PRIVATE SCHOOLS AND THE PUBLIC GOOD : THE EFFECT OF PRIVATE EDUCATION ON POLITICAL PARTICIPATION AND TOLERANCE IN THE TEXAS POLL

The development of the public school system was prompted in part by the fear that private education would not adequately socialize students to the values required to function in a democratic system. Private, especially Catholic, schools were thought not to be well suited to instilling norms of participation and tolerance in the waves of immigrants an-iving from Ireland and Italy. Much of this anxiety was nothing more than thinly disguised antiCatholicism and xenophobia. Nevertheless, the belief that public schools are better at imparting desired civic values persists despite conscious efforts on the part of Catholic and other private schools to provide a quality civic education and despite a lack of an empirical basis for this belief. It is still widely held that public goals in civic education are best served by public schools, while private schools operate for the benefit of parochial interests. In this paper, the researchers test these hypotheses by examining the effect of public and private education on political participation and tolerance. Specifically, survey data are drawn from a representative sample of Texas residents to determine whether type of education infiuences political attitudes

T he development of the public school system was prompted in part by the fear that private education would not adequately socialize students to the values required to function in a democratic system.Private, especially Catholic, schools were thought not to be well suited to instilling norms of participation and tolerance in the waves of immigrants an-iving from Ireland and Italy.Much of this anxiety was nothing more than thinly disguised anti-Catholicism and xenophobia.Nevertheless, the belief that public schools are better at imparting desired civic values persists despite conscious efforts on the part of Catholic and other private schools to provide a quality civic education and despite a lack of an empirical basis for this belief.It is still widely held that public goals in civic education are best served by public schools, while private schools operate for the benefit of parochial interests.
In this paper, the researchers test these hypotheses by examining the effect of public and private education on political participation and tolerance.Specifically, survey data are drawn from a representative sample of Texas residents to determine whether type of education infiuences political attitudes and behaviors.The Texas poll, a state-wide survey administered annually to 1,000 Texas residents, contains a compendium of questions submitted by state agencies and academic researchers on a wide range of subjects.The survey results presented in this study were gathered in the fall of 1997.This analysis is limited to one state; therefore, caution should be exercised when inteipreting the results.Nonetheless, the findings are generally consistent with other research done in this area (Coleman, Hoffer, & Kilgore, 1982a;Glenn, 1988;Greene, 1998).In comparison to students who attended public school for all 12 years of primary education, those who spent some time in private school demonstrate greater levels of political participation and political tolerance, even after controlling for other factors.
Interestingly, attending private school for all 12 years of education does not seem to produce the same effect as attending for only some of the time.This leads us to hypothesize that there is a substantive difference, one we have not controlled for, between families who choose to send their children exclusively to private schools and those who choose to switch their children from public to private school.Further research is necessary to explain these differences more fully.This evidence, therefore, suggests that promoting civic values is of central concern for private school educators who are often more successful than their public school counterparts.The attention that Catholic and other private schools have paid to serving public ends appears to have produced successful results.

PROMOTING DEMOCRATIC VALUES: PUBLIC VERSUS PRIVATE SCHOOLS
The notion that government should be concerned with education has existed in this country since Thomas Jefferson and Noah Webster (Glenn, 1988).It was Horace Mann, however, who first articulated the idea of the "common school," a public school where children of different cultural and ethnic backgrounds would be taught the basic political values and virtues of U.S. citizenship.Mann's common school was in part a response to the growing numbers of Irish and Italian immigrants.Fearing that the new immigrants' Catholicism would undermine their loyalty to this country and its social and political establishments, many embraced the common school as an important socializing agent, one necessary for inculcating proper allegiances and respect for the values of participation and tolerance.Writing in the impassioned tone of the time, Boston officials in 1850 stressed the consequences for the state should they fail to educate children of foreign-bom parents: ...In our Schools they must receive moral and religious teaching, powerful enough if possible to keep them in the right path amid the moral darkness which is their daily and domestic walk.... Unless we can reclaim this popu-lation in their childhood by moral means, we must control them by force, or support them as paupers, at a maturer period of life.(Glenn, 1988, p. 84) Those who promoted the concept of the common school maintained that if loyalty to the state and acceptance of civic norms were paramount, then the government should naturally be responsible for operating the institutions of value transmission.This idea continues to fiourish.As Secretary of Education Richard Riley has argued.
The "common school"-the concept upon which our public school system was built-teaches children important lessons about both the commonality and diversity of American culture.These lessons are conveyed not only through what is taught in the classroom, but by the very experience of attending school with a diverse mix of students.(1997, p. 1) What is significant in many of the arguments about the superiority of public schools for promoting democratic values is that the source of their advantage cannot be pinpointed with precision.In Riley s words, it is the "v^/T experience of attending school with a diverse mix of students" [italics added].
Implicit in Riley's statement is the suggestion that pubhc schools are more integrated, thus offering a better replication of the polity and teaching through exposure to the values of diversity and tolerance necessary for democratic life.While it is certainly true that for democracy to fiourish in a heterogeneous society the integration of different groups of students is to be desired and promoted, there is very little evidence that public schools are successfully performing this function.In fact, research suggests that, nearly 50 years after the landmark Brown v. Board of Fducation decision, public schools are still highly segregated-and in fact increasingly so (Orfield, 1996).
Perhaps even more surprising, a growing body of research is finding that private schools might be better integrated and more egalitarian than their public school counterparts (Coleman et al., 1982a;Coleman, Hoffer, & Kilgore, 1982b;Greene, 1998;Greene & Mellow, 1998).Public school student bodies are drawn primarily from neighborhood attendance zones which typically replicate existing segregated housing patterns.Public schools, therefore, often simply mirror the distinct racial and class makeup of the area in which they are located.In their landmark comparison of public and private schools, Coleman, Hoffer, and Kilgore (1982b) describe the results of postwar segregated housing patterns by noting that: This stratification has in effect produced a "public" school system which not only no longer integrates the various segments of the population of students, but appears no more egalitarian than private education, and considerably less egalitarian in outcome than the major portion of the private sec-tor in America-the Catholic schools,(p. 196) Since private school attendance is a function of voluntary association, housing is a constraint on attendance only to the extent that transportation becomes a difficulty.For this reason private schools are likely to be equally if not better able to bring together diverse groups of students for the purpose of learning the values of mutual tolerance and respect for difference.
Some have argued, however, that public schools are better conduits for promoting democratic values because they are democratically governed.From this perspective, public schools teach students through example the desirability and efficacy of democracy (Gutmann, 1987).Collective goals and interests are articulated through the democratic process which governs schools with positive results.First, it ensures that schools adhere to the preferences and policies voiced by the majority and so most closely approximate the ideal of common interests in a democratic society.Second, it prepares students to embrace and practice democracy as they mature into tomorrow's citizens.Conversely, private schools are thought to promote only the narrow interests of their sponsoring group (Mann, 1957).Following this line of reasoning, and since these schools typically are not governed by democratic processes, the parochial interests imparted by the school go unchallenged.Moreover, because they govern according to authoritarian principles, private schools shut off even the demonstration of democracy, and this, it is feared, teaches students the values associated with authoritarian rule.
These arguments are theoretical and have yet to be supported by empirical evidence documenting the link between governance and students' political values.What evidence exists in this area suggests that the reverse may be true.Democratic governance has been shown to make public schools particularly unwieldy bureaucracies, a reality that stymies their teaching effectiveness (Chubb & Moe, 1990).It is possible that students who are exposed to the cumbersome nature of bureaucratic action are just as negatively impressed by this aspect of democratic politics as they may be positively impressed by the democracy that produced it.
Similarly, a growing amount of empirical research demonstrates that private schools may be better equipped to impart the values of democracy.For example, the ability to select the school that one's children will attend has been positively correlated with increased levels of parental involvement (Schneider, Teske, Marschall, Mintron, & Roch, 1997).Since private schools are voluntary associations that tend to be much less encumbered by large bureaucracies, it is likely that they more closely represent a true polis, with active community and parental involvement in the life of the school.In this regard, just as the typically smaller, more autonomous structure of private schools aids in their teaching effectiveness, it may also facilitate an informal democratic process.
Finally, there is evidence to refute directly the concern that private schools promote only parochial interests.Catholic schools, which constitute the majority of all private schools, have been found to devote significant attention to the teaching of political values of inclusiveness, individual responsibility, and tolerance for the purposes of encouraging a just and harmonious democracy (Bryk, Lee, & Holland, 1993;Greeley, 1982), while public schools have recently been criticized for their failure to teach adequately the desired civic values (Final Report of the National Commission on Civic Renewal, 1998).It is reasonable to speculate that because private schools offer an alternative education to the public school system, they make extra efforts to impart the types of political values that public schools are expected to teach.
Clearly, the assumption that private schools cannot or will not promote the necessary public virtues is a matter of theoretical dispute that lacks empirical support.Nor is there evidence for the proposition that public schools successfully convey democratic values.Nonetheless, public schools have been widely extolled as "institutions where we learn what it means to be a public and start down the road to common national and civic identity" (Barber, 1997, p. 1).As this quote demonstrates, public schools are sometimes credited with near-mythical ability to forge one from many, yet the mechanisms by which they are able to do so are not clearly indicated.The results of our empirical study contribute to this growing body of contradictory evidence about the advantages of public over private education at promoting civic values.

THE TEXAS POLL: VARIABLES IN THE ANALYSIS
The Texas Poll is an annual survey conducted by the University of Texas.In 1997, the year in which the data were collected, respondents were asked a host of demographic questions as well as a wide range of questions about their political knowledge, interests, attitudes, and behaviors.One thousand people from across the state of Texas participated in the survey.Table 1 presents descriptive statistics, the mean, standard deviation, minimum and maximum value, and N for all of the variables used.
The first dependent variable examined was political participation.Subjects were asked whether they had registered to vote.While this question did not indicate whether a respondent had participated directly in the democratic process (through voting in an election, working for a campaign, etc.), this was a good indicator of a general interest in participation.First, it is a necessary prerequisite to voting, and those who register to vote have indicated some interest and intention to participate in democratic governance.Second, it is a broadly inclusive indicator of a respondent's attitude about participation.
The second dependent variable considered was political tolerance.To construct this measure, the researchers combined survey responses from several different questions that are commonly used to measure this concept (Stouffer, 1955;Sullivan, Peirson, & Marcus, 1979).Respondents were first asked to identify their least liked group from a list that was provided.Groups included the Nazi party, the Ku Klux Klan, gay and lesbian groups, atheist organizations, and environmentalists.They were then asked whether their least liked group should be allowed to hold a public rally in their city.Answers ranged from 1 to 5, with a 1 indicating that they strongly disagreed and a 5 indicating that they strongly agreed that the group should be allowed to rally.Finally, respondents were read one of several statements designed to make them reconsider their original answer.For example, if they indicated that they felt the group should not be allowed to hold a rally, they were read the following statement, "Suppose someone said it would not be fair to allow some groups to demonstrate while denying the right to others.Would you still oppose the rally or would you change your opinion and support the rally being allowed to take place?"A similar type of statement designed to produce reconsideration was read to those originally in favor of allowing the group to hold a rally.
From these questions, we constructed a variable with a nine-point scale.On this scale, a 1 indicated that the respondent strongly disagreed that the group should be allowed to hold a rally and refused to change his or her mind, while a 9 indicated that the respondent strongly agreed that the group should be allowed to hold a rally and refused to change his or her mind.The more a person either is willing to be persuaded out of opposing the rally or is staunch in his or her continued support of the group's right to hold a rally, the more tolerant that person is considered.
The independent variables collected in the Texas poll allow us to differentiate among three types of educational experience.Respondents who attended public school for all of their primary and secondary education were compared to two groups of private school attendees: those who attended only private school for their entire primary and secondary education and those who attended private school for part of their primary and secondary education.
Because it could be argued that the effect of private education on respondents' political values is attributable to other factors, we controlled for a variety of background characteristics that might be associated with private school attendance and also with the two dependent variables of political participation and tolerance.For example, we controlled for the respondent's gender and age.Gender is important because men and women do not necessarily have the same educational opportunities, nor do they necessarily have similar political behaviors.Because older people tend to be more politically active and have been out of school for a longer time, controlling for the possible effect of age is also important.
Similarly, because groups can have different educational opportunities and different political experiences, we controlled for respondents' race and ethnicity with dichotomized variables for African-American, Asian, Hispanic, and other ethnicity.In our analysis.White was a default category against which the other groups were compared.Place of residence affects people's access to private education, and it also shapes political influences.For this reason, we controlled for non-urban versus urban dwelling, and we also controlled for whether the respondent lived in one of the two major metropolitan areas in Texas, Dallas-Fort Worth and Houston.
We also controlled for respondents' family income and the number of years they had lived at their place of residence.Both are measures of socioeconomic status in that low-income families tend to move at higher rates than high-income families.These variables are potentially problematic in that higher income and lower mobility may partially be the product of private education if it is true that private schools tend to cause better educational outcomes (Chubb & Moe 1990;Coleman, et al., 1982aColeman, et al., , 1982b;;Greene, Peterson, & Du, 1998;Hoxby, 1998;Neal, 1997).Whether private schools do, in fact, tend to produce better educational outcomes and therefore higher socioeconomic status later in life, however, is a matter of dispute (Cookson, 1994;Levin, 1998;Smith & Meier, 1995).If private education does contribute to academic and then financial success, then controlling for these two variables may partially control for, and depress, the estimated effect of private education.
The final control variables used in analyses of both dependent variables were a series of dichotomous measures of religion; i.e., whether the respondent identified himself or herself as Catholic, Baptist, traditional Protestant (e.g.. Episcopalian, Congregationalist), or other Protestant.These categories represented the primary religious affiliations of all respondents in the survey.Controlling for the effects of religious identification is important because people of different religious groups may have different attitudes about politics at the outset.Isolating the effect of the type of education from the attitudes which subjects' religious affiliation may have predisposed them to hold measures results produced directly and independently by private schooling.
Standard political behavior models often include a number of additional controls lor items such as respondents' ideological leaning or party identification.We chose not to control for these, because while they may infiuence tolerance and participation, they are also likely to be outcomes of educational experience.Controlling for these items would therefore potentially bias our results.At the same time, these items are not likely to have affected the type of education the respondent received as a child; since the effect does not predate the effect of the primary independent variable with which we are concerned, omitting these types of controls should not pose a problem.We believe that we have controlled for most important factors which are related both to type of education (our primary independent variable) and to participation and tolerance (our dependent variables).Thus we can be reasonably confident of the estimated effect produced by our analyses.
While all of the independent variables were included in both analyses, two additional variables were used in the tolerance model.Respondents were asked to rate, on a scale of 0 to 100, how threatening they felt their least-liked group was, with a 0 representing not threatening and 100 representing maximally threatening.This variable is important because a respondent who is willing to let an opposed group hold a rally yet believes that the group is harmless is not necessarily exhibiting the same level of tolerance as someone who is willing to let an opposed group hold a rally and finds that group to be extremely threatening.(Whether we, as a society, desire that degree of tolerance is a different question.)The second variable added to the tolerance model is a feeling "thermometer" which measures how strongly respondents feel about their least liked group, again, on a scale of 0 {maximally opposed) to 100 (maximally favorable).Similar to the threat variable, the thermometer is designed to capture intensity of feeling, which is important because people who feel neutrally about other groups may not have the same political response as people who feel strongly opposed to other groups.

RESULTS: POLITICAL PARTICIPATION
People who received some of their education in a private school setting are more likely to be registered to vote than those who received all of their primary and secondary education in public schools (see Table 2).Even controlling for other factors, the effect is significant.To illustrate the significance of this relationship, we generated predicted percentages that are registered to vote from our logit model of those who went to some private school and those who went only to public school.These predicted percentages are computed from the logit results by setting the value of all independent variables (besides the primary independent variables) to their means.
When compared to all public school attendance, some private schooling increases the likelihood of participation by 9%.Specifically, we would expect 84.8% of those who attended only public schools to register to vote.In comparison, 93.8% of those who had some years of private schooling are expected to register to vote.Both voter registration percentages appear high, which suggests that there is some degree of over-reporting among respondents (Shaw, de la Garza, & Lee, 1998).However, there is no reason to believe that this over-reporting is systematically biased toward one type of education over the other, and thus it is safe to assume that the difference observed between those who attend public school and those who attend some private school still holds.In other words, even taking into account some degree of over-reporting, private schooling still has a positive and significant effect on political participation.
Interestingly, however, the beneficial effects do not hold for people who received all of their education in private schools.These people are not significantly different from those who spent all of their primary and secondary years in public schools.In fact, while it is not a significant difference, attending only private schools has a negative estimated effect on voting registration.Again, this effect is not significantly different from zero impact, and so the negative effect observed should not be interpreted as anything more than the result of chance.While additional research (including a larger sample size) is necessary to firmly establish the causal linkages, it is clear that there is some substantive factor that differentiates those who attend only private schools and those who attend a mix of public and private schools.We will return to this point in greater detail in the conclusion of this essay.The effects of the other independent variables were consistent with what we would generally expect.Older citizens are more likely to participate than younger citizens.People with higher socioeconomic backgrounds, those whose families had higher incomes, and those who have stayed in one residence for longer periods of time are more likely to be registered to vote.These effects are positive and significant and are completely consistent with general political behavior research, which finds a high degree of correlation between socioeconomic status and participation and between age and participation (Verba, Schlozman, & Brady, 1995).Also consistent with the litera-ture in this area.Latinos and Asians are significantly less likely to register to vote when compared to White citizens.African-Americans are politically indistinguishable from Whites in this regard once income is controlled.
With regard to participation, there appear to be no differences between urbanites and non-urbanites.It is possible that some of these effects are absorbed by other variables such as race and income; Dallas and Houston (and Texas's other urban centers in general) tend to have greater concentrations of poor people and of African-Americans and Hispanics, and so controlling for these items may negate the independent effects of urbanicity.Gender also does not seem to matter in terms of voting registration.
Perhaps most interesting, religion appears to have no independent effect on participation.None of the categories in our analysis (Catholic, Baptist, traditional Protestant, and other Protestant) appear significantly different from each other or from our default category which includes non-religious and other religions (Jews, Muslims, etc.).This is important in that it suggests that the effect observed among those who attended some private school is not an artifact of their religious training.Catholic schools, for example, may encourage political participation beyond the level found in Catholics educated in public schools.

RESULTS: TOLERANCE
As with participation, attending some private school has a strongly positive effect on tolerance.In an Ordinary Least Squares (OLS) analysis of the ninepoint tolerance scale, receiving some private school education increases tolerance by .604when compared to attending only public school (see Table 3).This represents an increase of .23 standard deviations on the tolerance scale.To put the magnitude of this benefit in perspective, we would expect that someone who started out more tolerant than 50% of the population would become more tolerant than 59% of the population if they had attended some private school.The strength of this effect is all the more compelling given that a broad range of factors was controlled.However, as with participation, attending only private school does not seem to have a significantly different effect on tolerance than attending exclusively public school.
Only a few of our control variables appear to have a significant effect on tolerance.Higher income is associated with a greater degree of tolerance, as is being male.Other Protestant groups are also more likely to be tolerant than other religious and non-religious groups.Because the data do not specify which denominational affiliations fall into this category, it is difficult to interpret this result.Neither age nor years in residence has a significant effect on tolerance.
Finally, the thermometer and threat ratings are both significantly correlated with degree of tolerance.Not surprisingly, the more favorably (or the

CONCLUSION
The evidence presented suggests that private school education can contribute to the development of such key democratic values as participation and tolerance.Moreover, because we were able to control for a number of factors which are often associated with type of education and political behaviors or attitudes, we can be reasonably confident that the observed effects of private education are not spurious findings.Controlling for religion, for example, allows us to differentiate between the effect produced by schooling which may take place in a religiously affiliated school and that produced by religious affiliation in general.This is also true for the income and socioeconomic measures; by controlling for these items, we can be fairly confident that we are not misinterpreting the effect of any possible self-selection by higher income families into private education for that of private education itself.These findings fly in the face of many conventional attitudes about public school's superiority and add to the growing body of research on the beneficial effects of private schooling.What is not clear from our research, however, is which attributes of private schools are responsible for their greater degree of effectiveness in promoting desired political values.Moreover, this study applies only to the state of Texas and therefore raises questions of external validity.This shortcoming notwithstanding, it is reassuring that the relationships between schooling and political values identified are consistent with the results of research conducted on a nationwide basis and among specific subsets of the population (Greene, 1998;Greene, Peterson, & Du, 1998).Further research is necessary in order to extend the results to other areas of the country and to gain a more precise sense of the mechanisms by which private schools better promote democratic values.
More research is also necessary to understand what differentiates those who attend a mix of public and private schools from those who attend private schools exclusively.For now, the researchers can only hypothesize about what clearly is a significant substantive difference.It is likely that those whose families chose to send them to private school for all 12 years did so for clear and purposeful reasons.They may have an ideological opposition to the type of educational experience they believe that the government-run public schools provide, for example.An entire private school education may sometimes not be as much an indicator of the educational preferences of the family as an indication of the political rejection of the civic values of participation and tolerance generally attributed to public schooling.In this case, it is possible that families that choose exclusively private over public schooling, as a principle, are interested in exposing their children to a different value system-one at odds perhaps with what is traditionally assumed to be offered by government-run common schools.The irony is that while 12 years of private schooling may represent a rejection of socially desired political values, significantly better civic outcomes are not produced by 12 years of public education.
In contrast, those who went to private school for a number of years, but not exclusively, exhibit a greater commitment to democratic values than either their purely public or purely private school counterparts.One possible interpretation of this is that these individuals grew up in households that were generally in favor of the ideals associated with public schools, but spent some years in private schools in response to specific circumstances.This type of private school experience likely does not represent an outright rejection of the public school system and the political values typically associated with government-provided education.Rather, it may be an isolated response to a perceived problem.Whatever the circumstances prompting the change between public and private schools, the partial exposure to private schooling appears to be associated with promoting democratic values.
While additional research is necessary to understand more fully the differences observed between all and some private schooling, this paper has presented important new evidence to contradict the assumption that private schools are unconcerned with civic education.Along with a growing body of literature in this area, we have found that private schooling can significantly increase public commitment to such democratic values as political participation and tolerance.Understanding how private schools are able to be more effective in this than public schools is the next step.More importantly, however, it is time to stop assuming the superiority of public schools for promoting civic values and begin to test these assumptions with systematic empirical research.