Force reproduction error grows sharply without feedback

What it is

Ask someone to reproduce a specific force purely from internal bodily sensation — with no display showing the actual force being applied — and the error between the reproduced force and the target will be considerably larger than when guided by feedback. This isn't because people have no sense of force at all; rather, that internal sense has limited precision on its own, and without external feedback to calibrate it, the resulting error grows to a degree that's easy to overlook in everyday use.

Why it happens

A person's sense of how much force they're currently exerting comes largely from the efference copy the brain generates when issuing a motor command — also called the sense of effort — rather than from a precise readout of actual force by muscle and tendon receptors, which are used mostly for reflexive regulation and rarely feed directly into a channel available for deliberate, fine-grained conscious comparison. This means a person's judgment of "how much force am I using right now" is essentially an estimate of the strength of their own command, not a direct reading of the actual output — an estimate inherently prone to drift from accumulating fatigue, postural change, or divided attention. Without external feedback telling the user how far the actual output has strayed from the target, that drift has no chance to be corrected, which is why reproduction error without feedback is markedly larger than with it.

Studying it

A common paradigm is a force-matching task: participants first experience a target force, then all feedback about the current force (numeric display, sound, haptic cue) is withdrawn, and they're asked to reproduce the target relying only on memory and their present sense of effort; reproduction error with and without feedback is then compared. The independent variables are the presence and form of feedback (visual numeric, auditory, haptic confirmation); the dependent variables are the magnitude and directional bias of the reproduction error (systematically under- or over-shooting, for instance).

Methodological note: fatigue accumulates over the course of the experiment and systematically shifts the mapping between effort and actual force output. In a long force-reproduction study, failing to control for accumulated fatigue means the observed increase in error may partly reflect fatigue rather than the absence of feedback itself — block rotation or a cap on single-session duration is usually needed.

Where it stops holding

This applies to tasks that require reproducing a specific absolute force value. If a task only requires judging relative magnitude — lighter or heavier than before — rather than reproducing an absolute value, people's ability to discriminate relative differences is considerably more reliable, and the degradation without feedback is much smaller. Also, people extensively trained on a specific type of force application (professional instrument playing, or a particular manual trade) develop a more stable mapping between effort and output through extensive calibration than untrained individuals — this entry describes the general case of someone without such specialized training.

Applying it

• Any interaction that relies on users applying a specific force purely by feel, with no screen or sound indicating the actual force (a blind pressure-sensitive key, a force knob hidden inside a device housing), should never require reproducing a precise force value — it should only distinguish coarse relative levels (light/medium/heavy).
• Interactions that genuinely need precise force must provide real-time external feedback (a numeric visual readout, a clear haptic confirmation point, an audible cue) — never assume users can self-calibrate purely from internal effort sensation.
• To verify: have users repeatedly attempt to hit the same target force with zero feedback, and measure how scattered these attempts are. If the scatter exceeds what the interaction can tolerate, external feedback must be added — a manual or a one-time demonstration isn't enough to build a stable sense of force.

Related

• Same group: A8.11.1 People have a limited ability to grade output force · A8.11.2 Relative force-control precision worsens as the fraction of maximum force rises · A8.11.4 The usable number of pressure levels falls far short of sensor resolution · A8.11.5 Sustaining a given force level is harder than reaching it
• Nearby: C10.12 Pressure input and force grading · A8.07 Closed-loop control
• Search terms: sense of effort · efference copy · force reproduction