Version without tables (for Lynx)

Was the Dunblane Inquiry Misled by the Scottish Office?

Following revelations by independent researchers that Home Office and Scottish Office statistics show a strong negative correlation between licensed gun ownership and armed crime in Great Britain,[1] a Home Office spokesman countered:

"that a ban on licensed guns was being considered not because the statistics showed or didn't show they caused crime, but because they were dangerous."[2]

The notion that handguns are specially dangerous was put forward in the main findings of a Scottish Office report to the public inquiry into the shootings at Dunblane primary school:

"Although handguns featured in only 9 (26%) incidents shots were fired from them proportionately more often, and they were responsible for more injuries and deaths, than any other type of firearm."[3]

This conclusion was widely reported in the press[4] and is repeated at paragraph 9.47 of Lord Cullen's inquiry report:

"In the study of serious crime in Scotland to which I referred in 9.10, although handguns featured in only 26% of the incidents (9 out of 34), shots were fired from them proportionately more often, and they were responsible for more injuries and deaths, than any other type of firearm."[5]

However, the study which Lord Cullen quotes was confined to one year only (1993) in which the use of pistols in homicide and attempted murder is higher than in any other year from 1986 to 1995 (see Table 1, based on Scottish Office statistical bulletins). Nine homicides and attempted murders were committed with pistols[6] during that year, which is more than twice the average figure for the ten-year period. Furthermore, 1993 is one of only two years in the period 1986-95 in which the fatality rate for homicide attempts with pistols is well above average while the fatality rate for other firearms homicide attempts is below the average for the ten-year period. Regrettably, the Scottish Office researchers did not draw attention to the fact that 1993 was such an exceptional year. To the contrary, they wrote:

"…there is no reason to assume that 1993 was significantly different from immediately preceding or subsequent years."[7]

At paragraph 9.55 of his report, Lord Cullen wrote:

An assessment of risk involves an examination of two elements in combination - (i) the chances of harm happening; and (ii) the nature and extent of the harm.[8]

By basing their widely-publicised report to Lord Cullen on an unrepresentative data set, the Scottish Office Central Research Unit exaggerated both elements of the assessment of risk posed by the use of pistols in crime. For this reason the relevant conclusions of the Dunblane Inquiry and the Government response thereto should be re-examined.


Footnotes

  1. P H Jackson, J A G Hawkins, Prof. A R Horrocks and R A I Munday, Was the Dunblane Inquiry Misled? (crimstat.zip, November 1996).

  2. J Booth, Cullen was Misled on Facts, say Researchers (The Scotsman, 5 November 1996).

  3. D Oag, H McKay & N Coghill, Cullen Inquiry: An Evaluation of Serious Incidents Involving the Use of a Firearm in 1993 (Scottish Office Central Research Unit, August 1996).

  4. Criminals `are more likely to fire handguns' (The Daily Telegraph, 28 August 1996).

  5. The Hon Lord Cullen, The Public Inquiry into the Shootings at Dunblane Primary School on 13 March 1996 (The Stationery Office, 16 October 1996).

  6. All of these pistols were illegally held, D Oag et al, op. cit.

  7. D Oag et al., op. cit.

  8. The Hon Lord Cullen, op. cit.

Table 1: Homicide and attempted murder by firearm type, Scotland, 1986-95
(Source: Scottish Office statistical bulletins)

Year Homicide Attempted murder Homicide & Attempts Fatality rate % of firearms
offences carried out
with pistol
  Shot
gun
Rifle Pistol Shot
gun
Rifle Pistol Shot
gun
Rifle Pistol Shot
gun
Rifle Pistol Homicide Homicide &
attempted
murder
1986 3 0 1 11 0 1 14 0 2 21% - 50% 25% 13%
1987 3 0 0 15 2 1 18 2 1 17% 0% 0% 0% 5%
1988 1 0 0 20 1 1 21 1 1 5% 0% 0% 0% 4%
1989 4 0 0 7 0 1 11 0 1 36% - 0% 0% 8%
1990 0 0 3 19 0 3 19 0 6 0% - 50% 100% 24%
1991 3 1 0 26 0 4 29 1 4 10% 100% 0% 0% 12%
1992 5 1 0 27 2 4 32 3 4 16% 33% 0% 0% 10%
1993 2 0 3 21 0 6 23 0 9 9% - 33% 60% 28%
1994 2 2 3 18 0 4 20 2 7 10% 100% 43% 43% 24%
1995 2 0 1 13 0 7 15 0 8 13% - 13% 33% 35%
1986-95
total:
25 4 11 177 5 32 202 9 43          
1986-95
average:
2.5 0.4 1.1 17.7 0.5 3.2 20.2 0.9 4.3 12.4% 44.4% 25.6% 27.5% 16.9%

Notes:

  1. 1993 is the year with the highest number (9) of homicides and attempted murders carried out with a pistol during the 10-year period 1986-95.
  2. In 1993 the number of homicides and attempted murders carried out with a pistol was more than twice the 10-year average.
  3. 1993 and 1990 are the only two years in which the pistol fatality rate is above average, while the rifle and shot gun fatality rates are below average.
  4. In 1993 the percentage of all firearms homicides which were carried out with a pistol was more than twice the 10-year average.
  5. In 1993 the percentage of all firearms homicides and attempted murders which were carried out with a pistol was 1.66 times the 10-year average.
  6. Any comparison of fatality rates for offences committed with different weapons should note that the annual rifle and pistol data sets are numbered in single digits and therefore fluctuate widely, whereas shot gun numbers are more substantial (see note 7 below). It should also be noted that the statistics published by the Scottish Office do not differentiate between full-length and sawn-off shot guns; the latter are known to be much less lethal at any range over a few feet due to imperfect internal and external ballistics. Allowing those caveats, the fatality rate over the 10-year period for offences committed with rifles is twice that for offences committed with pistols, and shotguns half.
  7. Over the 10 year period, more than five times more attempted murders and nearly two and a half times more homicides were carried out with shot guns than with pistols.

P H Jackson
+44 (1644) 470223
(revised 11 December, 1996)

The author invites critical review. The full set of data and calculations summarised in this paper is published as crimstat.zip

Send mail to Peter at:
Peter H Jackson <pjackson@forge.demon.co.uk>

Posted: Dec 1996