Passage Summary: Scientists fix errors in their data, but for some reason, these 'fixes' almost always make the results look better for Jones's theory than the original data did.

Reasoning: Researchers standardly correct data errors, but in this specific field, those corrections consistently move the data closer to the predictions of Jones's theory.

Analysis: The mystery here is the lack of neutrality in the 'corrections.' If errors were truly random, corrections should move data toward or away from a theory with equal frequency. The fact that they favor Jones suggests a bias in the error-detection process itself. Perhaps researchers are more likely to double-check (and thus 'correct') data that contradicts Jones than data that supports him. Look for an answer that explains why the researchers' methods or expectations would lead to this skewed result.