Incremental Change

“Silvia Abbato, the district’s superintendent, said she could not pinpoint any one action that had led to the better scores. . . . 

“It’s not something you can do overnight,” Ms. Abbato said. “We have been taking incremental steps everywhere.”

—Motoko Rich, Amanda Cox, and Matthew Bloch, “Money, Race, and Success: How Your District Compares” in The NY Times

School Climate Matters

A classroom in Guipuscoa

Chalk up more research confirming what-we’ve-been-saying-all-along here at Schools & Ecosystems: a school’s learning environment impacts student learning.

In case you don’t know, NYC has been collecting what folks call “school climate” data via surveys administered to teachers, parents, and students since 2007. It’s important information to have about a school–arguably more important, to my mind, than test scores (I believe both should be considered).

Last July, I had quoted Match Education’s Mike Goldstein asking an important question about all this data:

Is anyone aware of scholars and reporters digging deep into this data set?  Is there any other data set in the USA just as good?

I think it’d be hugely productive to identify NYC schools which have made progress in “Total Climate” — and then study why.

Well, Mike, you’ve got your answer.

NYU’s Research Alliance for New York City Schools published a study using NYC’s school climate information that demonstrates that a school’s learning environment not only impacts student learning, but furthermore teacher retention. As Chalkbeat NY’s Alex Zimmerman reports:

Each measure, the report found, is independently linked to decreases in teacher turnover. And gains on two of those measures, high academic expectations and school safety, were directly connected to better scores on state math exams.

The study found that if a school improved from the 50th percentile across the study’s four measures of school climate (leadership, expectations, relationships, and safety) to the 84th percentile, teacher turnover would decline by 25 percent, or 3.8 percentage points.

A similar percentile increase in measures of school safety and high academic expectations alone boosted math scores enough to account for an extra month and a half of instruction. (Improvements in school climate also boosted language arts scores on state tests, but those gains weren’t statistically significant.)

It’s important to note that this study confined its focus to the following aspects of school climate:

  • safety and order
  • leadership and professional development
  • high academic expectations
  • teacher relationships and collaboration

Missing in such an examination (and mostly from these surveys themselves) is a focus on the physical environment of a school. There are questions pertaining to cleanliness and conditions of a school, but as we’ve also been arguing on this blog, the actual design, and the incorporation (or absence) of access to natural light and greenery, colors, furniture, etcetera (all largely subconscious factors), all have an impact on learning and relationships in a school.

If your school is interested in collecting school climate data, the US Department of Education is sharing free surveys and information for collection of data similar to NYC’s. Check it out and share.

My current views on testing, in answer to my past views on testing

While up in Albany a few weeks ago, I was interviewed by someone from NYSED about what I might say to parents who are considering “opting out” their child from state testing. You can view the video here*.

Someone on Twitter, “WiffleCardenal,” voiced a critique to me regarding the video, in contrast to things I’ve said in the past on testing. In fact, they even tweeted quotes of my own words! I deeply appreciate that someone out there is actually listening, and willing to take the time and effort to hold me accountable to them. I have elected to respond here, since Twitter isn’t the greatest venue for nuanced discussion, especially at the end of a long day, and I also hate typing things on my phone.

This is in reference to a live chat I did back in 2012 on The Nation‘s website with journalist Dana Goldstein and educator Tara Brancato. Have my views shifted since then? I would say they have in some ways.

You know, honestly, they’re not as terrible as I thought back then. I proctor these tests each year and go through the experience of answering the questions along with my students. The questions are often cognitively demanding and require multiple reappraisals of the text in question. A few of them are duds, certainly, but having tried to write many of my own text-dependent questions since then, I’ve come to appreciate a well-written multiple choice question. Check out this post from Joe Kirby (UK educator) on the rationale for using multiple choice questions for assessment.

Unfortunately, this continues to hold true. In reaction to this, the Center for American Progress recently created a “testing bill of rights” to advocate for better aligning tests with a more meaningful purpose.

This doesn’t mean, however, that I’m opposed to having test scores factor into my own evaluation or my school’s evaluation. When scores are considered over multiple years, I think they can be an important and useful measure of teacher effectiveness. But they are extremely variable, so I would only want them to be considered alongside other data that can provide adequate context.

One of the things I’ve become more aware of over time is that while our testing and evaluation schemes are extremely problematic, if we look at the big picture, accountability and testing do bring transparency to serving populations of students that were traditionally ignored. No Child Left Behind was certainly faulty and overzealous policy — but it also brought attention to holding school districts accountable to serving students with disabilities and other underserved populations based on data. This was entirely new, and it has raised awareness.

This is why the NAACP, the National Disability Rights Network, and other national civil rights groups oppose anti-testing movements.

Yes, I continue to believe this. Test measures are only one source of data that need to be coupled with qualitative observational data and other forms of understanding. Fortunately, I do feel like our focus, at least in NYC, has shifted to better match this understanding.

To give further context on my statements on the NYSED video, I was speaking about how I use testing data, which I do every week when developing IEPs for my students with disabilities. I compile all information I have on a student, including multiple years of state test data, in-house assessment data, such as reading, writing, and math scores, GPA, attendance, psychoeducational evaluations, social histories, etc. When viewed all together, in tandem with teacher observations and student and parent interviews, I find aggregate state testing data useful!

So it’s important to understand I’m not advocating now and never have advocated for a state test score as a singular reference point to judge myself or a student. But when viewed with appropriate context, I do find state testing data to be useful. (More on how I use that to develop IEPs here.)

No, unfortunately. While I do think that test scores should factor into an account of an individual teacher’s effectiveness (only in aggregate and when considered in terms of growth, not proficiency), we’re creating incentives for competition, rather than collaboration.

If I could set the rules for how we use test scores for accountability, I would do something kind of radical: I would hold all grade-level teachers accountable for student scores on literacy tests. And I’d stop labeling them “ELA” tests and call them “literacy” tests. Why? Because if we are honest about what we’re really testing, we’d acknowledge that the knowledge required to understand complex texts comes not solely from ELA, but furthermore from science, social studies, music, art, and so forth. (More on my argument on this here).

Furthermore, I’d try to better level the playing field for all students by requiring test makers to broadcast one year in advance which texts would be tested (not specific passages, just the general title/author). I would allow parents and educators an opportunity to vote on which texts they wanted tested that year as well to make it more reflective of current interests. The reason I would do this is that this would provide an opportunity for all students to build up the requisite vocabulary and background knowledge to access a text. Right now we just give them random texts, as if every child will be bringing equivalent knowledge and vocabulary to them, which is false.

Yes, unfortunately this continues to hold true in too many schools. But this is also why I have been a consistent supporter of Common Core standards, which have become synonymous with testing in some people’s minds. Yet the Common Core standards provided us an opportunity to move away from test prep, because they are fundamentally about building student knowledge and academic vocabulary through engagement with rich and complex texts — this is the exact opposite of test prep!

This speaks to the problem of making state tests so high stakes, and why we need multiple measures, such as direct observation, to hold schools accountable. It also is the reason for why I would advocate for the seemingly radical measure, as per above, of communicating which texts would be assessed that year so that “test prep” instead would simply be about reading and studying and discussing the rich texts that were selected for that year’s assessment.

Yes, it can be inhumane when a student is several years behind in reading ability or struggles in coping with anxiety and stress.

While computerized testing brings a whole new set of problems, I do believe we should move in this direction, because with computerized testing, we can use adaptive testing that can better scale to meet a student where they are. Otherwise we end up punishing students who are struggling, for whatever reason. Unfortunately, the needs of students with disabilities never seem to be factored into test design except as a final consideration, rather than from the ground up.

But there’s another side to this, too. I think we have to ask ourselves, as a teacher, a school, and a system, how do we prepare all of our students to be able to engage with a challenging text independently? And in what ways are we sequentially building their knowledge and skills and vocabulary in order to prepare them for doing so? It is the failure to do so systematically and adequately that we are failing students who most need those skills and knowledge.

Pearson is out of the picture, in case you didn’t know. I have no idea what Questar tests will be like, though I imagine they will be comparable.

From what I’ve heard, PARCC assessments are far superior to the cheaper assessments NY decided to get from Pearson. I think we get what we pay for, and if we want better test design, we have to be willing to fund them.

Personally, I think if we’re going to just use tests for accountability purposes, then we could make them every 2 or 3 years instead of every year to save money, and they could still continue to be used for that purpose.

What would be awesome is if we could move more towards performance based assessment. There’s a great article on them in the most recent American Educator. This seems like the right direction to go in if we truly interested in assessing the “whole child.”

Well, don’t know if all of this fully says everything I would like to say about testing, but I’m seriously tired after a long week, so this will have to do.

WiffleCardenal, whoever you are, thank you holding me accountable and I welcome continued critical dialogue on these issues.

* This was after a long day of a train ride from NYC and meetings with legislators, so I apologize for my shiny face. Won’t apologize for the winter beard, however. And no, I was not paid for that interview nor given a script. As ever, I speak my own mind (or so I like to think. Certainly let me know if it ever seems like I don’t).

Accountability for the Long-Term

time_distance

I receive a monthly newsletter from bcg.perspectives that I scan for any relevant connections to school systems. Their work often centers on business policy, but sometimes their work has either a direct or indirect connection to the education sector.

A recent post,”Gauging Long-Term Impact in the Social Sector” on developing a system of long-term evaluation for a large international nongovernmental organization (INGO), there are lessons well worth considering in developing systems of long-term accountability for schools.

The INGO discussed, named SOS Children’s Villages, works towards “improving the situation of children who are at risk of losing, or who have already lost, parental care” across 134 countries.

The assessment methodology that SOS Children’s Villages developed jointly with BCG evaluates two elements of the programs’ long-term impact: the nonfinancial (or all-around) impact on the individual program participants and the community and the financial impact on society. The determination of the long-term impact on individual participants is based largely on information gleaned from interviews of former program participants by external researchers. This is supplemented by qualitative research conducted through focus group discussions with former child participants and their caregivers. . . 

The programs’ long-term financial impact on society is gauged by the programs’social return on investment (SROI), a comparison of the programs’ total costs and benefits to society. . .  The calculation of societal benefits is based on easily quantifiable elements.

This combination of intensive qualitative and correlated quantitative data gathering seems to make great sense when considering systems for school accountability. Many school systems have been relying primarily on isolated testing data — but why not go straight to the source, and interview the ones we most seek to impact? The students and families and community. And then correlate that with longer-term impacts via “social return on investment”? What are the long term outcomes of students after they graduate?

Raising test scores is wonderful. But enriching one’s community and society over the long-haul is the true goal of education. Developing better combinations of quantitative and qualitative evaluation of our school systems that can help us determine long-term impact is key to not losing sight of that higher purpose.

Responsible Regulation, Not Ideology

“it’s time for school choice advocates to dispense with ideology, engage regulators, and get serious about a policy environment that promotes measurable quality, scale, and access.”

—John White, “America’s most disadvantaged students need real accountability, not ideology” on Flypaper