Evidence Engine

Method & weights

Nothing here is a black box. Every weight that goes into a grade is published below. Back to the engine.

1 · Design weight

Study designs differ in how much they can tell us. Each design has a base weight:

Randomized controlled trial1.00
Cohort study0.60
Observational study0.35
In-vitro / lab study0.15

2 · Sample-size factor

Bigger studies count for more, with diminishing returns (log scale). Final study weight = design weight × size factor.

n = 1×0.40
n = 10×0.60
n = 100×0.80
n = 1,000×1.00
n = 10,000×1.20

3 · Result direction

A study's result signs its weight — supporting evidence pushes the grade up, null results push it down, mixed results add weight but no direction (which dilutes support).

Supports the claim+1
Mixed / partial0
Null result (no effect)-1

4 · How the score is built

score.formula
score = 100 × support × strength × consistency
support   = (signedWeight / totalWeight + 1) / 2
strength  = 1 − e^(−totalWeight / 1.0)
consistency = 0.55 + 0.45 × agreement
single-study cap = 84 (one study can never reach A)

5 · Grade bands

A90 – 100
A-85 – 89
B+80 – 84
B75 – 79
B-70 – 74
C+65 – 69
C60 – 64
C-55 – 59
D+50 – 54
D40 – 49
D-0 – 39