Employing Matching to Alleviate Bias in Survey Data
Bibliographic record
Abstract
There are two main obstacles that impede the ability of political scientists to evaluate the effect of campaigns and mass communication. First, it is dif-ficult to measure media and campaign exposure reliably. Surveys often rely on self-reporting from a participant, which is unreliable and contributes to the difficulty of measuring exposure to a message. Second, it is difficult to find significant media biases that can act as natural experiments. Public opinion in the aggregate is reasonably stable, even in the short term, and movement of opinion is expected when there is a change in the balance of messages across time or individuals, which is rare (Ladd and Lenz, 2007). Also, as Bartels em-phasizes, media coverage is an equilibrium condition, and opposing messages mask each others effects (1993). Newer surveys incorporate time dimensions, such as rolling cross-sections, in addition to the computer assisted and weighted internet survey designs. Political scientists are hoping to use these improved methods to create causal inferences, which are inherently difficult using obser-vational data, instead of correlation inferences. I examine what techniques are best for analyzing these new designs and whether these methods are sufficient to allow for causal inferences using the 1988 Canadian National Election Studies data as a test case. Employing a Genetic Matching method, I find that in the 1988 election, there was a significant debate effect regarding perceptions of the candidate that had no significant effect on the vote decision of the election.∗† ∗The author would like to thank Jas Sekhon and Henry Brady for advice and guidance throughout this project and Rocio Titiunik for answering every programming and data question encountered.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.194 | 0.426 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.003 | 0.003 |
| Bibliometrics | 0.006 | 0.017 |
| Science and technology studies | 0.003 | 0.003 |
| Scholarly communication | 0.003 | 0.004 |
| Open science | 0.005 | 0.005 |
| Research integrity | 0.003 | 0.002 |
| Insufficient payload (model declined to judge) | 0.011 | 0.003 |
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 itClassification
machine, unvalidatedMachine predicted; the direct Gemma label and the distilled Codex classifier agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".