EAC Meeting Review - 02/11/2013
The EAC met last night to get more information from BPS and to further discuss the proposed plans. The new information from BPS included breakouts of several statistics by race, ethnicity and free lunch status; an analysis of a new 11 zone plan, and further analysis of the impact of various ways of implementing walk zone priority. These reports are available on the EAC meeting materials page.
The breakouts by race and ethnicity show that all groups have better minimum access to quality under all the new plans except for Asians. However, Asians have by far the highest minimum access in the current plan and would still have the highest minimum access in the new plans, but by a smaller margin. Those who qualify for free lunch and those who don't qualify for any lunch support also fare better under all of the new plans. Students who receive reduced lunch did worse, but it's important to note that this is quite a small group. When MIT analyzed their demand model for accuracy, they did not do a good job of predicting this group's choices because they didn't have enough data, so I don't feel that it's useful to look at the data for this group.
The next section of the BPS report covers their analysis of a new 11 zone plan that the EAC asked them to look at. This is a modified version of the 10 zone plan released on January 23rd. EAC members were concerned that the zone containing pretty much all of Dorchester and Mattapan was too big. It offered something like 14 school choices which was significantly more than other zones. They asked to see a similar plan with the Dorchester/Mattapan zone split into two zones. The modified plan doesn't look significantly different as far as overall equitable access to quality or distance traveled. The new northern Dorchester zone would have a smaller percentage of quality seats than the new southern Dorchester/Mattapan zone.
Laura Perille asked EAC members who had asked for the plan if the committee could go ahead and make a decision on whether they preferred the 10 or 11 zone version of the plan. She would like to go forward with three plans to choose from rather than adding a fourth. There was some discussion over which was better. The consensus favored the 11 zone version, though there was some concern that it splits Dorchester in half. I actually think that many members of the committee didn't have a strong opinion because they favor one of the home-based plans.
Carolyn Kain asked if they could refer to it as a modified 10 zone plan to be sure that people understood it was a modified version of the 10 zone plan from January rather than the 11 zone plan that BPS proposed in September. Ruthzee Louijeune and others felt like it was a bad idea to call the plan a 10 zone plan when it has 11 zones and worried it would create more confusion. She was critical of how BPS explained the plans at the meeting at Orchard Gardens last week and wanted to be sure they were explained more clearly at the next community meeting. In the end, I think they decided to call it an 11 zone plan but to make clear that it was based on the 10 zone plan from January.
The last part of the BPS report is on different the impact of different walk zone policies. For each plan they looked at median distance traveled with walk zone priority ranging from 20% to 80%. They also broke it out by using the current processing order versus the new proposed processing order (details on different processing orders in this post). Median distance ranged from 1.01 miles in the 10 zone plan and the home-based plan A using the current processing order to 1.24 miles for home-based plan B using the current processing order. If you look at keeping the percentage at 50% and just changing processing order, the impact on distance is tiny. The biggest change is for home-based plan B where the new processing order would reduce the median distance by .03 miles (158 feet).
BPS also showed minimum access to equity for each plan with various walk zone percentages and both processing orders. Here the differences are quite large. They range from 0% in the 10 zone with 80% walk zone priority and the current processing order to 28.1% for home-based plan B with less than 50% walk zone and the and the current priority.
Based on how the small reduction in distance traveled and how quickly equity drops off after 50 or 60% in most plans, it didn't seem like there was any appetite among committee members to increase walk zone priority beyond 50%. It seemed like the committee was leaning towards keeping the walk zone as it is. They still need to weigh in on whether to change processing order. At 50% it seems like the new processing order would make almost no difference except in home-based plan B where it would reduce minimum access to quality from 27.9% to 25.5%. There was also some discussion of whether it might be best to just keep walk zone priority exactly as it is. The idea being that if they remove it, that will get a lot of attention, while if they leave it as is it doesn't have much impact but it won't become a distraction from the changes in the new plans.
BPS has scheduled two more community meetings to explain the plans and get more feedback from the community. Tomorrow, Feb. 13, 6 pm at the Trotter School, 135 Humboldt Ave, Dorchester, and next Thursday, Feb. 21 at 6pm in East Boston at a location to be determined. I strongly urge everyone to attend one or both of these meetings if you can. These may well be the last chances for the community to give their feedback on these plans before the EAC makes a recommendation.
In addition, the committee is holding meetings to discuss their recommendations on Wednesday, 2/20, 6 pm. at Suffolk University, 73 Tremont St., 9th flr. and Saturday morning, 2/23 - exact time and location TBD. The Saturday meeting is "if needed," but it's hard to imagine that they will get through all the details of their recommendations without that meeting.
They are scheduled to vote on their final recommendations on Monday, 2/25 - 6 pm. at Suffolk University, 73 Tremont St., 9th flr.
Comments
Presenting the results
The way that the results of the analysis are being presented is interesting.
Among other issues - it's hard to make comparisons between plans. For example, I'm here with two pages from the main report - 30 and 42 - so I can hold them up to the light and superimpose Figure 15(d) and Figure 27(d) and try to understand the differences. That's the data behind the 1-30-13 BPS memo. There are a few comparisons are in the graph appendix, comparing the results for the various plans at a high level. But this data needs more comparisons.
It also needs more interpretation. Maybe some of this will happen when the School Committee reviews this.
It will be interesting to see how this is presented tomorrow.
I've had the same problem
I found that I had to print the report out so I could hold up pages next to each other or even superimpose them. It's particularly hard to try to compare things like the neighborhood-by-neighborhood charts among the different plans.
I'm not particularly happy with the way all of this has been presented to the EAC, let alone the public. For example, I think they're over-simplifying by only showing the lowest child's access and not showing the full box plots where you can see 25th and 75th percentiles as well.
I certainly hope they give more detail tonight. At a minimum, they need to describe how the home-based plans work and why.
Show more than one measure of quality
Yes, given its importance and the difficulty of measuring it - using one summary measure for access to quality doesn't make sense. They should present 3 or 4 ways of looking at the impact on access to quality.
Close to home is easier to present, as the hard measures like distance and bus coverage area seem to capture it fairly well.
Access to Quality?
One of these reports makes me question if the "Access to Quality" number is what I think it is.
I thought it represented the chance of a kid getting into one of the top 50% of seats in the system, MCAS wise. However, it seems impossible to me that any plan could increase access to quality for all three groups that represent 92% of students (excluding Asians since their number goes down).
I mean if the number of quality seats is X, and access to quality for group A goes up by 10%, that would have to result in those seats being less available for group B, right? I guess all the seats could be coming from Asians, but that does not seem like it would add up. By my numbers there are 23k Hispanic children in the system, so if 8% more of them are going to quality seats (as in Home B) that would mean they would be occupying 1,800 more seats. The only group that is going down is Asians, and if the 5k or so of them are going to be 8% less likely that only clears up about 400 seats.
This implies that the supply X is not fixed. Is there an assumption that there are going to be more quality seats or that a lot are going unfilled today?
Also, even more deeply if quality seats is defined at the top 50%, then the number would be fixed by definition as 50% of enrollment, would it not? It would be mathematically impossible to add more people to the top 50% without adding more people to the bottom 50%, otherwise it would not be 50% any more.
Does access to quality not mean what I think it does, or is the supply not fixed in these numbers?
This is MINIMUM access to quality
The key distinction is that this is minimum access to quality. This is looking at the probability of the child with the lowest access getting in to a top 50% school. So what's really happening is that the range of access has been shrunk so that whoever has the least chance has a better chance. As you say 50% of all kids will get a quality seat by definition.
Why Choose this Number
This seems like an interesting number to pick. Since it is the absolute minimum it is easily effected by outliers.
I guess I am a little suspicious that this number is being used exclusively because it conveys the impression that everyone is a winner (except Asians I guess).
I could not find a way to get the median or 75-25 percentile in tabular form from the MIT report since they only seem to appear in really small little graphs, but from a median standpoint it seems like the main effect from a racial standpoint is to reallocate quality seats from Black to Hispanic children, drawing a few away from Whites and Asians as well.
I am kind of surprised that the Black community is OK with this, or maybe they are not and I just have not heard anything about it yet.
Compare the charts
If you want a more complete answer, you have to compare charts in the main report. You'll see the impact on the average and the distribution by neighborhood or race.
That is, compare the charts on page 30 of the main report to their counterparts for each of the new plans.
Charts
Yeah, I did figure that out. That is where I got my numbers above about Black access going down and Hispanic going up.
Call me cynical here, but I find it rather convenient that the politically expedient number which shows access for the absolute minimum person, including all possible outliers, is presented to everyone and is easily accessible to the public in large font on a presentation that is one click from the EAC page.
If you want to get the median, which represents effect on ALL students, you need to visually compare a 1in x 1in bar chart on pages 30 and 42 of a report that is 5 clicks from the EAC page on a different website.
Any other view of the data is going to show the winners and losers. If people want to make the case that raising up the lowest percentage is more important then a significant drop in the overall average for some groups that is certainly a valid case to make. However, I think people have a right to see this number, in large text and in a tabular form, right next to the minimum number.
I suspect it is because showing a large racial or neighborhood group with a negative number next to their access to quality seats would go over like a lead balloon. But maybe I am wrong, the only way to find out is to show people the numbers.
I was surprised
It may have to do with the specific measure you're looking at. If the measure is the access of the person who has the lowest percentage - than yes, that could go up for all groups.
I was surprised to see that the assignment plan could in fact change the average access to quality. (I would have thought the average would always be the average.) There's some discussion in the main report. It's a little unclear, but see the first bullet on p. 16.
At some level the takeway may be - if we can reduce transportation by 40% and still improve equality of access to quality, and have a flexible plan that can adapt as the system changes - maybe that's pretty good.
But the complexity of what's going on is yet another reason to show multiple measures of access to quality. And to be very clear about what they are.