AB
1
COMPILATION OF THE DATA
2
ASSUMPTIONS AND PERTINENT DETAILSThe system-level data, in its original raw form exported directly from ACER, can be viewed in the tab at the bottom of the page.
Please note, however, that compilations have not been done for 2018, as only one school was included in the data export for 2018, and the inclusions were from March of that year - outside of the norm sample collection point and so, therefore, misleading data.
3
In compiling the data, the primary references for system achievement that have been used are:
- Average Percentile rank
- Percentage of students above the 50th percentile
- Average Scale Score (compared to norm sample group)
- Achievement distribution across percentiles, for each test (1-8), vs normative distribution.
Other references available (not in graph form) include average stanine, and lowest and highest percentile ranks and rate of incidence of these across a cohort.
These are presented for each calendar year in the tabs at the bottom of this sheet, as well as a 2019-2021 comparative tab, highlighting achievement over time.
4
TESTS TAKEN VS YEAR LEVEL OF STUDENTS
5
Given that there are 36000 test results to compile, and limitations existed in the exported data from PAT, there is no way of knowing if the tests taken by students at every school are those that are congruent with their year level. Given that it is now the most common practice to give all students their year-level test, the assumption has been made that Test 1 = Year 1. This means that the percentile rank and stanine information that have been used to measure achievement are those at the year level of the test in question.
For this reason, error bars have been used in the graphs to indicate that there is likely to be a small possibility for variance in acknowledging that a small percentage of schools give tests that are not congruent to a child's year level.
When the raw data was delivered, it came with percentile ranks and stanines for the year level congruent with the test, as well as those one year above and below. These one-year-above and below indications have been removed so as to only count the percentiles and stanines congruent to the year level. If all were left in, these measures would have skewed all data for all tests and been counted three times for the one test taken.
6
In the full export, entries were seen from many schools multiple times in a calendar year, meaning that PAT R had been delivered more than once.
For validity reasons, any test taken outside the norm reference window (October - November) has been removed, as the percentiles and scale scores are not applicable when a test has been taken, for example, in May. We have provided a two week grace period either side of the norm reference window, to allow for students leaving to go on holiday or students returning from absence.
7