Just this month, OSPI released a new kind of data: Student Growth Percentiles (SGP). What are student growth percentiles? In short, SGPs describe a student’s growth in state test scores as compared to other students with similar prior test scores. Here’s a five minute video:
You can find Student Growth Percentiles for your specific school or district here: http://data.k12.wa.us/PublicDWP/Web/WashingtonWeb/PublishedReports/PublishedReports.aspx
or http://bit.ly/1lE2Pi9
What are student growth percentiles for? Teacher evaluation is one potential use, and will be an issue in the upcoming legislative session. Washington state recently received a high risk warning from the federal government regarding teacher evaluation. The issue? Whether state test scores “can” or “must” be used in teacher evaluation—the U.S. Department of Education is saying that state test scores must be used in order for Washington state to continue to receive a NCLB waiver. We’ve written extensively about this waiver on our blog—see posts from Mark, Kristin, Tom, and myself.
One issue with including state test scores in teacher evaluations? Very few teachers in Washington state even teach classes associated with a state test! The number of teachers with state test data has been estimated at 16% at the most by OSPI—see the chart.
How do you evaluate teachers with state tests when these teachers don’t even teach courses that are tested? In Tennessee, teachers without test scores were able to choose a test for their evaluation, leading to some unusual conversations, “The P. E. teacher got information that the writing score was the best to pick,” said the art teacher. “He informed the home ec teacher, who passed it on to me, and I told the career development teacher. It’s a bit like Vegas, and if you pick the wrong academic subject, you lose and get a bad evaluation.” In Florida, teachers have been evaluated using school wide test averages, meaning that some teachers are evaluated based on test scores from students they have never taught. North Carolina attempted to test students of all teachers in all subject areas with 52 different standardized tests. All these approaches have proved problematic.
Here in Washington state, a different idea has been proposed. Senate Bill 5960, sponsored by Senators McAuliffe and Rolfes, would limit the teachers evaluated by state tests to those who teach courses with federally mandated state tests. This would essentially create two side-by-side evaluation systems—English and math teachers would be evaluated with state tests and other teachers would not. The bill states, “For teachers who teach reading or language arts or mathematics in a grade in which the federally mandated statewide student assessments are administered, one of the multiple measures of student growth must be the student results on the relevant assessment.” (Science tests are not considered as part of Adequate Yearly Progress under NCLB, and so would be exempt.)
In addition to the larger issues involved with evaluating teachers with state tests, there are two very specific problems with this bill if growth were to be determined with SGPs. First, student growth percentiles are designed to be calculated from one year to the next. However, there is a two year time span between eighth and tenth graders taking their state tests, meaning that tenth grade English and math teachers would have highly questionable student growth percentile ratings.
A second major issue with this bill is that it would be implemented in 2014-2015, the first year of the new Smarter Balanced Assessment Consortium (SBAC) tests, based on new standards. This means that student growth percentiles would be calculated using scores from the last year of the current state tests and the first year of the new SBAC tests. Growth would be measured between two very different tests and two very different sets of standards. Teachers should not receive a brand new high stakes component of their teacher evaluation in the very first year of a new student test—this must not happen. OSPI itself recommends waiting until the 2016-2017 school year.
Another concern with student growth percentiles is the timing—the plan is for the data for each school year to be available in September or October of the following school year. Teachers receive their evaluations in May, meaning that those evaluations would be based on work from a previous school year instead of the current one.
It is worthwhile to note that even if student growth percentiles are adopted for teacher evaluation, they are just one slice of the total teacher evaluation pie. The exact size of that slice would be determined by a series of state and local decisions. So why is this such a concern? Teacher evaluations are slated to be used for a wide range of human resource decisions, far beyond renewal/nonrenewal of contracts. These decisions may include preference in assignments, leadership and training opportunities, probation, and layoffs. Yes, SGPs would be just one slice of pie for one slice of teachers. But for those teachers affected, this is a huge deal.
Also, these calculations are complex, and mistakes are made! Washington DC public schools reported this week that calculation mistakes were made in 10% of the teacher evaluations with a value added measurement (different than SGPs). Some evaluations were too low, some were too high, and one teacher was mistakenly fired! The district is attempting to reinstate the teacher’s contract and compensate the person for lost wages.
Student growth percentiles involve attributing specific students to specific teachers, which raises a number of questions:
- How long does a student need to be in a particular teacher’s classroom before that student's scores are attributed to them? 1 month? One semester? The full school year?
- Is a student’s score attributed to the teacher they had for most of the year, or the teacher they had at the time of the test?
- What if a student has extremely severe attendance problems? Will that student’s scores still count towards that teachers' evaluation?
- Who will be in charge of “cleaning up” the data before it is submitted to the contractor or the state for calculations? Someone at the school or district level will need to be double-checking that students are attributed to the correct class; to the classroom teacher instead of the teacher who proctored the exam; and so on.
- Will teachers have access to the raw data used in the calculations?
We have been developing a strong teacher evaluation system in Washington state over the past few years, grounded in improving instructional practice. This is the first year that all districts in the state are implementing the new system—it is not the time for major changes.
Thanks for the great information and the video is just awesome.
To get SGPs for your school or district, try this link: http://data.k12.wa.us/PublicDWP/web/Washingtonweb/PublishedReports/PublishedReports.aspx or http://bit.ly/1lE2Pi9
If that doesn’t take you straight there, click on “static data files” at the top, then “assessment” over to the left hand side, then “SGP School Growth Summary Reports” also on the left. Finally, select your school district.
Thank you very much for a thorough, well detailed article. Now if I could just figure out how to retrieve the SGP scores from OSPI’s website. I have been doing a little digging, and the research on using SGPs to evaluate teacher effectiveness seems to strongly indicate that this practice is not supported. This concerns me immensely. In fact, most of what I read indicates that using SGPs is even more problematic than using the value added measures that have come into question in so many other states. If determinations of teacher effectiveness are clearly questionable with SGPs, then why are we proposing to use them?
If the state decides to use SGPs to evaulate teachers – which I have no doubt they will – I see two potential implications:
The first regards student placement. Students are not currently placed randomly into classrooms. Especially at the elementary level. They are placed according to what administrators see as a “good match” or in response to parent request. That will change abruptly if SGPs are used for evaluative purposes. Teachers will want – and will get – their fair share of students who have a history of high academic growth.
The second regards teacher placement. For obvious reasons, an elementary principal will feel pressure to place the most effective teachers in 4th, 5th or 6th grade. And the least-effective teachers in 3rd grade, as that’s the “base-line grade.”
These shifts have nothing to do with actually improving education, and everything to do with making it look like we’re improving education.
Thanks, Maren, for helping to get the word out to other educators in our state. I feel many are unaware of the decisions our legislators will be making and the ramifications of their decisions.
As always, thoughtfully considered and excellent summarization of where we find ourselves at the moment. Our legislators would likely be loathe to consider scrapping this ill-considered plan, but I don’t believe there is a “fix” that would be fair or wise. These measures should not be part of ANY teacher evaluations. Or HR decisions, for that matter. At the State level, we must educate our legislators on the folly of traveling one more step down this dangerous path for the students of our state. “MUST” MUST not be in the law, unless to say that teachers and districts MUST be free to create, collaborate, and decide on the measures that help us ensure the incorporation of appropriate student growth data; our own choice of authentic, appropriate assessment as part of the cycle of excellent teaching and learning. As I’ve noted elsewhere, a teacher’s ability to design appropriate assessment must be demonstrated as part of TPEP as well as PRo-Cert and NBCT certification. We need to trust our teachers with the opportunity to make these vital choices for our students, as allow us to demonstrate competence in using our data in our professional practice.
This is great information.
My big worry is that student growth percentiles LOOK like a very simple solution to anyone not directly involved with education. And since those are the people who make decisions about education policy, we need to inform them.
Thanks for such a thorough overview of the situation, Maren. It’s a mess.
First, what is all of this doing for the children who are easily performing at standard? It’s driving the curriculum, it’s eliminating more risky creative and critical thinking work, and it’s making math and literature – two areas with huge potential for creative work – all about the test. And I think our state test is not that great. It gives a sort-of reliable snapshot of where a child is, but I don’t think it’s consistent, or objectively scored, or all that reliable. Plus, teaching to the test means teaching to the bottom tier of meeting standard, and I would like education to push students to go beyond that.
The second biggest problem is that we’re using one lens – growth on a not-that-great-test – to measure some teachers, and we’re using a different lens – how well do you teach – for all the other teachers. I’m not worried about fairness, but I am worried about what this does to the philosophical unity in a school.
Because students who live in poverty or who have special needs are not meeting standard on the state test, the WHOLE SYSTEM is being changed to boost them – but it’s not being changed in ANY WAY that really matters, it’s just being totally overloaded with gathering data that students who live in poverty or who have special needs are not meeting standard. Frustrating.