CENTRE FOR CRIMINOLOGY

UNIVERSITY OF HONG KONG

Monitoring for What: Varying Expectations of CCTV users?

By

Professor Paul Wilson and Helene Wells

Bond University, Golden Coast, Queensland, Australia

Seminar summary (7August2006)

Recent development

Growth is enormous all over the world.

Used in car-parks, public transport, commercial and shopping centres.

Welsh and Farrington’s evaluation showed 8% reduction in crime in exp areas and 9% increase in control areas.

More successful in Britain than USA.

Not effective in violence, alcohol crimes

Most effective in car parks: also in detecting offenders and court actions

Problems

Extremely difficult to Evaluate – hard to get controls, police introduce new patrols, street lighting etc

Displacement major problem and hard to study

Results differ from location to location – one major reason is that the design and installation varies enormously

Street lighting maybe more effective – often introduced simultaneously

Crime and CCTV in Australia: Understanding the Relationship

ARC Linkage Grant: Safeguarding Australia, protecting Australia from terrorism and crime. $ Australian 100,000+ grant, financial and in-kind contributions from QPS, QR, GCCC and Dept of Communities (total about $300,000)

Phase 1 - Public space (GCCC)

Phase 2 - Public transport (QR)

Phase 3 – Surveys (businesses, commuters, residents)

Phase 4: Assessment of the Operation and Management of CCTV systems (QR and GCCC)

Today’s presentation focuses on the Gold Coast Safety Camera Network (GCCC)

http://www.bond.edu.au/hss/criminology/arc_linkage_grant.htm

Gold Coast Safety Camera Network (GCCC)
Total operational hours per day / 24
Number of cameras / 66
Type of Cameras / Overt and Semi-overt cameras
PTZ Colour / Fixed Colour
Number of monitors / 9
Operator/s per shift / 1 on “regular” shifts, 2 on “busy” shifts
Number of cameras per operator / 38 cameras for “local” operator
28 cameras for “remote” operator
Ownership / Local council
Operators employed by / Private security company
(sub-contracted by local council)
Areas monitored / Surfers Paradise, Southport, Broadbreach, Coolangatta
Communication with police / Police radio (one-way) and telephone
Recording technology utilised / Analogue and digital
Funding / Local business levy

Expectations of CCTV systems – Research areas of interest

What was the original purpose/s for the installation of public space CCTV cameras on the Gold Coast?

Is the original purpose/s consistent with the monitoring procedures currently undertaken by camera operators?

What are the current expectations of the end users involved?

Are the expectations of various end users in conflict?

If so, what are the policy implications from these conflicting expectations?

Original purpose of CCTV

Public safety

Policy and Procedures Manual

Alcohol-related violence

“followed pressure from local businesses concerned that the image of Surfers Paradise as a family friendly tourist resort was being undermined by alcohol-related violence” (Wilson and Sutton, 2003, p. 31).

Anti-social behaviour

Community feedback, local police

Original purpose consistent with current monitoring?

To determine this, an observational study of 100 hours was undertaken

Methodology: Ethnography / process and outcome evaluation

100 hours of observational analysis, document analysis of monthly incident reports, interviews with operators, surveys and interviews with police

23 shifts between Sept-Dec 2005 (including Indy Week and Schoolies Week)

Hypothesised …

1. The majority of an operator’s shift is spent identifying anti-social behaviour and alcohol-related violence

2. Camera operators identify highly visible behavioural incidents (i.e. assaults) compared to less visible, ‘discrete’ incidents (i.e. drug deals)

3. Most incidents captured on CCTV are initiated by camera operators rather than police

4. Camera operator surveillance leads to more arrests than if CCTV cameras were not implemented

Observational Period / Minutes / %
Total time spent actively monitoring and searching for incidents / 986 / 16.43%
Actively searching / 117 / (1.95%)
Actively monitoring / 869 / (14.48%)
Total time spent engaging in “other activities” (admin; back-searching etc) / 5014 / 83.57%
Total observational period / 6000 / 100.00%

Hypothesis 1

The majority of an operator’s shift is spent identifying anti-social behaviour and alcohol-related violence

The majority of the observational period (83.57%) was dedicated to activities other than the active searching/monitoring of footage

Routine (‘manual tour’) surveillance of camera network

Administrative duties (log book, paperwork, phone calls)

Recording and ‘back searching’ for surveillance

Communication with police and external agencies (226 phone calls, 67 via police radio)

However, the majority of actual surveillance was dedicated to the identification of anti-social behaviour and alcohol-related violence (86.19%)

Reason for Initial and Continuing Surveillance / Incidents / % of overall incidents / Minutes / % of overall minutes
Crime / 110 / 60.77% / 673 / 68.25%
Good order / 31 / 17.13% / 146 / 14.81%
Safety issue / 12 / 6.63% / 87 / 8.82%
Local law issue / 3 / 1.66% / 13 / 1.32%
No obvious reason / 9 / 4.97% / 20 / 2.03%
Other / 16 / 8.84% / 47 / 4.77%
Total / 181 / 100% / 986 / 100%

Types of Suspicion

Based on Norris & Armstrong (1999) and Goold (2004) analysis of UK CCTV systems

Behavioural – i.e. fighting and drunkenness, running

Categorical – based on sex, age, race, appearance

Locational – location of individual (car park at night)

Personalised – previous knowledge of target (suspect)

Protectionalised – fear of person’s safety (lone woman)

Routine – part of a set surveillance routine (i.e. ATMs)

Transmitted – initiated by external source (police)

Voyeuristic – personal interest, sexual or otherwise

Deployment Data

Types of Suspicion / Police presence at the incident / No. of incidents with arrests / No. of arrests
Yes / No / Total
Behavioural (easy to recognise) / 21 / 21 / 42 / 13 (30.95%) / 19 (37.25%)
Categorical / 3 / 3 / 6 / 2 / 2
Locational / 3 / 4 / 6 / 0 / 0
Personalised / 1 / 4 / 5 / 0 / 0
Protectional / 0 / 4 / 4 / 0 / 0
Routine / 8 / 8 / 16 / 2 / 2
Transmitted / 61 / 38 / 99 / 25 (59.52%) / 28 (54.90%)
Voyeuristic / 1 / 2 / 3 / 0 / 0
Total / 98 / 83 / 181 / 42 / 51

Hypothesis 2

Camera operators identify highly visible behavioural incidents (i.e. assaults) compared to less visible, ‘discrete’ incidents (i.e. drug deals)

YES. 137/181 can be classified as highly visible behavioural incidents (75.69%) – operators would have seen them anyway

Types of Suspicion / Incidents
Behavioural (42) / 42
Categorical (6) / 4
Locational (6) / 4
Personalised (5) / 2
Protectional (4) / 4
Routine (16) / 10
Transmitted (99) / 68
Voyeuristic (3) / 3
TOTAL / 137

Hypothesis 3

Most incidents captured on CCTV are initiated by camera operators rather than police

99 (of 181) incidents surveilled were due to an external source (54.70%)

Approximately half of all targeted surveillances is due to an external source, thus camera operators identify an equivalent number of incidents

i.e. 2 motorcyclists driving erratically throughout Surfers Paradise, police notified and males fined

Hypothesis 4

Camera operator surveillance leads to more arrests than if CCTV cameras were not implemented

Of the 51 arrests, we can assume 44 arrests would have occurred regardless of the camera network (i.e. 86% of arrests)

7 arrests were the result of the detection of an incident by a camera operator (14% of arrests during observational period)

Arrests attributable to external communication (i.e. police transmitting information)

7 arrests – as a direct (or indirect) result of the camera network

Incident A: naked male running along the beach

Incident B: heavily intoxicated male outside a nightclub

Incident C: youths drinking and urinating in a park*

Incident D: male exposing himself to onlookers, additional male videotaping (police radio: “we have it on camera”)

Incident E: Serious assault (ambulance required) – local law officer informed operator simultaneously

Incident F: male insulting police officer behind his back (police radio: “got good footage of that”, “it’s recorded on the monitors”)

Incident G: Suspected shoplifter trying to hide, located via camera network and security guards simultaneously

Getting back to the Expectations

Is it consistent with the original purpose?

YES. 86.18% of all incidents related to anti-social behaviour, alcohol-related violence and promoting public safety

Current Expectations of Public and Police?

Detect, deter, prevent crime

“Eyes” / Guardians of public space users (yes)

24 hour surveillance (but in reality 24 recording)

Collection of evidence for police investigations (yes – the real impact)

Are these expectations in conflict?

Locally owned camera network used by local police (most CCTV in Australia owned by council but used by Police – a free service for the police?)

Conclusions

Need to contend with varying degrees of tension and/or expectations with dual users (some but not large)

General increase in surveillance – need for specified training (i.e. detection deception – Mark Frank and Paul Ekman research)

National codes of practice and guidelines being developed)

Evaluations have to consider the installation and management issues (not done so generally)

The cost effectiveness of this crime prevention tool – is it as effective as proponents would suggest? (part of our research).