ESRA logo

ESRA 2023 Glance Program


All time references are in CEST

Measuring crime in times of change: survey developments and challenges 2

Session Organisers Dr Billy Gazard (Office for National Statistics)
Ms Catherine Grant (Office for National Statistics)
TimeWednesday 19 July, 09:00 - 10:30
Room U6-01b

Measuring crime in times of change: survey developments and challenges

Decision makers in politics and policing rely on survey studies for reliable trend data on crimes against populations at a national and local level as not all crimes are reported to the authorities. While the complex nature of classifying crimes has meant many crime surveys have relied on face-to-face surveying, disruption caused by the COVID-19 pandemic has resulted in a number of challenges to long standing survey designs and data collection methods. This has resulted in a number of developments in crime survey measurement, such as transformation of sample designs and exploration of alternative data collection modes, including online methods. However, crime surveys face a number of unique challenges in transformation, such as how to respond to the evolving nature of crime and respondent safety and well-being in answering questions on sensitive crimes.

This session will explore innovative approaches to crime survey measurement, including how survey methods and key measures of crime have been adjusted or developed following changes in society and in individuals’ experience of crime, as well as papers discussing experiments or pilot studies on any of the following topics:

• Changes to crime survey design, including multi-mode and online survey design experiments,
• Changes to sample design, including “push to web” experiments
• Discussion of differences in mode impact at different points in data collection, including sample bias, respondent safety and barriers to inclusion
• Measurement of new and emerging crime types and violence against women and girls

Keywords: crime, victimisation, survey design

Papers

How does survey mode influence how respondents answer victimisation questions? Findings from an RCT

Mr Joel Williams (Kantar Public) - Presenting Author
Mr Samuel Sullivan (Kantar Public)

One of the UK's longest-running and highest profile surveys is the Crime Survey of England & Wales (CSEW). It is a critical source for the regular production of crime statistics and is a valued time series in its own right (the first survey was carried out in 1982). Traditionally, all the data has been collected via in-home interviews, but the Covid pandemic forced a near-18 month switch to telephone interviewing as well as the acceleration of existing Office of National Statistics (ONS) plans to trial different data collection modes for the CSEW.

One of those trials was designed to identify the measurement differences between textually matched versions of the CSEW core questionnaire: one suitable for telephone interview and one for online self-completion. To do this, a large-scale randomised controlled trial was carried out using Kantar Public's random sample panel, Public Voice. In addition, a subset of online respondents was asked to complete the telephone interview version a couple of weeks later using the same 12 month reference period.

This trial has produced rich information about how the survey mode directly affects the distribution of victimisation data, controlling for questionnaire structure and respondent profile. It tells us a lot about the compatibility of telephone interviews and online self-completion questionnaires with respect to this topic. This paper will present these findings in depth as well as evaluating the suitability of the trial and its follow-up survey as methodological tools for extracting measurement effects due to mode.


Sponsor framing effects in crime victim surveys: An experiment in the context of Finnish Crime Victim Survey (FCVS).

Mr Ilari Kolttola (Institute of Criminology and Legal Policy / University Of Helsinki) - Presenting Author
Professor Janne Kivivuori (Institute of Criminology and Legal Policy / University Of Helsinki)
Dr Matti Näsi (Institute of Criminology and Legal Policy / University Of Helsinki)
Mr Petri Danielsson (Institute of Criminology and Legal Policy / University Of Helsinki)

One of the challenges of survey studies in recent years has been the steadily declining response rate. Advance letter is one of the tools which can be used to try increase response rate. It can build trust and motivate the respondent to participate. It has been noted that the identity of the survey funder or sponsor can influence survey framing and therefore also patterns of responding. Drawing on this, we conducted a methodological experiment as part of the nationally representative 2017 Finnish Crime Victim Survey. As part of the 2017 data collection experiment, we added further information in the advance letter stating that the Finnish Police also served as a collaborator in the survey, as well as including a logo of the Police University College in the postal envelope. The aim was to see how this would affect the survey response rate and non-response bias. The same questionnaire was sent to two randomly selected groups. One group received the survey with both University of Helsinki and police logos on the envelope, as well as both institutions being mentioned on the advance letter. The other group did not have any reference to the police in their material and instead of the police logo on their advance letter, they had the Ministry of Justice logo on it. The response rate of the two groups were significantly different. In the University and police group, the response rate was 49% (592), while on the other group the response rate was 40% (482). In this paper we examine the impact of the sponsor mentioned in the advance letter on (1) response rate, (2) respondent composition (3) victimization prevalence and (4) socioeconomic correlates of victimization. The findings can be relevant for interpreting international comparison because the sponsor of national crime victim surveys varies across countries.


Using attitude surveys to measure attitudes towards, and experiences of, violence against women and girls. Learning for KPI development

Professor Dirk Schubotz (ARK - QUB) - Presenting Author
Dr Paula Devine (ARK -QUB)
Dr Martina McKnight (ARK - QUB)

The devolved government in Northern Ireland is working on a policy to tackle the ongoing violence against women and girls. As part of this policy, government aims to develop key policy indicators (KPIs) that they intend to regularly monitor using two annual social attitude surveys run by ARK. ARK is Northern Ireland’s Social Policy Hub which was set up in 2000 by academics from Queen’s University Belfast and Ulster University. ARK has collected and disseminated social science information on Northern Ireland since then.

In this presentation we report on methodological issues and challenges of using cross-sectional attitude surveys to develop and monitor KPIs on violence against women and girls. The two surveys used for this project were the Northern Ireland Life and Times (NILT), a survey of adults aged 18 years and over which has run since 1998, and Young Life and Times (YLT) – a survey of 16 year olds undertaken since 2003. Explorative question modules to capture experience of, and attitudes towards, violence against women and girls were placed in NILT and YLT in 2022.

In this presentation we report the main descriptive survey findings from NILT and YLT. In addition, we will discuss gender and other differences in relation to the completion rate of the surveys, and some of the methodological challenges faced. We will then report on some of the learning from the surveys with regard to capturing sensitive information of this nature.


Emerging Forms of Gender-based Violence and Their Measurement in Higher Education Context

Ms Claudia Schredl (GESIS - Leibniz Institute for the Social Sciences ) - Presenting Author

Gender-based violence is a complex, persistent feature and force, often unspoken, of many organisations, including higher education institutions. It includes physical, psychological, economic, sexual and online forms of violence, violations, and harassment and has serious, sometimes deadly, consequences (Strid et al., 2021).

Until now, it continues to be difficult to analyse the problem of gender-based violence in higher education institutions due to the dearth of evidence. The few existing studies on gender-based violence in higher education context focus on sexual harassment or sexual violence and include little or no data on emerging forms of violence, such as psychological or economic violence. One reason for this is the lack of validated measurements.

To address this knowledge gap, the UniSAFE survey (Lipinsky et al., 2021) adapted and further developed measurements of psychological violence and economic violence for the higher education context, based on a mapping of measurements of gender-based violence (e.g. FRA’s EU-wide Survey on Violence against Women) and cognitive pretest results.

This paper contributes to developing survey measurements on emerging forms of gender-based violence in higher education context by (1) providing an overview of existing measurements of psychological violence and economic violence from previous surveys, (2) presenting cognitive pretest results and factor analyses results of the UniSAFE survey data on the measurements of these emerging forms of gender-based violence, and (3) elaborating directions for their future development.

References
Lipinsky, A., Schredl, C., Baumann, H., Lomazzi, V., Freund, F., Humbert, A. L., Tanwar, J., & Bondestam, F. (2021). UniSAFE D4.1 Final UniSAFE-Survey Questionnaire. Zenodo. https://zenodo.org/record/7220636#.

Strid, S., Humbert, A. L., Hearn, J., Bondestam, F., & Husu, L. (2021). UniSAFE D3.1 Theoretical and conceptual framework. https://zenodo.org/record/7333232#.


From telephone to web in the Swedish Crime Survey

Mrs Sanna Wallin (The Swedish National Council for Crime Prevention) - Presenting Author

The Swedish Crime Survey (SCS) is a survey about exposure to crime, fear of crime, confidence in the criminal justice system, and crime victims' contact with the criminal justice system.
The SCS has been conducted since 2006 and is based on a nationally representative stratified simple random sample of people aged between 16 and 84. Since 2017 the survey has included a sample of 200,000 people while in the SCS for 2007–2016 had a sample size of 20,000. The method for the SCS was revised in 2017, whereby the collection procedure was changed from mainly telephone interviews to web questionnaires and postal questionnaires. The sample population was also expanded in terms of age and some of the questions were reformulated and new questions were added. Since one main purpose of the SCS is to be able to make comparisons over time, a method has been developed to enable the results for the period 2007–2016 to be compared with 2017–2022. Every effort has been made to ensure that all the descriptions of development over time are unaffected by the change in method. It is important to emphasize this as the purpose of the survey is to study development over time and compare different groups in the population, rather than to estimate exact levels.
In SCS 2022, almost 65 000 people participated in the survey. Young people have been oversampled, and are thus overrepresented in the survey sample. When analyzing the material, cases are weighted both to account for this over-representation and also to adjust for differences in levels of non-response among different groups within the sample.