Education's 'big data' problem - and how to fix it
Monday, August 18, 2014 5:00 PM
Every year, American public schools spend $2.2 billion replacing teachers who drop out of the profession.
But the true cost to the system is even bigger. When teachers stick with their jobs, they build up valuable experience and improve their instructional techniques. When they leave, that valuable experience vanishes. Administrators have to expend time and money - both usually in short supply - training replacements.
Retaining teachers, therefore, could save billions and improve education quality.
Fortunately, preserving the intel experienced teachers have isn't that difficult. It's as simple as mimicking the stat geeks who are revolutionizing baseball and the political operatives who are winning elections.
They're harnessing the power of data. And they're succeeding.
For the first time ever, schools have detailed information on the strengths and weaknesses of individual educators - and the computing power to analyze it. They can also track student performance, rates of learning growth, demographic trends, even the effectiveness of various teaching methods across subjects.
Used correctly, such information could fuel teachers' professional development, empowering them to cultivate their craft and get the most from their students.
Unfortunately, that's not how many school officials are using this invaluable data.
Forty-one states require regular teacher assessments. In most cases, administrators collect data on teacher performance, offer detailed descriptions of their deficiencies - and yet do little to help them improve.
Instead, schools pack their teachers into one-size-fits-all "professional development" programs that are wholly insensitive to the challenges of individual classrooms - and to the learning needs of individual teachers. Students enjoy a half-day off, while teachers are subjected to stock lectures on, say, what regulations they must follow when administering standardized tests.
Every year, the government spends more than $4 billion on such professional development. That's about $1,000 per instructor.
The returns on that investment have been non-existent. American student achievement has remained flat for thirty years. Less than one-third of American high-schoolers can read proficiently. Over a quarter read significantly below grade level. And just a quarter are proficient in math.
Even when teachers get data on their performance - or their students' - they receive little to no instruction on what to do with it and on what will help them strengthen their practice. As one teacher from Chicago said, "They always look at the ... data but never talk about the next step; someone has to show us how we can use it to improve."
In many cases, that data can be used for punitive ends. The Los Angeles Times, for instance, publicly ranks more than 11,000 teachers in the Los Angeles Unified School District based on student test scores.
It's no wonder that many teachers have soured on the idea of data-driven school reform altogether. In their experience, it serves solely to embarrass or punish them.
There's a better way. By using education data to drive personalized professional development for each educator, school officials can put the data they're required to collect to good use and raise the skill level of their teachers - without alienating them in the process.
Consider a teacher whose evaluations show that she's having trouble keeping her students focused. Instead of boring her with a three-hour lecture, her administrators could set her up with teacher -development programs keyed to the exact size of her class and the specific nature of the distractions.
Attention-grabbing techniques that work for a group of 12 students won't necessary work for 30. The same is true for single-sex versus integrated classes - or those with varying numbers of non-native English speakers.
Likewise, technology can make evaluations less painful - and more meaningful - for both principals and teachers. A laptop loaded with data from previous observations can allow an administrator to hone in on only those teaching indicators that merit the most attention - and provide instantaneous feedback accompanied by videos of best practices from master teachers.
The stakes are high. Research shows that the number-one predictor of student success is the quality of his or her teacher. For instance, a math teacher with even slightly above-average skills increases student achievement scores by nearly six percent. That alone speaks volumes.
In order to produce better students, we must produce better teachers. The "Big Data" revolution that's taken hold in other parts of our society can do just that.
Jason Lange is the CEO and Co-Founder of BloomBoard (www.bloomboard.com).