Maybe a good trick-or-treat location should be defined less by the available candy and more regarding its design:

Great neighbourhoods for trick-or-treating also tend to be great neighborhoods for families everyday:

• Tree-lined streets designed for walkers more than speeding cars.
• Enough density and community completeness, to activate what I call “the power of nearness” – everything you need, nearby.
• Good visual surveillance through doors and stoops, windows (and I don’t mean windows in garages), porches and “eyes on the street.”
• Connected, legible streets that let you “read” the neighbourhood easily -grids tend to be good for this, but other patterns work too…

If kids ARE being driven in, that can mean it’s a great neighbourhood from a design perspective (or perhaps just that it’s a more affluent community, with “better candy”) — but having too few local kids can show that there isn’t enough housing diversity, new infill, and family-friendly “infrastructure” to keep kids in the neighbourhood. In fact, in many beautiful, tree-lined neighbourhoods popular on Halloween, the number of local kids may be actually dropping, with resulting pressures on local schools to close. This as household sizes decrease, and new density and “gentle infill” that could stabilize the population and keep kids in the neighbourhood, is often locally resisted.

From this point of view, good neighborhoods promote walkability and ultimately sociability. There are few times of years where this matters as much as Halloween as many Americans do not regularly walk down their streets to visit a number of neighbors at once.

More broadly, the practice of trick-or-treating is closely tied to social trust. Even with no documented cases of poisoned candy, parents want to know that their kids are safe. And with declining social trust in the United States, again, there are limited numbers of opportunities where Americans ritually interact with physical neighbors as opposed to seeking out people they whom they share an identity or interests.

It sounds like there is an empirical question to be answered here: do neighborhoods with (1) more traditional design and (2) higher levels of social trust (which may or more not be related to the neighborhood design experience more satisfying trick-or-treat experiences (measured by numbers of children trick-or-treating, percent of households providing candy, and perceptions of whether the neighborhood is a good place for this)?