Summer research for MNPS: Early Childhood Policy!

Well, graduation has come and gone, and it's time to put my MPP to work. I'm grateful that I decided to spend so much time while at Peabody to data handling, analysis, program evaluation experience. Grateful, too, for the practicum placement and ongoing volunteer work with Martha O'Bryan Center, which has really demanded that I understand the complexities (and boundaries) of high-quality, useful evaluation and data work. But now I get to build my own projects: from nebulous start, and planning out each step, tracking down the data I need for each piece of the project... through the analysis of data, making judicious decisions about how to handle demographic trends, all the way to writing findings/recommendations and reporting out -- presumably to the superintendent, if not also to the mayor's office and the folks at PRI who are at the table.

So I'm excited to get started. It is a delight to spend this summer working on such concrete policy questions. In particular, I get to spend time and mental energy on the planning and strategy that will inform expanding access to pre-kindergarten for four-year-olds in Davidson County! On Monday morning, I spent a wonderful five minutes in the front lobby watching a four-year-old and her (very slightly) older sister run around and quiz each other about colors, and it was just fantastic to know that I get to be part of a process to serve these lil' ones.

There are already new initiatives under way for the expansion of PK4 in Nashville: the early learning centers at Ross, Bordeaux, and Casa Azafrán, as the district announced and as The Tennessean reported in February, in May, and just earlier this week. Certainly the upcoming initiatives are the main focus for  the director of early learning innovation, but my role will mostly be future-focused... as in, not primarily aimed at phase one year one but informing the strategy for smart decisions in later years and/or phases. 

Sharing my process for answering pesky strategic questions for PK4

I know from research I did last year (for coursework, under Joe Murphy, and for additional practicum credit, under CG Loss) that it is HARD to find some of the answers to questions that must inform any smart new policies affecting early childhood. For example, how many students are currently in some kind of formal learning environment before they enter kindergarten? We could tell you how many kids are in the MNPS system, but that still leaves a big pocket of "unknowns" - and are they at home, or are they in some other environment? In simplest terms, what is the need vs. the available resources? Currently I know of no simple way to get a good answer. Another example question: Which parts of the district are objectively most in need of PK4 capacity? (Wait lists can present a skewed perspective, depending on how they are tracked, handled, etc.) How will that capacity need to be distributed, five years from now? What about ten years?

These are the kinds of questions that will become increasingly important as states and districts everywhere consider the possibility of expanding access to PK4, in anticipation of new federal resources to do just that. (Assuming, of course, that the House decides to pass the legislation, that it gets through a budget process, etc., etc.) And they are complex enough questions that they cannot be answered just through a quick Google. (Believe me, I tried.)

Consequently I would like to make the process of tracking down these answers as transparent as possible. This will hopefully help other policymakers and/or data nerds to avoid the roadblocks I will inevitably encounter in the process... or at least, to see how different strategies can lead to different answers. 

What do we need to know for PK4 planning? 

Specifically, I am tasked with research to address the following:

1. What is the "seat" capacity for four-year-olds in Davidson County? This includes public MNPS pre-K, private schools, for-profit providers, nonprofit providers, Head Start, charters if any, etc. And of course, there are follow-ups: are all those seats full? is there a waiting list everywhere?
2. Where in the county are these families with 4YOs living? In particular, are some neighborhoods experiencing more over-subscription (i.e., wait list) than the others?
3. There are districts and states where voluntary pre-K is available to every family who chooses to enroll their child. What do these programs look like? How are they paid for, how are they administered, how are they held to high standards that are also developmentally appropriate, etc.? I am thinking particularly of somehow "comparable districts" -- Denver, for example, has enrollment numbers close to MNPS. Also, cities like San Antonio, Tulsa, Boston, even NYC have done a lot for PK4 access -- what can be learned from those models, and what just cannot "translate" to the Nashville context? Perhaps most important from a practical perspective, how do comparable districts pay for universal (voluntary!) PK4 access?

Surely there will be other tasks, but this is where I'm starting.

Ideally I would only use public-access, free resources for some of these tasks... but frankly I will happily take advantage of any useful data resources I can still access through the Peabody library. (Just for a couple more months, until Vandy shuts down my ability to login with my ID.) So making it replicable isn't exactly the goal, since some of the data won't be accessible to the general public. But I would at least like to make the process transparent. 

Signal, noise, transparency, software

Hopefully the multiple strategies I have in mind for these several projects will allow me to triangulate estimates that are actually useful for planning purposes. Given the inherently "fuzzy" or "noisy" nature of population estimates, I do not think it is possible to get exact answers for some of these questions ... but it would be great to at least get better parameters than I have so far been able to access. For example, I recall an estimate from a few months ago, that I did together with the whipsmart Bryn at MOBC: in one particular geographic area, census estimates for "how many four year olds live in our target service area?" came back as something like "One hundred children, plus or minus 80 children." Statistically accurate, I'm sure, but totally useless from a planning perspective. So I will probably need to combine multiple estimates, rather than rely on "one best system". (See what I did there, ed-pol nerds?)

I plan to use both Stata and R, as the relevant tasks demand. I am not yet fluent enough in R to scrap my use of Stata, even though the justice side of me says "Use The Open Source Software, So There Aren't Cost Barriers To Your Data or Project!" But I'm learning. Maybe by the end of the summer I will be exclusively in R.

So, here I am at the beginning. Let's see where we end up, shall we?