Tuesday 16th July
Wednesday 17th July
Thursday 18th July
Friday 19th July
Download the conference book
Attrition in Panel Surveys - Prevention and Correction 1 |
|
| Convenor | Mr Ulrich Krieger (MEA, Max Planck Institute for Social Law and Social Policy) |
| Coordinator 1 | Dr Peter Lugtig (Department of Methods and Statistics, Utrecht ) |
| Coordinator 2 | Dr Galit Gordoni (Israel Central Bureau of Statistics) |
Focus:
In recent years, longitudinal surveys have become increasingly popular for studying change and stability in a wide variety of phenomenon within society. The availability of accurate and valid data, therefore, on change and stability is paramount. This session focuses on one of the most important sources of error and bias: drop-out or attrition.
Correction:
Differential non-response in panel surveys can lead to large errors over the course of the study as the survey estimates may become biased. Research into the nature and causes of attrition in various panel surveys typically finds that attrition often coincides with major events in the lives of respondents. In addition, fieldwork procedures, survey innovations, social and psychological characteristics of the respondents as well their survey experiences are also thought to be factors generating longitudinal non-response error. Knowledge of these correlates can serve as a basis for correction methods to correct for attrition.
Prevention:
Rather than correct for attrition, it is better to prevent attrition from occurring. Evidence for what procedures are successful in limiting attrition is however scarce. We invite contributions that assess different ways to prevent attrition, and evaluate their consequences on attrition rates, and attrition bias.
Session Details:
Examples of contributions sought for this session include but are not limited to:
• Experimental studies contrasting different survey procedures and their effect on attrition
• Examples of best practices in preventing attrition from occurring.
• Examples of studies where attrition has been corrected by post-survey adjustments
• Papers that propose new statistical methods for correcting or ameliorating the effects of attrition.
We particularly invite methodological papers that incorporate experimental data.
4th co-organizer: Dr. Jon Miller, jondmiller@umich.edu, University of Michigan
The french longitudinal survey on the integration of first-time arrivals (ELIPA) has two main objectives : the path integration knowledge within the three years following the first residence permit in France and the assessment of the welcome arrangement for new migrants (CAI). The first wave was conducted in 2010 with a representative sample of 6,107 migrants aged 18 or older, nationals of countries outside the EEA and Switzerland. These migrants were reinterviewed in 2011 and the final wave will happen in 2013.
Some other countries also carried out such longitudinal surveys. In France, ELIPA follows a similar survey (PPM), conducted in 2006 and 2007.
Firstly, the methods used to control attrition between the first two waves will be presented. These methods meet two needs : limit attrition due to moves and due to refusals. In wave 2, ELIPA's attrition rate is 22 %. This is considered to be a good result, partly because the survey is not mandatory, but also because the target audience may be hard to reach (language barrier, vulnerable population, mobile and sometimes reluctant to answer this kind of survey).
The second part will deal with the determinants that might explain the ELIPA's lower attrition compared to PPM (38 %), the only comparables surveys.
Finally, the non-response in wave 2 will be analysed. This analysis will enable to set up some guidances to interview potentially hard-to-reach people and/or people who are possibly more likely to refuse to answer the third and final wave.
It is well known that some interviewers achieve better response rates than others. The skills of these 'better' interviewers are often drawn upon towards the end of fieldwork when individuals who have refused are re-issued to them.
However, a key advantage of longitudinal studies is the ability to use data from previous sweeps to identify sample members who may be more reluctant to respond, and then to target resources at this group from the outset.
This paper describes an experiment conducted on the Age 42 survey of the 1970 British Cohort Study, the most recent round a longitudinal study following around 17,000 people who were all born in Britain in a particular week in 1970 in which potentially reluctant respondents were allocated to interviewers with particular skills.
Prior response patterns were used to identify the group considered most likely to refuse. Where it was geographically possible these cases were allocated to a sub-group of interviewers who were identified as being particularly successful at achieving interviews with people who had previously refused on other surveys. Where this was not possible, cases were allocated to the remaining interviewers, creating a natural experiment.
This paper will evaluate the success of targeting particular interviewers at these potentially difficult cases by comparing the response rates achieved by the two groups of interviewers and examining para data containing call records to assess any impact on fieldwork efficiency.
Missing observations due to panel attrition or nonresponse can cause severe biases in survey data if the missings are not random. Therefore, research on nonresponse and panel attrition focuses primarily on the missing observations. The aim is to identify factors that foster survey refusals and panel drop-outs. An obstacle to these analyses is the limited information about nonrespondents.
The data of almost 30 waves of the Socio-Economic Panel (SOEP) allows changing the perspective and to focus on those incredibly loyal panel members that took part in every wave. About 25% of the first-wave participants in 1894 dropped out in the first 3 waves. 50% stayed for up to a maximum of 11 waves. Anyhow, 25% stayed for at least 23 waves. This distribution affirms the idea of a continuum of resistance which ranges from complete survey cooperation ("hardcore respondents") to total survey refusal. We will compare these subgroups according to their characteristics and reasons for nonparticipation.
The second part of the paper focuses on the incredibly loyal members that stayed for 28 waves in the panel (11.85%). What kind of persons are they? Do they differ from the "ordinary" respondents? As recent research on panel attrition has revealed the impact of major life events on the probability of panel drop-outs, we intend to figure out whether loyal panel members' lives are free from such life events. The analysis obtains its particularity due to the rich data basis derived from almost 30 waves of the SOEP.
Response rate in online panels is affected by several factors. During 6th ISM Workshop 2012 in Ljubljana Caroline Jarrett's contribution (Effortmark Ltd) explored with an empirical approach and from a respondent perspective the reasons of response and of the declining trend. Saturation effect on over-surveyed people was emphasized as a determinant leading to lower willingness to respond. In order to shed more light on this phenomenon a special experiment was carried out by the CASI Research Center at the University of Bergamo (Italy) within the PAADEL project (PAnel Agro-food and DEmographic in Lombardy). We analyzed response rate of a panelists sub-population previously saturated by the quick administration of 3 surveys waves in few weeks. Findings show that elapsing time could reduce the negative saturation effect on response rate more in control group, where an incentive wasn't provided, than in the experimental group benefited by an incentive. Results seem to point out that, in a context of over-surveyed people, inner motivations are more effective in prevent saturation effect on response rate than external inducements given in exchange for participation. Nevertheless our research with the reported online experiment is not without limitations. In fact we are aware that having recourse to all the members of a sub-population and not to a probability sample of the overall population may weaken the chance of generalize the findings. Moreover using different amount of questionnaires and diverse intervals between waves and reminders could affect the results, so outcomes seek