean ranking of autonomy or compassion relative to the mean ranking of hard work to examine value changes in the survival versus self-expression values dimension. As shown in Figure 2, with some fluctuations, the mean ranking of preferences for autonomy relative to hard work gradually declined from 0.71 in 1986 to −0.02 in 2018. The difference between 1986 and 2018 was statistically significant (p < .01, t-test). Notably, until 2008, the numbers were positive, meaning that on average Americans ranked autonomy higher than hard work among the five traits. In the 2010s, however, the numbers shifted to be negative (except for 2012), which indicates that Americans in the last decade ranked hard work about the same or higher than autonomy as an important trait for children to learn. For compassion, the other trait indicating self-expression values, the mean ranking of compassion to the mean ranking of hard work declined slightly from −0.01 in 1986 to −0.17 in 2018. Although statistically significant (p < .05), the difference in the mean ranking of compassion relative to hard work between 1986 and 2018 was much smaller than the differences in the mean ranking of autonomy relative to hard work across the same time period.

Figure 2.

Trends in Americans’ Mean Ranking of Autonomy and Compassion Relative to Hard Work Among the Five Traits for Children Asked in the GSS: 1986 – 2018

Thus far, we saw a notable shift away from autonomy and toward hard work and a smaller shift from compassion to hard work, supporting the precarity thesis. To what extent were these changes in values derived from population turnover? As shown in Table 1, the distribution of Americans in several characteristics associated with values for children changed from 1986 to 2018. For three key variables we reviewed in the background section, there was much change. First, there was an increase in U.S. residents with a college degree from 20 percent to 33 percent. Second, the shares of those with professional or non-professional white-collar occupations increased, while the share of those with non-professional blue-collar occupations decreased. Finally, the percentage of “other race” adults increased a great deal over time from 4 percent to 12 percent. Notable among other changes are that the percentage of those aged 55 or older increased from 28 percent to 34 percent. We see more never-married or divorced respondents in 2018 than in 1986. The percentage of respondents who were living with a child under 18 declined a great deal, from 43 percent in 1986 to 27 percent in 2018. Americans who identified themselves as politically liberal increased from 23 percent to 30 percent. The frequency of religious service attendance decreased from 4.07 to 3.14 with a range from 0 to 8. To what extent have these population changes contributed to the shift in Americans’ preferences for hard work for children?

Table 1.

Weighted Means (Std.) for Independent Variables in the Analysis

	1986		2018		Differences 2018 – 1986
Education
College degree	0.20		0.33		0.13***
Occupation
Professional	0.28		0.37		0.09***
White-collar	0.36		0.35		−0.01
Blue-collar	0.36		0.28		−0.08***
Race
White	0.86		0.72		−0.14***
Black	0.11		0.16		0.05**
Other race	0.04		0.12		0.09***
Age
Age 18 – 34	0.36		0.33		−0.02
Age 35 – 54	0.37		0.32		−0.04
Age 55+	0.28		0.34		0.07**
Gender
Women	0.55		0.54		−0.01
Marital Status
Married	0.64		0.48		−0.16
Widowed	0.07		0.05		−0.02
Divorced	0.11		0.16		0.05***
Never married	0.18		0.31		0.12***
Parental Status
Childless	0.23		0.28		0.05***
Parents	0.43		0.27		−0.16***
Parents w/ adult kids	0.34		0.45		0.12***
Employment
Non-employed	0.40		0.39		−0.01
Part-time	0.11		0.13		0.02
Full-time	0.49		0.48		−0.01
Family income	44.45	(33.1)	53.34	(44.8)	8.89***
Political views
Conservative	0.36		0.33		−0.03
Independent	0.41		0.37		−0.03
Liberal	0.23		0.30		0.07***
Religious service attendance [0 – 8]	4.07	(2.7)	3.14	(2.8)	−0.93***
N	732		1,542

^*p < .05;

^**p <.01;

^***p <.001 (two-tailed t-tests)

To answer this question, we used Firebaugh’s (1997) general model for decomposing the aggregate change in Americans’ preferences for autonomy or compassion relative to hard work for children, respectively, between 1986 and 2018 into four components including (a) changes in the intercept, (b) changes in the slopes (i.e., effects), (c) changes in levels (i.e., means), and (d) joint changes in the slopes and the levels. Results for preferences for autonomy relative to hard work are presented in Table 2. Columns 1 and 2 present b coefficients and standard errors from OLS regression models examining the association between independent variables and preferences for autonomy relative to hard work in 1986 and 2018 respectively. As expected, Americans with college degrees were more likely than those with lower levels of education to prefer autonomy to hard work. Also as predicted, Americans with professional jobs were more likely to prefer autonomy to hard work compared with Americans with blue-collar jobs, although this association was significant only in 1986. Consistent with our prediction, although significant only in 2018, Americans who identified themselves as “other race” were less likely than Americans who identified themselves as White to emphasize autonomy over hard work. Some other results, for which we did not make predictions, are worth mentioning. Younger Americans (aged 18 to 34) were less likely to prefer autonomy to hard work than Americans aged 35 to 54. Women were more likely than men to prefer autonomy to hard work for children. Those who had never been married were more likely than the married to prefer autonomy to hard work, albeit significant only in 1986. Also significant only in 1986, parents living with children under 18 were more likely than the childless to prefer autonomy to hard work. Americans who identified themselves as politically liberal were more likely than Americans who identified themselves as independent to prefer autonomy to hard work. Finally, those who attended religious services more often were less likely to prefer autonomy over hard work, although significant only in 2018.

Table 2.

Decomposition Analysis of 1986 to 2018 Change in Americans’ Preferences for Autonomy Relative to Hard Work for Children

	(1) 1986		(2) 2018		(3) Δb	(4) 1986	(5) 2018	(6)$\mathrm{\Delta }\overline{x}$	(7) Slope	(8) Level	(9) Joint	(10) Sum
	b1	SE1	b2	SE2	b2-b1	$\overline{x_1}$	$\overline{x_2}$	$\overline{x_2}-\overline{x_1}$	$\mathrm{\Delta }b\overline{x_1}$	$b_1\mathrm{\Delta }\overline{x}$	$\mathrm{\Delta }b\mathrm{\Delta }\overline{x}$
College degree	.137	.081+	.138	.052**	.001	−.115	.133	.249***	.000	.034+	−.001	.034*
Professional	.174	.085*	.044	.064	−.130	−.070	.104	.174***	.009	.031+	−.024	.017
White collar	.102	.077	.023	.059	−.078	−.037	−.030	.007	.003	.000	.000	.003
Black	.071	.071	−.030	.044	−.101	−.078	.095	.172**	.009	.013	−.019	.003
Other race	−.006	.104	−.105	.041*	−.099	−.197	.108	.306***	.020	−.001	−.032	−.012
Aged 18 – 34	−.199	.077*	−.125	.062*	.075	.082	−.092	−.174	.006	.034*	−.012	.028*
Aged 55+	−.036	.093	.011	.058	.047	.008	.233	.225**	.000	−.009	.010	.002
Women	.156	.073*	.231	.050***	.075	.060	.006	−.054	.005	−.008	−.005	−008
Widowed	.054	.058	.017	.045	−.037	.167	.077	−.090	−.006	−.005	.003	−.008
Divorced/separated	−.018	.072	.037	.045	.055	−.012	.160	172***	−.001	−.004	.010	.005
Never married	.230	.097*	−.020	.059	−.250	−.129	.134	.263***	.033+	.061*	−.066*	.027
Parents with children < 18	.197	.095*	−.088	.066	−.285	.109	−.233	−342***	−.031+	−.066	.097*	.000
Parents with adult children	.116	.101	−.085	.066	−.201	−.116	.163	.278***	.024	.032+	−.057	−.001
Nonemployed	.026	.079	−.019	.053	−.045	.098	.081	−.017	−.004	.000	.000	−.004
Part-time employed	−.024	.078	.060	.049	.084	−.103	−.010	.093	−.008	−.002	.007	−.002
Family income	.140	.096	.045	.053	−.095	−.223	−.047	.176***	.018	.024	−.014	.027
Conservative	.044	.070	−.103	.053+	−.146	−.006	−.062	−.055	.001	−.002	.008	.007
Liberal	.210	.073**	.139	.049**	−.071	−.064	.084	.148***	.004	.031*	−.010	.025*
Religious service	−.072	.067	−.176	047***	−.104	.151	−.205	−.356***	−.016	.026	.036	.046**
Intercept	.822	.073	−.090	.051
R2	.066***		.070***					Sum	.067	.189*	−.068	.189***
N	732		1,542
Summary of Decomposition Equation					Aggregate Change		Intercept	Slope		Level	Joint
					$\overline{y_2}-\overline{y_1}=-.722$		Δα + −.910***	$\mathrm{\Sigma }\overline{x_1}\Delta b$+.067		$\mathrm{\Sigma }\Delta \overline{x}{b}_1$+.189*	$\mathrm{\Sigma }\Delta \overline{x}\Delta b$ −.068

⁺p <.10;

^*p < .05;

^**p <.01;

^***p <.001.

Note. Omitted reference groups are: blue-collar occupation, White, aged 35–54, married, childless, full-time employed, and independent. Explanatory variables were mean-centered.

Column 3 in Table 2 shows differences in the coefficients between 1986 and 2018, which would indicate changes in the effects of particular variables on Americans’ preferences for autonomy relative to hard work for children. Although some variables that were significant in 1986 were not significant in 2018 and some variables that were not significant in 1986 were significant in 2018, according to calculations using formula suggested by Hardy (1993), none of the differences in the coefficients between 1986 and 2018 were statistically significant. Columns 4 and 5 in Table 2 present centered means for each variable in 1986 and 2018 respectively; and Column 6 presents differences in centered means between 1986 and 2018. The same information was shown in Table 1, but here, all independent variables were mean centered.

Columns 7 to 10 show components of the aggregate change in preferences for autonomy to hard work between 1986 and 2018 based on demographic turnover. Column 7 shows the extent to which the aggregate change in preferences can be attributed to changes in the slopes (i.e., effects) of the independent variables. There were no significant changes (at p < .05 level) in the effects that contributed to the aggregated level of change in preferences. Column 8 shows the extent to which the aggregate change in preferences can be attributed to changes in the levels (i.e., means) of the independent variables. The decrease in Americans aged 18–34 in the population (.034) and the increases in Americans who had never been married (.061) and who identified themselves as politically liberal (.031) significantly contributed to changes in preferences for autonomy relative to hard work. Column 9 shows the joint effects of the changes in the slopes and the levels. The joint effects between the increase in the level of Americans who had never been married and the change in the association of this group with preferences negatively contributed to the shift in preferences for autonomy relative to hard work (−.066). In contrast, the joint effects of the decrease in the level of Americans living with children under age 18 and the change from the positive to the negative association of this group with preferences for autonomy relative to hard work contributed the shift toward autonomy from hard work (.097). Column 10 is the sum of the components from columns 7 to 9, which indicates the overall contribution of the changes in independent variables to the aggregate change in Americans’ preferences for autonomy relative to hard work. The increases in Americans with college degrees (.034) and those who identify as politically liberal (.025) and the decreases in Americans aged 18–34 (.028) and those who attend religious services regularly (.025) contributed to an increase in preferences for autonomy over hard work from 1986 to 2018.

The summary decomposition equation is shown at the bottom of Table 2. The overall contribution due to demographic turnover (“level”) was significant, and the sign was positive (.189), indicating that demographic turnover contributed to the shift toward autonomy relative to hard work. The aggregate change was negative (−.722) which means that the overall shift occurred away from autonomy to hard work. And this change was driven by the change in the intercept (−.910), which indicates a general cultural shift—a change in Americans’ values for children unexplained by demographic change. In fact, the change in the intercept explains 126 percent of the aggregate change, suggesting that if there were no changes in the U.S. demographic composition, such as the increases in people with college degrees and people identifying themselves as politically liberal, who tend to prefer autonomy over hard work, the increase in preferences for hard work relative to autonomy would have been even greater.

Table 3 presents the results for preferences for compassion relative to hard work. OLS regression models for 1986 and 2018 (Columns 1 and 2) show that Black Americans were less likely than White Americans to prefer compassion relative to hard work for children, albeit significant only in 2018. Younger Americans (aged 18–34) were less likely than middle-aged Americans (aged 35–54) to prefer compassion to hard work although significant only in 2018, whereas older Americans (aged 55 or older) were more likely to prefer compassion to hard work although significant only in 1986. Women were more likely than men to prefer compassion to hard work (significant only in 2018). Parents who have adult children only were less likely than people who did not have children to prefer compassion to hard work. Americans who identify as politically liberal were more likely than those who identify as independent to prefer compassion relative to hard work. Even though some variables show significant associations with values for children only one of the two years, there were no significant differences in the coefficients for each variable between 1986 and 2018 (Column 3). Most decomposition components were not significant (Columns 7 to 10). None of the components in the summary decomposition equation, which are presented at the bottom of Table 3, were statistically significant. These results suggest that the change in Americans’ preferences for compassion relative to hard work was negligible.

Table 3.

Decomposition Analysis of 1986 to 2018 Change in Americans’ Preferences for Compassion Relative to Hard Work for Children

	(1) 1986		(2) 2018		(3) Δb	(4) 1986	(5) 2018	(6) $\mathrm{\Delta }\overline{x}$	(7) Slope	(8) Level	(9) Joint	(10) Sum
	b1	SE1	b2	SE2	b2 – b1	$\overline{x_1}$	$\overline{x_2}$	$\overline{x_2}-\overline{x_1}$	$\mathrm{\Delta }b\overline{x_1}$	$b_1\mathrm{\Delta }\overline{x}$	$\mathrm{\Delta }b\mathrm{\Delta }\overline{x}$
College degree	−.038	.079	.032	.046	.069	−.115	.133	249***	−.008	−.010	.017	−.001
Professional	.078	.083	−.029	.056	−.106	−.070	.104	174***	.007	.014	−.018	.003
White collar	.082	.075	−.035	.052	−.117	−.037	−.030	.007	.004	.000	−.001	.004
Black	−.057	.069	−.125	039***	−.068	−.078	.095	.172**	.005	−.009	−.012	−.016+
Other race	−.014	.102	−.013	.036	.001	−.197	.108	.306***	−.002	−.006	.003	−.005
Aged 18 – 34	.079	.075	−.183	.055***	−.262	.082	−.092	−.174	−.021	−.014	.046*	.011
Aged 55+	.211	.090*	.024	.051	−.187	.008	.233	.225**	−.002	.048*	−.043+	.003
Women	.022	.071	.088	.044*	.066	.060	.006	−.054	.004	−.001	−.004	−.001
Widowed	−.022	.056	.063	.040	.085	.167	.077	−.090	.015	.002	−.008	.009
Divorced/separated	.027	.070	.020	.040	−.007	−.012	.160	172***	.000	.006	−.002	.003
Never married	.045	.094	.011	.052	−.034	−.129	.134	.263***	.005	.013	−.010	.008
Parents with children < 18	.080	.092	−.052	.058*	−.132	.109	−.233	−.342***	−.015	−.028	.045	.003
Parents with adult children	−.023	.099	−.134	.058	−.111	−.116	.163	278***	.013	−.006	−.032	−.024
Nonemployed	.068	.077	−.003	.047	−.071	.098	.081	−.017	−.007	−.001	.001	−.007
Part-time employed	−.149	.076*	.048	.043	.197	−.103	−.010	.093	−.020+	−.013	.017	−.016+
Family income	−.098	.094	−.072	.046	.026	−.223	−.047	.176***	−.006	−.017	.005	−.019
Conservative	−.047	.068	−.052	.047	−.005	−.006	−.062	−.055	.000	.002	.000	.003
Liberal	.121	.071+	.093	.044*	−.028	−.064	.084	.148***	.002	.018	−.004	.016*
Religious service	−.017	.066	.048	.041	.066	.151	−.205	−.356***	.010	.005	−.022	−.008
Intercept	−.026	.071	−.177	.045
R2	034***		034***					Sum	−.015	.003	−.022	−.034
N	732		1,542
Summary of Decomposition Equation					Aggregate change		Intercept	Slope		Level	Joint
					$\overline{y_2}-\overline{y_1}=-.186$		Δα +−.152	$\mathrm{\Sigma }\overline{x_1}\Delta b$+−.015		$\mathrm{\Sigma }\Delta \overline{x}{b}_1$+.003	$\mathrm{\Sigma }\Delta \overline{x}\Delta b$−.022

⁺p <.10;

^*p < .05;

^**p <.01;

^***p <.001.

Note. Omitted reference groups are: blue-collar occupation, White, aged 35–54, married, childless, full-time employed, and independent. Explanatory variables were mean-centered.

In sum, decomposition analyses indicate that there was a shift from autonomy to hard work that was not derived from population turnover, suggesting a newly revealed cultural turn from self-expression values to survival values. We note some demographic shifts—e.g., the increase in college degrees and the decrease in religious attendance—have contributed to a shift from hard work to autonomy, which suggests that had the demographic and other changes not occurred, change in values from autonomy to hard work would have been more dramatic. In contrast, we saw little shift in preferences away from compassion, another value indicating self-expression values, to hard work. These results suggest that although there has been an increase in emphasis on hard work to survive in an increasingly precarious and insecure economy, the tendency toward self-expression values has also continued, constituted in the area of compassion toward others.

The Continued Decline of Valuing Obedience: Supplemental Analysis

While our focus in this study was on the survival versus self-expression values dimension, we conducted supplemental analyses which examined the traditional versus secular-rational values dimension focusing on three traits that represent secular-rational values—hard work, compassion, and autonomy (Inglehart and Welzel 2010)—relative to one trait that represents traditional values: obedience. Figure 3 presents descriptive trends in Americans’ preferences for hard work, compassion, and autonomy relative to obedience from 1986 to 2018, all of which increased over time. The mean score of Americans’ ranking for hard work minus ranking for obedience increased from 0.22 in 1986 to 1.31 in 2018. Both numbers were positive, meaning that more Americans ranked hard work higher than obedience in 1986 already, but the difference became larger steadily over time. Similarly, the mean score for compassion minus obedience increased from 0.21 in 1986 to 1.14 in 2018. The mean score for autonomy minus obedience, which prior research has studied extensively, also increased from 0.93 in 1986 to 1.29 in 2018, although the increase was less steep than increases in the preferences for hard work or compassion relative to obedience.

Figure 3.

Trends in Americans’ Mean Ranking of Hard Work, Autonomy, and Compassion Relative to Obedience Among the Five Traits for Children Asked in the GSS: 1986–2018

We conducted decomposition analyses for changes in these three pairs of traits; and the summaries of the decomposition equations are presented in Table 4. For hard work relative to obedience, changes in demographic shifts contributed to the increase in preferences for hard work relative to obedience, although it was only 13.1 percent (i.e., .133/1.012) of the aggregated change. The change in the intercept explained 86.9 percent of the aggregated change. Similarly, the increase in preferences for compassion relative to obedience was also driven more by changes in intercept (88.2 percent, i.e., 0.727/0.824) than the change in demographic shifts (11.8 percent). In contrast, the shift toward autonomy compared to obedience was driven by demographic shifts (112 percent, i.e., 0.321/0.288). In additional supplemental analyses (not shown), we found that the increases in Americans with college degrees, those with professional jobs, and those who were politically liberal and the decrease in religious attendance most contributed to the shift from valuing obedience toward autonomy. In sum, we saw Americans’ preferences for values to instill in children change from obedience to autonomy, hard work, and compassion, all of which suggest a further shift from traditional to secular-rational values, as Inglehart and colleagues have contended (Inglehart 1997; Inglehart and Baker 2000; Inglehart and Welzel 2010). The shift from obedience to autonomy was due to population turnover such as the increases in college degrees, professional jobs, and liberal political views and the decrease in religious service attendance. The shifts from obedience to hard work and compassion were not derived from population turnover.

Table 4.

Summary of Decomposition Analysis for Shifts in Americans’ Preferences for Hard Work, Compassion, and Autonomy Relative to Obedience for Children: 1986–2018

	Aggregated Change $\overline{y_2}-\overline{y_1}$	(1) Intercept Δα +	(2) Slope $\mathrm{\Sigma }\overline{x_1}\Delta b$	(3) Level $\mathrm{\Sigma }\overline{x}\Delta b_1$	(4) Joint $\mathrm{\Sigma }\overline{x}\Delta b$	Sum of the Components (2)+(3)+(4)
Hard work relative to obedience	1.012	879***	.007	.101	.025	.133*
Compassion relative to obedience	.824	727***	−.009	.098	.009	.098+
Autonomy relative to obedience	.288	−.033	.070	.285**	−.035	321***

⁺p <.10;

^*p < .05;

^**p <.01;

^***p <.001.

Calculations are done in supplemental analyses that are not shown.

In another set of supplemental analyses, following Allison and Christakis (1994), we conducted exploded logit models to examine whether Americans’ preferences for traits for children changed over time using the interaction terms between each trait and year. Results are shown in the Appendix Table. Note that because we reverse-coded the ranking, higher values indicate lower ranks. Results are consistent with what we showed in Figures 2 and 3 above. Compared with hard work, preferences for autonomy and compassion (also obedience and likability) went down over time. Compared with obedience, preferences for autonomy, compassion, and hard work went up over time.

CONCLUSION

Knowing what adults believe is important for children and how this changes over time is a compelling window into American culture and society. Prior research has shown that adults’ values for children changed from obedience to autonomy from the 1920s to the mid- 1980s, reflecting economic development of the nation (Alwin 2001; Inglehart and Welzel 2010; Lesthaeghe 2010). In this article, we extend this line of research to investigating changes in Americans’ values for children from 1986 to 2018, during the period when the nature of work in the U.S changed and the sense of economic insecurity and inequality among Americans rose. This has not received much research attention. The results indicate some significant shifts in terms of what adults think are important for children today, compared with more than a quarter-century ago.

Autonomy, or to think for oneself, remains the number one trait Americans view as ideal to develop in children, among the five traits ranked, but by a much smaller margin than in the past. Consistent with the precarity thesis, we see a shift to survival values—there has been a steep rise in Americans’ beliefs in the importance of children working hard to succeed in life. There is some indication of the continued emphasis of self-expressivity, however. The value of compassion, the other trait that indicates self-expression values, did not decline much relative to hard work; and the percentage of Americans who chose compassion as most important among the five traits has increased steadily since the mid-1980s, although not as steeply as the increase for hard work. All in all, the most remarkable shift since the 1980s is the rise in valuing working hard for children.

Decomposition analyses show that these shifts are not explained by demographic changes in the U.S. population. Rather, we find that the increases in the share of college graduates and people with liberal political views as well as the decreases in the share of younger people (aged 18–34) and people who attend religious services regularly in the U.S. population from 1986 to 2018 contributed to an increase in preferences for autonomy over hard work during the period. Thus, had the shifts in these characteristics of the population not occurred, changes in adults’ preferences for hard work would have been greater than we observed. Note that occupation and race did not contribute to shifts in adults’ values for children, although we predicted they might have.

The increase in Americans who rank hard work as a more important trait than autonomy for children to learn to prepare them for “life” may reflect a heightened perception of economic insecurity and the need for individual efforts for paving one’s way, presumably economically (Cooper 2014; Nelson 2010)—a retreat back to survivalist values in an insecure world. Much research has noted that today’s parenting culture emphasizes parents’ efforts to push children to work hard academically and in developing talents for their futures (Lareau 2003; Milkie and Warner 2014). Though the question asks about what is important for children’s future lives in general, the perceived instability of the workplace, and the competitiveness in educational institutions are likely relevant factors to imagining what children need for success. We urge future research to investigate macro-level indicators of the college premium, economic insecurity, globalization, and other large scale changes that may have altered what adults think children need to do in order to prepare for their futures.

Past research on Americans’ socialization values for children focused almost exclusively on the comparison between obedience and autonomy (e.g., Alwin 2001). Though not central in this paper, we show that the status of diligence and compassion as values to inculcate in children, both of which were rarely examined in past research, now greatly surpass obedience. Obedience has had a decades-long decline as a value viewed for children to ready themselves for their adult lives, marking a further move away from traditionalism. As the United States has moved far into a postindustrial society and its economic and cultural climate continues to change, research in this field should investigate new patterns of challenges that parents and others perceive for children.

Although the GSS provides an invaluable opportunity to track trends in adults’ socialization values for children across three decades, there are some limitations. First, there was a very limited number of traits that respondents were asked about, and those that are potentially important or rising in importance in U.S. society are not necessarily included. For example, the value of the trait of leadership may have grown over time if asked, given its emphasis in recent years by colleges, which often desire a demonstration of young applicants’ leadership skills from school or extracurricular programs. Similarly, the desire to instill the trait of flexibility for children could be rising in importance in times of economic uncertainty (Pugh 2015), but is not measured here. Second, respondents were unable to choose two values as having the same levels of importance. On the basis of postmodern theory, Ovadia (2003) argued that numerous values might increase in importance over the same time period. Given this logic, if people were allowed to select more than one as “the most” important, different patterns could be observed over time. It also suggests that even as the relative ranking might change, overall, it is possible that more people believe that it takes more of several different qualities for children’s success in life. To conclude, the social and cultural backdrop for children’s ideal qualities has continued to change. What traits should children acquire as they forge a future in an economically uncertain post-industrial world in the 2020s and beyond? Compared to the 1980s, working hard is viewed as relatively more important, and thinking for oneself relatively less important. Moreover, compassion is seen as the most important quality more often than it was in the past, surpassing obedience, which has had a decades long decline in America as a top value views as important for children to ready themselves for their adult lives. These changes are not explained by changes in demographic compositions in the U.S. population. Macro-level changes, such as the economic landscape and larger cultural orientations of U.S. society, likely have shifted the climate for assessing children’s futures, and point to one in which more adults view children’s preparation as requiring hard work. Notably, changes may be bi-directional, as this research underscores fluidity in a shift “back” to survival values. Future research to investigate macro-level indicators of economic insecurity and perceptions of the college premium as sources of changes in adults’ values for children will continue to extend knowledge about the complexities of the structure-values nexus.

Appendix Table. Results from Exploded Logit Models Predicting Americans’ Preferences Regarding Traits for Children by Year (N = 23,109).

	Compared with Hard Work			Compared with Obedience
	b	SE		b	SE
Autonomy	−.822	.025***	Autonomy	−.993	.027***
Compassion	.122	.023***	Compassion	−.050	.024*
Obedience	.171	.024***	Hard work	−.171	.024***
Likability	2.097	.030***	Likability	1.925	.029***
Autonomy x year	.019	.001***	Autonomy x year	−.008	.001***
Compassion x year	.003	.001*	Compassion x year	−.024	.001***
Obedience x year	.027	.001***	Hard work x year	−.027	.001***
Likability x year	.025	.002***	Likability x year	−.002	.002

^*p <.05;

^**p < .01;

^***p < .001.

Model fit: Wald χ² (8 df) = 38383.73. The ranks were reverse-coded (1 = lowest rank to 5 = highest rank); thus, lower values indicate higher ranks. Positive signs for the interaction terms mean that the ranking for the given trait went up in more recent years. Negative signs for the interaction terms mean that the ranking for the given trait went down in more recent years.