Pressure and load interact in an inverted-U — both ends of the curve hurt performance

What it is

Pressure's effect on performance is not linear — it follows an inverted-U curve: performance suffers when pressure is too low (insufficient motivation, low alertness), peaks at moderate pressure, and drops again when pressure is too high. This curve shape has been widely described on its own; the point worth stressing here is that it isn't fixed across tasks — its optimal point shifts with the task's own load. For a low-load, simple task, the optimum sits toward the higher-pressure end — a simple task can absorb more time pressure before performance drops. For a high-load, complex task, the optimum sits toward the lower-pressure end — a complex task starts sliding down the curve at a pressure level that wouldn't even register as excessive for a simpler one. This shift is the real substance of "load interacting with pressure": whether a given amount of pressure counts as moderate is itself a function of load, not something judged independently of it.

Why it happens

The shift happens because a high-load task has already claimed most of the available processing capacity, leaving little slack for absorbing extra pressure — a small additional draw from pressure (monitoring the clock, worrying about consequences) is enough to push an already-thin resource pool into shortfall. A low-load task claims little capacity to begin with, leaving ample slack that pressure can consume without hurting task performance, so it takes a much higher pressure level before performance starts to suffer. Pressure and load compete for the same resource pool, and whichever fills that pool first is what pushes the system into the curve's downward segment — the more load has already claimed, the less room pressure has left to add, which naturally pulls the peak toward the lower-pressure side.

Studying it

Verifying the shift requires setting task load at multiple levels (not just a high/low split), then manipulating several pressure intensities within each load level and observing whether the performance-versus-pressure curve shifts systematically left or right as load level changes. Beyond task performance, dependent measures should include participants' subjective reports of what pressure level felt optimal — if the high-load group reports feeling "too tense" at a lower objective pressure level than the low-load group, that cross-validates the objective performance data and confirms the shift is real rather than an incidental fluctuation in the performance numbers.

Where it stops holding

This curve shift is only reliably detectable when the load difference between conditions is large enough; if the two tasks being compared have similar load to begin with, the difference in curve position may be too small to distinguish from ordinary measurement noise, and shouldn't be over-interpreted. The curve's overall position is also shaped by experience and training: an expert working on a high-load task they're skilled at effectively claims less of their own capacity for that task, so their optimal point sits further toward the higher-pressure side than a novice's would. That means how much pressure counts as excessive can't be judged independently of the operator's skill level — the same deadline can sit in different segments of the curve for a novice versus an expert.

Related

• Same group: A9.15.1 moderate time pressure raises alertness and engagement, but excessive pressure crowds out processing capacity · A9.15.3 high-load tasks are more prone to strategy degradation under time pressure · A9.15.4 under pressure, actively strip extraneous load from the interface
• Nearby: A5.09 Cognitive tunneling
• Search terms: Yerkes-Dodson law · inverted-U · arousal-performance