Springtime in Alberta…

“Springtime in Alberta” is one of my favourite Ian Tyson songs and it’s proving to be somewhat prescient in light of the current election campaign.

Just like spring time in Alberta
Warm sunny days endless skies of blue
Then without a warning
Another winter storm comes raging through

Although the polls are show a remarkable lead for the Alberta NDP, something most people would have considered to be impossible just three weeks ago, many people also seem to think that another winter storm may yet blow through this campaign, just as it did in the dying days of the 2012 campaign. This time, though, it might not be the Wildrose losing its lead, but the NDP.

I won’t go out on a limb to make a prediction, but I do think that the lead in the public opinion polls should be taken more seriously than most people currently are and that a change in government is possible. The graph below shows the results of each public opinion survey published in the 2012 and 2015 elections. Obviously, the NDP has a big, big lead in current public opinion surveys and it’s getting bigger in the last day or two. Continue reading

These results do not count for undecided voters, and that will certainly be something to watch in the next 4 days.

As people point out, the Wildrose Party had a substantial lead in 2012, and that evaporated. This is true. But it is worth pointing out that polls can differ from final vote intention for two reasons. Not to put it too bluntly, but they can get public opinion wrong or public opinion can change between the publication of poll results and the casting of ballots.

The fear with most polls these days, and particularly in this campaign, is that because they are online surveys of voluntary panel participants or automated “robo-polls,” they can overestimate those who have the strongest motivation to stick around and participate. IN this case, it seems plausible to think that the polls could be overestimating opposition party support, given how long in the tooth the PC dynasty is and some of its more spectacular instances of “foot in mouth disease.” Are the 2015 polls overestimating support for the NDP and underestimating support for the PCs? Maybe. But it’s worth it to cast a close eye on the final data point on the 2012 graph. The field work for that poll was done the day before the election and the results were published that night. It showed more than a 10% point jump in support for the PCs and a corresponding drop for the Wildrose. And there were still another 12-24 hours until people interested in casting a ballot, did so. There are a lot of voters who tell survey researchers that they make up their mind on who to vote for the day of the election. To me it seems at least as plausible to suggest that the 2012 polls were actually accurately gauging public opinion over the course of the campaign but that many voters made a final switch at the last minute back to the PCs. Polls weren’t necessarily wrong; they were just measuring decided vote intention, which, it perhaps bears emphasizing, can change once undecideds make up their mind.

The other reason that some people are hesitant to believe that the NDP could win are because it is Alberta and, in the words of Premier Prentice, “Alberta is not an NDP province”. Essentially, this the belief that Albertans are fundamentally opposed to government intervention.

This argument is one of the more common ones but is it far too simple to accept and rule out an NDP victory on Tuesday. The graph below shows the difference between Alberta and Canadian public opinion on a more or less random set of questions from the Canada Election Study. For each item, I took the percentage of Canadians that selected the most liberal option and subtracted from it the percentage of Albertans that selected the same item. The y-axis shows the difference in percentage points between Canadians and Albertans selecting the most liberal option. Positive values suggest that more Albertans selected the most liberal option than Canadians; negative values suggest that more Canadians selected the most liberal option.


Two things stand out to me in this graph: First, on most of the items, the gap between Albertans and Canadians has been getting smaller. Perhaps this is a product of demographic change in Alberta, as people have flocked to the province to participate in the growing economy, perhaps not. Either way, the gaps between Albertans and Canadians don’t seem to be that large. Keep in mind there is an margin of error associated with each measurement of 2-3 percentage points. There is a slight tilt toward conservatism, but it’s not so dramatic that we should think an NDP win on Tuesday to be impossible. In fact, in 2008 and 2011 there were more Albertans expressing a willingness to increase personal income taxes, than in the rest of Canada. Moreover, the gap between people who strongly disagree that job creation should be solely left to the private sector has been dropping.

And if public opinion isn’t convincing enough, government actions also paint a more complex picture. The PC governments that have governed Alberta have been far more flexible than their caricature suggests. In the 1970s, the Lougheed government purchased an airline and directly subsidized the nascent oil sands industry. Despite a turn to the right under Premier Klein, the government of Alberta remains the only government in Canada with a Crown corporation dedicated to retail financial operations (i.e. it owns a bank). Currently, Alberta has the highest per capita expenditures on health care , but they have the lowest taxes.

But to me, this is not the mark of an ideologically conservative population, but quite a normal one, one that wants things good things without paying for them. Up until now, the Progressive Conservatives have had the fiscal resources to provide both those for Albertans. To me, that is the bigger reason for the PCs’ longevity: not any kind of deep, ideological commitment to right-wing governance in the population. And the combination of a tough recent budget that raised taxes and fees, an early election despite fixed-election legislation and frankly, a brutally, inept campaign mean that the polls should be taken seriously.

It is, in fact, springtime in Alberta.

Mentors and Giants: An Interview with Christopher Achen

Christopher Achen is Roger Williams Straus Professor of Social Sciences and Professor of Politics at Princeton University. According to his bio, he “was the first president of the Political Methodology Section of the American Political Science Association, and is a member of the American Academy of Arts and Sciences. He has received fellowships from the Center for Advanced Study in the Behavioral Sciences, the National Science Foundation, and Princeton’s Center for the Study of Democratic Politics. He received the first career achievement award from The Political Methodology Section of The American Political Science Association in 2007. He is also the recipient of an award from the University of Michigan for lifetime achievement in training graduate students. Recent academic placements of graduate students for whom he was the principal dissertation advisor include Stanford, Duke, and the London School of Economics.”

During my first year at Laurier, I was appointed colloquium officer. We had a tiny budget, but I, being fresh out of grad school, was feeling ambitious and was determined to try and bring to Laurier a big name in American political science to spend the day with us. My hope was that this individual would give a public lecture and host a smaller workshop with political science graduate students and faculty members. There was also, at the time, a strong push to help develop LISPOP, and so I thought I would attempt to bring in someone who was a giant in public opinion and/or methodology.

Continue reading

One of the first names that came immediately to mind was Chris Achen. I remember reading his monograph, Intermediate Regression Analysis (Sage: 1982), at UofT, which, although dated, really helped me get a handle on the logic and math underpinning regression analysis. As well, although I’m sure there were others, at the time I thought he was one of the few “big names” in political science who was rallying against a certain methodological trend of “dumping” as many variables as one could into regression models and magically finding statistical significant relationships. And so I really wanted to meet him!

Happily, Chris accepted my invitation and his public talk and workshop were amazing. As my colleague Loren King mentioned the other day, he was a pioneer in getting to know your data and figuring out how to do matching to establish causality long before matching became a trend in recent days. An added bonus was that Chris was such a nice, humble, and encouraging guy. I made a lot of rookie mistakes during my first year at Laurier, including taking Chris to a bit of a “dumpy” bar instead of a fancy restaurant (darn budget!). But rather than complain, he happily had a beer and burger with the rest of us and told me he preferred the bar to the fancy restaurant (even though I’m sure that’s not true)!

Even though I haven’t spent very much time with Chris in person, I count him both as a giant and a mentor to me. His work and his visit to Laurier had a profound effect on how I have pursued my academic career so far.


I wish someone had told me at the beginning of my career

How much time faculty spend on committees and administration, and how important it is to learn to manage those obligations while making sure that teaching and research get the time they need.

The individual I admire the most academically

I have a long list of predecessors I greatly admire, but Harold Gosnell, founder of political methodology, is a personal favorite. He did the first field experiments in the 1920s, he used statistical techniques in the 1930s that didn’t come into common use for another 30 years, and he pioneered among students of African-American politics. I had one memorable lunch with him when he was already in his nineties. Alas, he is no longer with us.

My best research project during my career

I always feel that my current one will be the best.

My worst research project during my career

I spent a summer before Bayesian software was invented, laboriously programming and analyzing a Bayesian model of the representativeness of Austrian mayors.

The most amazing or memorable experience when I was doing research

I wrote a paper about rational party identification in 1989 and published it in 1992. The original draft included a footnote saying that if the argument of the paper was correct, the Republicans would become the majority party in the House of Representatives in the not-too-distant future. At that point, the Democrats had controlled the House for nearly all of the last 60 years. The footnote seemed crazy, and I lacked courage. I took it out before publication. Of course, in the 1994 elections, the GOP took over the House, and they have controlled it all but four years since then. The moral: stick to your guns.

The one story I always wanted to tell but never had a chance

One year my APSA paper with Duncan Snidal collapsed completely on August 15, two weeks before the convention. We had to work hard and quickly on a new paper, worrying that the argument was all wrong, and hoping that no one would attend the panel. Instead, it struck a nerve and, after considerable revision, became the lead article in World Politics. We were lucky. But there is a moral here, too: sometimes not worrying about crossing t’s and dotting i’s can free the mind.

A research project I wish I had done

Using political science tools to understand the Weimar elections that led to Hitler. The electoral patterns are quite complex and varied across German subdivisions, as Weimar historiography makes clear. Just mushing the electoral units together statistically at the national level was a very helpful starting point twenty-five or thirty years ago, but it has long been clear that something more locally informed is needed in the twenty first century. A serious command of German and of regional history and politics, a good deal of time in archives, and many years of patient investigation would all be needed, but the result would be a tremendous contribution. I hope someone will do it.

If I wasn’t doing this, I would be

retired from playing middle linebacker for the Oakland Raiders in their glory years. Alas, I am small, slow, and talentless, so I had to go into poli sci.

The biggest challenge in American politics in the next 10 years will be

managing the growing specialization into subfields—political behavior, institutions, American political development, public law, race and politics, public policy, and much else. The important problems and the most interesting intellectual challenges cut across those divisions.

The biggest challenge in political science in the next 10 years will be

making experimentation and other forms of causal inference become as fruitful on the big, longstanding theoretical issues in the study of politics as they have been in political psychology and public policy.

My advice for young researchers at the start of their career is

listen to wise advice, but follow your heart.

Public Opinion, Precaution and Toxicology

There is news today that a coroner in New Zealand formally ruled that the excessive consumption of Coca Cola led to the untimely death of Natasha Harris three years ago. Harris, it seems, developed a nasty habit of drinking up to 10 litres a day. “I find that when all the available evidence is considered, were it not for the consumption of very large quantities of Coke by Natasha Harris, it is unlikely that she would have died when she died and how she died” is what the coroner determined.

So what does this have to do with a blog about Canadian, politics and public opinion? Because this tragic incident illustrates a bedrock principle of toxicology that citizens, journalists and politicians often forget when our attention turns to issues of concern with possibly toxic substances. That principle is “the dose makes the poison.” It was formulated in the 16th century by the Swiss-German chemist Paracelsus and it holds that all substances (Coca-Cola included) are toxic in some quantity. Even something as seemingly innocuous as water can be toxic when enough has been consumed. This poor soul died from the excessive consumption of water during a brutal hazing ritual in New York. From this observation, the traditional role of toxicology has developed which is to ascertain at what level of exposure any given substance causes adverse effects.

With the rise of concerns about toxic substances in the post-World War II era, scientists have discovered some substances that deviate from this principle (see tamoxifen, an important drug in chemotherapy). But these effects are rare, difficult to demonstrate and the scientific debates that surround them are heated and complex. But when issues shift into the public and media spheres, the complex science behind these assertions, sadly, usually gets lost in the shuffle.
Continue reading

There are multiple reasons for why this is. One reason is our own inability to perceive risks rationally. There is so much literature on the psychology of human risk perception that it is difficult to know where to start, but I might recommend this thispaper or this book. The long and short of the field of risk perception is that humans get very frightful about lots of things that actually present very little danger. After September 11th, Americans turned away from planes for their long trips and hopped into their cars, leading to a sharp increase in traffic fatalities in the months following.

Chemicals are one of those things that people fear a great deal. Ask people if they think the colourless, odourless chemical dihydrogen monoxide (the ingestion of which is the cause of thousands of deaths each year) should be banned and you’d be surprised (and a little concerned) at the number of people who would agree.

So when the public and the mass media turn their attention to issues of possible harm from exposure to chemicals, fears can be stoked, perspective can be lost and can lead, ultimately, to over- and unnecessary regulation of products and activity. For example, I was able to show in this paper that within-state newspaper coverage of Bisphenol A was positively linked to the chance that a US state legislature would introduce or adopt legislation banning products made with BPA the following year. It showed (again) the important role that increased media salience can have on policy changes. This, even though, scientific information suggests that exposure to BPA is hundres, if not thousands of times lower than the level at which there is a consensus that it causes adverse effects in rats (see here).

This research area is rife for investigation and it’s become a long-term project I’m working on, so I’m going to have a lot more to say about this in the future. I thought the news about Ms. Harris’ tragic death due to the excessive consumption of Coca-Cola was a useful way to set the stage for more thoughts and research to come.