Canadian provinces and territories on 30 questions before the 2011 federal elections

[someone123]

Originally Posted by ciudad_del_norte

Hmm. Well these are kind of interesting, the colours and shading are kind of useless and misleading. It gives an idea, but "agree more than average" and "agree less than average" aren't very helpful as applied. How much more than average? You can get kind of an idea with the spread that is vageuly shown with each map, but its hard to meaningfully interpret when black vs. deep yellow actually mean. The way people interpret colour scales like this is not well translated to this sort of pseudo statistical analysis.

I'm not sure of a better way to present this stuff in such a user friendly format, I just know if I tried to present this sort of information in this format at school I would have been ripped apart.

Yup. These may look neat to some but they don't provide much information. You can't even see most of the ridings and it's not clear how large the differences are or how big the sampling error was in relation to the differences measured. It looks like they deliberately chose a scale here to exaggerate the differences between ridings instead of representing the actual distributions of responses.

The visualization with the dots is interesting. You still need to be careful about how you generate your colour mappings though. Usually the colourings imply vastly greater differences than what are measured.

[Acajack]

Originally Posted by eternallyme

I find the boundary along the Ottawa River between the upper Ottawa Valley and the regions across the river the most pronounced differences.

Yeah, the one slight exception to that is the riding of Pontiac which is just northwest of Ottawa-Gatineau. It's about one-third anglophone, so on many issues it's similar to the rest of Quebec, but in a lighter shade.

[MolsonExport]

Originally Posted by Acajack

Interesting you should mention that.

There was a study a few years ago (maybe by Jack Jedwab - he likes that stuff) that showed that on almost all issues except independence and language (and even on this last one it's not entirely clear-cut), Anglo-Quebecers were much closer in mindset to francophones in Quebec than they were to other Canadians.

Yeah, I saw that study too. Probably by J Jedwab. And I agree that most Anglos in Quebec are closer to the Francophone viewpoint on many issues, compared to the ROC viewpoint.

[Dr Nevergold]

After reading that, I'm closer to Quebec on most issues. Save for the role of private health care.

[ciudad_del_norte]

Originally Posted by someone123

Yup. These may look neat to some but they don't provide much information. You can't even see most of the ridings and it's not clear how large the differences are or how big the sampling error was in relation to the differences measured. It looks like they deliberately chose a scale here to exaggerate the differences between ridings instead of representing the actual distributions of responses.

The visualization with the dots is interesting. You still need to be careful about how you generate your colour mappings though. Usually the colourings imply vastly greater differences than what are measured.

Yeah, it's definitely set up to exaggerate differences. It's one of the troubles with nuances of using colour scales. Typically in a case like this people will associate anything near whyte to be more 'average' and assume any sort of noticeable contrast represents a dramatic deviation. The colour scale progresses quickly and polarizes the sides when if you look at the actual distributions things are typically much closer to the average than the colours imply. A more realistic portrayal would show a much slower progression to the black/yellow with the more dramatic shift happening towards the extreme, not immediately outside of the average. That portrayal, however, would make the country look much less dramatic (and likely more realistic).

[Acajack]

Originally Posted by ciudad_del_norte

Yeah, it's definitely set up to exaggerate differences. It's one of the troubles with nuances of using colour scales. Typically in a case like this people will associate anything near whyte to be more 'average' and assume any sort of noticeable contrast represents a dramatic deviation. The colour scale progresses quickly and polarizes the sides when if you look at the actual distributions things are typically much closer to the average than the colours imply. A more realistic portrayal would show a much slower progression to the black/yellow with the more dramatic shift happening towards the extreme, not immediately outside of the average. That portrayal, however, would make the country look much less dramatic (and likely more realistic).

Not saying you are putting it this bluntly, but if you look at who Vote Compass are, I somehow doubt they have a vested interest in showing Canada being dramatically divided, or in taking sides in the great SSP Canada debate over which parts of the country are more progressive than others.

[ciudad_del_norte]

Originally Posted by Acajack

Not saying you are putting it this bluntly, but if you look at who Vote Compass are, I somehow doubt they have a vested interest in showing Canada being dramatically divided, or in taking sides in the great SSP Canada debate over which parts of the country are more progressive than others.

I don't think they have some hidden agenda in trying to make somebody look better or worse. But they would still be interested in producing interesting and engaging communications material. That would involve playing up differences.

These scales aren't set by accident. I'm not even in communications but my education has enough communications worked into it to know that you have to take into account the way people interpret things. Nothing about what they are presenting is a lie, but do you honestly believe that the majority of people looking at these maps actually understand what they mean? If a person takes the time to see what scale they are using, and what they are actually showing then it means something completely different than most people who skim through the images and look for colour contrast - which I would assume would be most.

[Drybrain]

Interesting that Nova Scotia is shown as the province most amenable to immigration increases. (Locally, we're absolutely deluged with a insistence that our province is aggressively xenophobic and unwelcoming to newcomers.)

It's pretty interesting as well to see Alberta's data confirm the province's reputation as socially conservative, but as others said below, it seems the regional differences are exaggerated due to the way the data is presented. I also wonder about the methodology of the data collection.

[1overcosc]

Wow. Was poking into this in more detail and found this heartwarming set of charts: Canadians response to the statement "Marriage should be between a man and a woman". Most demographics were overwhelmingly at 'strongly disagree' and even the more right-wing demographics (the old, CPC voters, etc.) were divided. No single demographic mostly agreed.

[eternallyme]

The colors are relative, not absolute. The deepest shade represents the 1st and 308th ranked ridings, not how strongly they share those views.

[eternallyme]

I think the biggest divisions are seen on military-based issues and gun control. In both cases, harsh feelings are found on both sides - relatively few take a "neutral" position.

[lio45]

Originally Posted by someone123

You still need to be careful about how you generate your colour mappings though. Usually the colourings imply vastly greater differences than what are measured.

Originally Posted by eternallyme

The colors are relative, not absolute. The deepest shade represents the 1st and 308th ranked ridings, not how strongly they share those views.

Yep, from what I gather, the calibration of the yellow->black spectrum is on a per map basis with the two extremes always represented. If on a given issue all Canadians happen to have nearly exactly the same position, you'll still see deep yellow to deep black shades on the map.

I have a problem with heat maps done that way... totally misleading... and it happens really way too often for my taste! So often that it's even happened on here in the past; I remember us discussing misleading heat maps on the subject of renewables on this very forum with Alberta's solar potential map; the heat map made the Lethbridge area seem very interesting, but on a continental-scale heat map (or a global one) the solar potential of the place is actually represented in a weak-ish color from the full spectrum while the strongest color (usually red on those) is reserved for places like the US Southwest (or Middle East, Sahara on a global one).

For the issues presented in this thread, IMO a better way to have done it would have been to calibrate the yellow->black scale in an absolute way using "realistic" global positions for the two extremes (i.e. for each extreme, you stick with the most extreme position that still has decent amount of support at least somewhere on a global scale) and then painting Canada that way.

For example, on gay marriage and rights, the entire map would be ranging from deep black to something like light grey at worst. (Scale calibrated so that places like Iran would be the ones in deep yellow.)

Obviously the maps would convey a different message then... they wouldn't be highlighting the differences within Canada any more, but rather be showing how different or how similar Canadians are on those issues in a global context.

[lio45]

Originally Posted by Drybrain

It's pretty interesting as well to see Alberta's data confirm the province's reputation as socially conservative...

Which it totally is, within Canada.

...but as others said below, it seems the regional differences are exaggerated due to the way the data is presented.

These maps are _designed_ to highlight regional differences. For each map, the color spectrum goes from the most extreme position found within Canada to the opposite most extreme position found within Canada.

In every case, there's no info on how large or how small the gap between deep yellow and deep black is.

[big W]

Originally Posted by lio45

Which it totally is, within Canada.



These maps are _designed_ to highlight regional differences. For each map, the color spectrum goes from the most extreme position found within Canada to the opposite most extreme position found within Canada.

In every case, there's no info on how large or how small the gap between deep yellow and deep black is.

Which is what they want to convey to create discussion. If you play up the differences it would allow for discussion. More importantly it actually highlights the differences in the rural areas as they make up the larger portions of the map and urban ridings are small so they barely appear.
Also compare the City maps (they appear to be based on ridings) where most people live the differences them become less so on a national basis. In other words downtown Edmonton or Calgary tend to be quite different than the rest of the province and more similar to say Montreal on many issues vs rural ridings in Alberta. The thing is urban Alberta is 80% of the population and 1% of the map.

[ciudad_del_norte]

Originally Posted by lio45

Which it totally is, within Canada.



These maps are _designed_ to highlight regional differences. For each map, the color spectrum goes from the most extreme position found within Canada to the opposite most extreme position found within Canada.

In every case, there's no info on how large or how small the gap between deep yellow and deep black is.

Yup.

It's also important to look at where the average is set. The natural assumption for most is that 'average' or white is neutral but the 'average riding' may not necesarily be neutral on an issue. There are a lot of moving peices here which is fine, but I've seen these maps posted around the internet a few times with some pretty strange conclusions given what the data actually shows.

[big W]

Originally Posted by 1overcosc

Wow. Was poking into this in more detail and found this heartwarming set of charts: Canadians response to the statement "Marriage should be between a man and a woman". Most demographics were overwhelmingly at 'strongly disagree' and even the more right-wing demographics (the old, CPC voters, etc.) were divided. No single demographic mostly agreed.

What is interesting in this scale is that Canadians born in Canada (French or English) are the same on the issue and the differences is the foreign born.

[ciudad_del_norte]

Originally Posted by big W

What is interesting in this scale is that Canadians born in Canada (French or English) are the same on the issue and the differences is the foreign born.

...also gender although not really suprising I guess.

[kwoldtimer]

Originally Posted by big W

What is interesting in this scale is that Canadians born in Canada (French or English) are the same on the issue and the differences is the foreign born.

Interesting, but not in the least surprising.

[big W]

Which indicates further dilutes the significant difference between English and French Canadians the larger map tends to show.

[Architype]

These maps, while interesting to look at, really just visually end up representing rural and small town Canada more than anything else. The urban areas where most people actually live can't be seen in any significant way, if at all.