The Atlantic Cities has a must-read about why people still fear being hit by New York subway trains, even though the subway is one of the safest ways to travel. The union representing New York subway workers is proposing a series of steps to reduce the risk of subway-person collisions, assisted by lurid graphics. It just so happens that their main ideas require hiring more unionized staff! This includes the proposal to slow down trains as they enter stations, which will slow down everyone's travel and increase the number of trains, and hence drivers, needed to maintain the current frequencies.
If subway-person collisions were common, these would be valid safety precautions. Transit agencies do take these expensive steps when an objective safety issue arises.
But as the article states, the facts are these:
And yet, subway deaths remain exceedingly rare. The fatality rate has not changed significantly over the last decade. Of the 55 fatalities on the subway tracks in 2012, 19 were suicides. The remaining 36 accidental deaths on the New York City subway in 2012 occurred on 1.66 billion subway rides. That’s one death for every 46 million rides.*
For infrequent riders, death on the rails is less likely than being hit by lightning. If you’re a twice-a-day commuter, you’re likely to be killed once every 100,000 years. …
A significantly more dangerous feature of city life is car traffic. Even the most dedicated mass transit commuters are twice as likely to be hit and killed by a car than a train. One in 50,000 New Yorkers is killed by a car each year, and one in five hundred is injured.
If your desire to continue living is quite clear in your mind, it's very easy not to be hit by a subway train. Stay behind the yellow line. If that doesn't feel safe, stay back still further.
The real question is: Why do we reward the media for giving us lurid details of every subway fatality but not for every road fatality? The Atlantic article has some ideas about that, though I think it dwells too long on the late-20c period when New York was much more objectively dangerous than it is today.
Let's also note that some subway systems are installing platform walls with doors (like these in Singapore) opening only when and where a train door is present. These further reduce risk and are useful in stations with very high crowding, but are very expensive (Over $1m per station) and technically difficult to fit into the already-compact New York platforms. The MTA appears to be considering these, and other technological options. The goal, however, would be to increase the feeling of safety, since actual safety is already extremely high. How infinitesimal does the risk need to be before we focus our investments on other things, like more useful service?