Advances in Business Statistics, Methods and Data Collection

19Recent Findings from Experiments in Establishment Surveys

Josh Langeland¹, Heather Ridolfo², Jaki McCarthy³, Kathy Ott³, Doug Kilburg³, Karen CyBulski⁴, Melissa Krakowiecki⁴, Larry Vittoriano⁴, Matt Potts⁴, Benjamin Küfner⁵, Joseph W. Sakshaug^5,6,7, and Stefan Zins⁵

¹U.S. Bureau of Labor Statistics, Washington, DC, USA

²U.S. Energy Information Administration, Washington, DC, USA

³U.S. National Agricultural Statistics Service, Washington, DC, USA

⁴Mathematica, Princeton, NJ, USA

⁵German Institute for Employment Research, Nuremberg, Germany

⁶Ludwig Maximilian University of Munich, Department of Statistics, Munich, Germany

⁷University of Mannheim, Department of Sociology, Mannheim, Germany

19.1 Introduction

As survey methodologists continue to grapple with increasing survey costs and decreasing response rates, it becomes more important to monitor and improve the efficiency of current data collection procedures. Experiments are useful tools for evaluating survey methodologies since investigators can isolate the effect of interventions in well controlled experiments. Whether survey managers wish to include a new element of data collection (mode of contact, timing of contacts, questionnaire design, etc.) or would like to evaluate existing methodologies, controlled randomized experiments can be utilized to measure the impact of specific interventions. While a wide assortment of literature exists for experiments in household surveys, the results are not easily transferable to establishment ...