Delay between action and feedback determines whether causal agency holds

What it is

Action-effect temporal contiguity is the closeness in time that lets someone attribute a perceptible outcome to their own action. It concerns perceived agency, not whether the system executed a command. A pressed control changing, a new list item appearing, or an object following a drag makes the relation legible as “I did that.” When the effect arrives after an unexplained gap, the same event can instead look as though the system changed on its own.

Why it happens

Temporal proximity is one cue people use to infer causal control, alongside an effect’s expected location and form. With increasing delay, intervening eye movements, attention shifts, animations, and other system events make the earlier input harder to retain as the cause. Direct manipulation is particularly vulnerable: when a finger, pointer, or controller moves without a corresponding visual update, continuous bodily action and discontinuous display events cease to feel coupled. An asynchronous operation need not finish immediately; what matters is that an empty interval does not sever attribution.

Studying it

Researchers vary visual, auditory, or haptic feedback latency after an otherwise identical action, then measure causal attribution, perceived control, duplicate input, and task error. Dragging and tracking tasks reveal sensorimotor asynchrony, while discrete confirmation tasks expose when a single action is misread. A laboratory mean threshold is not a design constant: reversibility, whether users watch the outcome, delay stability, and an earlier acknowledgement all shift acceptance. Interface studies should log end-to-end latency and frame-to-frame jitter separately, since steady slowness and irregular delay produce different experiences.

Where it stops holding

Immediate does not mean latency-free. Rendering, networks, and physical devices impose unavoidable waits, and users may expect a computation, upload, or cross-device synchronization to take time. A prompt, action-specific acknowledgement can preserve the causal link while the eventual result remains pending. Conversely, consequential actions should not display an unverified success merely to feel instant; a truthful in-progress state is safer than a premature commitment.

Applying it

• Map acknowledgement, start of work, and completion separately for every trigger; put evidence of recognition close to the operation.
• Prioritize low, stable display lag for dragging, zooming, and sliders; do not present delayed control as real-time manipulation.
• When waiting exceeds a brief moment, acknowledge first and then continue visibly through progress and completion.
• Review recordings against input timestamps. Repeated taps and remarks such as “did it register?” identify a broken attribution interval, not merely a slow final load.

Related

• Within the group: D1.01.2 Acknowledge input when an immediate result is unavailable · D1.01.3 Missing immediate feedback leads to repeated action
• Adjacent: D1.11 Distinguishing input receipt from result production · D1.14 Feedback duration
• Search terms: action-effect temporal contiguity · sense of agency · latency