Control-Systems Behaviour Model · Interactive Simulator

COM-B as a fluid system: pushing capability, opportunity & motivation

A live illustration of the capability–opportunity–motivation–behaviour (COM-B) framework rendered as a control-systems fluid analogy. Drive the intervention levers and watch the construct "tanks" fill, feedback loops kick in, and physical-activity and sedentary-behaviour outputs respond over time.

Read this first. The underlying paper is a conceptual model — it specifies the structure (tanks, flows, first-order dynamics) but leaves the coefficients symbolic. Here the relationship strengths between constructs are anchored to published COM-B→activity path coefficients, while time constants and feedback gains are illustrative and adjustable. This is a teaching / hypothesis tool for exploring dynamics — not a validated, person-level predictor.

day 0 · speed

Behaviour outputs

Live level (0–100). Steady state = where the system settles if levers are held.

Physical activity

±0 vs baseline

steady state → 30

Sedentary behaviour

±0 vs baseline

steady state → 70

Trajectories over time

Physical activity and sedentary behaviour as the system evolves. Disturbances and feedback show up here.

Physical activity Sedentary behaviour Capability Motivation

Intervention levers

Each lever is a behaviour-change input feeding one construct. Push it up and watch downstream.

Baseline Capability-focused Motivation-focused Opportunity-only Multi-component Relapse scenario

Capability input

Skills, know-how, self-regulation training0

e.g. action planning, self-monitoring, problem-solving (behavioural-regulation domain)

Opportunity input

Social support + environment restructuring0

e.g. active workstation, supportive peers, prompts/cues (social & physical opportunity)

Motivation input

Goal setting, feedback, incentives, identity0

e.g. goal-setting, reward, self-identity work (reflective & automatic motivation)

Unexplained variability & disturbances

The COM path coefficients explain only ~50% of activity variance. These controls make the missing half tangible — each maps to a control-systems failure mode.

Intention–behaviour gap46%

Attenuates the motivation→behaviour path (a "leaky valve"). ~46% of intenders don't act.

Habit / learning feedback0.15

Gain of the behaviour→capability loop. High = virtuous cycle; low = snaps back after a shock.

Chronic environmental drag0

Persistent disturbance (desk job, weather, caregiving) pushing activity down / sitting up.

Behavioural noise6

Day-to-day randomness (unmeasured context). Real behaviour is never a smooth curve.

Re-draws each path coefficient from its published 95% CI — the population average is not any individual.

Model & parameter provenance

Steady-state relationship strengths (β) come from a published prospective COM-B path model; dynamics are illustrative.

Path	β	Source
Capability → Motivation	0.81	published ^[1]
Opportunity → Motivation	0.12	published ^[1]
Capability → PA (direct)	0.27	published ^[1]
Motivation → PA (direct)	0.46	published ^[1]
Opportunity → PA (direct)	~0.02	published ^[1]
Mediation via Motivation	structure	published ^[4]
C / O / M → Sedentary	−0.30 / −0.20 / −0.35	directional* ^[2]
Intention–behaviour gap (46%)	valve	published ^[3]
PA displaces sitting	−0.15	illustrative
Time constants τ (days)	3–8	illustrative
Feedback gains	tunable	illustrative

*Sedentary coefficients are directional: the evidence identifies behavioural-regulation (capability), social-influence (opportunity) and habit (motivation) as the drivers of reduced sitting, but standardised betas aren't reported, so magnitudes here are plausible placeholders. PA and sitting are modelled as separate outputs (they are not simple inverses).

How to read the tanks: pipe thickness ∝ coefficient strength; animated flow shows an active pathway; the dashed upward pipe is the learning/habit feedback loop. Try Opportunity-only to see why opportunity barely moves activity directly (it works through motivation), then Relapse scenario with a ⚡ Life shock to watch weak habit feedback let everything decay back to baseline.

Parameter provenance: COM-B→PA path coefficients from a prospective validation study (N=186); sedentary drivers from the companion sitting-time validation; intention–behaviour gap from meta-analytic work on non-translators. Coefficients are surfaced here for transparency — apply your usual de-identification before any public ReD Advise deployment.

References
[1] Howlett N, Schulz J, Trivedi D, Troop N, Chater A. A prospective study exploring the construct and predictive validity of the COM-B model for physical activity. J Health Psychol. 2019;24(10):1378–1391. doi:10.1177/1359105317739098
[2] Howlett N, Schulz J, Trivedi D, Troop N, Chater A. Determinants of weekly sitting time: construct validation of an initial COM-B model and comparison of its predictive validity with the Theory of Planned Behaviour. Psychol Health. 2021;36(1):96–114. doi:10.1080/08870446.2020.1763994
[3] Rhodes RE, de Bruijn G-J. How big is the physical activity intention–behaviour gap? A meta-analysis using the action control framework. Br J Health Psychol. 2013;18(2):296–309. doi:10.1111/bjhp.12032
[4] Willmott TJ, Pang B, Rundle-Thiele S. Capability, opportunity, and motivation: an across contexts empirical examination of the COM-B model. BMC Public Health. 2021;21:1014. doi:10.1186/s12889-021-11019-w