Base120 · Decomposition

Variable Control & Isolation

DE · Decomposition

Hold factors constant to measure single variable's causal impact

Model code
DE14
Transformation
DE (Decomposition) — Break problems into factors, layers, and independent subproblems.
Priority
14 of 20
Framework
HUMMBL Base120 — 120 mental models across 6 transformations

When to use

"Use when you need to measure the causal impact of a single variable. Hold all other variables constant, vary only the target, and measure the difference. The foundation of experimental science."

Example

"To test whether a new onboarding flow improves retention, hold everything else constant (same traffic source, same time period, same user segment) and vary only the onboarding flow. The difference in retention is the causal impact of the flow."

Common misuse

"Failing to hold variables constant and attributing the result to the variable you varied. If traffic source changed at the same time as the onboarding flow, the retention difference could be due to either."

Also known as

Controlled ExperimentVariable Isolation

Related models

DE9DE17RE11