Single-Subject Experimental Design

Description

Establishes causal effects within individual users by repeatedly measuring behaviour across baseline and intervention phases (e.g. ABAB, multiple baseline).

Single-subject experimental designs establish causality at the level of the individual by repeatedly measuring a behaviour and systematically introducing and withdrawing an intervention. In an ABAB reversal design, baseline (A) and intervention (B) phases alternate; in a multiple-baseline design, the intervention is introduced at staggered times across participants, behaviours or settings. If behaviour changes reliably each time the intervention is applied, alternative explanations become implausible. In UX this suits accessibility work, highly individual users and behaviour-change features. Rigour comes from stable baselines (at least three to five data points per phase), visual analysis of level, trend and variability, non-overlap effect sizes, randomisation of phase starts where possible, and replication across participants.

Attributes

Loss type
Siloed
Stage
Evaluation
Effort
Weeks

Steps

  1. AB design for UX behavior change:
  2. Baseline phase (A): measure the target behavior (e.g., task completion time, error rate) across multiple sessions without the design change.
  3. (2) Treatment phase (B): introduce the design change. Continue measuring.
  4. (3) Analysis: plot the data. Visual inspection: is there a change in level (immediate shift), trend (gradual change), and/or variability at the phase change point?
  5. (4) ABAB extension: withdraw the treatment (return to A) and reintroduce (B). Reversal of the behavior confirms the causal effect of the design change.

Sources

  • Research Methods in Applied Behavior Analysis — (various)

Also known as

  • Single-case experimental design
  • N-of-1 design
  • ABAB reversal design

Categories

  • Analytics & Behaviour
  • Experimentation & Causal Inference