Estimating sample size for complex surveys: building consensus in an environment of multiple hypotheses, multiple stakeholders and limited budgets
|Product:||Statistics Canada International Symposium Series: Proceedings|
|Authors:||Victor, J. Charles|
Diemert, Lori M.
Brown, K. Stephen
How does one efficiently estimate sample size while building concensus among multiple investigators for multi-purpose projects? We present a template using common spreadsheet software to provide estimates of power, precision, and financial costs under varying sampling scenarios, as used in development of the Ontario Tobacco Survey. In addition to cost estimates, complex sample size formulae were nested within a spreadsheet to determine power and precision, incorporating user-defined design effects and loss-to-followup. Common spreadsheet software can be used in conjunction with complex formulae to enhance knowledge exchange between the methodologists and stakeholders; in effect demystifying the "sample size black box".
health surveys, survey design, survey methodology, survey sampling.
- Date modified: