ESRA 2027 sessions by theme
Back to Overview of Sessions
Handling Methods and Examining for Handling Bias from Mode Effects in Analyses of Mixed-Mode Survey Data |
| Coordinator 1 | Professor Richard Silverwood (University College London) |
| Coordinator 2 | Mr Liam Wright (University College London) |
| Coordinator 3 | Dr Georgia Tomova (University College London) |
Session Details
Surveys continue to adopt mixed-mode designs, both within and between waves of data collection. Unaccounted for, mode effects – systematic differences in measurement between modes– can be a problem for obtaining unbiased estimates from analyses of mixed-mode survey data. For instance, mode effects can induce spurious associations between variables subject to mode effects, while changes in survey mode between waves can bias estimates of temporal change.
The development of methods that assess or account for the bias from mode effects is essential, given prior and projected use of mixed-mode survey data. But existing methods generally make assumptions unlikely to be achieved in practice (e.g., no mode selection, i.e. no differences in respondent composition between modes). This session will invite submissions on: (1) new methods that can robustly account for bias from mode effects post hoc; (2) research providing evidence (empirical or simulation) on bias from mode effects in practice, especially in data analyses estimating associations between variables; and (3) frameworks helping data users reason about the likely effect of mode effects in particular analyses.
