Surveying Children and Young People 3 |
|
| Session Organisers |
Ms Kate Smith (Centre for Longitudinal Studies, UCL Institute of Education, London ) Dr Emily Gilbert (Centre for Longitudinal Studies, UCL Institute of Education, London ) |
| Time | Thursday 18th July, 14:00 - 15:30 |
| Room | D18 |
Many large-scale surveys successfully collect a variety of distinct types of data from children and young people (up to the age of 25). However, there is relatively little methodological evidence in this area. Much of the literature relating to children and young people’s participation in research focuses on small-scale qualitative studies and tends to concentrate on ethical issues relating to the rights of children and young people in research. This session will cover the challenges and experiences of including children and young people in surveys as they move from childhood to adulthood, and related survey design issues. A major challenge when interviewing teenagers is that while children’s participation in surveys is often mediated by and involves their parents, teenagers and young people make autonomous decisions, bringing challenges particularly in terms of engagement. The session aims to explore a variety of methodological issues around surveying young people. Submissions are particularly welcomed on:
- designing questionnaires for children and young people, including question testing methods
- collecting data on sensitive topics from young people, including methods for ensuring privacy and encouraging accurate reporting
- collecting different types of data from children and young people including physical measurements and cognitive assessments
- using different methods of data collection, including the use of innovative technology such as the web and mobile phones
- inclusivity in data collection methods, including facilitating the participation of children and young people with lower literacy levels
- assessing the reliability and validity of children and young people’s self-reports
- preventing non-response by engaging young people in research, including designing survey materials to appeal to young people and using new technology and digital media for participant engagement
- the challenges of retaining young people’s contact and interest in surveys over time
- ethical issues in involving children and young people in surveys, including gaining informed consent and protecting young people’s rights and well-being
Keywords: Children, young people, surveys
Mr Robert Hoffmann (Robert Koch Institute)
Mr Michael Lange (Robert Koch Institute)
Mr Sebastian Hinck (Robert Koch Institute)
Mr Robin Houben (Robert Koch Institute) - Presenting Author
Dr Antje Goesswald (Robert Koch Institute)
Background
The second wave of the "German Health Interview and Examination Survey for Children and Adolescents" (KiGGS) aims to collect representative data on the health status of the target population aged 3-17 in Germany. The following procedure was used to recruit participants: Written invitations and reminders were followed by phone call attempts and personal home visits if no feedback was received. The success of the different measures is to be analyzed separately.
Methods
The success of the four measures will be assessed (1) on the basis of the number of participants recruited at each phase, (2) the sample composition with regard to the educational level of the household compared to available data from "microcensus" (official federal statistics), conducted in 2013, and (3) via response, cooperation and contact rates following the standards of the American Association for Public Opinion Research (AAPOR). The latter also act as indicators for the efficiency of the measures taken.
Results
A total of 3,567 children and adolescents took part in the cross-sectional part of the examination survey of KiGGS Wave 2. This corresponds to a response rate 2 (according to AAPOR) of 41.5%. Without home visits, the response rate would have only been 30.3%, without telephone advertising 26.0%. The composition of the sample improves considerably only after home visits. The cooperation rate declines while the contact rate increases.
Conclusion
The additional measures taken during KiGGS Wave 2 lead to a strong increase in response rates. Furthermore, an improvement in the sample composition has been found. Personal contact with the target persons during home visits plays a major role in this. The results indicate that contact rates ought to be increased in future surveys. It would be desirable to keep the amount of cooperation on a steady level. Future analyses will tackle this question.
Dr Jessup Karena (Australian Institute of Family Studies) - Presenting Author
Dr Galina Daraganova (Australian Institute of Family Studies)
Ms Diana Smart (Australian Institute of Family Studies)
Ms Renda Jennifer (Australian Institute of Family Studies)
Growing Up in Australia: The Longitudinal Study of Australian Children (LSAC) was designed to provide data that enables a comprehensive understanding of children’s development within Australia’s current social, economic and cultural environment. Since 2004, two cohorts of 5,000 children and their families have been interviewed every two years. The “Baby” and “Kinder” cohorts are being interviewed for the eighth time in 2018 and are aged 14-15 and 18-19 respectively. Response rate have declined gradually over the waves, however Wave 7 saw a significant drop in response for both cohorts. In addition, the study faces new challenges in retaining respondents as the children move into early adulthood.
With a view to improve or at least maintain response rates for future LSAC waves, substantial work has been undertaken to identify and implement strategies for engaging participants and encouraging ongoing study participation. This paper will begin by summarising findings from a review of sample retention and participant engagement strategies used in other longitudinal studies. The review, along with extensive consultation, has helped inform planning for LSAC’s response to attrition. Careful consideration has been given to methodological factors, including who LSAC’s primary participants should be, how often data should be collected, the length of the survey instruments and the inclusion of innovative data collection techniques. Methods for maintaining contact with LSAC participants, maximising the effectiveness of incentives and re-engaging participants have also been thoroughly explored.
Mr Robert Lipp (Frankfurt University of Applied Sciences) - Presenting Author
Mr Sven Stadtmüller (Frankfurt University of Applied Sciences)
Ms Andrea Giersiefen (Frankfurt University of Applied Sciences)
Linking data over the course of a panel study can be a tedious task. This is particular true when respondents’ anonymity has to be ensured and no list of names can be employed. In this case, researchers have to rely on ID-codes that respondents generate anew in every panel wave, consisting of elements that they can a) easily recall and b) are individual enough to not produce duplicates. While using the right ID-components (like the first letter of the mother’s first name) can greatly improve the quality of matches, mistakes are bound to happen resulting in panel cases not being recognized as such (or cases being matched wrongfully). When surveying young people, even more of these mistakes are to be expected.
The contribution uses data from the first four annual waves of the study “Health Behavior and Injuries in School Age” (GUS), a longitudinal study of ~10.000 German pupils which started in their first year of secondary education (5th grade). It highlights the challenges associated with using self-generated ID-codes in panel studies with children. Furthermore, it introduces record linkage methods (as suggested by Schnell et al. 2006) as a way to increase the number of successful matches. This post-processing technique uses a fuzzy-string-merge to match IDs that do not match perfectly but are very similar.
In the GUS study, this was accomplished using the Stata Ado “reclink”. This way, approx. five percent of previously unmatched cases could be paired up. Time-constant variables in the dataset were used to verify the matches.