Findings

Student and Household Characteristics by Public and Private School Types
Data from the National Household Education Surveys Program (NHES) show that the percentage of children enrolled in public, assigned schools for 1st through 12th grades decreased from 80 percent in 1993 to 76 percent in 1999. The decrease in public, assigned school enrollment was almost completely offset by an increase from 11 to 14 percent in public, chosen school enrollment. Enrollment in private, church-related schools remained relatively stable at 7 or 8 percent between 1993 and 1999 and enrollment in private, not church-related schools stayed at 2 percent. Most of the overall change in enrollment took place between 1993 and 1996 (Figure 1).

The general patterns and trends in enrollment in the four types of schools studied here were broadly reflected across a wide range of students from many different backgrounds. Table 1 and Table 1 Continued shows the estimates and standard errors for these data.¹⁷ The following discussion considers differences in the type of school attended by students’ grade, sex, race and ethnicity, disability status, household income, parents’ education, family structure, region, and urbanicity of students’ residence.

Students' grade and sex
In 1993, there were few differences in enrollment in types of schools among grade groups. However, by 1996 and 1999 there were differences in enrollment. In 1996 and 1999, a higher percentage of students in grades 6 to 8 attended public, assigned schools than did students in grades 1 to 5 while a higher percentage of students in grades 1 to 5 and grades 9 to 12 attended public, chosen schools than did students in grades 6 to 8. In the private school types, in all survey years, students in grades 1 to 5 were more likely to attend private, church-related schools than were students in the upper grades, 9 to 12. There were no differences detected among grade groups and years for students in private, not church-related schools.

Overall, girls and boys followed the same patterns of enrollment in school types as the overall population with a decrease in public, assigned school enrollment over the survey years and an increase in public, chosen school enrollment. Slightly more girls attended private, church-related schools in 1993 (8 percent vs. 7 percent) and slightly more boys attended private, not church-related schools in 1996 (3 percent vs. 2 percent).

Race and ethnicity
In each survey year, a smaller percentage of Black students than White students attended public, assigned schools. In 1999, the rate of assigned school enrollment for Black students (71 percent) was 6 percentage points lower than the rate of enrollment for White or Hispanic students (77 percent each). In each year, Black students had a higher rate of enrollment in public, chosen schools than did Hispanic students, and Black and Hispanic students had higher enrollment rates in public, chosen schools than White students. Black students were enrolled in public, chosen schools at a rate of 19 percent in 1993 and 23 percent in 1999 while Hispanic students were enrolled at a rate of 14 percent in 1993 and 18 percent in 1999. Nine and 11 percent of White students were enrolled in public, chosen schools in 1993 and 1999, respectively (Table 1 and Table 1 Continued). A larger percentage of White students attended each type of private school than did Black or Hispanic students in all three years.

Disability status
In 1999, about 75 percent of students with a disability attended public, assigned schools, 16 percent attended public, chosen schools, 6 percent attended private, church-related schools, and 2 percent attended private, not church-related schools. The percentage of students with disabilities attending private, church-related schools (6 percent) was two percentage points lower than the percentage of students with no disabilities attending private, church-related schools (8 percent). There were no differences detected in public, assigned and public, chosen school enrollment or in private, not church-related school enrollment for students with and without disabilities. Comparable data on students with disabilities are not available for 1993 and 1996, therefore it is not possible to report on a trend in the enrollment of these children.

Household income
The most noticeable trend in public school enrollment was the movement away from public, assigned school enrollment and toward public, chosen school enrollment by the lowestincome students (see Figure 2). In 1993, a higher percentage of students living in households with incomes of $10,000 or less attended public, assigned schools (83 percent), than did students living in households with incomes of more than $75,000 (72 percent). Differences were rare or not detected in the middle-income categories, compared to the extremes in each year, for students in public, assigned schools. By 1999, there was no detectable difference in the percentage of students enrolled in public, assigned schools between the $10,000 or less and more than $75,000 income ranges.¹⁸ The decrease in public, assigned school attendance for students living in households with incomes of less than $10,000 was mirrored by an increase in the proportion of these students attending public, chosen schools (14 percent in 1993 to 22 percent in 1999). Note that the categories for household income are in current dollars, which have not been adjusted for inflation.¹⁹

The rate of enrollment in private schools stayed the same for most income groups across the six-year period. Students in households with incomes of more than $75,000 had a higher enrollment rate in private schools in 1993, 1996, and 1999 than did students in other income groups. Students in the highest income group (more than $75,000) were enrolled in private, church-related schools at about five times the rate of enrollment for students in the lowest income group ($10,000 or less) in 1993 and 1999. Comparing the highest income groups and middle income groups, 14 percent of students living in households with an income of more than $75,000 were enrolled in private, church-related schools in 1993 and 1999, compared to eight or nine percent of students living in households with an income of $35,000 to $50,000. Enrollment in private, not church-related schools had a pattern of enrollment similar to private, church-related schools between the highest and lowest income groups and the highest and middle income groups, but there were no differences detected between the lowest and middle income groups.

Parents' highest level of education
Parents who are more educated are also more likely than other parents to seek out different educational opportunities and choices for their children (Nord, Brimhall & West 1997). As expected, in each survey year, students’ whose parents highest education level was a bachelor’s degree or a graduate or professional degree had a higher rate of enrollment in private schools, both church-related and not church-related, than students whose parents obtained at most a high school diploma, GED or less.²⁰ Unlike income, differences in public, assigned school enrollment among different education levels persisted from 1993 to 1999. Students whose parents were relatively well educated were less likely to be enrolled in a public, assigned school than students whose parents were relatively less well educated. The difference between the highest and lowest education levels in 1993 in the percentage of students enrolled in public, assigned schools (73 vs. 84 percent) was still present in 1999 (68 vs. 80 percent).

During the three-year period between 1993 and 1996, the percentage of students enrolled in public, assigned schools decreased and the percentage enrolled in public, chosen schools increased only for students whose parents attained an education beyond a high school diploma or GED. By 1999, students whose parents’ highest level of education was a high school diploma or GED also had a decreased rate of public, assigned school enrollment and an increase in public, chosen school enrollment. There was no change detected in the rate of enrollment in public, assigned schools across the 3 time points for students whose parents had less than a high school diploma.²¹

Family structure
Students who lived in households with non-parent guardians (e.g., grandparents or godparents) experienced a decrease of 11 percentage points in public, assigned school attendance between 1993 and 1999 and an 8 point increase in public, chosen school attendance. By 1999, 22 percent of students living in households with non-parent guardians attended public, chosen schools—10 percentage points higher than students living in two-parent households. In addition, a greater percentage of students living in one-parent households attended public, chosen schools (18 percent) than did students living in two-parent households (12 percent) in 1999. The pattern of attendance for students living in households with non-parent guardians is expected because many students living with non-parent guardians were Black, non-Hispanic students, lived in urban areas, and had low household incomes (data not shown in tables). Data presented in this report show that Black students, students living in urban areas, and students living in families with a low household income were more likely than other students to attend public, chosen schools. Given the relationships found in the data, it is likely that the relationship between family structure and public, chosen school attendance can be explained, at least in part, by the influence of these other common factors—race/ethnicity, urbanicity, and household income. A regression analysis of the 1999 data shows that the relationship between students living with non-parent guardians and public, chosen school attendance is not significant once other factors (urbanicity, income, and race) that are associated with attending public, chosen schools are controlled.

Students from two-parent families were less likely to be in public, chosen schools than were other students. However, students living in two-parent households were more likely to be enrolled in private, church-related schools than students with one-parent households in all three of the survey years. For example, 8 percent of children with two-parent households attended private, church-related schools, compared to 5 percent of children with one-parent households in 1999.

Region and Urbanicity
Overall, the data suggest that in areas of the country which have been shown to have greater availability of school choice in the public school system, namely in the western region of the country and in urban areas (U.S. Department of Education 1996a), there is greater use. In 1999, students living in the West had a higher rate of enrollment in public, chosen schools (18 percent) than students living in any other region (13 to 14 percent). Students living in the Northeast and Midwest had a higher rate of enrollment in private, church-related schools than students living in the South or West regions. Although the percentage of students enrolled in public, chosen schools was higher in the West in 1999, the percentage of students enrolled in public schools of choice increased for students in all regions from 1993 to 1999. In general, enrollment rates for private, church-related and private, not church-related schools remained relatively stable within each of the regions from 1993 through 1999.

Students living inside urban areas were enrolled in public, chosen and private, church-related schools at a higher rate than students living just outside urban areas or in rural areas in all three survey years.²² Between 1993 and 1999, the rate of enrollment in public, assigned schools decreased (four and five percentage points, respectively) and the rate of enrollment in public, chosen schools increased (three and four percentage points respectively) for students living inside and just outside urban areas. While the rate of enrollment in public, chosen schools for students in rural areas did increase between 1993 and 1999, there was no detectable decrease in their public, assigned enrollment rate. Approximately 85 percent of students in rural areas continued to be enrolled in public, assigned schools in 1999.

Characteristics of Homeschooled Children
Some parents turn to homeschooling as an alternative to sending their children to public or private schools. In 1999, approximately 850,000 students, ages 5 to 17 with a grade equivalent of kindergarten to grade 12, were homeschooled either full or part-time—accounting for 1.7 percent of the school-age population (Bielick, Chandler and Broughman 2001). This section uses data from the 1999 Parent Survey to compare the characteristics of students who were primarily homeschooled to the characteristics of students who attended public or private schools in 1999.²³ The population analyzed in this section on homeschooling differs from the population in the previous sections of the report. Because questions were only asked about homeschooled children ages 5 to 17 years old, the analysis in this section is restricted to just those 1st through 12th graders who are 5 to 17 years old.²⁴

Table 2 and Table 2 Continued shows the percentage of students with various demographic characteristics within the four types of public and private schools and homeschoolers. The similarities and differences among students in the four types of public and private schools described in the previous section are also reflected in the estimates in Table 2 and Table 2 Continued. The estimates in the table suggest that homeschoolers differ from students in public and private schools on a number of characteristics.

Homeschoolers differed from students in public schools in that their parents tended to be better educated. Homeschoolers were more likely to be White and to live in two-parent households than were students in public, assigned or public, chosen schools. With the exception of two income groups ($10,000 or less for public, assigned and chosen, and $50,001-75,000 for public, chosen), no other statistically significant differences in household income were detected between homeschoolers and students in public schools.²⁵ Geographically, homeschooled students were more likely to reside in the South and in rural areas than students in public, chosen schools and were less likely to reside in the Midwest than students in public, assigned schools.

Homeschoolers differed from private school students in fewer ways than they differed from public school students. Homeschoolers were less likely than private school students to live in households with annual incomes over $75,000. They were also less likely to live in the Northeast and inside urban areas and more likely to live in rural areas compared to private school students. Looking at the different types of private schools, homeschoolers were more likely than were students in private, church-related schools to have a disability.

Parent Involvement and Satisfaction, and Student Plans for Postsecondary Education
Results prior to this section showed the relationships between key student and family characteristics and the use of school choice. This section shows how the kind of school children attended relates to several important behavior and attitudinal variables. In particular, the relationships between the type of school students attended, and their parents’ involvement in and satisfaction with the students’ schools are examined. The analysis focuses on three questions: 1) Are students in schools of choice more likely to have involved parents than students in assigned schools? 2) Are students in schools of choice more likely to have parents who are satisfied with their children’s schools than students in assigned schools? 3) Do parents of students in schools of choice have greater expectations for their children’s postsecondary education than parents of students in other schools?

The population considered in this section differs from the population in the previous sections of the report. Because questions about parent satisfaction and involvement were only asked for students in 3rd through 12th grade in 1993, the analysis in this section is limited to just those students for each survey year. Parents of homeschoolers were not asked about their level of satisfaction nor their involvement in public or private schools.²⁶

Parent satisfaction Parents who chose the schools their children attended were more satisfied with the schools their children attended than were other parents. In both 1993 and 1999, higher percentages of parents of students in grades 3 to 12 who attended private school (both church-related and not church-related) were very satisfied with their children’s schools, teachers, academic standards, and order and discipline than parents whose children attended public school (both assigned and chosen).²⁷ Parents of children in public, chosen schools were more satisfied than were parents of children in public, assigned schools, but were less satisfied than private school parents. Looking only at the 1999 data, parents of students who attended public, chosen schools were more likely to be very satisfied with the various aspects of their children’s schools than were parents of students who attended public, assigned schools (Table 3, Figure 3). Roughly 8 percent more students in public, chosen schools had parents who were very satisfied with the teachers at their school than students in public, assigned schools. This difference widened to 14 percentage points for parent evaluations of the students’ schools in general.

Parent Involvement
Not only did students in private schools have the most satisfied parents, they were also more likely than other students to have parents who were involved with their schools (Tables 4a and 4b). Approximately 93 percent of parents of students in private, church-related and 87 percent of parents of students in private, not church-related schools said they attended a general meeting at the school in 1999 compared to 75 percent of parents of students in public, assigned and public, chosen schools. Fewer parents of students in each school type reported that they went to a parent-teacher conference. More parents of students in private schools, both church-related and not church-related reported attending a school event (84 and 76 percent) in 1999 than did parents of students in public, assigned or chosen schools (63 and 59 percent). Parents of children in private, church-related schools were twice as likely as parents of students in public, assigned or chosen schools to volunteer or serve on a committee (65 percent versus 30 percent and 31 percent, respectively). They were also more likely to volunteer or serve on a committee than were parents in not church-related private schools (53 percent). While parents of students in public, chosen schools were more satisfied with their children’s schools than parents of students in assigned schools, they were generally not more involved.

Plans for postsecondary education
One key measure of academic success given the importance of a college education in the current economy is whether students attain additional education after high school. While the NHES does not measure actual attendance, it provides data on an important predictor of a student’s probability of obtaining education after high school—parents’ expectations for their children’s educational attainment (Ingels et al. 2002). Parents of at least nine out of ten 6th through 12th grade students thought their children would attend postsecondary education after high school regardless of the type of school their child attends. As the 1999 data show, the percentage of students attending private, church-related schools whose parents thought they would attend school after high school was 5 and 6 percentage points higher than public, assigned and public, chosen school students. The percentage of students attending private, church-related schools whose parents thought they would graduate from a 4-year college was 10 percentage points higher than for public school students (Tables 5a and 5b). Similar differences were not detected for parent expectations between public, assigned and public, chosen school students or between students in private, not church-related schools and other schools.²⁸

Multivariate Analysis of Parent Satisfaction and Level of Involvement
The previous analyses showed that several household, parent, and student characteristics are related to the type of school students attend and that school type is related to parents’ satisfaction and involvement. Because these household, parent, and student characteristics are related to the type of school attended by the child, it is possible that these characteristics are what actually drive the relationship between school-type on the one hand, and parent satisfaction and involvement on the other, or both. Although a complete analysis of parent satisfaction and involvement is beyond the scope of this report, a multivariate analysis using data from 1999 examines whether the relationships between school type and parent satisfaction and involvement remain when other household, parent, and student characteristics are controlled. In addition, the multivariate analysis considers the effects of parent involvement on parent satisfaction and the effects of parent satisfaction on parent involvement to illustrate how the two variables are related.

The multivariate analysis was conducted using ordinary least squares regression analysis. The results of the analysis are shown in Tables 6 and 7. Table 6 shows the results for parent satisfaction and Table 7 shows the results for parent involvement. In this analysis, parent satisfaction is the simple average of each of the 4 satisfaction items. Similarly, level of involvement is a sum of the number of activities parents’ reported they were involved in. Readers should refer to the Survey Methodology and Technical Notes section for detailed information about the regression analysis procedure.

In each table, the multivariate analysis includes three analyses. Analysis 1 shows the bivariate relationship between school type and the level of parents’ satisfaction (Table 6) and involvement (Table 7) with their children’s schools. Analysis 2 shows the relationships between several household, student, and parent characteristics and the level of parents’ satisfaction and involvement with their children’s schools. The third analysis contains the school type variables and the household, student, and parent variables and shows how school type relates to parents’ satisfaction or involvement after controlling for household, student, and parent characteristics. The results of the regression analysis are summarized below.

Parent satisfaction
The regression results indicate that a number of household, parent, and student characteristics are related to parent satisfaction (Table 6). The results in analysis 3 show that the relationship between school type and parent satisfaction remained significant even when household, parent, and student characteristics were considered. Parents who chose the schools their children attended, both public and private, were more satisfied than parents whose children attended public, assigned schools. Students in private schools tended to have parents who were more satisfied than students in public schools irrespective of public school type. Those parents who were most involved with their children’s schools tended to be more satisfied with them.

Aside from school type, the results show that students’ sex, race/ethnicity, grade, and disability status were associated with parental satisfaction. Parents of girls were more satisfied than the parents of boys and parents of White students were more satisfied than those of Black students or other, non-Hispanic students. In addition, the parents of elementary school students (grades 3 to 5) were more satisfied than parents of high school students (grades 9-12) and parents of students with no disability were more satisfied than parents of students with a disability.

Certain characteristics of these students’ parents were also related to parental satisfaction. Even when controlling for various student, household, and school characteristics, parents with a high school education or higher tended to be less satisfied than parents with less than a high school education. Parents whose primary home-language was English were less satisfied than were other parents.²⁹ Apart from school type, school size as reported by parents also seems to be related to parental satisfaction. Generally speaking, students in schools with less than 1,000 pupils had parents who were more satisfied than students in schools of 1,000 or more.³⁰

Parent involvement
The multivariate analysis 1 in Table 7 supports earlier statements that parents whose children attended private schools were more involved than parents whose children attended public, assigned schools or public, chosen schools.³¹ However, the relationship between private school type and parent involvement was weakened when other household, parent, and student characteristics were considered. Parents who were more satisfied with their children’s schools tended to be more involved with their children’s schools.

Apart from the type of school the student attended, other student, parent and family, household, and school characteristics were related to parental involvement. Looking at the characteristics of the 3rd through 12th grade students, only one student characteristic seemed to be related to parent involvement and that was grade of enrollment. Students in grades 6-8 were more likely to have involved parents than students in grades 9-12 and students in grades 3-5 were the most likely to have involved parents. Parents of students with a race or ethnicity other than White, Black, or Hispanic were less involved compared to parents of White students.

Considering students’ parents and families, students whose parents have completed at least high school and speak English at home and students who live in two-parent families have parents who were more involved than students with less educated parents, students with non-English speaking parents, or students who have family structures that do not include two parents. Looking at household characteristics, students from households making over $20,000 and from households in rural areas are more likely to have involved parents than students from poorer households and households inside urban areas.³² In addition to the type of school the child attends, the size of the school as reported by parents also seems to be related to parental involvement. Students attending schools with fewer than 1,000 students tend to have more involved parents.³³

______________________________
Notes:

¹⁷ For additional details about this population, please see table A1 in appendix A.
¹⁸ The data also show a slight increase in enrollment in private, not church-related schools among students in the lowest income category between 1993 and 1996. While this might have occurred, since the data reported here are based on parents? reports of their children?s school enrollment and the change is very small (2 percentage points), the increase in low-income enrollment in private, not church- related schools may be an artifact of reporting error. A 1999 field test of this item showed that with the increase in the number of magnet and charter schools, some parents have difficulty distinguishing between private, not church-related schools and public, chosen schools because they consider public magnet or charter schools to be private schools.

¹⁹ Income data are categorical and thus were not adjusted for inflation. Independent analyses not shown here indicate that the patterns found for unadjusted income are the same as those found using a measure of poverty, which adjusts for inflation.

²⁰ Statistically, there was one exception, in 1996 there were no differences detected between parents with a bachelor?s degree and parents with less than a high school degree.

²¹ Apparent differences for students whose parents had less than a high school diploma were not statistically significant, possibly because of the relatively large standard errors for those estimates.

²² The definitions for these categories are taken directly from the 1990 Census of Population. An urban area comprises a place and the adjacent densely settled surrounding territory that together have a minimum population of 50,000 people. Households are considered to be urban if they are located inside this urban area?referred to in this report as inside urban area? or outside but adjacent to the urban area where the population is at minimum 2,500 people?referred to in this report as outside urban area. Places not classified as inside or outside urban areas are rural.

²³ For this analysis, students who were primarily homeschooled were defined as students who were homeschooled and who did not attend public or private schools for more than 8 hours per week. Students who were homeschooled and who also attended public or private schools for 9 or more hours per week were considered to be students in public or private schools.

²⁴ For additional details about this population and how it compares to the full 1st ? grade 12 population, please see table A3 in appendix A.

²⁵ Note that standard errors for many of the homeschool estimates are relatively large. Results for these estimates should be interpreted with caution.

²⁶ Students who were homeschooled and who also attended public or private schools for 9 or more hours per week were considered to be students in public or private schools in 1999. Data about these part-time homeschoolers is not available for 1993 or 1996.

²⁷ In 1993, only parents of children in grades 3 to 12 were asked questions about their satisfaction with their children?s schools.
²⁸ Apparent differences between students in private not church-related schools and other schools were not statistically significant, possibly because of the large standard errors associated with the estimates for students attending private, not church-related schools.

²⁹ Parents? English usage may also be a factor that influences the level of parent involvement in schools in the United States. It is included in the multivariate analysis to help explain the relationship between Hispanic ethnicity and parents? satisfaction and involvement with schools found in previous research (De La Luz Reynoso & Tidwill 1996).

³⁰ School size is measured by parents? reports of the number of students who attended their child?s school. Parents were asked to report a broad category of school size. On average, parents somewhat underestimated the size of the child?s school when compared to administrative records. Additionally, a separate analysis of highest and lowest grades in children?s schools was undertaken to determine if the effects of school size were confounded by a correlation between school size and the grade levels at children?s schools. There were no meaningful changes in the regression coefficients for school size as a result of this analysis.

³¹ An OLS regression, not shown, using ?public, chosen? schools as the comparison category provided the result that compares the private schools to a public, chosen schools.

³² Statistically, no differences were detected between students in households with incomes of $10,000 or less and households with incomes of $10,001 and $20,000.

³³ A separate analysis of highest and lowest grades in children?s schools was undertaken to determine if the effects of school size were confounded by a correlation between school size and the grade levels at children?s schools. There were no meaningful changes in the regression coefficients for school size as a result of this analysis.