Round One 2022-23 Kindergarten Placement Analysis
This blog post is an analysis of placements for 2022-23 kindergarteners through the Intentional Integration Initiative in the South Orange-Maplewood School District. This analysis includes dimensions that Michael Alves did not cover in the Year 1 Recap Report, such as how the III impacted each neighborhood and school.
I was able to get the distances for the 423 first-round 2022-23 kindergarten students, from May 2022, from Greg Mortenson, who was able to get this data through the discovery process of his case. The SOMSD rejected two OPRA requests I made for this data, but they evidently decided they did not have grounds to reject a request made in a lawsuit because they provided the data, in PDF format, to Mr. Mortenson.
My data is not 100% complete for three reasons:
- It only contains Round 1 students, so it doesn't have assignment data for later registrants.
- Every school has some reserved seats and I do not know the final disposition of those seats.
- It includes students who registered but who did not enroll in September.
- There are probably errors made where some students got algorithmic placements when they should have gotten sibling placements. The Administration would have then switched these students into their older siblings' schools.
For the above reasons, I hope readers understand that many numbers, percentages, and averages in this post are not authoritative & final and thus will differ from what the SOMSD officially presents. However, certain tendencies which I show here, such as
- There was a low response rate for educational and income questions.
- Low-income, African-American, and Seth Boyden-area kids travel the longest distances; relatively few students at Boyden live near Boyden.
- Students who are super-close to a school are no more likely to be assigned there than students who don't live that close to any school. IE, there is no "proximity weight."
ought to persist into the final, complete data.
Background on Enrollment and Variance
There were 423 students assigned in Round 1. They were 30% low-SES, 32% medium-SES, and 38% high-SES.
Due to the +-5% variance allowed between the elementary schools, the following are the demographic floor and ceiling for each socioeconomic stratum, ie, the "seat control."
 |
| Source, 2022-23 SOMSD Variance |
Although the 2022-23 must have gotten a few dozen late-registrants, 423 kindergarteners is well below capacity and what pre-Covid enrollment was. The district's ideal kindergarten enrollment was set at 504 students.
This is the racial breakdown of 2022-23 First Round kindergarteners.
FYI, the students who live nearest to Seth Boyden were 46% African-American, which continues the trend of a declining African-American population in Hilton.
The 2022-23 kindergarten is 17% multiracial, which is a large jump from the 2021-22 kindergarten, which was only 7% multiracial.
1. There was a Low Response Rate to Questions about Education and Income
Only 55% of parents answered the questions about income and educational attainment, 10% answered the question for education but not income, and 35% didn't answer either.
Of the 55% that did answer, 92% were high-SES, so there is an obvious skew towards high-SES people being willing to divulge income and educational information.
Since so many people did not reveal income and educational information, I expect that there are a lot of inaccurate SES classifications.
The high non-response rate requires that the Admin/BOE clarify how SES categories are assigned when parents do not divulge income & educational data.
2. Distribution of Assignment Types
The 423 kids were placed in the following way: Six ELL assignments (1.4%), 96 sibling assignments (22.7%), 34 sped assignments (8%), and 287 algorithmic assignments (68%)
3. Distribution of Distances
Of the 423 kids, the average distance to their assigned school was 1.17 miles, though the median was only 0.86 miles.
Here is the distribution of distances:
| 2022-23 Kindergarten Distances | ||
| Number of Students | Percentage of Students | Distance |
| 14 | 3% | <.25 miles |
| 64 | 15% | 0.25-0.49 miles |
| 151 | 36% | 0.5-.99 miles |
| 77 | 18% | 1.0-1.49 miles |
| 42 | 10% | 1.5-1.99 miles |
| 65 | 15% | 2.0-2.99 miles |
| 10 | 2% | 3.0 and over |
The range of distances was 0.11 miles up to 3.66 miles. The 3.66 mile child lives near the border with Irvington, southeast of DeHart Park, and is assigned to South Mountain. Since South Mountain starts at 8:00 AM, we should be concerned about how early this child and other kids in the same situation have to wake up to catch their bus.
The longest possible distance would be 5 miles, from Van Ness Terrance-area to the South Mountain Annex, but no students from that part of Maplewood are assigned to the Annex, nor did anyone have this assignment last year.
There is a large negative skew against low-SES kids again. Of the 10 kids with the longest trips, SEVEN are low-SES, two are mid-SES, and one is high-SES.
Of the 75 kids who are 2.0 miles and above, 38 are low-SES (51%), 8 are middle-SES (11%), and 29 are high-SES (39%).
| 2022-23 Average Distance Traveled by Racial Classification | |
| Category | Avg. Distance |
| Asian | 1.15 miles |
| Black | 1.56 miles |
| Hispanic | 1.22 miles |
| Multiracial | 1.37 miles |
| White | 1.03 miles |
| All Students | 1.17 miles |
I do not agree with any argument that if the SOMSD had a transfer policy that white and/or high-SES people would disproportionately use it.
As you can see from this chart and the SES distance charts, African-American and low-SES students already are more likely to be sent to distant schools. If distance is what motivates people to want transfers, African-American and low-SES students would be more likely to request transfers. Under a waitlist transfer process, disenrollment of students would enable those transfers to be granted.
Of kids in the "high-stress" 1.0-1.99 mile range, 32 are low-SES (27%), 48 are mid-SES (40%), and 39 are high-SES 32%). Low-SES students are slightly underrepresented here.
There were only 14 kindergarteners in the 1.75-1.99 mile range, which is the same as last year. I continue to believe that Taylor, Burnside, and Joshua are badly incorrect when they say it is impossible to even bus the 1.75-2.0 mile kids.
Since the SOMSD is preparing to bus 1.0-1.99 mile kids, I will present a more detailed picture of how many students are in each distance range.
Note: there are about 30 students who live over 1.0 mile from any elementary school and whose distant placement is not the result of the III.
4. Kids in Which Neighborhoods are the Most Likely to Be Assigned to the Local School versus a Distant School
Kids who live near Marshall (85% to Marshall), Tuscan (60% to Tuscan), and South Mountain (55% to South Mountain) tend to be assigned there, but kids who live near Clinton (40% to Clinton) and especially Seth Boyden (only 31% to Seth Boyden) are scattered. As you will see below, there are almost as many Seth Boyden-proximate kids being sent to Marshall (13) as there are to Seth Boyden (16). Very few Seth Boyden kids even got to go to Tuscan (9) and Clinton (7), even though Tuscan and Clinton are the second and third closest schools to Boyden, respectively.
Notice, the above categories are not the old Marshall, Tuscan, Boyden etc official sending zones. My analysis uses students who live the closest to a particular school, not what old zone their residence is in.
| Assignments of Kids Who Live Closest to Marshall | ||
| Assigned School | Number of Kids | Percentage |
| Marshall | 52 | 85.2% |
| South Mountain | 7 | 11.5% |
| Clinton | 1 | 1.6% |
| Boyden | 1 | 1.6% |
| Tuscan | 0 | 0.0% |
| Assignments of Kids Who Live Closest to Tuscan | ||
| Assigned School | Number of Kids | Percentage |
| Tuscan | 59 | 60.2% |
| Marshall | 14 | 14.3% |
| Boyden | 13 | 13.3% |
| South Mountain | 6 | 6.1% |
| Clinton | 6 | 6.1% |
| Assignments of Kids Who Live Closest to South Mountain | ||
| Assigned School | Number of Kids | Percentage |
| South Mountain | 64 | 55.2% |
| Marshall | 30 | 25.9% |
| Tuscan | 13 | 11.2% |
| Boyden | 6 | 5.2% |
| Clinton | 3 | 2.6% |
| Assignments of Kids Who Live Closest to Clinton | ||
| Assigned School | Number of Kids | Percentage |
| Clinton | 38 | 40.4% |
| Marshall | 29 | 30.9% |
| South Mountain | 13 | 13.8% |
| Boyden | 9 | 9.6% |
| Tuscan | 5 | 5.3% |
| Assignments of Kids Who Live Closest to Seth Boyden | ||
| Assigned School | Number of Kids | Percentage |
| Boyden | 16 | 30.8% |
| Marshall | 13 | 25.0% |
| Tuscan | 9 | 17.3% |
| South Mountain | 7 | 13.5% |
| Clinton | 7 | 13.5% |
5. Few Boyden Kindergarteners Live Near Boyden
To put it another way, very few Seth Boyden kindergarteners live near Seth Boyden.
69.1% of Clinton kindergarteners live closest to Clinton, 68.6% of Tuscan kids live closest to Tuscan, 66% of South Mountain kids live closest to South Mountain, 37.7% of Marshall kids live closest to Marshall, but only 35.6% of Boyden kids live closest to Boyden.
Only a small percentage of Marshall students live near Marshall, but Marshall has the biggest grades by far, with 138 kids in Round 1 assignments, versus only 45 for Boyden. 85.2% of students whose closest school was Marshall got to go there, versus only 30.8% of students whose closest school is Seth Boyden.
| Geography of Seth Boyden Round 1 Kindergarteners, 2022-23 | ||
| Most Proximate School | Number of Students | Percentage |
| Boyden | 16 | 35.6% |
| Tuscan | 13 | 28.9% |
| Clinton | 9 | 20.0% |
| South Mountain | 6 | 13.3% |
| Marshall | 1 | 2.2% |
Partly this is due to Seth Boyden having fewer kindergarten seats than the other elementary schools, but it only had 52 students for whom Seth Boyden is their closest school, so it would naturally have the smallest enrollment even if all students were sent to their closest school. The school with the least room for its area-students is Clinton.
Low-SES, African-American, and Boyden-area students travel long distances because Clinton and Tuscan are also schools that have more proximate-students than space for them, so Clinton and Tuscan fill up quickly with their own neighborhood students and Boyden-area students end up being assigned to South Mountain (2.1 miles away) and Marshall (3.0 miles away).
The disparity is also due to the nature of integration, since the low-SES population of Seth Boyden is not allowed to exceed the district's overall low-SES population by 5%, and, in fact, the BOE, Michael Alves and Ronald Taylor kept the variance much tighter than +-5% (see below).
 |
| Click to Enlarge. |
Of the 52 students for whom Boyden was the closest school, 42 (80.7%) were low-SES, 4 (7.7%) were middle-SES, and 6 (11.5%) were high-SES.
Since a school cannot exceed 35% low-SES (18 students), 24 low-SES Boyden students must be sent to other schools. (42-18=24)
By point of contrast, South Mountain is "only" 70.7% high-SES. Since the +-5% variance allows for South Mountain to be up to 43% high-SES, not nearly as high a percentage of South Mountain's high-SES students must be sent to other schools as Boyden's low-SES students.
6. How Many Students Get Proximate Versus Distant Assignments
Of the 287 kids who got algorithmic assignments, only 144 (50%) were assigned to their nearest elementary school. The other 143 were assigned to more distant elementary schools.
In 2021-22, 60% of algorithmically-assigned kids were assigned to their nearest school, so there has been a 10 point drop. Perhaps the percentage of students assigned to their nearest school will turn out to be higher once the later-registrants are included or if the Admin transferred anyone, but if it is 50% of students at their proximate school, it is a significant change from 2021-22. I suspect the reason for fewer kids being assigned to their nearest school is that the district used even tighter seat control than last year.
THERE IS NO "PROXIMITY WEIGHT"!
Prior to the summer of 2022 I was under the misimpression that being closer to a school made a child more likely to be assigned there, but that is NOT how the placement system works. What determines if a child will be assigned to their proximate school is their random number, since their random number determines when they will be assigned relative to other students.
Of the 84 kids who live at or below 0.5 miles from their nearest school and were placed by the algorithm, 43, 51.2%, were assigned to their nearest school, which is the same as students who are beyond 0.5 miles. This is important to know because it underscores that, contrary to what Taylor and the BOE have said, there is no "proximity weight." Students who are super-close to a school are no more likely to be assigned there than those who live 1.0 mile to their nearest school.
Algorithmically-placed students who are above 0.5 miles have a 50.5% chance of being placed at their nearest school.
7. Where Kids of Each SES-Tier are Assigned When they Do Not Go to their Nearest School
What I found for placement tendencies is what one would predict, except that the fact that Marshall's grades are so enormous means that it receives non-neighborhood students of all three socioeconomic tiers.
All students analyzed below received algorithmic placements.
Comments
Post a Comment