By Aaron Hepburn
This is the second of two posts on reading proficiency and growth together. Part one, “What Proficiency Alone Will Never Tell You,” explains how the charts for reading these data points together are built.
The chart goes up on the screen and the room gets quiet for about eight seconds.
Then somebody finds their grade level. That is always the first move, and it is worth expecting, because what happens in the next two minutes decides whether the rest of the day is useful. There are two paths from here. If the conversation turns into an explanation of why one grade or subject area's number looks the way it does, the day is gone. If it turns into curiosity about why the grades are separated at all, you get a plan out of it.
I run these reviews with district leadership teams, and the single thing that most reliably produces the second path is showing three years instead of one.
A single-year number invites a defense. Somebody in the room taught that grade, hired that team, or built that schedule, and a number on a screen with no history attached functions as a verdict. The instinct to explain it is not defensiveness for its own sake. It is a reasonable response to being judged on one data point.
Part 1 makes the case for reading two measures side by side. Proficiency is a status measure: what share of students met the standard. A student growth percentile, or SGP, says how far a student moved compared with students across the state who started in the same academic place, where 50 is a typical year.
Reading three years of both is what changes the sentence. “Grade 6 is low” is an accusation. “Grade 6 has drifted down in both proficiency and growth for three consecutive years” is a description of a system, and no single person owns three years of drift. It also survives the most common and most legitimate objection in the room, which is that one year was unusual. Of course it was. That is why we are looking at three.
One of the districts below gives a clean example. In the middle of a three-year window, one grade level posted a median SGP of 25. On its own that number starts a bad meeting. Read across three years and it sits between 57 the year before and 44 the year after, which makes it a single hard year rather than a pattern. The question shifts from what is wrong with that grade to what happened that year, and those two questions send a team to completely different places.
Two shapes are worth telling apart. A collapse is a sudden drop. Drift is a slow decline across several years. They usually have different causes, different fixes and different owners, and a single year cannot tell you which one you are looking at.
The other thing the two-measure view does is spread the finding around. Plot proficiency against growth and the strongest grades and the struggling grades both end up with a question attached.
In one district the highest-scoring grade in the building, fifteen points above the state average, carried the sharpest question of the day: is the work demanding enough for students who already arrive capable? In another, a grade sitting below the state average turned out to be growing faster than typical, which is evidence of something working that no proficiency report would ever have surfaced.
Once everybody has a question, nobody is on trial.
We ask three questions of every point on the chart, and we ask them in this order.
The third question is where the real work starts, because it converts a hunch into something observable in a classroom.
One team said they were surprised that the strongest grade in the building was growing the slowest, and offered that the work might not be demanding enough for the students in it. So what would have to be true? Text at or above grade-level complexity. Students writing at length about what they read. Tasks that ask them to transfer rather than repeat. All three are things you can walk into a room and look for. That is the difference between a theory and a next step.
What follows is the same read applied to two districts that look nothing alike. One tests above its state average and one below. Both came out of the day with four priorities, and the priorities are not the same.
District A is the district Part 1 is built on:
District A had a genuine improvement story. District-wide proficiency had climbed in each of the last five years. Leadership said so, publicly and accurately.
Staff described something else. In survey comments and focus groups, the same theme kept surfacing: materials that did not connect across grade bands, gaps individual teachers were filling on their own, professional learning that had changed how people thought without changing what happened in every classroom. One comment summed up the pattern: teachers were expected to fill the gaps themselves.
Two narratives, both supported by evidence, apparently contradicting each other. That contradiction is where most planning conversations stall, because there is no way to settle it with proficiency data alone. Rising proficiency is exactly what you would expect from a district where committed teachers are compensating individually for a system that is not coherent. Those gains are real. They are also fragile, and they are not evenly distributed.
Growth data broke the tie. The grade levels where proficiency was rising fastest were not the grades with the strongest growth. The clearest case was the highest-scoring grade in the district, fifteen points above the state average, whose growth fell below typical for the first time in three years after two years above it. The pattern was consistent with the staff account rather than the leadership account, and once the leadership team saw that, the argument stopped being an argument.
What came out of that day was four priorities, in a deliberate order:
None of these are new ideas in education. What made them stick is that each one traced back to a specific pattern in the district's own data that the room had looked at together.
District B looks nothing like District A on the surface. Its district-wide proficiency runs about six points below the state average, and most of its grade levels sit on the left half of the chart. That is the more common starting point, and it is why this is the chart most districts will recognize.
Read it the way Part 1 describes. The hollow marker is where the grade level sat three cycles back. The solid marker is the most recent year. The line between them is the story. Read the solid markers first to see where each grade stands now, then come back for the lines.
Read that way, across the same three years, the district separates into four groups that each need something different.
The idea that those four patterns point to distinct underlying causes comes from the Cohesive Learning framework developed by We Do Instruction, also known as Cohesive Schools, which names access to core instruction, the strength and rigor of core content, and whether learning gets solidified over time. Part 1 credits it at more length. We use it as one lens among several.
One grade sat at 90 percent proficient with a median SGP of 75.5. High on both measures. The response there was protect and replicate: figure out what is happening in that grade and whether any of it travels.
Three consecutive grades sitting at the transition between the elementary program and the secondary one were all growing in the low forties, and two of them carried the lowest proficiency in the district. They became the district's coherence priority for the year.
One grade finished at 72 percent proficient, mid-pack for the district, with an SGP of 31, the lowest anywhere in the building. That one went on the list as an open question rather than a priority. Was this a strong cohort arriving, or a transition effect the district was creating? Nobody in the room knew, and saying so was the right answer.
Three more grades were growing faster than typical while their proficiency stayed below the state average. Students were moving, and the movement was not yet accumulating. That points at retention and cumulative review, which is a different challenge with a different fix. It is also the group a proficiency-only review writes off for not being proficient, when the growth is the strongest evidence in the building that something is working.
A subgroup view carried the sharpest finding. For students with individualized education programs, proficiency had risen from 13 to 19 percent over five years, which reads as progress. Median growth for that group began and ended those five years at 37, and never once reached typical. Rising proficiency alongside five years of below-typical growth is not progress in the way it first appears, and it moved the district's response from intervention design toward access to core instruction.
The cleanest argument for reading both measures sits in two dots at the same horizontal position. Grades 7 and 9 both finished at 72 percent proficient. Grade 7's median SGP was 62. Grade 9's was 31. On a proficiency report those two grades are identical and neither draws a second look. They are not the same grade level, they do not have the same challenges, and they should not get the same response.
Adding the three-year tails sharpens two of those readings and surfaces a grade the single-year view missed completely.
Grade 8 did not arrive at 90 percent. Three years earlier it sat at 75 percent with growth of 52, which is to say typical. Its growth has climbed every year since and its proficiency is fifteen points higher. That is the district's clearest evidence that something specific is working, and it is worth understanding while the people who built it are still in the building.
Grade 9 is the one that changes most under a three-year read. Its proficiency has risen fourteen points, which is the kind of number a district puts in a newsletter. Its growth across the same three years ran 42, then 24, then 31, below typical every year. Rising proficiency with persistently low growth is the same signature District A found in its own grade 8. In both districts it stayed invisible until the two measures sat on one page.
Grade 5 is the quiet one, and it is the drift case. Growth fell from 63 to 53 to 44.5 while proficiency slipped five points. No single year would have triggered a conversation. Three years running, both measures point down.
Four groups, four different instructional responses, and a room that could see why the responses had to differ.
Two moments from District B are worth naming, because they are the kind of thing that only happens once a team can see the whole picture at once.
The first was subtraction. Laying out the year's planned work on one page surfaced three separate classroom observation tools, built by two different teams and due on three different dates. The team's own question was the right one: is this one tool built once, or three tools with three purposes? They chose one.
The second was sequence. The original plan called for a multi-tiered support system, an observation system, new secondary materials, and a vertical standards alignment, with most of it starting at once. The team re-dated each initiative deliberately, on the reasoning that a tiering system built before Tier 1 instruction is consistent will spend its life sorting students rather than teaching them. Sequencing was the workload protection, and re-dating was treated as permission rather than failure, as long as a reason was recorded.
Neither decision came out of the data directly. Both came out of a room that had stopped defending the data and started questioning it.
Named owners, in writing. Every priority carries one person who owns it from start to finish, and one person responsible for telling the right people once it is finished. “Complete” means finished and communicated, not finished alone.
Honest dates. A date nobody believes is worse than a later date everybody does. When a date moves, the reason gets recorded. That single habit is what keeps a plan from quietly becoming fiction by January.
A short list. Four priorities, not fourteen. Every district we work with can name more than four real problems. The districts that make progress pick the ones their own data ranked highest and let the sequence carry the rest.
Part 1 of this series, “What Proficiency Alone Will Never Tell You,” walks through the mechanics: what a growth percentile is, why one year of data is not enough to act on, and how to build the two-measure chart from your own state report card in about an hour. It is worth doing before you bring anyone else into the room, so you know what you are looking at.
Then bring your leadership team in, and start with what they notice, not with what you concluded.
If you would rather have someone facilitate that first conversation, that is work we do with districts. Send me a note (aaron@compass-pd.com) and we will look at your data together.
Betebenner, D. W. (2009). Norm- and criterion-referenced student growth. Educational Measurement: Issues and Practice, 28(4), 42–51.
City, E. A., Elmore, R. F., Fiarman, S. E., & Teitel, L. (2009). Instructional Rounds in Education: A Network Approach to Improving Teaching and Learning. Harvard Education Press.
Cohesive Learning framework, including Academic Momentum and the Three Factors: We Do Instruction (Cohesive Schools), wedoinstruction.com. Used with attribution. The framing of access to core instruction, strength of core content, and solidifying learning over time is theirs.
Aaron Hepburn is Co-Founder of Compass PD, LLC, where he leads the analytical data reviews Compass PD conducts with partner districts.