Attention resources aren't one pool — they split into pools by channel, stage, and response

What it is

Multiple resource theory (Wickens) holds that the attention or processing resources available when someone handles several tasks at once aren't a single, undifferentiated pool — they're spread across several relatively independent dimensions: perceptual modality (visual vs. auditory), processing stage (perceptual/cognitive encoding vs. response execution), processing code (spatial vs. verbal), and response modality (manual vs. vocal). This theory answers a question about how resources get allocated across tasks and channels in dual-task or multitasking situations — a different level from the one cognitive load theory addresses. Cognitive load theory is about how a single task's working-memory budget gets split among intrinsic, extraneous, and germane load; multiple resource theory is about whether two or more concurrent tasks draw on the same resource dimension or different ones. Both frameworks use the word "resource," but they're answering entirely different questions.

Why it happens

If attention really drew from one single capacity, the interference between any two tasks should depend only on their combined resource demand — overall difficulty — regardless of which specific channels they use. But experiments repeatedly find that two demanding tasks placed on different channels (a purely visual judgment paired with a purely auditory one) interfere far less than two moderately demanding tasks sharing the same channel. That pattern only makes sense if resources themselves are split by dimension, with dimensions not competing against each other: each dimension has its own independent supply ceiling, so two tasks on different dimensions largely run in parallel without queuing, while two tasks on the same dimension genuinely compete for one limited supply.

Studying it

The paradigm is a dual-task combination experiment: systematically vary whether two tasks overlap in perceptual modality, processing stage, and code, while using single-task performance to control each task's difficulty level, then compare dual-task cost across overlap combinations — often plotted as a resource-demand-versus-difficulty curve to check whether the effect of raising difficulty is clearly smaller than the effect of dimensional overlap. Methodological note: single-task baselines must first confirm the two tasks are matched in difficulty, or a difficulty gap will be confounded with the effect of channel overlap, making it impossible to tell whether interference comes from "harder" or from "more conflicting."

Where it stops holding

The theory only names a handful of broad resource dimensions (modality, stage, code, response) — it doesn't claim every specific task has its own exclusive resource. Even when two tasks don't overlap on any of these listed dimensions, residual interference can still appear from shared central coordination or sequencing demands (frequently deciding which task to handle first); interference doesn't automatically drop to zero just because dimensions are separated.

Applying it

Before designing two information streams or tasks a user must handle at the same time, "score" each of them against this dimensional framework — tag which perceptual modality, code, and response modality each occupies. That's the diagnostic starting point for judging whether they can run in parallel and how to allocate resources, not a design decision by itself. Verification: for a given pair of tasks, measure each one's single-task baseline and the drop in performance when they run together; if the degree of dimensional overlap identified in the diagnosis tracks the size of the measured interference, the framework is a valid diagnostic tool for that situation.

Related

• Same group: A9.10.2 tasks sharing a resource pool interfere more than tasks using different pools · A9.10.3 vision and hearing are separate perceptual resources, and speech and manual response are separate response resources · A9.10.4 multiple resource theory explains why dual-task interference can't be predicted from task difficulty alone · A9.10.5 designing parallel tasks should deliberately assign them to different resource pools to reduce interference
• Nearby: A9.06 intrinsic, extraneous and germane load (how three load types share one working-memory budget inside a single task — a different level from this group) · A5.02 divided attention and dual-tasking (the observed cost and behavior of dual-tasking)
• Search terms: multiple resource theory · Wickens · dual-task interference