The argument tells us that most bicyclists aged 18 or older have lights on their bike. 18 + most -> lights
and concludes that most bicycles that have lights are at least 18 years old. lights most -> 18+
This is a flawed reasoning because the only proper inference we can make from a most statement is if we reverse it and replace most with some:
Some bicyclists that have lights on their bicycles are at least 18 years old.
Since we have no information about the overall group size of the bicyclists that have lights on their bike, we cannot conclude that most of them are at least 18 years old. It is possible that overall there are more bicycle riders that are under 18, in this scenario it is possible for most of total bicyclists with lights to be under 18 but at the same time, for most of the 18+ riders to have lights on their bicycles. Thus, we can only infer that some (at least 1) bicyclist with lights must be over 18.
Let's look at (C):
Most of the residents who voted are on the Conservative party list.
Voted most-> Conservative party
Thus most of the residents on the Conservative party list voted in the same election.
Party most -> voted
This is the exact same flawed inference as we see in the stimulus. We can only infer that some of the people on the Conservative party list voted in the last election:
Party some -> voted
Let me know if this makes sense and if you have any further questions.