Sage Journals: Discover world-class research

Abstract

For the past half century, debate over income-targeted and universal approaches to American preschool policy has divided advocates, policymakers, and practitioners. This is the first paper to inform the debate with evidence from public opinion. It begins with the design and fielding of a nationally representative poll of preferences for targeted and universal preschool (N = 1,000). This poll yields rich data with which to assess the causes and correlates of support for each approach. Results indicate that preschool preferences are conditioned by financial self-interest and egalitarian values and that a savvy policymaker should not necessarily endorse universal over targeted preschool. In fact, Americans facing the possibility of tax increases to fund public preschool and those who prioritize equality of opportunity prefer a targeted approach.

Keywords

early childhood educational policy experimental research policy politics survey research

Public preschool has become a staple of education policy agendas. Forty-five states, the District of Columbia, and several cities fund preschool, with additional investment from the federal government. The resulting programs benefit both children and families, but their focus on early learning and school readiness distinguishes them from investments in childcare as a work support. Following contraction during the Great Recession, public preschool has resumed growth in recent years. States reported record enrollment and spending in the 2015–2016 school year, serving 1.5 million 3- and 4-year-olds with expenditures totaling $7.4 billion (Barnett et al., 2017).

Within the context of rapid preschool expansion, stakeholders debate whether programs should be targeted to low-income students (and others deemed at risk) or universally available to all (Barnett, Brown, & Shore, 2004; Finn, 2009; Zigler, Gilliam, & Barnett, 2011; Zigler, Gilliam, & Jones, 2006). Targeted programs are available in most states, although eligibility criteria and program enrollments vary substantially (from less than 5% of children in eight states to nearly half of all 4-year-olds in Texas; Barnett et al., 2017). Proponents of targeting cite its efficient use of resources and capacity to address gaps that arise before kindergarten entry (Finn, 2009; Zigler et al., 2011). Advocates of the universal approach tout benefits for all children and greater durability among programs that serve more, and more advantaged, families (Barnett et al., 2004; Zigler et al., 2011, 2006). To date, 11 states, the District of Columbia, and cities like New York, Boston, and Seattle are working toward universal provision. Although their efforts have generated vastly different programs, the perceived political appeal of universalism has been influential in itself: Universal initiatives launched in rapid succession between 1995 and the present, compared to targeted programs started largely in the 1960s through 1980s (Barnett et al., 2017). Yet, no study has compared public support for targeted versus universal preschool or assessed whether citizens do, in fact, have a preference.

This study offers the first empirical evidence on expressed preferences for targeted and universal preschool. A nationally representative public opinion poll (N = 1,000) also probes the values and considerations that underlie these preferences using direct questioning and randomized experiments. In all, findings demonstrate moderate support for public preschool and no significant preference for its targeted or universal forms, on average. Results suggest that financial self-interest and beliefs in equal opportunity drive the politics of preschool, informing the targeted-versus-universal debate and illuminating public preferences for education policy, more broadly.

Relevant Literature and Theoretical Framework

Public opinion has a frequent and often substantial impact on policy and appropriations (e.g., Burstein, 2003; Erikson, Wright, & McIver, 1993; Kingdon, 2011; Page & Shapiro, 1983). As measured through polling, public opinion helps define electoral incentives for public officials in democratic governments. Although special interest groups, political parties, and influential individuals also play a role, public opinion is an independent and important determinant of the scope and content of policymaking, especially for issues salient to the public or those that have undergone recent shifts.

Existing survey research finds that most Americans support the public provision of preschool (e.g., Bobo & Kluegel, 1993; Bostrom, 2002; Kahn & Barron, 2015; Page & Jacobs, 2009; Sniderman & Carmines, 1997; Sylvester, 2001). As taxpayers, they favor spending existing revenues on early learning and would be willing to pay more in taxes to support preschool programs (Page & Jacobs, 2009). In one state, respondents even favored funding preschool over tax and spending cuts (Nagle & Goidel, 2007). To date, however, polls have not shed light on the question of targeted-versus-universal approach. As a result, those engaged in the debate draw on more general reasoning.

Preschool Support and Self-Interest

Several scholars have argued that self-interest drives preferences for public policy, including preschool policy (Barnett et al., 2004; Skocpol, 1991a, 1991b; Zigler et al., 2006). Self-interest operates through two separate channels: considerations of who pays (and how much) and requirements around who is eligible to attend. These channels have conflicting implications for preschool preferences.

Regarding the first channel, self-interested taxpayers will seek to maximize wealth and favor public programs that do not increase their personal tax burden—or do so less than available alternatives. Accordingly, those choosing between a small, targeted program and a larger, universal one are likely to pick the former. This principle may be especially relevant for higher-income Americans, who pay a disproportionate share of public investments in preschool.

The second channel of self-interest, who is eligible to attend, contrasts universal programs offering preschool to all families with targeted programs that serve only a select few. Many scholars have argued that public support depends on program access—even though expanded access will cost those who pay (Barnett, 2011; Barnett et al., 2004; Bobo & Kluegel, 1993; Skocpol, 1991a, 1991b; Wilson, 1991; Zigler et al., 2006). Their arguments adopt the view, articulated by Skocpol (1991a), that “Americans will accept taxes that they perceive as contributions toward public programs in which there is a direct stake for themselves, their families, and their friends, not just for ‘the poor’” (p. 432). As a result, personal eligibility will affect preferences for preschool, making universal preschool policies more successful than targeted ones in communities with middle- and higher-income residents.

Preschool Support and Egalitarian Values

Evidence from policy areas beyond preschool suggests caution in assuming that self-interest is the only—or even the most important—motivation for citizens’ preschool policy priorities. For example, political and personal values might dictate these priorities (cf. Lau & Heldman, 2009; Sears & Funk, 1990, 2001). Americans may conflate poverty and race and, due to racial bias, oppose preschool programs narrowly targeted by income (Greenstein, 1991a, 1991b; Wilson, 1991). The adjudged deservingness of program beneficiaries and changing economic and political context may affect support for both targeted and universal programs (Greenberg, 1981; Sampson, 1981; Skitka & Tetlock, 1991).

Although proponents of targeted preschool tout its economic efficiency and high returns on investment, their primary arguments often center on values regarding equality of opportunity. Specifically, they argue that targeted programs can narrow racial and socioeconomic disparities that arise before the start of formal schooling by helping low-income families identify and access high-quality early learning opportunities (Finn, 2009; Zigler et al., 2011). This framing reflects the history of American public investments in preschool, including Head Start, which have long been associated with equalizing opportunity (Bobo & Kluegel, 1993; Page & Jacobs, 2009; Sniderman & Carmines, 1997).

Proponents of the universal approach appeal to values, as well. They argue that universal programs can facilitate economic integration in the classroom, expand access among low-income families by raising program visibility and reducing stigma, and improve offerings for lower-middle-income families. They prioritize the growth and development of all children—both through the direct provision of preschool and by raising the quality of other early care and education options through market-based competition (Barnett et al., 2004; Zigler et al., 2011, 2006).

Whether targeted or universal preschool better fulfills egalitarian values likely hinges on who benefits, in addition to who is eligible to attend. Targeted programs of small and large scale have been shown to improve children’s academic and social-emotional readiness for school and later life success. Because they serve only children from low-income families (or those otherwise deemed at risk), and because several studies suggest they are most effective for the least advantaged children, targeted programs are an efficient way of narrowing gaps (e.g., Bitler, Hoynes, & Domina, 2014; Campbell, Ramey, Pungello, Sparling, & Miller-Johnson, 2002; Miller, Farkas, Vandell, & Duncan, 2014; Wong, Cook, Barnett, & Jung, 2008).

There is considerably less evidence on the effects of universal preschool on gaps, as universal programs are newer and measuring gap effects is statistically and financially demanding. Yoshikawa and colleagues (2013) synthesize two studies and find that programs in Boston and Tulsa, Oklahoma, benefit low- and middle-income children, with impacts “statistically significantly larger on some assessments for children from low-income families.” Those studies and several others yield mixed results at kindergarten entry, with some demonstrating significant gap narrowing (particularly in math and executive function), a few indicating gap widening, and most indistinguishable from chance (Bartik, Gormley, & Adelstein, 2012; Ceci & Papierno, 2005; Fitzpatrick, 2008; Friedman-Krauss, Barnett, & Nores, 2016; Gormley, Gayer, Phillips, & Dawson, 2005; Henry et al., 2005; Peisner-Feinberg, Schaaf, LaForett, Hildebrandt, & Sideris, 2014; Weiland & Yoshikawa, 2013). Children from more and less advantaged backgrounds experience roughly equivalent academic gains from universal preschool, on average, and although those gains are relatively larger for low-income children, given lower baseline scores (Halle et al., 2009), they are generally not large enough in an absolute sense to address early disparities. Given existing research, who is eligible to attend and who benefits from preschool are inextricably linked.

Based on the literature summarized above, I designed novel items for a public opinion poll to answer the following research questions:

What is the national level of support for targeted and universal preschool?

To what extent are preferences for preschool affected by self-interested considerations of who pays and who is eligible to attend?

To what extent are preferences for preschool associated with egalitarian values surrounding who is eligible to attend and who benefits?

Data

This study uses unique public opinion data collected from a nationally representative Internet survey. Items analyzed here were part of a broader omnibus survey developed through the Laboratory for the Study of American Values at Stanford University (see Valant & Newark, 2016, for analyses of other items included on the survey). The survey was fielded between February 16 and 25, 2013, and distributed by YouGov, an online polling firm. YouGov receives a B rating from FiveThirtyEight based on the historical accuracy and methodology of its polls (including a rating penalty for its Internet-only approach; Silver, 2014).

In brief, YouGov uses a nonrandom (voluntary) panel of respondents and complex weighting algorithm in an attempt to provide valid estimates of national patterns (Ansolabehere & Schaffner, 2014; Rivers, 2007). Respondents are matched to a nationally representative sampling frame constructed from several census-type sources, including the 2010 American Community Survey, the 2008 and 2010 Current Population Survey, and the 2007 Pew U.S. Religious Landscape Survey. Matching characteristics include age, education, gender, race-ethnicity, predicted voter registration, news interest, and party identification. Matching processes iterate until each case in the target sample, here specified to include 1,000 respondents, is represented in the survey sample. According to standard YouGov procedures, respondents receive small incentives (points redeemable for gift cards) upon survey completion.

Table 1 summarizes characteristics of the final survey sample with and without weights generated through YouGov’s matching procedure. This sample includes residents of the District of Columbia and 49 states, including all states with targeted preschool and the 10 states where leaders had expressed a goal of universal provision at the time of the survey (Alabama, Florida, Georgia, Illinois, Iowa, Massachusetts, New York, Oklahoma, West Virginia, and Wisconsin). The weighted sample matches characteristics of the U.S. adult population based on recent figures collected by the U.S. Census Bureau (2013a, 2013b, 2013c) and the Pew U.S. Religious Landscape Survey (Pew Research Center, 2014).

Table 1

U.S. Population and Survey Sample Characteristics

	U.S. population	Full sample		Tax experiment		Race and values experiment
Variable	Adults ages 18+	Weighted means	Unweighted means	Tax-free prime	Tax prime	Black prime	White prime
Male	49%	48%	43%	42%	44%	45%	41%
Age (mean)	47	48	48	48	48	49	47
White	66%	70%	78%	78%	78%	78%	78%
Black	12%	11%	8%	7%	8%	8%	8%
Latino	15%	12%	7%	6%	7%	7%	6%
Asian	5%	2%	3%	3%	2%	3%	2%
Democrat	44%	34%	38%	40%	36%	38%	38%
Republican	37%	25%	26%	26%	26%	27%	25%
Independent	—	35%	32%	32%	32%	32%	31%
Liberal	24%	26%	31%	31%	30%	30%	32%
Conservative	36%	33%	36%	36%	35%	36%	35%
Moderate	33%	32%	26%	28%	25%	28%	25%
High school degree or below	42%	46%	42%	37%	38%	35%	40%
Some college or 2-year degree	29%	31%	34%	34%	34%	32%	36%
BA or above	29%	23%	29%	29%	28%	33%	24%**
Low income (below $30,000 per year)	21%	34%	32%	31%	33%	30%	34%
Middle income ($30,000–$70,000 per year)	33%	41%	41%	42%	41%	44%	38%†
High income (above $70,000 per year)	47%	25%	27%	27%	26%	25%	28%
Parents of school-aged children	19%	23%	23%	25%	21%	24%	23%
Egalitarianism (mean score)	—	7.5	7.4	7.4	7.4	7.4	7.4
n	—	1,000	1,000	490	510	528	472

Note. Data on sex, age, race-ethnicity, education, parent status, and income come from the U.S. Census Bureau (2013a, 2013b, 2013c). Data on political party affiliation and political ideology come from the U.S. Religious Landscape Study (Pew Research Center, 2014). The Religious Landscape Study asked respondents to identify themselves as Democrat, Republican, or Independent. Those who identified as Independent were then asked whether they lean to the Democratic or Republican Party and have been classified according to their reported lean.

†

p < .10. *p < .05. **p < .01. ***p < .001.

The full text of key survey items is detailed in the next section, and all items appear in Appendix A. Novel items were iteratively pilot- and field-tested prior to inclusion on the survey instrument using Amazon’s Mechanical Turk, an online marketplace for “Human Intelligence Tasks” shown to be valid and reliable in social science research (Buhrmester, Kwang, & Gosling, 2011). Results of these tests helped refine question wording, randomization patterns, and question length based on respondent feedback regarding interpretation and cognitive burden. In addition, the survey includes a three-item measure of egalitarianism standard in the American National Election Studies (see Feldman, 1983). The scale showed good reliability in this study (coefficient alpha of 0.66), in line with prior research (e.g., Feldman & Steenbergen, 2001). Finally, YouGov provided information on respondents’ demographic, socioeconomic, and political characteristics, including age, gender, party identification, political ideology, race-ethnicity, educational attainment, annual family income, and home state.

Methods

This study relies on a central survey experiment to gauge support for targeted and universal preschool. Supplemental questions, including additional survey experiments and assessments of respondent background characteristics, are detailed in Appendix A. Together, these items provide a rich data set with which to assess whether and how self-interest and egalitarian values motivate preschool preferences.

The central survey experiment uses a novel approach to test the effect of financial self-interest on support for targeted and universal preschool. All respondents begin the experiment with the following description:

Most states and the federal government offer public preschool programs for 3-year-old and 4-year-old children. These programs are free for families who use them.

Next, respondents are randomly assigned to see one of two tax primes. The first prime induces their consideration of the personal cost of public preschool, whereas the second is a tax-free prime, in which costs are paid through other sources. (Existing preschool programs rely on a variety of funding streams, including lotteries and sin taxes in states like Georgia and Arizona and cities like Philadelphia; this prime is consistent with several of them.) Roughly half of respondents receive the tax prime:

Most experts agree that if the government is going to pay for preschool, taxes may have to be increased on households like yours.

The remaining half of respondents receive the tax-free prime:

Most experts agree that the government can pay for preschool without increasing taxes on households like yours.

Finally, all respondents view descriptions of targeted (Type A) and universal (Type B) preschool programs and are asked to express their preferences for both:

There are two types of preschool programs:

Type A:
Programs are offered to children from families earning less than $30,000 each year.
These programs:
• Help low-income children do better in elementary school.
• Don’t serve higher-income children.
• Help close the gap between low- and higher-income children in school and later on in life.

Type B:
Programs are offered to all children, regardless of family income.
These programs:
• Help low-income children do better in elementary school.
• Help higher-income children do better in elementary school.
• Don’t close the gap between low- and higher-income children in school and later on in life.

Would you support or oppose the government funding programs like these?

Respondents rate Type A programs on a 5-point scale ranging from strongly support (2) to strongly oppose (–2), with neither support nor oppose (0) at the midpoint, and then rate Type B programs on a second identical scale. In pilot and field tests, random ordering of these descriptions showed no order effects, so Type A was always presented first.

This study does not seek to force a choice between targeted and universal preschool. Respondents may prefer one approach to the other, but some may show strong preferences and others weak ones. Alternatively, respondents may demonstrate equal levels of support (or opposition) to both forms of preschool. To illustrate the full range of opinion, respondents express their level of support for both approaches, and preschool preferences are computed as differences in support between Type A and Type B programs.

In addition to the central survey experiment, a second novel experiment examines the role of egalitarian values in motivating preferences for public preschool. Specifically, this experiment assesses the question of whether targeted preschool programs are racialized in ways that decrease their viability among the electorate (Greenstein, 1991b, 1991a). If respondents associate targeted programs with a particular racial or ethnic group, and they believe that groups differ in the rights and opportunities they deserve, then racialization is likely to shape preferences for targeted and universal preschool.

All respondents begin with a generic description of a targeted program:

In a neighboring state, public preschool is available for poor children. Some studies show children who attend the program do better in reading and math when they get to elementary school. Other studies suggest there are some benefits, but they are very short-lived. Parents say they like the program, though a few report that it was hard to find a spot for their children. One classroom is shown below.

Respondents then view a photograph depicting the program described. Half of respondents are randomly assigned a picture of a White teacher with two White students; the other half is assigned a picture of the same teacher with two Black students. The pictures are identical except for the race of the students (see Appendix A). Respondents are then asked,

Would you support or oppose your state funding programs like this one?

Answers are expressed on a 5-point scale ranging from strongly support (2) to strongly oppose (–2).

Together, these two survey experiments demonstrate how considerations of financial self-interest and egalitarian values affect support for public preschool and preferences for its targeted and universal forms. As demonstrated in Table 1, randomization was successful in the tax experiment. In the race and values experiment, randomization produced baseline equivalence on most observable characteristics, although treatment and control groups differ on educational attainment (having a BA or above) and income (earning between $30,000 and $70,000 per year, a difference that is marginally significant). Uncontrolled analyses in these cases should be interpreted with caution.

Analytic approaches include weighted t tests and weighted least squares (WLS) regression. Given the large number of t tests performed, results are corrected for multiple hypothesis testing using the Benjamini-Hochberg procedure. The procedure is performed using a false discovery rate of 0.10, as recommended by Benjamini and Yekutieli (2005). The full specification of WLS models and results of robustness checks using ordered probit models are included in Appendix B.

Despite several robustness and specification checks, this study is not without limitations. First, data collection was designed to support inferences about national preferences for preschool rather than to generate state or locally representative samples. I am unable to examine how the experience of specific targeted and universal programs affects public support and what that experience may imply for future programming. However, analyses suggest that preferences are not shaped by home state policy context (Appendix C), providing confidence in national patterns of preschool preferences and their variation by individuals’ demographic, socioeconomic, and political characteristics. Second, the survey sample differs from the American adult population on a small number of characteristics, notably, income (Table 1). Analyses attempted to correct for this difference by disaggregating subgroups and including these characteristics as control variables in multivariate regression models, but they may not have done so completely. Third, data collection occurred in 2013. Public preschool has expanded and changed in important ways since then, including growth in funding and enrollment overall and the addition of universal programs in several cities (Barnett et al., 2017). These changes have the potential to shift public opinion on the targeted-versus-universal debate. Fourth, key survey items included in this study are novel. Although they reflect intentional design and iterative pretesting, they have not benefited from multiple survey administrations across different time periods and policy contexts. In particular, descriptions of the effects (and gap effects) of targeted and universal preschool reflect a simplified summary of recent findings; these descriptions do not significantly alter respondent preferences (Appendix C) but may differ from arguments made in contemporary policy debates. Responses may also be subject to order effects, given that targeted and universal programs were always described in the same order, although these effects were not evident during pretesting.

Results

Support for Targeted and Universal Preschool

This is the first study of public opinion relevant to the targeted-versus-universal debate. As such, the joint distribution of preferences can inform policymakers, practitioners, and researchers focused on public preschool policy. Table 2 displays this distribution across the full sample, pooling across both tax primes. Percentages on the diagonal indicate equal levels of support for targeted and universal preschool; those above the diagonal capture a preference for the universal approach, whereas those below reflect a preference for targeted.

Table 2

Distribution of Support for Targeted and Universal Preschool (in percentages)

Variable	Strongly oppose universal preschool	Oppose universal preschool	Neither support nor oppose universal preschool	Support universal preschool	Strongly support universal preschool	Total
Strongly oppose targeted preschool	6.1	0.5	0.5	0.5	3.8	11.5
Oppose targeted preschool	0.7	3.5	0.7	4.9	3.5	13.3
Neither support nor oppose targeted preschool	1.2	2.3	10.9	6.0	3.8	24.2
Support targeted preschool	1.0	4.3	8.0	9.9	4.9	28.1
Strongly support targeted preschool	2.0	4.1	5.5	6.1	5.3	23.0
Total	11.0	14.7	25.6	27.5	21.3	100.0

Note. Distributions are weighted using the survey weights provided by YouGov and rounded to the nearest 0.1%.

Public opinion on targeted and universal preschool appears more variable, and less positive, than previous literature might suggest. More than one third (35%) of respondents favor targeted over universal preschool, whereas 29% favor universal over targeted. The plurality of respondents, 36%, has no preference. These respondents are more likely to favor than oppose both approaches (43% compared with 27%), and many are ambivalent about preschool in general (30%).

Support for both forms of preschool is moderate, on average. On a 5-point scale ranging from −2 (strongly oppose) to 2 (strongly support), mean support for targeted preschool is 0.38 points, and mean support for universal is 0.33 points. The difference between them, 0.05 points, is not statistically significant (p = .499). That is, a nationally representative sample of Americans demonstrates no distinguishable preference for targeted or universal preschool, on average.

Preschool Support and Self-Interest

As described in the Relevant Literature and Theoretical Framework section, there are two ways in which self-interest may affect support for targeted and universal preschool. One is the financial self-interest of those who pay for public preschool, as mediated by program costs. A second is program eligibility and access, defined by family income. Table 3 documents both. The first and second columns summarize the mean levels of support for targeted preschool under each tax prime, followed by a weighted t test comparing the two. An equivalent set of results is then presented for universal preschool. Together, they assess the causal effect of the tax prime on preferences for public preschool.

Table 3

Effects of Tax Priming on Support for Targeted and Universal Preschool

	Support for targeted preschool				Support for universal preschool
Variable	Tax-free prime (reference)	Tax prime	Effect of tax priming	SE	Tax-free prime (reference)	Tax prime	Effect of tax priming	SE
Panel 1
Full sample	0.34	0.41	0.08	(0.09)	0.50	0.20	−0.30	(0.10)**
Panel 2
Low income (below $30,000/year)	0.44	0.66	0.22	(0.18)	0.41	0.16	−0.25	(0.18)
Middle income ($30,000–$70,000 per year)	0.30	0.38	0.08	(0.15)	0.56	0.26	−0.30	(0.17)†
High income (above $70,000 per year)	0.23	0.32	0.09	(0.19)	0.54	0.22	−0.32	(0.19)†
Male	0.15	0.41	0.26	(0.15)†	0.42	0.05	−0.37	(0.15)*
Female	0.49	0.41	−0.08	(0.12)	0.57	0.34	−0.22	(0.12)†
White	0.28	0.21	−0.07	(0.10)	0.53	0.14	−0.39	(0.10)***
Black	1.14	0.85	−0.30	(0.26)	0.42	0.46	0.04	(0.35)
Hispanic	0.15	0.86	0.71	(0.34)*	0.30	0.26	−0.04	(0.39)
Asian	0.30	−0.44	−0.74	(0.50)	0.59	0.55	−0.04	(0.46)
Democrat	0.92	0.96	0.04	(0.14)	0.77	0.41	−0.35	(0.16)*
Republican	−0.25	0.13	0.38	(0.21)†	0.40	0.20	−0.20	(0.18)
Independent	0.20	0.10	−0.10	(0.16)	0.33	−0.05	−0.38	(0.17)*
Liberal	1.00	1.13	0.13	(0.14)	0.64	0.37	−0.27	(0.17)
Conservative	−0.31	−0.05	0.26	(0.17)	0.21	0.03	−0.18	(0.19)
Moderate	0.43	0.28	−0.15	(0.18)	0.68	0.19	−0.49	(0.17)**
High school degree or below	0.19	0.44	0.25	(0.15)^†	0.43	0.16	−0.28	(0.16)^†
Some college or 2-year degree	0.41	0.50	0.10	(0.15)	0.48	0.28	−0.20	(0.16)
BA or above	0.47	0.23	−0.24	(0.19)	0.62	0.19	−0.44	(0.18)*
Parents of school-aged children	0.30	0.49	0.20	(0.19)	0.69	0.54	−0.14	(0.18)
Not parents of school-aged children	0.35	0.39	0.04	(0.11)	0.43	0.09	−0.34	(0.11)**
Low level of egalitarianism (3–5)	−0.57	−0.60	−0.03	(0.18)	0.16	−0.38	−0.54	(0.22)*
Middle level of egalitarianism (6–9)	0.50	0.59	0.10	(0.12)	0.55	0.30	−0.25	(0.12)*
High level of egalitarianism (10–12)	1.09	1.14	0.05	(0.18)	0.79	0.58	−0.21	(0.20)

Note. The effect of tax priming is the difference in support (relative to a reference group shown the tax-free prime) for targeted and universal preschool. Means, standard errors, and t tests are weighted using the survey weights provided by YouGov. Support is measured using a 5-point scale ranging from −2 (strongly oppose) to 2 (strongly support).

†

p < .10. *p < .05. **p < .01. ***p < .001.

Across the full sample of respondents (Panel 1), the possibility of higher taxes has no statistically distinguishable effect on support for targeted programs (0.08 points, SE = 0.09, p = .412). Americans feel equally favorable toward public investments in low-income preschoolers whether or not they may have to pay more in taxes to fund them. By contrast, the threat of higher taxation substantially decreases support for universal preschool (0.30 points or nearly one quarter of a standard deviation, SE = 0.10, p = .002). As mentioned above, these results may reflect self-interested considerations of program size and cost. Alternatively, or perhaps additionally, they may suggest the importance of values. After all, targeted programs may be less expensive than universal ones, but they are not free.

The second panel of Table 3 demonstrates heterogeneity in the effects of self-interest on support for targeted and universal preschool. Specifically, I disaggregate the sample into demographic, socioeconomic, and ideological subgroups and assess the effect of the tax prime within each group. Because some groups are relatively small (see Table 1), interpretation of results for these groups should proceed with caution.

Overall, the threat of higher taxation has little effect on support for the targeted approach. Estimates are generally small and vary in their precision. Among Hispanics, this threat significantly increases support for targeted preschool (0.71 points, SE = 0.34, p = .05). Results suggesting increased support among males, Republicans, and those with a high school degree or below are marginally significant (p < .10). None of these findings retains significance when corrected for multiple hypothesis testing using the Benjamini-Hochberg procedure.

Support for universal preschool is generally more susceptible to considerations of financial self-interest than support for targeted programming. Among males (0.37 points, SE = 0.15, p < .05), Whites (0.39 points, SE = 0.10, p < .001), Democrats (0.36 points, SE = 0.16, p < .05), Independents (0.38 points, SE = 0.17, p < .05), ideological moderates (0.49 points, SE = 0.17, p < .01), those with a BA or above (0.43 points, SE = 0.18, p < .05), those without school-aged children (0.34 points, SE = 0.11, p < .01), and those with low and moderate levels of observed egalitarianism (0.54 points and 0.25 points, SE = 0.22 and SE = 0.12, p < .05), the threat of higher taxation decreases support for universal preschool. These effects are robust to adjustment using the Benjamini-Hochberg procedure. Declines in support are also observed among respondents with low levels of formal education, females, and those of middle and high family income (p < .10), although these declines lose significance after adjustment for multiple hypothesis testing.

Finally, a small number of subgroups appears consistent in level of support for both targeted and universal preschool regardless of the possibility of higher taxation. These subgroups include Black and low-income respondents, parents of school-age children, respondents with moderate levels of formal education, and those who identify as highly ideological—either as liberals, conservatives, or strong egalitarians. (Likewise, Asian American respondents show no detectable change in opinion due to the tax prime, although this is likely due to insufficient sample size.) For these groups, financial self-interest does not appear to shift beliefs about either form of public preschool.

Overall, respondents shown the tax-free prime demonstrate a marginally significant preference for the universal approach (0.16 points, p < .10), whereas those exposed to the possibility of higher taxation show a more substantial preference for targeted preschool (0.21 points, p < .05). Disaggregating the sample by program eligibility, respondents who qualify for targeted preschool (based on a self-reported family income of $30,000 per year or less) indicate no significant preference for either preschool approach under the tax-free prime, whereas those ineligible favor universal preschool given the same priming (0.26 points for middle-income respondents, p < .05, and 0.31 for high-income respondents, p < .10). Under the tax prime, an opposite pattern emerges: Low-income respondents show a large and significant preference for targeted preschool (0.50 points, p < .01), whereas middle- and high-income respondents have no significant preference.

These results both uphold and counter Skocpol’s (1991a, 1991b) argument that Americans will be willing to pay more in taxes for programs they and others like them can use. Findings suggest a nuanced interaction between the self-interests of who pays and who is eligible to attend. They also suggest that personal values may underlie preferences for public preschool. I turn to these values next, focusing on egalitarian notions of equal opportunity.

Preschool Support and Egalitarian Values

Egalitarianism may function in two ways with respect to support for public preschool (see Cook & Hegtvedt, 1983). First, respondents may associate egalitarianism, and specifically the belief in equality of opportunity, with uniform, universally available services. Second, they may associate it with narrower public investments in children from low-income families. The descriptions of targeted and universal programs allow for both interpretations.

Egalitarianism is associated with support for preschool, generally, and higher levels of support for its targeted form. Figure 1 plots the smoothed Lowess curves of average support for targeted and universal preschool by respondents’ observed egalitarian values. A strong positive relationship between egalitarianism and both preschool approaches is immediately visible: One additional point on the egalitarianism scale corresponds to an increase in support for targeted preschool of nearly one fifth of a standard deviation (p < .001) and a smaller increase in support for universal preschool of roughly one tenth of a standard deviation (p < .001).

Figure 1.

Mean support for targeted and universal preschool by respondents’ egalitarianism.

In a second test of the relationship between egalitarian values and preferences for public preschool, I examine whether targeted preschool programs are racialized in ways that decrease their viability among the electorate (Greenstein, 1991b, 1991a). Table 4 shows no significant difference in support for targeted preschool by the race of children attending (SE = 0.09, p = .624). Furthermore, racialization does not significantly affect support for targeted preschool among any of the subgroups identified in Table 3. Estimates are generally close to zero and vary in their precision. These results document the strength of values underlying preferences for public preschool, particularly, egalitarian values regarding equality of opportunity.

Table 4

Effects of Racial Priming on Support for Targeted Preschool

Variable	Support for targeted preschool with White prime (reference)	Support for targeted preschool with Black prime	Effect of racial priming	SE
Panel 1
Full sample	0.55	0.60	0.04	(0.09)
Panel 2
Low income (below $30,000 per year)	0.71	0.73	0.02	(0.17)
Middle income ($30,000–$70,000 per year)	0.54	0.58	0.04	(0.13)
High income (above $70,000 per year)	0.42	0.56	0.14	(0.17)
Male	0.51	0.47	−0.05	(0.14)
Female	0.59	0.71	0.12	(0.10)
White	0.48	0.44	−0.04	(0.09)
Black	0.94	1.22	0.27	(0.26)
Hispanic	0.56	0.95	0.38	(0.32)
Asian	0.79	0.59	−0.20	(0.48)
Democrat	1.08	1.23	0.15	(0.12)
Republican	0.21	0.06	−0.15	(0.18)
Independent	0.32	0.32	0.00	(0.18)
Liberal	1.25	1.28	0.03	(0.12)
Conservative	0.01	−0.14	−0.15	(0.16)
Moderate	0.63	0.65	0.03	(0.14)
High school degree or below	0.60	0.50	−0.10	(0.14)
Some college or 2-year degree	0.47	0.69	0.22	(0.14)
BA or above	0.56	0.67	0.11	(0.17)
Parents of school-age children	0.74	0.70	−0.04	(0.18)
Not parents of school-age children	0.49	0.56	0.07	(0.10)
Low level of egalitarianism (3–5)	−0.41	−0.40	0.00	(0.17)
Middle level of egalitarianism (6–9)	0.72	0.79	0.08	(0.10)
High level of egalitarianism (10–12)	1.26	1.24	−0.02	(0.13)

Note. The effect of racial priming is the difference in support (relative to a reference group shown the White prime) for targeted preschool. Means, standard errors, and t tests are weighted using the survey weights provided by YouGov. Support is measured using a 5-point scale ranging from −2 (strongly oppose) to 2 (strongly support).

†

p < .10. *p < .05. **p < .01. ***p < .001.

Preferences for Targeted and Universal Preschool

Synthesizing findings in the previous two sections, I use multivariate regression to disentangle demographic, socioeconomic, and ideological predictors of preschool preferences. The results appear in Table 5. In each panel, column 1 reports findings from a basic model predicting preschool preferences with indicators of self-interest and egalitarian values. Demographic and socioeconomic controls are added in column 2. Column 3 adds political ideology, and column 4 adds party identification; because these variables are highly correlated with each other (raw correlation of 0.65) and with egalitarianism (0.64 and 0.54, respectively), I add them separately to minimize problems of multicollinearity. The full model and related robustness checks are detailed in Appendix B.

Table 5

Weighted Least Squares Estimates of Preferences for Targeted and Universal Preschool

	Panel 1.Dependent variable: Support for targeted preschool				Panel 2.Dependent variable: Support for universal preschool				Panel 3.Dependent variable: Preference for universal over targeted preschool
Variable	(1)	(2)	(3)	(4)	(1)	(2)	(3)	(4)	(1)	(2)	(3)	(4)
Focal independent variables
Tax prime	0.09	0.07	0.06	0.07	−0.32**	−0.29**	−0.29**	−0.28**	−0.41**	−0.36**	−0.35**	−0.35**
	(0.09)	(0.09)	(0.09)	(0.09)	(0.10)	(0.10)	(0.10)	(0.09)	(0.13)	(0.12)	(0.12)	(0.12)
Low income	0.11	0.12	0.1	0.12	−0.17	−0.16	−0.16	−0.16	−0.27†	−0.28†	−0.26†	−0.27†
	(0.10)	(0.10)	(0.10)	(0.10)	(0.11)	(0.11)	(0.11)	(0.11)	(0.15)	(0.15)	(0.15)	(0.15)
Egalitarianism	0.25***	0.25***	0.22***	0.23***	0.12***	0.11***	0.11***	0.11***	−0.13***	−0.14***	−0.11**	−0.12***
	(0.02)	(0.02)	(0.02)	(0.02)	(0.02)	(0.02)	(0.03)	(0.02)	(0.03)	(0.03)	(0.03)	(0.03)
Controls
Male		0.01	0.03	0.02		−0.17†	−0.17†	−0.16		−0.18	−0.2	−0.18
		(0.09)	(0.09)	(0.09)		(0.10)	(0.10)	(0.10)		(0.14)	(0.14)	(0.14)
Race (reference is White)
Black		0.2	0.2	0.12		−0.06	−0.06	−0.1		−0.26	−0.26	−0.21
		(0.15)	(0.15)	(0.15)		(0.20)	(0.20)	(0.20)		(0.25)	(0.25)	(0.25)
Hispanic		0.18	0.21	0.15		−0.1	−0.09	−0.14		−0.28	−0.31	−0.29
		(0.19)	(0.19)	(0.19)		(0.20)	(0.20)	(0.20)		(0.26)	(0.27)	(0.26)
Asian		−0.48	−0.38	−0.44		0	−0.02	0.04		0.48	0.37	0.48
		(0.33)	(0.32)	(0.32)		(0.30)	(0.31)	(0.31)		(0.41)	(0.41)	(0.41)
All other races		0.28	0.31	0.28		−0.45†	−0.45†	−0.42†		−0.73*	−0.77*	−0.7*
		(0.19)	(0.21)	(0.19)		(0.24)	(0.25)	(0.24)		(0.32)	(0.34)	(0.32)
Education (reference is some college)
High school diploma or below		−0.12	−0.12	−0.11		−0.08	−0.08	−0.08		0.03	0.04	0.03
High school diploma or below		(0.10)	(0.10)	(0.10)		(0.12)	(0.12)	(0.12)		(0.15)	(0.15)	(0.15)
BA or above		0.07	0.03	0.06		0.06	0.07	0.06		−0.01	0.04	0.01
		(0.11)	(0.11)	(0.11)		(0.12)	(0.12)	(0.12)		(0.16)	(0.16)	(0.16)
Age		0	0	0		0	0	−0.01		−0.01	−0.01	−0.01
		(0.00)	(0.00)	(0.00)		(0.00)	(0.00)	(0.00)		(0.00)	(0.00)	(0.00)
Parent		−0.04	−0.01	−0.04		0.15	0.13	0.15		0.19	0.15	0.19
		(0.13)	(0.13)	(0.13)		(0.13)	(0.13)	(0.13)		(0.19)	(0.18)	(0.19)
Public preschool use		0.12	0.07	0.11		0.3	0.32	0.3		0.17	0.24	0.19
		(0.19)	(0.19)	(0.19)		(0.20)	(0.20)	(0.19)		(0.27)	(0.27)	(0.27)
Political ideology (reference is liberal)
Moderate			−0.34**				0.12				0.46**
			(0.11)				(0.11)				(0.16)
Conservative			−0.39**				0.01				0.39†
			(0.14)				(0.15)				(0.21)
Party identification (reference is Democrat)
Independent				−0.3**				−0.24*				0.06
				(0.10)				(0.11)				(0.16)
Republican				−0.24†				0.04				0.28
				(0.13)				(0.13)				(0.19)
Constant	−1.57***	−1.59***	−1.13***	−1.26***	−0.33†	0.02	−0.02	0.14	1.25***	1.62***	1.11*	1.39***
	(0.14)	(0.24)	(0.29)	(0.26)	(0.18)	(0.27)	(0.32)	(0.29)	(0.21)	(0.36)	(0.44)	(0.38)
R ²	0.25	0.26	0.27	0.27	0.07	0.1	0.1	0.11	0.06	0.09	0.1	0.09
N	864	864	864	864	864	864	864	864	864	864	864	864

Note. Dependent variables in Panels 1 and 2 are measured using a 5-point scale, ranging from strongly support (2) to strongly oppose (–2), with neither support nor oppose (0) at the midpoint. Dependent variable in the third panel is the difference in support for preschool by approach (universal vs. targeted), measured using a 9-point scale. Column 1 reports findings from a basic model predicting each dependent variable with indicators of self-interest and egalitarian values. Column 2 adds demographic and socioeconomic controls. Column 3 adds political ideology, and column 4 adds party identification; these are included in separate models to minimize problems of multicollinearity. Analytic sample for all models is restricted to the respondents with sufficient information to be included in columns 3 and 4. All models are weighted using the survey weights provided by YouGov. Robust standard errors in parentheses.

†

p < .10. *p < .05. **p < .01. ***p < .001.

Panel 1 documents support for targeted preschool. There, as in Table 3, the possibility of higher taxation has no significant effect on public opinion (0.06 to 0.09 points, SE = 0.09). Likewise, although low income (a key condition for targeted program eligibility) is positively correlated with support for the targeted approach, this relationship is not distinguishable from chance (0.10 to 0.12 points, SE = 0.10). Instead, support for targeted preschool appears to be driven by ideological values. Egalitarianism maintains a strong and significant relationship with support for the targeted approach: Across columns 1 through 4, a one-unit increase in egalitarian beliefs is associated with an increase in support of nearly one fifth of a standard deviation (0.22 to 0.25 points, SE = 0.02, p < .001). This relationship is robust to the inclusion of political ideology and partisanship (columns 3 and 4).

Panel 2 suggests that the politics of universal preschool depart from those of targeted preschool in three important ways. First, egalitarianism continues to predict support for the universal approach, but the coefficients are half as large as those in Panel 1 (a difference of roughly 0.08 standard deviations). This finding confirms the relationship displayed in Figure 1. Second, as in Table 3, support for universal preschool is susceptible to the threat of higher taxation. The tax prime decreases support for universal preschool by 0.3 scale points (nearly one quarter of a standard deviation, SE = 0.10, p < .01). Third, universal preschool appears less partisan and ideological than the targeted approach: Neither of the political ideology variables in column 3, nor the Republican indicator in column 4, predicts support.

Panel 3 compares the politics of targeted and universal preschool directly. Here, the dependent variable is the difference in support for universal and targeted preschool (measured using a 9-point scale that ranges from −4, among respondents who strongly oppose universal preschool and strongly support targeted preschool, to 4, among respondents who strongly support universal preschool and strongly oppose targeted preschool). Here, a negative coefficient implies a decrease (or association with a decrease) in support for universalism relative to targeting. The threat of higher taxation induces such a loss (0.35 points in the most robust models, SE = 0.12, p < .01). Departing from the first two panels, low income marginally predicts a loss of support for universal relative to targeted preschool (0.26 to 0.27 points, SE = 0.15, p < .10). Egalitarian values are associated with a similar change (0.11 to 0.12 scale points, SE = 0.03, p < .001, in most specifications). The only predictors that increase relative support for universalism are moderate and conservative ideology (p < .01 and p < .10, respectively).

In all, both self-interest and values condition support for public preschool. The lesser tax burden of targeted preschool appeals to those who pay, targeted access appeals to those eligible to attend, and egalitarian values enacted by serving low-income children appeal across the board. As detailed in Appendix C, these findings are replicable and robust. A second survey (N = 1,000), fielded between June 24 and July 5, 2013, duplicates findings on support for targeted and universal preschool from the main survey. The second survey also assesses the impact of framing on preferences for preschool. Neither descriptions of the effects (and gap effects) of targeted and universal programs nor information about the costs of these programs produces any distinguishable change in support, on average. Respondents who believe that one form of preschool does a better job of educating young children or yields greater benefits to society are more likely to favor that form, but the plurality of respondents find targeted and universal programs equally effective and beneficial.

Discussion

For much of its history, American public preschool has been delivered through targeted programs. These programs were motivated by a particular set of goals and subjected to shifting social, political, and economic concerns. Over the past two decades, however, states have made major investments in universal provision, and cities have followed suit, with New York City, San Antonio, Seattle, and Philadelphia launching new programs in recent years. State and local efforts benefited from substantial federal support under the Obama administration, including $1.25 billion in funding through initiatives like Race to the Top–Early Learning Challenge, Preschool Development and Expansion Grants, and the Every Student Succeeds Act. These initiatives reflect the status of preschool for all as a national priority (Mead, 2017; Obama, 2013). Yet, their legacy remains uncertain under the Trump administration (Loewenberg, Bornfreund, Lieberman, & Loewenberg, 2016). Within a changing policy context likely to be defined by leadership from states and localities, this paper offers the first public-opinion study of preferences for targeted and universal preschool.

I conducted a nationally representative survey and found moderate support for targeted and universal preschool and no significant preference for either approach, on average. More than one third of respondents equally support (or oppose) both. Preferences for each approach can be explained to some extent by demographic characteristics and income eligibility. But the most consistent and meaningful differences in preferences are defined by financial self-interest and personal values. The possibility of higher taxation significantly decreases support for universal preschool but has no effect on support for targeted preschool. Self-identified Democrats, liberals, and egalitarians favor targeted preschool, whereas Republicans, conservatives, and inegalitarians favor the universal approach. These variations explain the lack of national preference, but they also suggest that public opinion may align clearly with one form of preschool or the other in some states, cities, and counties. In others, voters may be sufficiently ambivalent to allow policymakers to make the case for targeted or universal preschool.

Taken together, the results of this study depart from previous thinking on the politics of preschool. They suggest that a savvy policymaker should not necessarily endorse universal over targeted preschool, as many have argued (Barnett, 2011; Barnett et al., 2004; Bobo & Kluegel, 1993; Skocpol, 1991a, 1991b; Wilson, 1991; Zigler et al., 2006); rather, preferences for preschool vary based on the economic, political, and demographic context in which programs are proposed and implemented. When public preschool can be financed through existing sources, Americans demonstrate a marginally significant preference for the universal approach. However, those facing the possibility of higher taxation, and those with strong egalitarian values, prefer program targeting. On average, targeted and universal preschool receive similar levels of support.

These findings likely stem from the particular nature of early childhood and the history of early childhood education in America. Previous research suggests that Americans allocate scarce public resources based on three primary considerations: Need among potential beneficiaries, efficiency of resource use, and deservingness—whether potential beneficiaries are responsible for their life circumstances and worthy of public help to improve them (Greenberg, 1981; Skitka & Tetlock, 1991). Here, the potential beneficiaries are 3- and 4-year-old children. Although it is difficult to determine how individuals may judge their need and efficiency, it is clear that no young child can reasonably be blamed for the low income of his or her family and its attendant hardships. Moreover, these hardships pose a direct challenge to that quintessential American value: equality of opportunity.

Americans favor many policy interventions that equalize opportunity (Page & Jacobs, 2009), but preschool policy has special appeal. It offers a chance for success in life before the development of an internal locus of control (in the terms of Skitka & Tetlock, 1991) that might warrant blame for poor decisions or personal inaction. Indeed, American values may even transcend notions of equality of opportunity to demand equality of outcomes for very young children (Satz, 2007). To the extent that Americans believe targeted preschool will equalize opportunity and outcomes more effectively than universal preschool, those with egalitarian values will support preschool targeting as public policy.

In addition to contemporary values regarding early childhood, past values have structured a system of preschool programs that may influence public opinion, as well. Beginning in the 1920s, major federal, state, and local investments in young children have taken a targeted approach. Head Start and many state preschool programs maintain the goal of school readiness for children from poor families—a combined metric of equal opportunity and outcomes. These programs create the possibility of path dependence in preschool policymaking: Given institutional barriers to policy reform (Karch, 2013), past targeted investments may condition current values and priorities for both targeted and universal preschool. Support for targeted preschool may be associated, to some extent, with the historical prevalence of this approach.

Public conceptions of early childhood and past investments in early childhood education can explain differences between the above findings and arguments commonly made in the targeted-versus-universal debate. Nevertheless, these differences need not imply that the politics of preschool are entirely idiosyncratic. Preschool policy preferences are strongly associated with partisanship, political ideology, and egalitarian beliefs; they are also conditioned by taxpayers’ financial self-interest. It is not the case, as Kirp (2007) argues, that “the allure of pre-K transcends ideology” (p. 4; see also Rose, 2012). Instead, preferences for targeted and universal preschool follow logically from American values regarding equal opportunity and citizens’ regard for their own taxation.

Federal, state, and local governments are currently considering and implementing a diverse array of public preschool initiatives. These initiatives have the potential to alter K–12 education, Head Start, subsidies for private childcare, and existing early intervention programs. If preschool policies are to reflect the public will, then polls of targeted and universal preschool preferences must be replicated over time and focused on states and municipalities with evolving policy landscapes. Future research might also investigate public opinion on different forms of preschool targeting (for example, income based or geographic) and hybrid policy options (for example, universally available preschool with fees for higher-income families). Understanding how self-interest and egalitarian values shape opinion on these issues can inform the efforts of policymakers at all levels of government—and improve the early educational experiences of young children in America.

Footnotes

Appendix A

Appendix B

Appendix C

Acknowledgements

This paper has benefited from generous financial support from several sources. The Institute of Education Sciences provided funding through Predoctoral Training Grant No. R305B090016. The I. James and Viola Lewis Quillen and Karr Family Fellowships at Stanford University provided additional funding. The Laboratory for the Study of American Values and the Institute for Research in the Social Sciences at Stanford offered both the means and the opportunity to conduct this research. Professional development support from the Urban Institute allowed its conversion into a journal-length article. Numerous advisors and colleagues have enriched this study over the years. In particular, I thank Sean Reardon, Susanna Loeb, Deborah Stipek, Paul Sniderman, and Michael Tomz, along with the editors and anonymous reviewers at AERA Open. Despite this generous support, I have undoubtedly made some errors, for which I alone am responsible.

Author

ERICA H. GREENBERG is a senior research associate in the Center on Labor, Human Services, and Population and the Education Policy Program at the Urban Institute, 2100 M Street NW, Washington, DC 20037; egreenberg@urban.org. Her research focuses on early childhood programs and policies.

References

Ansolabehere

Schaffner

B. F.

(2014). Does survey mode still matter? Findings from a 2010 multi-mode comparison. Political Analysis, 22(3), 285–303.

Barnett

W. S.

(2011). Four reasons the United States should offer every child a preschool education. In Zigler

Gilliam

W. S.

Barnett

W. S.

(Eds.), The pre-K debates: Current controversies and issues (pp. 34–39). Baltimore, MD: Brookes.

Barnett

W. S.

Brown

Shore

(2004). The universal vs. targeted debate: Should the United States have preschool for all? Preschool Policy Matters, 6, 1–15.

Barnett

W. S.

Friedman-Krauss

A. H.

Weisenfeld

G. G.

Horowitz

Kasmin

Squires

J. H.

(2017). The state of preschool 2016: State preschool yearbook. New Brunswick, NJ: National Institute for Early Education Research.

Bartik

T. J.

Gormley

Adelstein

(2012). Earnings benefits of Tulsa’s pre-K program for different income groups. Economics of Education Review, 31(6), 1143–1161.

Benjamini

Yekutieli

(2005). Quantitative trait loci analysis using the false discovery rate. Genetics, 171(2), 783–790.

Bitler

M. P.

Hoynes

H. W.

Domina

(2014). Experimental evidence on distributional effects of Head Start (Working Paper No. 20434). Cambridge, MA: National Bureau of Economic Research. Retrieved from http://www.nber.org/papers/w20434

Bobo

Kluegel

J. R.

(1993). Opposition to race-targeting: self-interest, stratification ideology, or racial attitudes? American Sociological Review, 58(4), 443–464.

Bostrom

(2002). A meta-analysis of opinion data concerning school readiness, early childhood and related issues. Washington, DC: FrameWorks Institute.

10.

Buhrmester

Kwang

Gosling

S. D.

(2011). Amazon’s Mechanical Turk: A new source of inexpensive, yet high-quality, data? Perspectives on Psychological Science, 6(1), 3–5.

11.

Burstein

(2003). The impact of public opinion on public policy: A review and an agenda. Political Research Quarterly, 56(1), 29–40.

12.

Campbell

F. A.

Ramey

C. T.

Pungello

E. P.

Sparling

Miller-Johnson

(2002). Early childhood education: Young adult outcomes from the Abecedarian Project. Applied Developmental Science, 6, 42–57.

13.

Ceci

S. J.

Papierno

P. B.

(2005). The rhetoric and reality of gap closing: When the “have-nots” gain but the “haves” gain even more. American Psychologist, 60(2), 149–160.

14.

Cook

K. S.

Hegtvedt

K. A.

(1983). Distributive justice, equity, and equality. Annual Review of Sociology, 9, 217–241.

15.

Erikson

R. S.

Wright

G. C.

McIver

J. P.

(1993). Statehouse democracy: Public opinion and policy in the American states. Cambridge, UK: Cambridge University Press.

16.

Feldman

(1983). Report on values in the 1983 pilot study. Ann Arbor, MI: American National Election Studies.

17.

Feldman

Steenbergen

(2001). Social welfare attitudes and the humanitarian sensibility. In Kuklinski

J. H.

(Ed.), Citizens and politics: Perspectives from political psychology (pp. 366–400). Cambridge, UK: Cambridge University Press.

18.

Finn

C. E.

(2009). Reroute the preschool juggernaut. Stanford, CA: Hoover Institution Press.

19.

Fitzpatrick

M. D.

(2008). Starting school at four: The effect of universal pre-kindergarten on children’s academic achievement. The B.E. Journal of Economic Analysis & Policy, 8(1), Article 46.

20.

Friedman-Krauss

Barnett

W. S.

Nores

(2016). How much can high-quality universal pre-k reduce achievement gaps? Washington, DC: Center for American Progress.

21.

Gormley

W. T.

Gayer

Phillips

Dawson

(2005). The effects of universal pre-k on cognitive development. Developmental Psychology, 41(6), 872–884.

22.

Greenberg

(1981). The justice of distributing scarce and abundant resources. In Lerner

M. J.

Lerner

S. C.

(Eds.), The justice motive in social behavior: Adapting to times of scarcity and change (pp. 289–316). New York, NY: Plenum Press.

23.

Greenstein

(1991a). Relieving poverty: An alternative view. Brookings Review, Summer, 34–35.

24.

Greenstein

(1991b). Universal and targeted approaches to relieving poverty: An alternative view. In Jencks

Peterson

P. E.

(Eds.), The urban underclass (pp. 437–459). Washington, DC: Brookings Institution Press.

25.

Halle

Forry

Hair

Perper

Wandner

Wessel

Vick

(2009). Disparities in early learning and development: Lessons from the Early Childhood Longitudinal Study–Birth Cohort (ECLS-B). Washington, DC: Child Trends.

26.

Henry

G. T.

Rickman

Dana K.

with Ponder

B. D.

Henderson

L. W.

Mashburn

Gordon

C. S.

(2005). The Georgia Early Childhood Study, 2001–2004, final report. Atlanta: Andrew Young School of Policy Studies, Georgia State University.

27.

Kahn

M. E.

Barron

(2015). The political economy of state and local investment in pre-k programs (Discussion Paper Series No. IZA DP No. 9337). Bonn, Germany: Institute for the Study of Labor.

28.

Kahneman

Tversky

(1984). Choices, values, and frames. American Psychologist, 39(4), 341–350.

29.

Karch

(2013). Early start: Preschool politics in the United States. Ann Arbor: University of Michigan Press.

30.

Kingdon

J. W.

(2011). Agendas, alternatives, and public policies (2nd ed.). New York, NY: Longman.

31.

Kirp

D. L.

(2007). The sandbox investment: The preschool movement and kids-first politics. Cambridge, MA: Harvard University Press.

32.

Lau

R. R.

Heldman

(2009). Self-interest, symbolic attitudes, and support for public policy: A multilevel analysis. Political Psychology, 30(4), 513–537.

33.

Loewenberg

Bornfreund

Lieberman

Loewenberg

(2016, December 8). What do DeVos and Price picks mean for early ed? New America. Retrieved from https://www.newamerica.org/education-policy/edcentral/devos-price-early-ed/

34.

Mead

(2017, January 13). Obama’s early childhood education legacy. U.S. News & World Report. Retrieved from https://www.usnews.com/opinion/knowledge-bank/articles/2017-01-13/obamas-leaves-a-lasting-legacy-on-early-childhood-education

35.

Miller

E. B.

Farkas

Vandell

D. L.

Duncan

G. J.

(2014). Do the effects of Head Start vary by parental preacademic stimulation? Child Development, 85(4), 1385–1400. https://doi.org/10.1111/cdev.12233

36.

Nagle

Goidel

(2007). The Louisiana early childhood public opinion survey. Baton Rouge, LA: Louisiana Partnership for Children and Families.

37.

Obama

(2013, February 12). Remarks by the president in the State of the Union address. Retrieved from http://www.whitehouse.gov/the-press-office/2013/02/12/remarks-president-state-union-address

38.

Page

B. I.

Jacobs

L. R.

(2009). Class war? What Americans really think about economic inequality. Chicago, IL: University Of Chicago Press.

39.

Page

B. I.

Shapiro

R. Y.

(1983). Effects of public opinion on policy. American Political Science Review, 77(1), 175–190.

40.

Peisner-Feinberg

E. S.

Schaaf

J. M.

LaForett

D. R.

Hildebrandt

L. M.

Sideris

(2014). Effects of Georgia’s pre-K program on children’s school readiness skills: Findings from the 2012–2013 evaluation study. Chapel Hill: University of North Carolina, FPG Child Development Institute. Retrieved from http://fpg.unc.edu/node/6831

41.

Pew Research Center. (2014). U.S. religious landscape study. Retrieved from http://www.pewforum.org/religious-landscape-study/

42.

Rivers

(2007). Sampling for web surveys. Paper presented at the Joint Statistical Meetings, Salt Lake City, UT. Retrieved from http://www.laits.utexas.edu/txp_media/html/poll/files/Rivers_matching.pdf

43.

Rose

E. R.

(2012). The promise of preschool: From Head Start to universal pre-kindergarten. New York, NY: Oxford University Press.

44.

Sampson

E. E.

(1981). Social change and the contexts of justice motivation. In Lerner

M. J.

Lerner

S. C.

(Eds.), The justice motive in social behavior: Adapting to times of scarcity and change (pp. 97–124). New York, NY: Plenum Press.

45.

Satz

(2007). Equality, adequacy, and education for citizenship. Ethics, 117(4), 623–648.

46.

Sears

D. O.

Funk

C. L.

(1990). The limited effect of economic self-interest on the political attitudes of the mass public. Journal of Behavioral Economics, 19(3), 247–271.

47.

Sears

D. O.

Funk

C. L.

(2001). The role of self-interest in social and political attitudes. Advances in Experimental Social Psychology, 24, 1–91.

48.

Silver

(2014). FiveThirtyEight’s pollster ratings. Retrieved October 27, 2017, from https://projects.fivethirtyeight.com/pollster-ratings/

49.

Skitka

L. J.

Tetlock

P. E.

(1991). Allocating scarce resources: A contingency model of distributive justice. Journal of Experimental Social Psychology, 28, 491–522.

50.

Skocpol

(1991a). Targeting within universalism: Politically viable policies to combat poverty in the United States. In Jencks

Peterson

P. E.

(Eds.), The urban underclass (pp. 411–436). Washington, DC: Brookings Institution Press.

51.

Skocpol

(1991b). Universal appeal: Politically viable policies to combat poverty. Brookings Review, Summer, 29–33.

52.

Sniderman

P. M.

Carmines

E. G.

(1997). Reaching beyond race. Cambridge, MA: Harvard University Press.

53.

Sylvester

(2001). Caring for our youngest: Public attitudes in the United States. The Future of Children, 11(1), 52–61.

54.

U.S. Census Bureau. (2013a). ACS demographic and housing estimates: 2009–2013 American Community Survey 5-year estimates. Retrieved from https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid=ACS_15_5YR_DP05&src=pt

55.

U.S. Census Bureau. (2013b). Current Population Survey 2014: Annual social and economic supplement. Retrieved from https://www.census.gov/cps/data/cpstablecreator.html

56.

U.S. Census Bureau. (2013c). Educational attainment: 2009–2013 American Community Survey 5-year estimates. Retrieved from https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?pid=ACS_15_5YR_S1501&src=pt

57.

Valant

Newark

D. A.

(2016). The politics of achievement gaps. Educational Researcher, 45(6), 331–346.

58.

Weiland

Yoshikawa

(2013). Impacts of a prekindergarten program on children’s mathematics, language, literacy, executive function, and emotional skills. Child Development, 84(6), 2112–2130.

59.

Wilson

W. J.

(1991). Public policy and the truly disadvantaged. In Jencks

Peterson

P. E.

(Eds.), The urban underclass (pp. 460–478). Washington, DC: Brookings Institution Press.

60.

Wong

V. C.

Cook

T. D.

Barnett

W. S.

Jung

(2008). An effectiveness-based evaluation of five state pre-kindergarten programs. Journal of Policy Analysis and Management, 27(1), 122–154.

61.

Yoshikawa

Weiland

Brooks-Gunn

Burchinal

M. R.

Espinosa

L. M.

Gormley

W. T.

. . . Zaslow

M. J.

(2013). Investing in our future: The evidence base on preschool education. New York, NY: Foundation for Child Development and Society for Research in Child Development. Retrieved from http://fcd-us.org/sites/default/files/EvidenceBaseonPreschooEducationFINAL.pdf

62.

Zigler

Gilliam

W. S.

Barnett

W. S.

(Eds.). (2011). The pre-K debates: Current controversies and issues. Baltimore, MD: Brookes.

63.

Zigler

Gilliam

W. S.

Jones

S. M.

(2006). A vision for universal preschool education. Cambridge, MA: Cambridge University Press.

Public Preferences for Targeted and Universal Preschool

Abstract

Keywords

Relevant Literature and Theoretical Framework

Preschool Support and Self-Interest

Preschool Support and Egalitarian Values

Data

Methods

Results

Support for Targeted and Universal Preschool

Preschool Support and Self-Interest

Preschool Support and Egalitarian Values

Preferences for Targeted and Universal Preschool

Discussion

Footnotes

Appendix A

Appendix B

Appendix C

Acknowledgements

Author

References