Sampling theory methodology applicable to data validation studies
In data validation studies, surveys are conducted to obtain information about the data collection process and the uses of the data. In many cases standard sampling techniques can be used. Two methods, stratified random sampling and cluster sampling, were used for surveys in the Form 4 data validation study. Form 4 is a data collection system on monthly generation and consumption of fuels by electric power plants. A description of those applications is given. Sometimes time and cost constraints make more sophisticated controlled sampling approaches necessary. One such approach using balanced incomplete block designs is described; an appendix surveys the existence results for these designs. Sequential methods which may prove to be more cost effective are discussed, as are sequential approaches to the problem of determining the size of a population. Problems requiring further research are also discussed. Some preliminary results on the problem of stratification with respect to more than one variable are included. The results were obtained for the Form 4 respondent population. The Form 4 study indicated that standard statistical sampling methods could be useful in data validation surveys. For example, at least 30 percent of the respondents do not report net generation as the instructions define it, and only 25 percent of the state regulatory agencies use the Form 4 data. Such inferences were possible only because statistical sampling procedures were used. 3 tables.
- Research Organization:
- Oak Ridge National Lab., TN (USA)
- DOE Contract Number:
- W-7405-ENG-26
- OSTI ID:
- 5784909
- Report Number(s):
- ORNL/TM-7084
- Country of Publication:
- United States
- Language:
- English
Similar Records
Boston residential NO/sub 2/ characterization study. I. Preliminary evaluation of the survey methodology
U.S. residential consumer product information: Validation of methods for post-stratification weighting of Amazon Mechanical Turk surveys