Author
Abstract
This paper evaluates the effect of switching an existing panel study from a unimode face-to-face design to a sequential mixed-mode design (web followed by face-to-face interviewing) on attrition. I use large-scale randomised experimental data from the Innovation Panel of Understanding Society. While the introduction of a mixed-mode design increases panel attrition one wave after the mode switch (IP6), the effect is eroded at subsequent waves (IP7-IP9). The offer of higher incentives to sample members in the mixed-mode group cancels the negative effect of mixed-mode on attrition one wave after the mode switch (IP6) and leads to higher response over time (IP7-IP9).