We built StudyPulse because we believed that instant, curriculum-specific feedback would help students improve faster than traditional methods. But belief is not evidence. So we ran the numbers.
Over the course of 2025, we analysed 28,279 graded submissions from StudyPulse users, all on authentic VCAA past examination questions spanning 7 years of real VCE exams (2018-2024), across 11 subjects. We compared student performance against official VCAA population data representing 50,000+ students per year.
Here is what we found.
The Headline Numbers
By the end of the study period (November 2025), StudyPulse users were performing at a level consistent with:
- +5 VCE study score points above the population average (from 30 to approximately 35)
- +10 ATAR points above the population mean (from 70 to approximately 80)
- 3x more likely to achieve a 40+ study score compared to the general student population (23% vs 8%)
These are not cherry-picked results. Every valid submission across all 11 subjects was included. No outliers were removed.
How We Measured It
The challenge with measuring student performance is that not all questions are equally difficult. A student scoring 3 out of 4 on a hard question is performing better than a student scoring 3 out of 4 on an easy one.
To account for this, we used z-score normalisation. For every question, VCAA publishes the population mean mark and standard deviation. We converted each student’s raw score into a z-score, which tells us how many standard deviations above or below the population average they performed, adjusted for the difficulty of that specific question.
A z-score of 0 means you performed exactly at the population average. A z-score of +0.77 means you performed 0.77 standard deviations above average, which translates to roughly the 78th percentile.
The Improvement Over Time
We tracked the weekly average z-score of all StudyPulse submissions from August to November 2025.
- Early period (August): average z-score of approximately 0.21 - slightly above the population mean
- Late period (November): average z-score of approximately 0.77 - well above the population mean
That is an improvement of +0.55 standard deviations over the observation period. In educational research terms, this is a medium effect size (Cohen’s d = 0.55), comparable to interventions that typically require significant resources and training to achieve.
The trend was consistent and statistically significant (p < 0.001).
What This Means in Real Terms
For a VCE student, here is what those numbers translate to:
| Metric | Early (August) | Late (November) | Improvement |
|---|---|---|---|
| Study Score | ~31.5 | ~35.4 | +4 points |
| ATAR equivalent | ~73 | ~81 | +8 points |
| Percentile | 58th | 78th | +20 percentile ranks |
An improvement of 4 study score points can be the difference between a B+ and an A, or between getting into your preferred university course and missing out.
The 40+ Question
Every VCE student wants to know: can I get a 40+?
In the general population, roughly 8% of students achieve a study score of 40 or higher (based on normal distribution modelling with mean 30, SD 7).
For StudyPulse users performing at the late-period average, that probability rises to approximately 23%. That is roughly 3 times more likely than the general population.
This does not mean StudyPulse guarantees a 40+. But it does mean that consistent practice with targeted, criterion-referenced feedback shifts the distribution meaningfully in the right direction.
The Subjects
The analysis covered 11 VCE subjects: Physical Education, Legal Studies, Business Management, Health and Human Development, Psychology, Geography, Economics, Religion & Society, Accounting, Biology, and Physics.
All questions were authentic VCAA past examination questions with published marking schemes and population statistics. No third-party or synthetic questions were included.
Why This Matters
Most EdTech products make claims about improving outcomes. Very few show their working.
We published this analysis because we think transparency matters. The full methodology, including our z-score calculation, aggregation method, confidence intervals, and effect size analysis, is available in our research paper (PDF).
The data tells a clear story: students who practise with real past paper questions and receive instant, criterion-referenced AI feedback improve measurably over time. Not because of any magic, but because they are getting the kind of targeted, high-frequency feedback that works - the same kind that Hattie and Timperley’s research identifies as one of the highest-impact interventions in education.
Try It Yourself
Every claim in this article is backed by the data in our published analysis. If you want to see what AI-powered feedback can do for your results, start practising for free.
If you are a school interested in the full research paper or in piloting StudyPulse with your students, get in touch.