Population Synthesis: Comparing the Major Techniques Using a Small, Complete Population of Firms
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Recently, disaggregate modeling efforts that rely on microdata have received wide attention by scholars and practitioners. Synthetic population techniques have been devised and are used as a viable alternative to the collection of microdata that normally are inaccessible because of confidentiality concerns or incomplete because of high acquisition costs. The two most widely discussed synthetic techniques are the synthetic reconstruction method (IPFSR), which makes use of iterative proportional fitting (IPF) techniques, and the combinatorial optimization (CO) method. Both methods are described in this article and then evaluated in terms of their ability to recreate a known population of firms, using limited data extracted from the parent population of the firms. Testing a synthetic population against a known population is seldom done, because obtaining an entire population usually is too difficult. The case presented here uses a small, complete population of firms for the City of Hamilton, Ontario, for the year 1990; firm attributes compiled are number of employees, 3‐digit standard industrial classification, and geographic location. Results are summarized for experiments based upon various combinations of sample size and tabulation detail designed to maximize the accuracy of resulting synthetic populations while holding input data costs to a minimum. The output from both methods indicates that increases in sample size and tabulation detail result in higher quality synthetic populations, although the quality of the generated population is more sensitive to increases in tabular detail. Finally, most tests conducted with the created synthetic populations suggest that the CO method is superior to the IPFSR method.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.007 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it