Adjustment and the Threshold of Materiality.
1. Purpose
This note sets out the basis on which the network distinguishes between adjustments requiring public notification and adjustments which do not. It is published in the interest of transparency and is not a consultation.
The note applies to changes in sensing, signalling, illumination, routing, access and environmental modelling made within ordinary operating tolerances. It does not alter the duties attached to exceptional events, nor does it replace any notification required by law.
2. Background
The network makes, on average, between four and six thousand adjustments per operating period. These range from single-unit recalibrations to revisions of area-wide flow models. The overwhelming majority are imperceptible to any individual person.
Early operating practice was to notify all adjustments. This practice was discontinued. Analysis of the notification archive from that period found that fewer than 0.3% of published notices were accessed by any member of the public, and that the volume of publication was itself identified as a barrier to attention. A person who receives six thousand notices receives none.
The discontinued practice also gave equal prominence to unequal effects. A correction to the clock of an isolated unit appeared beside an area-wide change to crossing durations. Readers were required to determine significance from technical data for which no ordinary comparison was available. Notification was comprehensive but did not produce understanding.
The present approach therefore treats attention as a finite public resource. Publication is reserved for information capable of supporting a decision. Records of adjustments below that level remain available to the network for audit and aggregate review.
3. The threshold
An adjustment is material if a reasonable person, informed of it, would alter their behaviour in response. Adjustments below this threshold are made without notification.
The threshold is applied conservatively. Where the effect of an adjustment on an individual cannot be estimated with confidence, the adjustment is treated as material and is published. In the preceding twelve operating periods, 3.1% of adjustments met the threshold.
Materiality is assessed against the likely effect of the adjustment, not the technical effort required to make it. A change applied to many units can remain immaterial where its observable effect is negligible. A change to one unit can be material where it affects a route or condition on which a person may reasonably rely.
The assessment is made before implementation where the effect is predictable. Where a result depends on observed conditions, it is made as soon as sufficient data is available. A later finding of materiality results in publication at that point; it does not make the original decision unreasonable.
4. Aggregate effects
It has been suggested that adjustments which are individually immaterial may, in aggregate, produce a material effect.
This is correct, and the network monitors for it. Aggregate effects across an area are assessed at the close of each operating period against the same threshold. Where an aggregate effect is found to be material, it is published as such. Three aggregate findings have been published to date.
Aggregation is considered across place, time and function. Changes to lighting and crossing durations in the same area are not treated as unrelated merely because they are maintained by different systems. Where the combined effect can reasonably be expected to alter behaviour, the combined effect is assessed as one adjustment.
It should be noted that the alternative — publishing every constituent adjustment on the possibility that some combination of them might one day become material — would return the network to the condition described in section 2, in which nothing is read and nothing is therefore known.
5. Consent
The network does not seek consent for adjustments below the materiality threshold, on the basis that consent cannot meaningfully be given to a change one cannot perceive. Where a person wishes to be informed of all adjustments affecting an area, this can be arranged on request.
No such request has been received.
Consent and notification are separate matters. Publication does not of itself create a choice to refuse an adjustment, just as the absence of publication does not remove a right that otherwise exists. The threshold concerns whether information is useful before or after a decision; it does not determine the legal basis on which the decision is made.
6. Review
The threshold is reviewed periodically. The review is internal. Its findings are published where they meet the threshold.
Review considers the proportion of adjustments published, the observed use of those publications and any difference between predicted and reported effects. A threshold that produces more notices but less useful attention is not considered more transparent by reason of volume alone.
Nothing in this note prevents publication below the threshold where a clear public interest is identified. Such publication remains exceptional because routine exception would reproduce the archive conditions that the threshold exists to correct.