Reader Comments

Case: Vancouver Transit Plebiscite

In 2015, Metro Vancouver residents voted on a proposed new 0.5% sales tax to pay for transit improvements. Despite a large coalition of business, union, environmental and other groups supporting the proposal, 62% of voters rejected the plan. Early in the campaign I conducted a brief analysis of the 357 reader comments that responded to a CBC article about the issue. I think they help understand why the plan failed.

A scientific poll at the start of the campaign found a that 52% supported the tax. Failure was not inevitable. Though it might seem obvious that voters would never vote for more taxes, a scientific poll of 601 residents at the start of the campaign found a slim majority of 52% supported the tax plan. Unless that poll was seriously flawed, something must have changed people's minds.

Comments in the discussion I looked at overwhelmingly opposed the tax. In a random sample of 30, 16 were against and only 3 for. This is way out of line with the article's poll. In general, I have found counting comments is an unreliable way to guage sentiment: it tends to be skewed by prolific and fringe commenters. Another unreliable indicator is the activity of particularly active commenters. Counting comments is an unreliable way to guage sentiment. Of the ten most prolific commenters, four were in favor (including me); we collectively wrote 64 comments. The other six were against, and wrote 93.

Votes, or "likes," are a better indicator, though there are difficulties assessing them. In this discussion, the 34 comments with the highest vote totals were all opposed. This is a powerful indicator that the No side was winning the arguments, but it does not indicate by how much. Up-voted comments tend to attract more votes, magnifying even small differences in popularity.

One way to deal with this is to look at exchanges between comments and replies, where the replies attracts more votes. Of comments in favor of the tax, the one that probably had the most Yes votes received 4 (one of which as probably mine; at that point I had no plan to analyze the discussion). A rebuttal to that comment attracted 8 votes. In general, replies are less visible and attract fewer votes. That makes this comparison revealing: despite its disadvantage, the No-side reply attracted more votes. If the comparison were representative, it would suggests that the No side is attracting support by a ratio of at least 2:1. Unfortunately, the numbers are too small to make any such claim. It is simply one more clue that the No side was winning.

In general, CBC vote totals are low. This is because by default the site shows new comments first, effectively spreading out the votes. Many sites show older comments first, allowing votes to accumulate and providing better indicators of sentiment. Clearly a poll is a better indicator of opinion than vote totals. What the poll cannot explain, however, is why readers are opposed to the tax. After all, the public had previously supported the idea.

Targets of criticism
Other elites5

To understand why readers were voting No, I analyzed what they had to say. Among the 34 top-rated comments opposed to the tax, targets of criticism included: TransLink, the local transit authority (16); the provincial government or the governing Liberal party (8); the mayors (1); other elites, including “smug Vancouverites” (5):

  • start rolling back those upper management high salaries and maybe taxpayers will be on board with this
  • I believe this is just an attempt at a money grab . . . nothing . . . to do with a better transit system
  • I am a bus rider . . . I am voting No. Translink cannot be trusted
  • Vote NO to send this back to the BC Liberals, where it belongs
  • The BC Liberals have brought us nothing but a higher cost of living with only worse services to show for it

Other assorted complaints included “hideous service,” “that scam called a zone system,” “jaunts overseas,” “the improvements dont go far enough,” “Port Mann bridge tolls,” a truncated Broadway subway for the wealthy. Elsewhere there are criticisms that it hurts the poor and that it subsidizes riders.

The No campaign, which lacked the institutional support and organizational clout of the Yes side, had been pounding on real and perceived weaknesses with TransLink (particularly the near half-million dollar salary of its CEO). Ultimately, the vote was widely seen as a rejection of TransLink, not of transit. The top comments reflect this pattern; this discussion relatively early in the campaign was an indicator that the No side's message was resonating. Voters saw the vote as a rejection of both TransLink and the Province. But I think the comments reveal something else important: the vote was also against the provincial government, and against elites in general.

Unlike in some U.S. jurisdictions, there is no law requiring the B.C. government to hold a referendum on new taxes. The mayors of the region had been lobbying for transit funding for years. They put together the transit plan, and they proposed a sales tax. They did not want a vote, which they knew would be risky, but the province insisted on it. From the perspective of transit advocates, the mayors supported transit expansion, TransLink supported transit expansion (of course): but the Province created a roadblock. Yet many commenters who said they were against the plan criticized the Province.

When I analyzed this discussion at the time, I learned something I had not previously understood: very few citizens distinguished between TransLink and the Province. Both were seen as birds of a feather: elite, highly-paid technocrats who wished to impose a new tax on ordinary citizens. Even though the Province demanded a referendum that most observers expected to fail, voters saw a No vote as a rejection of both TransLink and the Province. Many of them said that they would vote Yes if TransLink was restructured (it has an appointed board), and believed a No vote would pressure the Province to make it more democratic (something the Province continues to refuse to do). In comment discussions, my argument that the Province was using TransLink as a scapegoat failed repeatedly. My analysis explained why this argument could not work.

The anger in these comments also this clarified why the Yes campaign was so ineffective. The Yes side boasted that they had the "biggest coalition in B.C. history," including over 100 groups in business, government, labor, environmentalim and education, from the Vancouver Board of Trade and the BC Federation of Labour to the David Suzuki Foundation and the Vancouver School Board: in other words, a laundry list of elites.

The Yes campaign confirmed they were a bunch of elites who knew what was best. The No campaign told voters that a bunch of fat-cat elites were picking their pockets. The Yes campaign confirmed they were a bunch of elites, that they knew what was best, and that voters should do as they were told. One prominent series of ads, unbelievably using the imperative and "you" instead of "we", told voters what to do: "You vote Yes." As the comments demonstrate, the Yes campaign played into their opponent's hands.

Though this is the only discussion I analyzed in detail, I read many others. I tried just one argument that resonated with angry commenters: the suggestion that politicians were using the vote as an excuse to "drown government in a bathtub" and reduce public services. Like the argument of the No campaign, this is one that targets elites.

Leading up to the Translink plebiscite, comments were a window into the minds of voters. An imperfect, not entirely trustworthy window no doubt, but one that went beyond the numbers in the polls to help understand why the vote was lost.