Executive summary
- 2,169 students scored across Foundation and Year 1 and grouped into EOY screener theta tiers within each year level: Lower 10% (
<=P10), next 15% (P10–P25), top 75% (>P25).
- In overlap samples, students in the lower two tiers (
<=P25) have higher odds of low outcomes: PAT <=P25 odds ratio 4.3–5.7 and teacher rating <=2 odds ratio 6.3–10.7 (each compared with the top tier >P25).
- External corroboration is supportive but coverage-limited: 179 students in the PAT overlap sample (8 schools) and 1,049 in the teacher overlap sample (25 schools).
EOY screener theta tiers are defined from the EOY screener theta distribution within each year level (P10 and P25 boundaries). PAT and teacher outcomes are used to check patterns in the overlap samples; they do not set the boundaries in this release.
EOY screener theta tiers (P10/P25 within year level)
EOY screener theta tier boundaries (P10 and P25)
| Foundation |
stan_pcm_v4_longitudinal |
Joint longitudinal IRT model; student EAP theta |
-1.1848 |
-0.5698 |
1099 |
| Year 1 |
stan_pcm_v4_longitudinal |
Joint longitudinal IRT model; student EAP theta |
-1.2260 |
-0.5044 |
1070 |
Notes:
- P10 and P25 are calculated within each year level. If the percentile falls between two students, the boundary is interpolated between their theta values.
- Tiering rule:
theta <= P10 (Lower 10%), P10 < theta <= P25 (Next 15%), else theta > P25 (Top 75%).
EOY theta distribution and tier boundaries
External corroboration (overlap samples)
External outcome data coverage
| PAT-only |
17 (2%) |
36 (3%) |
| Teacher-rating-only |
463 (42%) |
388 (36%) |
| Both |
78 (7%) |
120 (11%) |
| None |
541 (49%) |
526 (49%) |
Of 47 EOY schools, only 8 provided PAT data and 25 provided teacher ratings. Missingness is school-driven (entire schools either report or do not), not student-driven within schools.
Within each school (>=5 students with outcomes and >=5 without), we compare mean BOY screener theta for students who do vs do not have EOY outcome data; the average absolute gap in these school-level means is 0.027 theta (small).
Flow from EOY tiers to outcome bands (overlap samples)
Foundation
PAT crossover table (Foundation; n (row%))
|
EOY tier
|
<=10
|
11–25
|
>25
|
Total
|
|
<=P10
|
0 (0%)
|
2 (29%)
|
5 (71%)
|
7
|
|
P10–P25
|
2 (20%)
|
2 (20%)
|
6 (60%)
|
10
|
|
>P25
|
2 (3%)
|
5 (8%)
|
56 (89%)
|
63
|
Year 1
PAT crossover table (Year 1; n (row%))
|
EOY tier
|
<=10
|
11–25
|
>25
|
Total
|
|
<=P10
|
2 (18%)
|
4 (36%)
|
5 (45%)
|
11
|
|
P10–P25
|
1 (6%)
|
9 (50%)
|
8 (44%)
|
18
|
|
>P25
|
6 (9%)
|
6 (9%)
|
58 (83%)
|
70
|
Foundation
Teacher rating crossover table (Foundation; n (row%))
|
EOY tier
|
1
|
2
|
3
|
4
|
5
|
Total
|
|
<=P10
|
8 (14%)
|
11 (19%)
|
36 (61%)
|
4 (7%)
|
0 (0%)
|
59
|
|
P10–P25
|
5 (6%)
|
11 (14%)
|
55 (69%)
|
8 (10%)
|
1 (1%)
|
80
|
|
>P25
|
5 (1%)
|
15 (4%)
|
272 (68%)
|
89 (22%)
|
21 (5%)
|
402
|
Year 1
Teacher rating crossover table (Year 1; n (row%))
|
EOY tier
|
1
|
2
|
3
|
4
|
5
|
Total
|
|
<=P10
|
9 (20%)
|
19 (42%)
|
14 (31%)
|
3 (7%)
|
0 (0%)
|
45
|
|
P10–P25
|
3 (4%)
|
33 (39%)
|
41 (48%)
|
8 (9%)
|
0 (0%)
|
85
|
|
>P25
|
3 (1%)
|
28 (7%)
|
225 (60%)
|
108 (29%)
|
14 (4%)
|
378
|
PAT norm-covered validation (Adaptive excluded)
PAT by EOY tier (overlap sample)
| Foundation |
<=P10 |
7 |
86.6 |
3.3 |
0.0% |
28.6% |
| Foundation |
P10–P25 |
10 |
92.5 |
15.1 |
20.0% |
40.0% |
| Foundation |
>P25 |
63 |
93.5 |
6.9 |
3.2% |
11.1% |
| Year 1 |
<=P10 |
11 |
93.5 |
7.6 |
18.2% |
54.5% |
| Year 1 |
P10–P25 |
18 |
95.5 |
9.3 |
5.6% |
55.6% |
| Year 1 |
>P25 |
70 |
100.7 |
9.3 |
8.6% |
17.1% |
In the PAT overlap sample, the lower-theta tiers have higher rates of PAT below ACER P25 than the top tier. For example, Foundation <=P10 is 28.6% (n=7) vs 11.1% in >P25 (n=63); Year 1 <=P10 is 54.5% (n=11) vs 17.1% in >P25 (n=70). Treat PAT results as directional because overlap is limited and some tier counts are small.
Teacher validation
Teacher ratings by EOY tier (overlap sample)
| Foundation |
<=P10 |
59 |
2.6 |
32.2% |
| Foundation |
P10–P25 |
80 |
2.9 |
20.0% |
| Foundation |
>P25 |
402 |
3.3 |
5.0% |
| Year 1 |
<=P10 |
45 |
2.2 |
62.2% |
| Year 1 |
P10–P25 |
85 |
2.6 |
42.4% |
| Year 1 |
>P25 |
378 |
3.3 |
8.2% |
Teacher ratings show a clear gradient by EOY tier in a larger overlap sample. For example, the share rated low (1–2) is 32.2% in Foundation <=P10 (n=59) versus 5.0% in >P25 (n=402); in Year 1 it is 62.2% (n=45) versus 8.2% (n=378).
Summary statistic (<=P25 vs >P25)
Odds ratio: lower two tiers (P25)
| PAT <=P25 |
Foundation |
4.3 |
| PAT <=P25 |
Year 1 |
5.7 |
| Teacher rating <=2 |
Foundation |
6.3 |
| Teacher rating <=2 |
Year 1 |
10.7 |
School-blocked robustness (LOSO)
LOSO stability summary (all source strata)
| pat_low25 |
Foundation |
3 |
3 |
33.3% |
| pat_low25 |
Year 1 |
4 |
4 |
100.0% |
| teacher_low |
Foundation |
19 |
18 |
88.9% |
| teacher_low |
Year 1 |
20 |
18 |
83.3% |
% (<=P25) > (>P25) near 100% means that when each school is held out in turn, the remaining schools still show the expected ordering: students in the lower two EOY tiers (<=P25) have higher outcome rates than students in the top tier (>P25).
- PAT LOSO covers 3–4 held-out schools (depending on year level) with 3–4 valid folds — treat as directional stress testing given the small school count.
- Teacher LOSO covers 19–20 held-out schools with 18–18 valid folds — broader coverage makes this more informative for transportability.
Visual checks: EOY screener theta vs outcomes
The figures below show the relationship between EOY screener theta estimates and external outcome measures in the overlap samples.
Decision implications
- This snapshot supports checking whether outcomes (PAT, teacher ratings) show the expected ordering across EOY screener theta tiers in overlap samples.
- Outcome coverage is incomplete and school-driven; treat overlap-based results as corroboration, not a basis for setting final thresholds.
- This snapshot does not support national prevalence/attainment claims or high-stakes individual decisions on its own.