The shorter the interval between two stimuli, the greater the delay in the second response

What it is

The psychological refractory period effect follows a specific quantitative rule: the shorter the interval between two stimuli (stimulus onset asynchrony, SOA), the longer the second response is delayed, in an approximately inverse relationship — every bit the SOA shrinks, the delay to the second response grows correspondingly, until the SOA gets short enough that the delay levels off at a ceiling close to the processing time the first response itself requires.

Why it happens

This inverse relationship is a direct reflection of what queuing means here: the second stimulus's central processing cannot start until the first stimulus's central processing releases the bottleneck, and the wait is exactly however much time is left on the bottleneck being occupied by the first task. The shorter the SOA, the more of that occupied time remains when the second stimulus arrives at the bottleneck, so the wait is naturally longer. When SOA shrinks toward zero, the second stimulus arrives at the bottleneck almost simultaneously with the first, and the wait approaches the full processing time the first response requires.

Studying it

By systematically setting a series of decreasing SOA values and measuring the corresponding second-response RT, researchers plot RT2 against SOA. The typical result is a curve that rises as SOA decreases and flattens out once SOA is large enough (at which point the two tasks' processing no longer overlaps and produces no extra delay); the slope of the curve and where it flattens can be used to infer where in the processing pipeline the bottleneck sits.

Where it stops holding

The specific slope and ceiling of this inverse relationship depend on how complex each of the two tasks is — the more complex the task, the greater the room for delay and the higher the ceiling — so there is no single value that applies across all task pairings. When one of the two tasks is highly automatized and barely occupies the central bottleneck, this inverse relationship between SOA and delay can weaken substantially or disappear.

Applying it

If a product genuinely needs to present two prompts in quick succession that both require a user response, this quantitative rule can be used to estimate how short the interval can get before the second prompt's perceived delay increases noticeably — working backward to a floor below which further compression stops making sense, avoiding the trap of squeezing two prompts closer together in pursuit of a "faster" interface only to make the second prompt's experience worse. How to check: measure the second response's RT under several different real interval lengths for this product scenario, plot delay against interval, find the inflection point where delay starts growing noticeably, and keep the interval above that point.

Related

• Same group: A8.04.1 When the first response isn't finished, processing of the next stimulus is postponed · A8.04.3 Rapid successive prompts queue in the central system rather than being processed in parallel · A8.04.4 This limit cannot be trained away
• Nearby: A8.01 The structure of reaction time
• Search terms: stimulus onset asynchrony · SOA · PRP curve