Make us your home page


Education news and notes from Tampa Bay and Florida

A weekend interview about graduation rates with Florida deputy education commissioner Jeff Sellers



Sellers_J Every year Education Week puts out its report on high school graduation rates that differs greatly from the rates that Florida releases. And every year Florida officials decry the Ed Week report saying it's an imprecise measure that does a disservice to Florida's more accurate data system. Jeff Sellers, Florida deputy education commissioner for accountability, spoke with reporter Jeff Solochek about the pros and cons of the different ways to look at graduation rates.

I'm curious about what Education Week does provide. Is there any value you see in their report?

Well, yes, there is value in their report. It provides a methodology that can help us estimate future graduation rates before all of our data have actually been submitted by the districts. Although we prefer our cohort-based grad rate, which actually takes into account individual students and tracks them over time to see how many graduate.

So the state does something completely different. Have you ever talked to Education Week about why they don't do it your way?

I don't know that we have ever talked to them directly about it. Overall this has been an issue from the perspective that when you're trying to calculate, or if you try to have a national measure of any kind, this one being a graduation rate, you have to accommodate the lowest common denominator. So based on states' capabilities, you have to find ways to calculate any type of measure. And so for something like Ed Week's graduation rate, that approach is probably one of the best to accommodate a common measure for all states.

But in Education Week, they actually point to Florida's method as a very strong method. And I know the president and his education team have also talked about Florida's data system as being the best. Is there a disconnect here? ... Why would people want to take the Education Week numbers over your own?

Well, right. The federal government has come out with a new standard for measuring graduation rates that are due to be implemented in 2010-11. And it is going to follow very much the Florida rate. It will be just like the way we calculate our rate, from the perspective of following a cohort of students over time to see how many actually graduate. Now, there will be differences in how we define a graduate vs. what they define as a graduate, and those who are kept in the cohort vs. those who aren't. But overall I think from a national perspective we're heading in a direction ... that is consistent with Florida's direction.

You said certain items will be left out. Are they the GED students, for instance, that Florida counts?

GED and also the adult ed transfers.

If they don't want to include them, why does Florida think it's important to include those?

Currently, we've had our cohort, the way we've been calculating our grad rate, since 1998. So we've been doing it longer than the federal government's involvement. Also, in statute, our state statute, we have a requirement to calculate it the way we do it, with standard diplomas, special diplomas, GEDs, and what we do with adult ed transfers, those types of things. ...

As I recall, it was the commissioner who said -- a couple of commissioners ago -- that the important thing was finishing and it didn't matter when or how.

Right. And the reason that we do it now is because it is in law. ... The feds have their requirements, that are due to be released in 2010-11. We will also be following those requirements. Because of the nature of the data system that we have, we have the flexibility to accommodate these requirements as they come down.

There's been some questions raised about the way all the transfers are accounted for, and about the ability of a school or a school district to play with it a little bit, to make the rate appear better than it is. Have you ever heard any questions like that, or is there a good answer to that type of concern?

I don't see how they can really do that. We base our rates on the data that the districts submit to us. We actually give the districts an opportunity to validate that information before we release it publicly. They have to account for all their students. Through that accounting mechanism is how we produce our grad rate.

So there's no room for flexibility in coding, or fudging it somehow?

We have well defined categories for how students withdraw and the diploma types ... and those are mandated for all our school districts.

How can you tell if they're doing it right?

We do audit from a data perspective.

So why does this become such a big issue each year? Why do you think we pay such attention to graduation rates as an important measure?

I think it's just the nature of getting students, determining success by their completion of high school. And also ... looking at their preparedness for both college and career and making sure that the standard diploma meets those requirements.

Are you satisfied with the numbers you see?

Well, we always want to see higher numbers. But what we are pleased with, regardless of how you calculate it, you see increases in the rate. ...

So for people in Florida reading Education Week and thinking the graduation rate is as they describe it, what's your message? Don't read Education Week?

I think there is value in it, again from understanding where they are coming from establishing a national model that can be applicable regardless of the differing technological capabilities that each state has in this area. But I think for Florida, what becomes an issue for us is this rate is not as accurate for highly mobile environments. So when you have a student population that tends to move around quite a bit, which Florida does have, it does not do as good a job of tracking individual students while they move around. So it's an estimate. It gives you an idea. But with this situation this was 2005-06 data, we have already released this data actually following these students. You can see the difference a cohort-based rate makes from using this type of index.

[Last modified: Tuesday, May 25, 2010 10:26am]


Join the discussion: Click to view comments, add yours