|Funding for:||UK Students, EU Students|
|Placed on:||11th October 2016|
|Closes:||11th November 2016|
Funding: £16,296 stipend per year, as well as having tuition fees covered
Investigating measurement error and non-response in longitudinal surveys
Applications are invited for a PhD studentship in the Centre for Longitudinal Studies at UCL Institute of Education. The studentship is funded for three years from January 2017 and is based on one of the following projects which relate to investigating measurement error and non-response in longitudinal surveys:
Project 1: Using para-data on interviewer calls to advance understanding of non-response in longitudinal surveys
The aim of this project is to generate important new knowledge about the trade-off between survey error and survey costs by using call record information i.e. ‘para-data’ from the Millennium Cohort Study (MCS) and/or the 1970 British Cohort Study. It will provide an opportunity for the successful student to deepen their understanding of the factors influencing non-response bias and to extend their experience of multivariate modelling applied to rich longitudinal sources of data. Additionally it will give a valuable insight into fieldwork practices on birth cohort studies and provide access to unique new paradata.
Project 2: The impact of dependent interviewing and event history calendars on measurement error in event history data
Many longitudinal studies aim to collect event history data from their study participants. It is well established in the methodological literature that these kinds of data are subject to measurement error in the form of biasing at the seam between data collection periods. The aim of this project is to allow the student to use data from the cohort studies to examine the extent of seam effects in event history data and whether the use of dependent interviewing and event history calendars have been effective at reducing seam effects.
Project 3: Interviewer characteristics and their impact on measurement and response in longitudinal surveys
There is considerable interest in the impact of interviewer characteristics on measurement and response in surveys. Evidence suggests that more experienced interviewers tend to be able to secure higher co-operation rates, interviewers who have a positive attitude and expectations secure higher co-operation rates and certain personality types are also positively associated with gaining co-operation. However, there is also evidence that interviewers who are adept at securing high response rates may not necessarily be those who achieve the highest data quality, for example in relation to item missing data. The aim of this project will allow the student to examine the impact of interviewer’s characteristics on response rates and measurement error on the cohort studies.
Project 4: Effects of mode changes and mixed-mode data collection on measurement in longitudinal surveys
Technological innovation has led to a growing interest in the use of the web for data collection, and in the potential for mixed-mode surveys to achieve higher response rates than uni-mode surveys, and to collect data more cost-effectively. This has led some longitudinal surveys to adopt a mixed-mode approach, different modes for different survey waves. There is considerable interest in the effect of mode changes over time and the introduction of mixed-mode data collection on the comparability of longitudinal survey data both over time and within wave.
Within the aims of this project, the student will examine the impact of mode changes and mixed-mode data collection on measurement error in the cohort studies.
Entry requirements: Applicants must be Home/EU students only for fee purposes and require at least an Upper Second degree and/or Master’s degree (or international equivalents); A Masters is strongly preferred, but is not a strict prerequisite for studying at the Bloomsbury DTC. A range of training programmes is available based on an assessment of the applicant’s prior learning and training needs. All studentships are offered on a full or part time basis.
A background in a relevant discipline/s will be required (e.g., quantitative research methods, epidemiology, psychology, medicine).
Further details about the project: https://bloomsburydtcacuk.files.wordpress.com/2016/07/studentship-description-and-further-information.pdf
Please state in your application which project you are interested in. Applications should be submitted by 11 November 2016. Interview dates are currently being finalised.
If you wish to discuss the studentship informally, please contact the primary supervisor, Dr Lisa Calderwood in the Centre for Longitudinal Studies: email@example.com. Enquiries about the application process can be made to Isabelle Jerome: firstname.lastname@example.org