Great education requires great teachers, but the existing system makes it too difficult to retain the best and replace the worst. Fixing this requires, among other things, more generous pay. Instead we face the profession’s persistent, declining productivity.
When I was the superintendent of an urban charter school district, we saw in our classrooms every day what our data confirmed and decades of research told us: The quality of the classroom teacher is the largest determinant of student outcomes. Great education requires great teachers. If we had been able to retain our best teachers each year and replace the least effective ones with educators of just average effectiveness, our student outcomes would have soared. Unfortunately, the existing system makes that impossible, even in charter schools.
Attracting the most talented people into education and retaining them when they become great teachers requires, among other things, generous pay. The best and brightest weigh the compensation prospects of different career options when deciding on a career path. So do top-performing teachers when deciding whether to remain in the field. Too often, we see smart, dedicated people who go into teaching out of a sense of purpose finding that the monetary rewards can’t support their lifestyle expectations, especially if they’ve also incurred student debt. They fall economically behind their peers and leave the classroom after just a few years.
Why is compensation a negative attribute of a teaching career? The answer is the system’s persistent, declining productivity as measured by the student-to-teacher ratio.
The recognition that public education has a productivity problem is not new. Economist William Baumol predicted rising education costs in the 1960s due to competition for labor with industries that are more productive. He coined the term “cost disease.” Yet education turns out to be both an imperfect and an accelerated example of Baumol’s theory. Imperfect in my experience because education’s relatively importable skills have held down competition for labor with other industries and therefore held down compensation growth. Accelerated because the marginally lagging productivity that Baumol envisioned became a tsunami of negative productivity as policy makers flooded the system with adults through smaller class sizes and specialized instruction. Public education suffers the worst version of Baumol’s cost disease, with perpetually rising costs but without the competitive salaries that Baumol assumed would cause it.
Today, the relationship between pay and productivity is both logical and well documented. If the productivity of an industry increases, there is more value available for compensation. Conversely, when productivity in an industry decreases, costs rise and resources available for compensation become scarce. In simple terms, if a school district’s productivity declines by 1 percent each year, its budget would have to increase by that same amount just to keep teacher compensation flat. Even small annual declines in productivity add up over the long term, and in education they have done exactly that.
Over the last sixty-five years, the productivity of teachers in the U.S. has declined 41 percent while the average American worker’s productivity increased fourfold. Specifically, worker productivity has increased 2.2 percent annually since 1955 while the number of students per K–12 teacher has declined from 26.9 to 15.9. Teachers work as hard as anyone, yet are trapped in a system that destroys their productivity and therefore their prospects for generous pay.
What about ways other than student-to-teacher ratio to measure school productivity? School districts are funded based on their number of enrolled students and their largest expense is personnel. Therefore, the capacity of a district to provide compensation is determined by the ratio of funded students to staff—its productivity. When talking about teachers, that means the student-to-teacher ratio. In other contexts, it would be correct to measure teacher productivity by a combination of quality and quantity of output, not quantity alone. For example, if states funded schools on a combination of enrollment and test scores, that would be the case. Yet they don’t. When the objective is to determine a district’s financial capacity for pay and funding is based on the number of students, the student-to-teacher ratio is the unavoidable productivity metric to examine.
So how do we do improve teacher productivity? Paul Hill and Marguerite Roza provided a framework for doing exactly that in their 2010 white paper on curing Baumol’s disease in education—and I’ll emphasize one of their points. Hill and Roza advise considering strategies used by other labor-intensive industries and leveraging learning techniques used outside of schools. One of the lessons from outside education that they specifically cite is production process innovation, which turns out to be something the education industry is sorely lacking.
In my professional career, I built and managed a process engineering function for a large, complex commercial organization. The task of improving productivity through process innovation is the kind of challenge that process engineers excel at. Teachers spend about 40 percent of their work hours instructing students. The first step in the improvement process should target increasing that percentage. It’s possible to get there with the help of engineering techniques.
All non-instructional work assigned to teachers should be evaluated with each task given a designation of eliminate, automate, delegate, consolidate, or keep, in that priority order. Any task not justified by the value it creates should be eliminated. What is justified should be automated, and what cannot be automated should be delegated to a lesser-skilled resource. Tasks that can’t be eliminated, automated, or delegated should be consolidated into non-teaching periods such as summer or other breaks in the academic calendar. Internal reports are great candidates for elimination; regulatory reports for automation, grading, and lunch or hallway monitoring can be delegated to assistants, apprentices, or even students. Professional development and planning can be consolidated into periods when students are not in school. Only the bare minimum of non-instructional tasks should be kept in the hands of teachers and done within the academic calendar. These tasks, ultimately designated as “keep,” should be engineered to ensure that they are optimized.
Leveraging technology in instructional tasks is the second re-engineering strategy for improving teacher productivity. Technology can extend the reach of highly effective teachers through remote synchronous and asynchronous content delivery. It can provide the extra help that struggling students need while giving advanced students incremental challenge. Technology can also be used to break content up into small chunks interspersed with practice problems allowing students to learn at their own pace. Computers can enable flipping the classroom, allowing students to engage with content in off hours and do “homework” in class when help is readily available. Properly leveraged technology will allow the time that teachers spend instructing to be extended across more students, thus amplifying their productivity.
To get the greatest benefit from gains in teacher productivity, the value created should be channeled into compensation, but into differentiated compensation not in the traditional credential and tenure swim-lane model. Eric Hanushek provides a robust, research-based roadmap for reforming educator pay, benefits, and retirement plans, to which I would only add: Raise starting teacher pay within understaffed disciplines. Generous and differentiated compensation should be the target for attracting the best and brightest into education and retaining the most effective teachers. Those within the profession that are less suited to teaching should be helped to find an alternative role within the system or an different career entirely.
I realize that the strategy of increasing the student-to-teacher ratio may seem counterintuitive to those outside education policy discussions. Much of the contemporary debate on this topic has leaned to the side of smaller class sizes. Making classes smaller requires more teachers in a country with an existing teacher shortage. It also takes the available funding and spreads it across more people, therefore leaving smaller slices for each. Lesser pay simply exacerbates the problem of attracting and retaining the best. Continually expanding the total level of funding, as we have done over the past sixty years, has not produced better outcomes and is unsustainable.
The solution to lagging student achievement is great teachers attracted and retained within education by generous compensation. That level of compensation will require improved productivity, which can be achieved by learning from others and through the sciences of process engineering and leveraging technology. The goal should not be a teacher in front of every student but rather every student in front of a great teacher.
In a recent Fordham article, Daniel Buck makes several thoughtful critiques of the practice of schools that have replaced the mark of zero on a 100-point scale with a minimum grade of fifty. In my article, “The Case Against the Zero,” I argued that the zero, when combined with the use of the average to calculate the final grade, leads to mathematically inaccurate distortions in determining the student grade. I appreciate Buck’s expectation for high levels of performance and his rejection that 60 percent should be deemed passing, whether in high school algebra or brain surgery. I also appreciate Buck’s consideration of the finite nature of teacher time and the futility of students bringing in a truckload of missing assignments on the last day of the semester and expecting teachers to absolve them for months of late work. Thanks to the Fordham Institute for this opportunity to clarify my views now, almost eighteen years after the original article was published.
First, students do need consequences for failing to submit work. The only question is what the most effective consequences should be. If F’s, zeroes, and point deductions were effective consequences, then after a few centuries of these experiments in student motivation, all student work in 2022 should be on time and perfect. I know of no teachers on the planet who make such a claim. The penalty for not turning in work should be doing the work, and schools around the nation have been very clever in creating consequences that are effective. These include constraints on student time and space, with the requirement to get work done when it is due. Some brilliant athletic coaches do not allow students to suit up until homework is done. In other schools, there is a “quiet table” in the lunch room (expanded in other cases to a “quiet room”) to finish work. Students crave freedom and independence. These are consequences that work when threats of F’s, zeroes, and point deductions are futile.
Second, the problem is not just the zero, but the default of electronic grading systems to the average. This is why the zero becomes the academic death penalty. Far from incentivizing students to be diligent, the zero, when combined with the average, tells the student that all the blather about resilience and perseverance that they hear from their teachers is just so much hot air. After a few zeroes early in the semester, it becomes impossible to pass the class. This is why the last few weeks of every semester are plagued by chronic absenteeism and disruptive behavior. When academic success is impossible due to the use of the average, why bother?
Buck’s reference to surgeons is quite appropriate. We don’t evaluate surgeons based on the average of their early mistakes on cadavers but rather on their proficiency with real patients. Indeed, we expect surgeons, pilots, and a host of other professionals to make mistakes, learn from those mistakes, and improve their skills as a result of that learning. The same emphasis on learning from mistakes, getting feedback, respecting and applying feedback, and ultimately improving performance should be the focus of classroom work in schools. When I have interviewed employers and college professors, they disdain the idea that everyone gets it right the first time. In fact, they seek employees and students who can make mistakes and accept feedback without calling their mother, therapist, and lawyer. Our grading systems should reward response to feedback and quit with the fantasy that every student gets things right the first time. Besides, we learned during the pandemic that “perfect” homework may be more likely credited to Google and Kahn Academy than to the individual work of students.
Third, I have no problem with the zero, as long as it is mathematically accurate. Let us return to the good old days in which report cards were not on the 100-point scale, but on a much smaller scale, sometimes with letters, and sometimes with Latin descriptions. Most people are familiar with simple A–F grading, in which A=4, B=3, C=2, D=1, and F=0. This is the way that most schools calculate grade-point averages, and it has much to commend it. I trust teachers to tell their students the difference between an A and a B, a B and a C, and so on. But no teacher can persuasively explain the difference between a 32 and 33, a 75 and 76, and—here is where the blood is spilled when grades are reported—an 89 and 90, or 59 and 60. These one-point variations are classic distinctions without a difference.
Buck and I want the same thing—students who respect teachers, organize their work, meet deadlines, and learn the habits of responsibility and diligence. My respectful suggestion is that we pursue these goals with techniques that actually work rather than persist in a system of zeroes and averages that undermines the very character traits we seek to instill.
As a conservative, part of my job is to stand athwart rapid education changes yelling “is this really a good idea?” I did just that in a recent piece for the Fordham Institute, and Douglass Reeves, a rightfully-respected educational theorist whose work I criticized, has responded. The debate is joined. Should we re-apportion grading scales, even going so far as to eliminate the zero?
After reading Reeves’s response, I’m more resolute than ever in my initial conclusion: Let’s consider a 0–4 instead of an A–F scale. But I’m stiffened in my conviction that today’s fad of simply cutting off the scale halfway is the worst of all possible worlds. If we’re going to revolutionize our grading practices across the country, it should be done with forethought and decision, not haphazardly and without teacher buy-in.
Reeves and I agree on a few essential points. Like him, and as just mentioned, I’d be fine with his proposed 0–4 grading scale; I’m merely skeptical and looking for more evidence of its effectiveness. And we both agree that there needs to be consequences for unfinished work.
That being said, we disagree on what those consequences should look like. No one contests that zeros are disproportionate. They are; I simply think that this disproportionality is a good thing. There’s a big difference between shoddily finished work (a D or 60 percent) and something given no attempt (a zero). The 60 percent gap between a D and a zero reflects that difference.
Disproportionality is neither inherently unethical nor unjust. We must look at the consequences that result from different approaches to grading to decide such questions.
Proponents of the no-zero approach often gesture toward motivation. If a student receives a zero, it tanks their grade and deflates their drive in the class. If they have a chance of recovery, however, because their school doesn’t give zeroes, they’re more likely to put in the extra effort to save their grade.
Is there any evidence of this? I suspect the exact opposite. If a student doesn’t have the dead weight of a zero hanging on their grade, won’t they be less likely to pursue corrections, retakes, or extra tutoring in a last-ditch attempt to save their grade? Won’t more students just accept mediocrity rather than fighting to correct an F?
Both these framings of motivation are, of course, speculative and theoretical. We would need evidence to prove that an alternative grading system would actually improve outcomes, and here we are sorely lacking. As I mentioned in my first article, leading proponents of no-zero grading confess that there is much literature but very little evidence on this topic. In other words, there are many glorified opinion pieces suggesting that no-zero grading could help but no evidence that it actually does.
If anything, I’ve witnessed evidence that it works in reverse. One teacher wrote about her experience in a no-zero school, saying that the change “didn’t encourage students who had failed. It resulted in formerly hard-working students turning slack.” Schools that tried it before the practice became in vogue during the pandemic have since moved away from such policies for similar reasons.
I have worked in a school that opted for a behavioral penalty instead of a zero. That didn’t encourage better student completion. If anything, students felt less inclined to do work because what did a demerit mean to them? Conversely, a zero sat on their report card, a beating heart under the floorboard, reminding them every time they checked the grade book that they needed to make up work. In the end, I opted to assign zeroes again, even against my school administration’s wishes, because I found them more effective.
Which returns me to my initial conclusion. Perhaps these new grading fads are worth trying out in a handful of districts wherein researchers can keep close tabs to see if there’s any demonstrable improvement. If learning really does improve in a measurable way or school climates gets a boost, I’ll change my opinion. But scrapping our current grading practices without a viable alternative is unwise, reckless even.
In the spring of 2020, a group of researchers from the University of California San Diego was engaged in a longitudinal study of changes in young children’s learning experiences during kindergarten and first grade at an anonymous, medium-sized, socioeconomically diverse school district in southern California. Two weeks after the initial suspension of schooling in March 2020, the caregivers in the larger study were invited to participate in a different project, which would use a remote daily diary protocol to collect information on family life, including their children’s educational activities, under lockdown.
Thirty-two primary caregivers with a child in kindergarten agreed to participate. The majority were mothers; 44 percent of the parent-child dyads identified as Hispanic, 34 percent as white, and 22 percent as mixed race. The children were six and a half years old on average. About a third of the dyads came from Spanish-speaking homes. Families came from diverse socioeconomic backgrounds, with 25 percent from the lowest income group (earning less than $50,000 annually), 22 percent from the middle-income group ($50,000–$75,000), and 44 percent from the highest income group (greater than $75,001). Fifteen percent did not fully answer the income questions and could not be classified.
Families were prompted via text message to fill out a survey instrument five times per day over the course of five days at two different time points. The first survey wave was given three weeks after the district launched its virtual learning platform via Google Classroom, and the second wave was given four weeks later, in the second-to-last week of the school year. Messages were sent at random times throughout the day from 7:30 am to 9:00 pm for five consecutive days. Each text message included a link to the researchers’ five-minute survey, and participants were asked to complete it as close as possible to the time they received the text message. The link expired after one hour to help boost timely compliance. Parents primarily responded to surveys outside of school hours and mostly while they were with their children, which aided the analysis.
While the analysts framed their research interests in further detail in the report, they boil down to questions regarding what children and families were doing all day, whether caregivers ascribed positive or negative value to these activities, and whether caregivers were happy with the activities. They hypothesized that socioeconomic status of the family would likely impact the answers.
The majority of the time, children were reported to be with at least one parent. About half the time, children were also with at least one sibling. When not with a parent, children tended to be with a sibling, playing or sleeping alone in a room while the parent was in another part of the house, or with another relative. Low-income families were more likely to report that they were with their child at the time of responding to the text survey than were middle- and high-income families. Primary activities reported were mealtime, chores, educational activities, sleeping/resting, play, outside time, and technology use.
Children’s technology use time was primarily occupied by remote schooling (41.4 percent)—non-district educational activities occupied another 18.3 percent of tech time—followed closely by watching TV (36 percent). Parents were significantly more likely to be involved in educational activities on weekdays when they were with their child and less likely to be involved in educational activities on weekends. While parental activities reported did not vary with family income, children’s activities did. Low- and middle-income children were more likely than high-income children to be engaged in mealtime activities. Middle-income children were marginally more likely to be engaged in chores than high-income children. High-income children had higher rates of being outdoors than middle-income children. And low-income children were more likely to be engaged in an activity that included technology than middle- or high-income children.
Overall, caregivers reported high levels of enjoyment in their own activities and even more in their children’s activities. They reported feeling positive emotions a majority of the time (66 percent), followed by overwhelmed emotions (18 percent). Parents only reported a negative emotion in three percent of their responses. Low-income parents reported having more positive emotions than did their high-income peers. In general, caregivers attributed their emotional states, both positive and negative, to unspecified “personal experiences” or to the activity they were participating in at the time of reporting. Their emotional states were attributed to their children’s experiences 21 percent of the time. Parental enjoyment was correlated with children’s enjoyment and with children’s activities deemed to be building social skills. Notably, neither parent nor child enjoyment were associated with activities appraised as building academic skills.
What are we to make of all this? While it is a limited snapshot of a handful of families in the earliest weeks of the pandemic, the positivity expressed almost across the board is very clear. The families surveyed appeared to be going about their lives with as much normality as they could muster given the upheaval around them. If anything, the picture is a somewhat quaint one of generations living and interacting together more often and more fully. If parents were fearful of short- or long-term employment or educational disruptions at the time, it did not register in these detailed surveys. Like most of us, they likely believed the situation to be temporary, even two months into the plague. But when it ceased being temporary—similar research could likely pinpoint the timing and the catalysts—the positivity ebbed as learning loss and economic disruption became obvious and quantifiable.
SOURCE: Shana R. Cohen, Alison Wishard Guerra, Monica R. Molgaard, and Jessica Miguel, “Social Class and Emotional Well-Being: Lessons From a Daily Diary Study of Families Engaged in Virtual Elementary School During COVID-19,” AERA (May 2022).
We know that most American students are suffering from unprecedented learning loss. Student achievement—already below expectations before the pandemic—has dropped to crisis levels.[1] More than ever, high-quality instructional materials (HQIM) are needed in the classroom to get kids back on track.
EdReports—a nonprofit organization that evaluates K–12 curricula—has released a new report that looks at the use of HQIM in the classroom. “State of the Instructional Materials Market 2021: The Availability and Use of Aligned Materials” combines EdReports’s reviews, copyright dates, and data from the RAND Corporation’s American Instructional Resources Survey on curriculum use. It shows that, while teachers broadly recognize the importance of HQIM, adoption in the classroom has lagged behind that recognition.
Why the gap? The first issue is availability of HQIM, which are defined as being aligned to state standards. EdReports finds that barely half (51 percent) of available English language arts materials on the market in 2021 met expectations for standards alignment. (Pre-pandemic—i.e., in 2018—it was 48 percent.) The situation was worse in mathematics: 44 percent of available materials were standards-aligned (up from 31 percent in 2018).
The availability of HQIM obviously limits their use in the classroom. Only 25.6 percent of teachers reported using materials from at least one aligned curriculum per week in ELA in 2021. In mathematics, it was 39.7 percent. As low as those figures were, they still represented improvements from 2018 (14.8 percent and 30.2 percent, respectively).
These findings varied across grades levels, and students were less likely to be taught with HQIM as they matriculated. Less than half (44.9 percent) of elementary school teachers reported using at least one standards-aligned material per week in mathematics, but that was more than double the percentage of high school teachers who reported doing so (21.3 percent). In fact, a clear majority of high school teachers (63.6 percent) reported using “unrelated materials.”
But these figures did not represent educators’ views of HQIM. The vast majority of teachers (73.3 percent) stated that it was “extremely important” that they use materials that are standards-aligned, and another 20.9 percent said it was “somewhat” important. Only 1.3 percent of teachers said it was “not important” at all.
Aside from their availability, what other factors could be holding teachers back from using HQIM? Leadership matters: Less than half (41.9 percent) of teachers reported that their school principal encouraged them to use recommended or required materials. That was less than the percentage of teachers (43.7 percent) who reported their principal telling them to use “whatever materials they thought best.”
There are other obstacles to the implementation of HQIM not mentioned by the report. Not all states have adopted learning standards, making standards-aligned curricula moot. Even if standards exist for ELA and Mathematics, they might not for science, history, and other subjects. Some districts are in “state adoption” locations, meaning that there is no local control over classroom materials. And there might be some unlucky educators who teach subjects for which there are state standards but no aligned curricula on the market. In a nation where education falls under the responsibility of state and local governments, the barriers to adopting HQIM are many and diverse.
Of course, HQIM might not be effective if teachers are not properly trained to use them. Here was another cause for concern, as 22.8 percent of teachers reported receiving no professional learning on how to implement classroom materials. Another 37.7 percent received just one to five hours. Less than a tenth of teachers (9.8 percent) received more than twenty hours of professional learning. And the perceived quality of this professional learning was low. Half of teachers stated that this training “did not prepare me at all” or to a “slight extent.”
EdReports concludes this study with three recommendations: (1) that districts invest in HQIM, which will ensure that all students have access to rich, rigorous instructional materials; (2) that districts dramatically increase the required professional learning in HQIM for both teachers and their principals; and (3) that teachers should participate in the selection of HQIM.
The report suggests that teachers widely recognize the importance of HQIM but are held back by their availability and the lack of institutional support from their schools. Tackling pandemic learning loss will require schools to get serious about adopting and implementing HQIM.
SOURCE: EdReports, “State of the Instructional Materials Market: The Availability and Use of Aligned Materials,” EdReports.org (May 2022).
[1] Just 35 percent of fourth graders were reading at or above a proficient level in 2019. Forty-one percent were proficient or above in mathematics.
On this week’s Education Gadfly Show podcast, Mike Goldstein, founder of Match Education in Boston, a college prep charter school with an embedded Graduate School of Education, discusses his prediction that learning losses will continue without a bold plan to reengage students, teachers, and parents, and address misbehavior. Then, on the Research Minute, Amber Northern reviews a study on how developmental outcomes are affected by gender differences in how children spend their time.
Recommended content:
Feedback welcome!
Have ideas or feedback on our podcast? Send them to our podcast producer Pedro Enamorado at [email protected].
Cheers
Jeers
