Fair points. Wikipedia'a Public Good article shows at the top a fireworks display as an example. This is interesting because most large fireworks displays I'v

Fair points. Wikipedia'a Public Good article shows at the top a fireworks display as an example. This is interesting because most large fireworks displays I've seen do offer a small number of viewers the ability to watch from nearby (i.e. to obtain more-privileged access) in exchange for money. So non-excludability may not be complete and absolute: a "public good" may be available to all comers, but also available in an enhanced form for pay. Similarly with outdoor concerts having VIP seating, parades, and so on. In the end this is a semantic issue and we can probably agree that the language is not precise enough to succinctly describe what's going on with the Survey now. But perhaps the examples above are illustrative enough, as I have rarely heard complaint about the visible existence of ticketed/VIP seating at public events. Some others here claim the Survey's dissemination would not be objectionable if it were done in a more transparent fashion (but still in tiers for pay). Finally, it's interesting that this particular news item has such a "long" advantage period for paying customers. Two seconds is a long time for computers, and even if the Survey were released "simultaneously" to everyone, those who paid to build better networks and systems to use it would still benefit. The end result might be similar, only without the extra rents paid to Reuters. If Reuters and UMich argued that without these rents the Survey could not be produced at all, I wonder if it would calm the critics.