At the US Energy Information Administration (EIA), for various establishment surveys, Official Statistics have been generated using model-based ratio estimation, particularly the model-based classical ratio estimator. Other uses of ratios have been considered at the EIA and elsewhere as well. Please see
https://www.researchgate.net/publication/261508465_Use_of_Ratios_for_Estimation_of_Official_Statistics_at_a_Statistical_Agency
At the bottom of page 19 there it says "... on page 104 of Brewer(2002) [Ken Brewer's book on combining design-based and model-based inferences, published under Arnold], he states that 'The classical ratio estimator … is a very simple case of a cosmetically calibrated estimator.'"
Here I would like to hear of any and all uses made of design-based or model-based ratio or regression estimation, including calibration, for any sample surveys, but especially establishment surveys used for official statistics.
Examples of the use of design-based methods, model-based methods, and model-assisted design-based methods are all invited. (How much actual use is the GREG getting, for example?) This is just to see what applications are being made. It may be a good repository of such information for future reference.
Thank you. - Cheers.
For instance, here is a book review for some methodology:
https://www.researchgate.net/publication/261755774_Book_Review_for_Journal_of_Official_Statistics_on_Estimation_in_Surveys_with_Nonresponse
Does anyone have an example of an implementation of a method from that book by Carl-Erik Särndal and Sixten Lundstrom?
.....
Note also that a classic book such as Cochran (1977), Sampling Techniques, 3rd ed, Wiley, would have the design-based forms for ratio and regression estimation, but Cochran also has a section on strictly model-based ratio estimation.
So, to start it off, here is a presentation which covers some implementations of model-based ratio estimation, generally model-based classical ratio estimation, for some energy establishment surveys:
https://www.researchgate.net/publication/319914742_Quasi-Cutoff_Sampling_and_the_Classical_Ratio_Estimator_-_Application_to_Establishment_Surveys_for_Official_Statistics_at_the_US_Energy_Information_Administration_-_Historical_Development
But there are many other estimators to consider, even just among ratio estimators. Consider the following:
https://www.researchgate.net/publication/319434665_Comparison_of_Model-Based_to_Design-Based_Ratio_Estimators
......
......
So, examples please. Which methods are being used where to do what?
Thank you.
In Canada they have a Monthly Miller’s Survey, and an Annual Miller’s Survey. This would be a potential application, if used as I describe in a paper linked below. As in the case of a survey at the US Energy Information Administration for electric generation, fuel consumption and stocks for electric power plants, they collect data from the largest establishments monthly, and from the smallest ones just annually. After the end of the year, for a given data item, say volume milled for a type of wheat, they could add the twelve monthly values for each given establishment, and with the annual data collected, there is then an annual census. To predict totals each month, the previous annual census could be used for predictor data, and the new monthly data would be used for quasi-cutoff sample data, for each data item, and with a ratio model, one may predict totals each month for each data item, along with estimated relative standard errors. Various techniques might apply, such as borrowing strength for small area predictions, adjustment of the coefficient of heteroscedasticity, and multiple regression when production shifts from say, one type of grain to another, as noted in the paper.
Here are the mill surveys:
Canadian Mill surveys:
Monthly Miller’s Survey:
https://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&SDDS=3403 https://www.statcan.gc.ca/en/survey/agriculture/3403
Annual Miller’s Survey:
https://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&SDDS=3443 https://www.statcan.gc.ca/en/survey/agriculture/3443
This survey information is found on page 25 of the paper below, as of this date. There will likely be some revisions to this paper. This was presented as a poster paper at the 2022 Joint Statistical Meetings (JSM), on August 10, 2022, in Washington DC, USA. Below are the poster and paper URLs.
Poster:
https://www.researchgate.net/publication/361024156_Poster_for_Application_of_Efficient_Sampling_with_Prediction_for_Skewed_Data
The paper is found at
https://www.researchgate.net/publication/362370770_Application_of_Efficient_Sampling_with_Prediction_for_Skewed_Data
.........................
If you can think of any other applications, or potential applications, please respond.
Thank you.
Right now I have a ResearchGate "Spotlight" on Knaub, J.R., Jr.(2021), "When Would Heteroscedasticity in Regression Occur?" Pak. J. Statist., Vol. 37(4), 315-367, https://www.pakjs.com/,
http://www.pakjs.com/wp-content/uploads/2021/07/37401-1.pdf,
https://www.researchgate.net/publication/354854317_WHEN_WOULD_HETEROSCEDASTICITY_IN_REGRESSION_OCCUR
Mid-page 111 of Brewer, K.R.W.(2002), Combined Survey Sampling Inference: Weighing Basu's Elephants, Arnold: London and Oxford University Press, Ken Brewer shows why there should be heteroscedastic in regression, and bounds how much there should be. This is strongly empirically evident in the case of ratio models, where there is strong correlation between x and Y. Thus it is a substantial consideration in repeated survey sampling with an occasional census for predictor data for each item, a situation found in Official Statistics. (See Cochran, W.G.(1953), Sampling Techniques, 1st ed, John Wiley & Sons, pages 205-206.) I have found this to be important for establishment surveys, having a great deal of experience with energy data.
(See my JSM 2022 paper,
https://www.researchgate.net/publication/362370770_Application_of_Efficient_Sampling_with_Prediction_for_Skewed_Data.)
Greater heteroscedasticity in regression should be associated with larger predicted-yi, as shown by Ken Brewer. But for more complex models, there may be moderating circumstances. Still heteroscedasticity should be considered the norm (and it has been important to my work in Official Statistics).
But often, especially with more complex models, heteroscedasticity may be less pronounced, perhaps because predicted-yi might be flawed. (Also, one might encounter an example where OLS applies, only to find that the example uses artificial data.)
- Badges
- Science topic
- Similar topics
- Statistics
- Survey