The Department for Transport (DfT) has warned that when interpreting its quarterly results anything appearing 'statistically significant' should be 'regarded with some caution' and admits it has not yet modified the statistical tests to allow for this uncertainty.
It has previously developed tests to check for statistical significance in changes in numbers of reported road casualties during a year, but it concedes that 'the tests do not make any allowances for uncertainty in the figures being used.'
DfT officials added the figures 'assume that the casualty figures are complete counts of all the records'.
Provisional data for the first quarter of 2015, it continues, 'are not complete records of all the reportable accidents during January to March 2015'.
Not all police forces have provided data for the whole period and most will send additional records at a later date.
'In theory, therefore, any statistical test needs to take this uncertainty into account,' officials stated.
New figures issued by the DfT suggest that road deaths in Great Britain fell by 1% over the year ending March 2015 as compared with the previous 12 months.
There were also 23,570 killed or seriously injured (KSI) casualties, a 3% fall, and 186,060 reported road casualties of all levels of severity, a 2% fall.
At the same time, motor traffic levels rose by 1.8% and the overall casualty rate per vehicle mile travelled fell by 2%.
Surveyor's Highway Management conference will take place on September 23rd and 24th in Manchester. To register go to: www.highway.surveyorevents.com