A review and evaluation of the use of longitudinal approaches in business surveys

Abstract

Business surveys are not generally considered to be longitudinal by design. However, the largest businesses are almost always included in each wave of recurrent surveys because they are essential for producing good estimates; and short-period business surveys frequently make use of rotating panel designs to improve the estimates of change by inducing sample overlaps between different periods. These design features mean that business surveys share some methodological challenges with longitudinal surveys. We review the longitudinal methods and approaches which can be used to improve the design and operation of business surveys, giving examples of their use. We also look in the other direction, considering the aspects of longitudinal analysis which have the potential to improve the accuracy, relevance and interpretation of business survey outputs.