The Effect of Response Measures in Business Surveys

Often asked questions concerning business surveys are:

• What will be the increase in response rates if we apply such-and-such measure(s)?
• What would be perfect timing for these measures? And,
• What will be the costs?

Basically these questions ask for an efficient strategy to get response, aiming for a cost-efficient survey design both for the survey organisation (like a National Statistical Institute) and businesses alike, not burdening and chasing businesses too much. The effects of measures to get response for business surveys have not been studied systematically as much as for social surveys. Obvious reasons for this may be the fact that business surveys are mandatory by law, and the costs involved in getting response are not as
high as for social surveys using CAPI or CATI. Nowadays however, with ever decreasing budgets, and the pressure to reduce response burden even more efficient business surveys designs are required. An overview of various measures has been presented by Snijkers et al. (2013), but quantitative information to answer the above mentioned questions was to a large extend still lacking. In a study conducted at Statistics Netherlands (Snijkers et al., 2018) the effects of various measures to get response have been analysed for a number of business surveys, without doing an experiment. These measures include the obvious measures, like sending advance letters to
businesses introducing the survey and soliciting survey response, sending pre-due data reminders, and after the due date sending one or more reminder letters. For one survey (the Survey on International Trade in Goods) we modelled the effects of these measures using survival analysis, to find out what would have happened without any of these measures. At the
lecture the results will be presented.

Date

Speaker

Venue