.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... controls.2
`Control' has a special meaning in the conduct of experiments, viz a
test case not subject to the treatment but otherwise, as near as
practicable, identical to the experimental case that will be treated.
This meaning has been misappropriated by the advertising industry in
describing television advertisements involving dangerous stunts as
`recorded under controlled conditions'. This is not helpful to
non-scientists when they are faced with a need to follow scientific
arguments.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... conditions3
Amateur and professional astronomers in countries affected by World
War 2 mostly found themselves busy with the national war effort. Those
who did get to do some astronomy at the time of lighting blackouts
afterwards remarked on the substantial reduction in artificial skyglow and
some consequent great advances this allowed in our understanding of the
universe. The known presence of worrisome skyglow before the WW2
blackouts is also helpful in reconstructing the time course of the amount
of ambient light outdoors at night in cities and towns.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... validity.4
For comparison outside the USA, crime survey results indicated that
nearly two-thirds of the total of assaults, burglaries, robberies, thefts
and vandalism occurred after dark in England and Wales
(Ramsay and Newton 1991 [96]).
The percentage of burglaries at night was 62 %, and ten years
later, another survey indicated it was about the same
(ONS 2000 [76]). In the
writer's home state of Victoria, Australia, it is understood that most
burglaries occur in daylight hours. Robberies and assaults in central
Sydney, NSW, Australia showed a peak after midnight, especially on
Saturday and Sunday nights (Jochelson 1997 [54]).
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... displacement5
Challinger (1991) [19]
is a useful and readily available account of five
main types of crime displacement, although the original identifier of
these (Reppetto 1976 [97]) is not cited.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... luminaires6
A luminaire is the complete lamp fitting including any housing,
mounting, shielding, reflector, lens, lampholder, electrical components
and wiring, and the lamp.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... luminaires7
LPS lamps are often called SOX in the UK literature, and likewise SON
is used to indicate HPS lamps.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...SOCRU2002).8
The full texts were not available online and hard copies proved
impracticable for the writer to obtain.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... crime.9
Care is required to avoid confusing the usage of `direct' in this
context with the usage, avoided in this document, of describing a
correlation as `direct' if it is positive. There is no problem in the
descriptions of indirect effects and inverse or negative correlations.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... population,10
Unfortunately, the available population data were for 2001 rather
than 1999. For the present purpose, a uniform scale error in the
calculated 1999 crime rates is immaterial. Residual differential errors
between the calculated and actual county crime rates for 1999 are expected
to be too small to make any practical difference to the outcome of this
exercise.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... change.11
For example, the calculation for the first pair is
(947/767)/(597/721) = 1.491 or a 49% increase in crime in the first county
relative to that in the second county.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... proffered.12
The police may even go so far as to veto the type of lamp!
(SOLS 2002 [105]).
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... terms.13
The titles of
Poyner and Webb (1987) [93],
Grabosky and James (1995) [42]
and
several other works have a comparable restriction, indicating that this is
a surprisingly common fault in the literature of Situational Crime
Prevention. It goes beyond a mere stylistic blemish: for instance,
Grabosky and James also called for more examples, subject to the
condition: ``Evidence of success as demonstrated in the results of
scientific evaluation is a prerequisite for inclusion.'' Given that false
indications, beneficial and `maleficial', are equally likely to result
from statistical analysis of properly designed and conducted experiments
in which there may be no true relationship, publishing collections
confined to apparently beneficial results is not only bad science but
misleads architects, urban designers, lighting and security professionals,
insurance risk assessors, police and members of the public who are
interested enough to read the material. In due course, this leads to
unwitting perpetuation of the bias when readers prepare summary articles
about supposedly well-established facts for further dissemination.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... written.14
Australian Standards on road and public lighting issued in 1997 and
1999 also imply that lighting prevents crime but give inadequate
justification for this.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... area.15
The odds ratio is the reciprocal of the expression for numerical
relative change in crime in the quasi-experiments in Section 3.4.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... need,16
Again there was no explicit reason or reasons given for the
relighting decision.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... interventions.17
Presumably this refers to CCTV.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... intensity18
The peak or mean illuminance ratio may be more appropriate. The
difference from an intensity ratio can be substantial, depending, inter
alia, on differences in how the light is distributed before and after the
intervention.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... treatments.19
Even four successive treatments might be enough to have people
dancing in the streets!
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... overall.20
The overall total crime value does not fall between the overall
values for violence and property crimes because the total for Milwaukee is
much larger than the values given for violence and property crimes and
because only the totals of these two categories was available or
acceptable for Fort Worth and Indianapolis. Some of the contributing
measures were also for different durations.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... ratio.21
A pair of possible weightings for the UK odds ratio and the US odds
ratio can be found from the combined odds ratio as the roots of a
quadratic equation.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... fear.22
Part 2 goes even further than this.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... output.23
All of the outdoor luminaires shown in the four illustrations of the
paper appear to be emitting an appreciable amount of light at and above
the horizontal, contrary to a company representative's views about
minimising light pollution and glare
(Baldrey 2000 [4]).
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... quality.24
Note that vertical plane illuminances of over 1000 lux are used for
live colour TV broadcasting of sporting events.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... survey25
The survey was conducted by members of the Outdoor Lighting
Improvement Section of the Astronomical Society of Victoria.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.