Educational Statistics

Central tendency, dispersion, correlation and the normal curve, as this paper asks them.

Where candidates lose marks on this

1. Averaging two group means without weighting by group size.
2. "Adding a constant changes the SD." It does not. Only multiplication rescales spread.
3. Reading skewness from the hump. The tail names the skew; the mean follows the tail.
4. Higher CV = more consistent. Inverted — LOWER CV is more consistent.
5. Squaring reliability / reading correlation directly. Reliability is read directly; correlation is squared (r²) for explained variance.
6. Treating r = −0.8 as weaker than r = +0.6. Strength is the absolute size.
7. P₉₀ = "scored 90%". No — 90% of the group lies below that score.
8. Using Wechsler SD 16 or Binet SD 15. It's the reverse: Wechsler 15, Binet 16.
9. AQ computed as MA/EA. It is EA over MA — achievement against ability.
10. Forgetting z can be negative — a score below the mean must produce a negative z; T handles it (T = 50 + 10z stays positive).

Where this is taught

Start your VP & Principal 2026 preparation → How the exam is marked