UNECE-UNODC Meeting on Crime Statistics, 25-27 January 2006

1
UNECE-UNODC Meeting on Crime Statistics, 25-27
January 2006

• Session 4Dissemination of crime victim data
• Results from the UNECE-UNODC questionnaire on
  victim surveys
• Philippe Lamon
• UNODC

2
Overview

• Comparability of results
• Margin of error
• Availability of data and dissemination

3
Were the results of the survey compared with data
on crimes reported to the police?
Surveys Surveys Countries Countries
Yes 22 28 17 53
No 50 64 26 81
No answer 6 8 5 16
Total 78 100 32
4
Rates per population of selected crimes

• 436 data are given on 1660 possible data (26)

5
Rates per population of selected crimes

• Problems
• - Period reference ?
• - Comparability ?
• - per household or per owner ?
• - 0.09 0.09 or 9 ?
• - etc.

6
If the survey results have been made public, what
margin of error was adopted in presenting the
results?

• 33 relevant information in 78 surveys (42)

7
Types of margin of error
Relevant information Relevant information of all surveys of all surveys
Over 25 are considered unreliable 6 18 8
Different according to the data(groups, territory level, etc.) 2 6 3
95 confidence interval 14 42 18
Not enough specific(published with S.E., etc.) 3 9 4
Other (0.03, 0.15, 1-2, 0.035) 8 24 10
N/A 57
Total 33 100 78 100
8
Margin error adopted by the ICVS
9
Margin error adopted by the ICVS
10
Is there a public use microdata file?
Surveys Surveys Countries Countries
Yes 31 40 16 50
No 43 55 23 72
No answer 4 5 3 9
Total 78 100 32
11
Are survey results available on internet?
Surveys Surveys Countries Countries
Yes 51 65 25 78
Not available yet 3 4 3 9
No 22 28 18 56
No answer 2 3 2 6
Total 78 100 32
12
Were printed reports published with the results
of the surveys?
Surveys Surveys Countries Countries
Yes 47 66 29 91
Yes, as part of the ICVS analysis 5 7 4 13
No 22 28 14 44
No answer 4 5 3 9
Total 78 100 32
13
Conclusion (1)

• Comparability of the data1. with police data
  28 of surveys but 53 of countries (63 of
  surveys in 88 of countries in dark
  figure/unreported crime, but only 28 of surveys
  compare with police data)2. with other surveys
  lack of comparison

14
Conclusion (2)

• Comparison of data collected between different
  surveys?Possible if at least- same
  methodology- same reference population, rates
  per owner/pop, etc.Possible in two ways for
  international comparison- ICVS- comparison of
  two surveys (complicated)

15
Conclusion (3)

• Margin of errorShould we propose a model?

16
Conclusion (4)

• Dissemination of results1. on internet(65 of
  surveys, 78 of countries) good2. printed
  version(67 of surveys, 94 of countries) ok
  for countries, lack of finalization for surveys

17

• Thank you