The Desegregation Efficiency of Magnet Schools Tables

Table 1 a Prevalence of Desegregation Plan Type Among Districts with Plans and Prevalance of Plans in U.S., 1991 PLAN TYPE b Voluntary M to M Magnet-Voluntary Districts > 5,000 With Plans 4% 8% (31) (37) 23% 21% Districts > 27,750 With Plans 5% (5) 29% (176) (98) (26) 27% 29% 34% (207) (135) (31) 17% 28% (130) (130) 52% 37% (399) (172) (11) ALL MANDATORY 69% (529) 65% (302) 61% (56) Controlled Choice 4% 6% 5% (31) (28) (5) 100% (767) 100% (465) 100% (91) ALL DISTRICTS >5,000 IN SIZE > 27, 500 IN SIZE DISTRICTS WITH c PLANS IN U.S. 12% (767) 29% (465) 59% (91) TOTAL DISTRICTS IN U.S. W/>1 SCHOOL AT A SCHOOL LEVEL* 6, 392 1,604 155 ALL VOLUNTARY Magnet-Mandatory Mandatory--No Magnets TOTAL a All Districts With Plans 49% (45) 12% Plans are weighted by their district weights to show the prevalence in the U.S. The percentages for plan type are calculated by dividing the frequency for a plan type by the total number of districts with a plan in that column. c The percentage having any plan is calculated by dividing the number of districts with plans by the total number of districts in the U.S. in that column. b Table 2 Relationship Between Change in Interracial Exposure 1968-1991 and School District and Desegregation Plan Characteristics Districts >5,000 in 1991 Mean b Beta Change in Interracial Exposure 1968-1991 0.001 Districts >27,500 in 1991 Mean b Beta 0.066 DEMOGRAPHIC VARIABLES % White 1968 82.188 -0.003 * -0.24 72.647 -0.003 * -0.21 South 0.327 0.074 * 0.14 0.383 a a Northeast 0.154 -0.060 * -0.09 0.087 a a West 0.280 -0.127 * -0.23 0.368 -0.090 * -0.18 County District 0.333 0.049 * 0.09 0.402 0.114 * 0.24 Suburban 0.457 -0.084 * -0.17 0.313 a a Urban 0.314 -0.089 * -0.17 0.670 578.994 -9.E-06 * -0.07 245.064 Normal % White Enroll. Change '68-91 23.381 5.E-04 * 0.21 5.863 0.001 0.32 % Free/Reduced Lunch 1991 30.118 -0.004 * -0.34 38.625 -0.002 -0.15 Income 1980 20815 -1.E-05 * -0.27 20514 a Log of Enrollment 1968 4.026 0.068 * 0.10 4.735 15.482 0.006 * 0.09 15.765 0.055 * 0.06 % Nonwhite Enroll. Change '68-'91 Years Since First Major Plan a a -1.E-04 * -0.32 a -0.19 * -0.28 a a 0.103 0.123 0.16 DESEGREGATION VARIABLES FOR CURRENT PLAN Formerly Had a Deseg. Plan 0.090 Formerly Mandatory Plan 0.035 -0.02 0.128 0.026 0.04 Voluntary M to M 0.025 0.222 * 0.14 0.034 0.241 * 0.19 Magnet-Voluntary 0.050 0.114 * 0.10 0.171 0.215 * 0.34 Controlled Choice 0.016 0.035 0.02 0.026 0.014 0.01 Magnet-Mandatory 0.081 0.100 * 0.11 0.291 0.192 * 0.37 Mandatory, No Magnets 0.103 0.117 * 0.14 0.077 0.132 0.15 Constant -0.031 0.280 * 1.101 * Adjusted R2 0.53 0.42 N (District Weights) 956 117 * Significant at .05 level or better. a Eliminated from equation. Note: -9.E-06, and similar notations in other cells, means the coefficient is -9 but the decimal point should be moved 6 digits to the left. In this case, the coefficient is -.000009. Table 3 Relationship Between Change in Interracial Exposure 1978-1991 and the Percentage of Schools that are Magnets Controlling for School District and Plan Characteristics Districts > 5,000 With Plans Equation 1 Mean Equation 2 Equation 3 b Beta b Beta -0.0002 -0.04 -0.0002 -0.05 -0.0006 * -0.05 0.45 -0.0878 * -0.50 -0.0889 * -0.51 -0.0870 * -0.53 Northeast [vs. Central] 0.16 -0.0791 * -0.33 -0.0837 * -0.35 -0.0608 * -0.35 West 0.18 -0.0855 * -0.38 -0.0874 * -0.38 -0.0622 * -0.40 County district 0.31 0.0246 0.13 0.0268 * 0.14 0.0628 * 0.14 Suburban [vs. Rural] 0.30 -0.0433 * -0.23 -0.0432 * -0.23 0.0019 -0.21 Urban 0.56 -0.0743 * -0.42 -0.0743 * -0.42 -0.0124 -0.43 % Nonwhite Enroll. Change '78-'91 49.76 -0.0002 * -0.27 -0.0002 * -0.29 -0.0002 * -0.27 % White Enroll. Change '68-'78 -0.17 -0.0112 -0.05 -0.0093 -0.04 -0.0068 -0.04 % Free/Reduced Lunch 1991 40.67 -0.0009 * -0.21 -0.0009 * -0.22 -0.0012 * -0.22 Income 1980 19173 -7.E-06 * -0.34 0.0000 * -0.35 -5.E-06 * -0.36 4.23 0.0025 0.01 0.0025 0.01 -0.0101 0.00 16.54 -0.0033 * -0.21 -0.0033 * -0.21 -0.0013 -0.19 Ever had a mandatory plan 0.62 0.0371 * 0.21 0.0424 * 0.24 0.0351 * 0.20 Voluntary plan only 0.14 0.0409 * 0.16 0.0583 * 0.23 0.0391 * 0.16 % of schools that are magnets 9.80 -0.0008 * -0.17 0.0002 0.05 % Magnets X Voluntary Plan Only 2.27 -0.0016 * -0.20 % Magnets X Ever Mandatory Plan 6.28 -0.0010 -0.17 < 5 % of schools are magnets 0.09 0.0096 0.03 5-10% of schools are magnets 0.10 -0.0094 -0.03 10-50% of schools are magnets 0.20 -0.0112 -0.05 >50% of schools are magnets 0.05 -0.0467 * -0.11 Change in Inter. Exposure 1978-1991 b Beta -0.07 SOCIAL, DEMOGRAPHIC VARIABLES % White 1978 South [vs. Central] [vs. Central] [vs. Rural] Log of Enrollment 1978 Years Since First Major Plan 64.94 DESEGREGATION VARIABLESa MAGNET VARIABLESb (Constant) 0.2481 * 2 3.43 0.2477 * 3.45 0.2628 * Adjusted R 0.36 0.36 0.34 N (district weights) 300 300 300 * Significant at .05 or better. a b The omitted variable is controlled choice only. The omitted variable for equation 3 is no magnet schools. Note: -7.E-06, and similar notations in other cells, means the coefficient is -7 but the decimal point should be moved 6 digits to the left. In this case, the coefficient is -.000007. Table 4 Relationship Between Change in Interracial Exposure 1978-1991 and the Percentage of Schools that are Magnets Controlling for School District and Plan Characteristics Districts with Magnet Schools Equation 1 Mean Change in Inter. Exposure 1978-1991 b Beta Equation 2 b Equation 3 Beta b Beta -0.08 SOCIAL, DEMOGRAPHIC VARIABLES Percentage White 1978 59.26 0.0005 0.10 0.0004 0.08 0.0004 0.08 South 0.26 -0.0865 * -0.39 -0.0914 * -0.41 -0.0967 * -0.44 Northeast 0.22 -0.0593 * -0.25 -0.0606 * -0.26 -0.0650 * -0.28 West 0.26 -0.0546 * -0.25 -0.0468 * -0.21 -0.0589 * -0.27 County district 0.16 0.0054 0.02 0.0118 0.04 0.0058 0.02 Suburban 0.14 -0.1455 * -0.53 -0.1495 * -0.54 -0.1354 * -0.49 Urban 0.83 -0.1942 * -0.74 -0.1868 * -0.72 -0.1914 * -0.73 % Nonwhite Enroll. Change '78-'91 34.33 -0.0009 * -0.43 -0.0009 * -0.43 -0.0009 * -0.44 % White Enroll. Change '68-'78 -0.28 0.0525 0.14 0.0600 0.16 0.0615 0.16 % Free/Reduced Lunch 46.14 -0.0012 -0.20 -0.0016 * -0.28 -0.0009 -0.16 Income 1980 19423 -9.E-06 * -0.30 0.0000 * -0.38 -8.E-06 * -0.28 4.43 0.0496 * 0.21 0.0482 * 0.20 0.0512 * 0.21 14.93 -0.0036 * -0.23 -0.0035 * -0.22 -0.0033 * -0.21 Ever had a mandatory plan 0.67 0.003 0.01 0.0224 0.11 Voluntary plan only 0.21 0.021 0.09 0.0854 * 0.36 21.95 -0.001 * 0.0000 0.00 -0.0021 * -0.35 -0.0005 -0.10 Log of Enrollment 1978 Years Since First Major Plan DESEGREGATION VARIABLESa MAGNET VARIABLESb % of schools that are magnets % Magnets X Voluntary Plan Only 5.02 % Magnets X Ever Mandatory Plan 13.91 -0.206 5-10% of schools are magnets 0.26 -0.0178 -0.08 10-50% of schools are magnets 0.44 -0.0199 -0.10 >50% of schools are magnets 0.10 -0.0595 * -0.19 (Constant) * 0.2409 2 1.77 0.2833 * 2.10 0.222 Adjusted R 0.40 0.42 0.38 N (district weights) 135 135 135 * Significant at .05 or better. a b The omitted variable is controlled choice only. The omitted variable for equation 3 is less than 5% magnets Note: -9.E-06, and similar notations in other cells, means the coefficient is -9 but the decimal point should be moved 6 digits to the left. In this case, the coefficient is -.000009. Figure 1 A Classification of Magnet Program Structures XXXSCOPE OF PROGRAMxxxxxx ASSIGNED NEIGHBORHOOD STUDENTS YES NO ENTIRE SCHOOL PORTION OF STUDENT BODY Whole-School Attendance Zone Magnet Program Within a School Magnet (PWS) 1 2 Dedicated Magnet 3 Figure 2 Effect of Magnet Structure on School % Black in Magnet Schools in Prince George's County, Md. School District, 1996-97 30% SCHOOLS >= 70 % BLACK SCHOOLS < 70 % BLACK 19.4% 19.0% 20% Change in School % Black with Magnet 12.7% 10% 4.1% Elementary Middle 0% Elementary -1.2% -10% -6.2% -7.6% -13.4% -20% -24.5% -30% Whole School Attendance Zone Program Within A School Dedicated Middle Appendix 1 Relationship Between % White Enrollment Change 1978-1991 and the Percentage of Schools that are Magnets Controlling for School District and Plan Characteristics Districts > 5,000 with Plans Equation 1 Mean % White Enroll. Change 1978-1991 b Equation 2 Beta b Beta -20.52 SOCIAL, DEMOGRAPHIC VARIABLES % White 1978 64.94 0.001 0.00 -0.005 0.00 0.45 -11.220 * -0.23 -10.993 * -0.23 Northeast [vs. Central] 0.16 -2.694 -0.04 -3.586 -0.05 West 0.18 6.180 0.10 5.922 0.09 County district 0.31 17.719 * 0.34 17.942 * 0.35 Suburban [vs. Rural] 0.30 6.345 0.12 5.336 0.10 Urban 0.56 1.429 0.03 0.790 0.02 0.23 0.036 * 0.21 South [vs. Central] [vs. Central] [vs. Rural] % Nonwhite Enroll. Change '78-'91 49.76 0.038 * % White Enroll. Change '68-'78 -0.17 26.241 * 0.41 26.531 * 0.42 % Free/Reduced Lunch 1991 40.67 -0.116 -0.10 -0.123 -0.10 Income 1980 19173 -0.001 * -0.24 -0.001 * -0.24 4.23 -9.890 * -0.16 -9.869 * -0.16 -0.090 -0.02 -0.082 -0.02 Log of Enrollment 1978 Years Since First Major Plan 16.54 DESEGREGATION VARIABLESa Ever had a mandatory plan 0.62 9.854 * 0.20 10.691 * 0.22 Voluntary plan only 0.14 12.516 * 0.18 17.081 * 0.25 Change 1968-78 -0.33 -21.609 * -0.21 -20.008 * -0.19 Change 1978-91 -0.04 8.677 0.05 11.890 0.07 -0.15 -0.003 0.00 RACIAL IMBALANCE VARIABLESb MAGNET VARIABLES % of Schools that are Magnets % Magnets X Voluntary Plan Only % Magnets X Mandatory Plan Only (Constant) 9.80 -0.181426 * 2.27 6.28 -0.359 * -0.16 -0.152 -0.10 37.651 * 38.313 * Adjusted R 0.58 0.59 N (district weights) 300 300 2 * Significant at .05 or better. a The omitted variable is controlled choice only. b Black-white index of dissimilariy. APPENDIX 2 CONTROL VARIABLES IN EQUATIONS 1) Reduction in Racial Imbalance. This is used in some equations to control for the scope of the school desegregation plan as measured by the change in the black-white dissimilarity index. 2) Formerly Had a Desegregation Plan. This is used in the equation with a five plan typology for the 1991 plan to control for a previous plan among districts that currently have no plan. 3) Formerly Mandatory Plan. This is used in the equation with a five plan typology for the 1991 plan to control for a previous mandatory reassignment plan. 3) Population and enrollment trend characteristics: % White 1968. Because interracial exposure--the percentage white in the average minority child's school--can be no higher than the percentage white in the school district, the pre-desegregation percentage white must be controlled for. Region (South, Central, West, Northeast dummy variables). There are regional variations in state policies, court orders, and demographic trends that influence the desegregation outcomes. County District (dummy variable). Controls for the possibility that countywide districts, because they include the city and the suburbs, will have less white loss and thus less decline in interracial exposure. Area (suburban, rural, urban dummy variables). Suburban and rural districts are gaining whites, and urban districts, regardless of having a desegregation plan, are losing them. Percentage Change in Nonwhite Enrollment, 1968-91 or 1978-1991. Interracial exposure is heavily influenced by the percentage white in the district and the percentage white in turn is influenced by two factors: change in white enrollment and change in nonwhite enrollment. Percentage Change in White Enrollment, 1968-78. This is the control for the change in white enrollment that occurred in a district before magnets were implemented. Normal Percentage White Enrollment Change, 1968-91. This variable controls for the underlying demographic trend affecting school districts independent of any desegregation. Because most school districts had no pre-desegregation trend data and no information on metropolitan trends, the “normal” demographic trend in white enrollment change was estimated by computing the mean proportional white 1 enrollment change from 1968 to 1991 among the districts that never had a desegregation plan for five different categories of 1968 percentage white crosstabulated by four different regions and further crosstabulated by size--above and below 27,750. This variable then became a control variable for the districts with desegregation plans in the same quintile of percentage white in 1968, the same region, and the same size category. For the non-desegregating districts, the normal demographic trend was simply their actual white enrollment change during this period. Years Since the First Major Desegregation Plan. Although desegregation plans have the greatest single annual negative effect on white enrollment and positive effect on interracial exposure in the first few years following a desegregation plan (see Rossell, 1990a, 1990b, 1995b), the greater the number of years since the first major desegregation plan, the greater the white enrollment decline. The year of the first major desegregation plan is the year of the first mandatory reassignment plan (regardless of subsequent iterations) or, if there was no mandatory reassignment plan, the first comprehensive districtwide voluntary or controlled choice plan (regardless of subsequent iterations). For districts that never had plans, years since desegregation was set to 1976, the average desegregation year. Enrollment 1968 (log). This an estimate of busing distance and the logistical difficulty in desegregating a school system. The greater the busing distance, the greater the white flight (Rossell, 1978, 1988). Social Class Characteristics. The higher the social class, the greater the white flight with the implementation of a desegregation plan. (Rossell 1978; Rossell 1990a, 1990b, 1995b). The social class variables are:% Free/Reduced Lunch, 1991—this is a measure of the poverty of the public school children; median Family Income, 1980--census data aggregated up to school districts. 2