Step 1: Understanding the Question:
Experts expect better safety training to make workplaces safer, but a survey found that after manufacturers improved their training programs in the 1980s, on-the-job accidents actually went up in the following months. We need the option that best explains this apparent contradiction.
Step 2: Key Formula or Approach:
An apparent contradiction like this is often resolved by questioning the assumed order of cause and effect, maybe the accidents were not caused by the new training, but were instead the reason the training was improved in the first place.
Step 3: Detailed Explanation:
Option 3 says manufacturers tend to improve their safety training only after they have already noticed a rise in on-the-job accidents. This flips the assumed direction: the accident numbers were climbing before the improved training was brought in, and the survey caught the tail end of that existing rise in the months right after training changed, not a new rise caused by the training. This removes the puzzle completely, training did not cause the increase, a pre-existing increase caused the training to be improved.
Option 1 talks about a different sector, transportation, where accidents stayed flat. This is just a comparison point and does not explain why manufacturing accidents rose after training improved.
Option 2 links improved training to a growing workforce. A bigger workforce could raise the raw accident count, but this does not by itself explain why the rise specifically follows soon after training improves, it is a weaker, less direct explanation than option 3.
Option 4 says the increase was probably not random. This makes the contradiction harder to dismiss, it does not resolve it, if anything it deepens the puzzle by ruling out chance as the explanation.
Option 5, dropped in the folded version, said other safety measures were already in place before training improved, this does not explain the timing of the increase in accidents right after training was improved.
Step 4: Final Answer:
The discrepancy is best resolved by the fact that manufacturers tend to improve training only after accidents have already started rising, option 3.