“People assume newly hired teachers come straight from preparation programs, but our data shows that isn’t really the case. . . There’s a whole pool of people that we don’t really think about, but are a big part of finding replacements when districts lose a teacher.” – Dr. Jeremy Landa
Worries about the pipeline of qualified new teachers are nothing new, but in the wake of the COVID-19 pandemic they rocketed to new levels. As districts prepared for the 2022-23 school year, headlines across the United States warned of teacher shortages after an exodus from the profession. Would preparation programs be able to keep up with growing demand?
The Texas Education Agency (TEA) was in a better position than some states to answer that question. Since the mid-2000s, TEA has kept track yearly of the newly hired teachers across all 1,200 districts and charter schools, as well as the newly issued teaching certificates by preparation programs.
The state’s rich teacher-preparation data had informed a prior SDP project, which tracked the demographics of aspiring teachers throughout the preparation process and identified observations as a speed bump for Black candidates. But the separate certification and hiring datasets couldn’t help leaders answer detailed questions about whether preparation programs were meeting districts’ hiring needs and where districts were finding their new hires. That was the goal of SDP Fellow Dr. Jeremy Landa, TEA’s director of educator data, research, and strategy for educator preparation, certification, and enforcement.
“We were trying to solve a broader landscape problem,” he said. “It’s very hard for either of those entities—an educator preparation program and a local education agency—to accurately describe what’s happening when it comes to the production of new teachers and where those new teachers wind up when they start working.”
Such insights would shed light on districts’ new hires, so they could understand and improve their recruiting and hiring practices. But with 43,000 new hires made in 2021-22 alone, the data would quickly overwhelm any one program or agency. Landa and his colleagues would have to investigate these data-driven questions, and the resources that could answer them, from the ground up.
A “Goldilocks” Problem
The first order of business was to reframe the questions by looking at the unit of analysis. From his graduate work, Landa knew that teaching was a fundamentally local profession.
“The numbers range, but as many as 80 percent of teachers in the workforce work within 40 miles of where they went to high school,” he said. “If teachers are indeed that localized, we need to think really carefully about how local pipelines are set up in every region of the state.”
To glean understanding from the state’s existing data, leaders needed to glimpse local conditions. Any region or community in Texas whose trends differed substantially from average would be hidden with a few large statewide numbers. Yet sorting information by 1,200 local education agencies would create so many small data points, trends would be tough to tease out. “It was a Goldilocks problem,” Landa said. “If you're looking only at the state level, you get almost no information beyond a basic descriptor. If you go to the local level, there's so much information that it's hard to think actionably about it.”
So Landa and his team settled on the “just right” size of their unit of analysis: they grouped the data based on the state’s 20 Education Service Center Regions. For each region, they created data illustrations showing the percentage of individuals who completed teacher-preparation programs who also were hired to work in that same region.
The analysis showed that across Texas on average, seven out of 10 newly certified and employed teachers were trained and taught locally. But it also revealed important differences between regions. For example, while more than 90 percent of new teachers in the El Paso area were locally trained, in the Beaumont region, that figure was just 21 percent. Some regions had strong local pipelines of new teachers. But others did not.
Gaps in Training and Hiring
A driving question remained: did Texas have enough newly certified teachers to meet district needs? Landa and his colleagues again capitalized on existing data to answer and illustrate this question: a simple clustered bar chart showed that new hires dramatically outpaced newly issued standard teaching certificates in every region and academic year from 1999-2000 until 2020-21. The differences ranged from 3,600 to as high as 22,000 more new hires than new certificates each year.
Where were districts finding these teachers? Landa presented the bar chart to colleagues throughout the department and asked questions like, “what part of the policy context are we missing?” These conversations helped identify various types of teacher certifications that were likely part of the differences between newly trained and newly hired teachers, while also helping Landa and his team build institutional knowledge.
At the same time, Landa also was building a new set of data that included all newly hired teachers in the state. When he merged the new hire data with the new certification data, he found two surprising results. First, a large group of new hires were not new teachers—they were experienced certified teachers who were returning to the classroom after a break.
“Reentering teachers accounted for about 30 percent of the new hires in any year in Texas for the last 23 years,” he said. “It’s a huge number, and it hasn’t really been part of the policy landscape. . . There’s a whole pool of people that we don’t really think about but are a big part of finding replacements when districts lose a teacher.”
Second, state laws that allow local districts to hire teachers without Texas certification were also having a major impact on staffing. The data shows that in the 2021-22 school year, 20 percent of all new hires were not certified, or 8,500 of 43,000 new hires.
“When we are looking at the sources that contribute to new hires, we observe LEAs hiring teachers from a variety of backgrounds,” Landa said. “This should raise questions for everyone about teacher preparation, teacher recruitment, and teacher hiring.”
These new insights are now available to the public through the Newly Certified and New Teacher Hires Dashboards, posted on TEA’s website. The dashboards were built by another SDP Fellow, Dr. Dina Ghazzawi, a data analyst at the department.
“It’s just another piece of our story as we continue having people with the SDP network doing really exciting work,” said Landa.
Lessons Learned
Landa identified three related major takeaways from his work as an SDP Fellow.
First, capitalize on existing data infrastructure to get started quickly and produce “easy win” projects. This supports his second and third takeaways: the importance of building relationships and mastering institutional dynamics.
Creating simple visual presentations of data that are already in-house can serve as strong starting points for candid conversations within an organization. Such conversations contribute to informal networks, institutional understanding, and trust. They also can identify which employees or offices have the ear of likely decision-makers—the sorts of context and insight that can help analysts present and explain a well-chosen dataset to the right audience for maximum impact.
“There are not many organizations where there isn’t someone who’s been there for a long time. Those people are beacons of knowledge,” he said. “It’s important to figure out as best you can who has the ability and the access to the people who are actually going to stand up in front of public external stakeholders."