Intention-to-treat analysis
Analyze everyone in the group they were assigned to, regardless of compliance.
What it means
Intention-to-treat is an analysis principle for randomized trials in which participants are evaluated according to the group they were randomized into, irrespective of whether they completed, switched, or adhered to the treatment. By preserving the original randomization, it protects against the selection bias that creeps in when dropouts and non-compliers are excluded, since those who deviate often differ systematically. The trade-off is that ITT estimates the effect of being offered or assigned a treatment in realistic conditions, typically a more conservative figure than the effect under perfect adherence. It is contrasted with per-protocol analysis, which restricts to compliant participants and thereby sacrifices the integrity of randomization.
Examples
Counting patients who quit a demanding diet as part of the diet arm, rather than dropping them, to keep the groups comparable.
A vaccine trial counts everyone assigned to the vaccine, including those who never came back for the second dose, so the result answers what happens when you offer the programme.
A tutoring evaluation keeps the pupils who never showed up in the tutoring arm; drop them and only the keenest remain, and the programme flatters itself.
First described in Clinical-trials methodology; codified in CONSORT guidelines.