Evaluating the Effect of Dependent Sampling on the National Compensation Survey Earnings Estimates

• Activity Number: 89
• Type: Contributed
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Omolola Ojo*+ and Chester Ponikowski
• Companies: Bureau of Labor Statistics and Bureau of Labor Statistics
• Keywords: sampling cells ; multi-stage sampling ; respondent burden

Abstract

The National Compensation Survey is an establishment survey that provides measures of occupational earnings and benefits. The private industry establishment sample is divided into five sample rotation panels, with the panels being fully replaced over a five-year period. Data are updated periodically for each selected establishment and occupation until the panel in which the establishment was selected is replaced. The current sampling method allows for establishments to appear in more than one rotation panel. This can increase the reporting burden if a new set of occupations is selected for an overlapping establishment. This research uses sample simulations to explore an alternative design in which establishments selected in a sample panel would not be eligible for selection in subsequent panels of the rotation and discusses potential pitfalls in employing such a sampling approach.