Other Topics of Interest for the NISS-EIA Program

Using the Internet to Collect and Edit Online Data From Establishments

EIA is currently using the internet to collection much of its data. This is particularly true in the electricity and coal areas. One of the features of these internet data collection systems in the provision of on-line edits for respondents. These on-line edits have already proven themselves as valuable tools in reducing the number of call backs required each survey cycle. This project would examine EIA’s current on-line edits and suggest additional or alternative edits that would increase the effectiveness of our data collection systems. It would also be interesting to test whether these edits improved the quality of the data submissions.

Survey Design for Establishment Surveys

Future data collection efforts for EIA’s energy consumption surveys will undoubtedly entail multimode data collection involving two or more of the following modes: mail, Internet, telephone and personal interviews. Each of these data collection modes has unique requirements for the various phases of survey design such as questionnaire development, interviewer training, and data processing. Research is needed in this area to assess how the individual modes of data collection can be integrated into a well-coordinated survey. Specific research work could focus on designing a set of questionnaires that capture the same information for mail, Internet, CAPI or CATI data collection. Cognitive testing with respondents would be a necessary component of this research. Additional work could focus on designing an integrated interviewer training program for personal and telephone interviewers. Work in this area could include the use of videos and the Internet to enhance the training.

Edit and Imputation for Establishment Surveys

For some of EIA's monthly and weekly surveys, time series techniques have been used to provide "forecasts" for the individual respondent data that are expected to be reported during the current survey cycle. These estimates are used for editing incoming data and also for imputing values for nonrespondents. EIA is currently engaged in redesigning its survey computer processing systems. As a result, this is the time to consider whether to adopt new approaches for editing and imputation. This project would involve evaluating the procedures that have been used within EIA, determining alternative methods, and evaluating the alternative methods based on performance with the appropriate data series and comparison with the current methods. Criteria for adopting new approaches include model stability, accuracy, significance of parameters, and ease of implementation.

Macroeconomic Linkages to a Linear Programming Formulation of World Energy Markets

EIA is developing a 15-region world energy market model that will be used to develop the international energy outlook and to analyze selected energy policy issues. Each world-region's energy market is formulated as a linear program (LP) that minimizes system cost from energy production, transmission/transportation, distribution, refining, and intermediate transformations, to the satisfaction of end-use service demands. Two options have been developed for the energy markets solution: the first uses a dynamic LP which provides a solution for optimal technology choices and stock turnover; the second provides a time-stepped, vintaged stock, representation of the energy market which sequentially generates an estimate for each period, one at at time, and moves forward, carrying forward the decision previously made. The purpose of this project is to investigate if simple, yet reasonably useful, regional macroeconomic models can be developed for each world region and solved within a linear or non-linear framework to capture energy-economic feedbacks and trade in an aggregate manner.

Forecast Models for Early Estimates

EIA releases some monthly data with a two to three month time lag. EIA uses time series or forecasting models with monthly data to predict the totals for specific petroleum products that will be reported two or three months ahead. These "early estimates" are published to provide users with up-to-date estimates of the energy situation. The purpose of this project is to analyze and evaluate the input data and model specifications, consider the potential alternative input data series and specifications for the models. Criteria for adopting new models include stability, accuracy, significance of parameters, and ease of use with respect to publication production. It is especially important for these models to perform well in times of change. Petroleum exports data is one area that is in particular need of such an analysis.