"Blame Stats Canada!"

This is interesting:

[Industry Minister Tony Clement] added that he took the privacy concerns to Statistics Canada and asked they be incorporated into the next census. “They gave me options and we chose one of those options,” he said.

“This is a methodology that Statistics Canada offered to us and if it's good enough for Statistics Canada, it should be good enough for some of our critics.”

Well, no, it isn't, for the following reasons:

1. It's still stupid.
2. We don't know what StatsCan offered, and they're not allowed to speak for themselves.
3. We don't know what else they offered.
4. We don't know if the option came with the proviso: "Warning: This Policy Is Too Stupid For Words." (I'm guessing it did.)
5. StatsCan routinely consults with the wider research community, and it didn't this time.
6. Where is the technical report?
7. More reasons that I will come up with shortly, and which readers will no doubt suggest.

On the other hand, maybe this is good news. If the Conservatives want to blame the whole fiasco on bad advice from StatsCan officials and use that as an excuse for reversing the decision to make the long form voluntary, then that's fine with me.

Comments

[Bob]

So where do I get this voluntary long form?

[tomslee]

I appreciate the way you and others are pushing back over this issue. It's one where academics' - particularly economists' - voices should carry some weight, and I have a feeling you may carry the day.

[Bouchecl]

More reasons that I will come up with shortly, and which readers will no doubt suggest

Maybe the minister is a personal friend of a well-known census specialist? Which brings us back to reason number 1.

[Paul Friesen]

It seems this is not the first thing this government has wrecked at StatsCan. Haroon Siddiqui reports in the Toronto Star that they have also negatively affected the measurement of inequality, but he is vague about it. Does anyone know anything about this?

[Jon]

Out of curiousity, who is the Deputy Minister of Stats Can and what is his/her background and experience (and how long have they been in their postion for)?

[Kevin milligan]

The Census Tweet heard round the world is still echoing. link.

[Stephen Gordon]

Hey! I'm prominent!

[Simon van Norden]

“They gave me options and we chose one of those options,”

If I were the Chief Statistician (Munir Sheikh) and the government asked me to give options for eliminating the compulsory form, I would do so....it's my job.

[maddog]

Since the Tories came in they have politicized Statistics Canada. The former Chief Statistician and other senior officials have left or been squeezed out by the government telling them how to do their jobs. Mr. Clement's suggestion that we 'trust the experts at StatCan' is hilarious, since it's their political appointee who's advising them. I have alot more faith in Dr. Felleghi, former Chief Statistician. Moreover, since when have the Tories ever been willing to listen to the experts (e.g., reducing the GST, numerous other issues).

Key issue: if those of us who support the Long Form want to keep it, we have to find a face-saving option for the Tories to adopt, and fast.

[Andrew F]

My suggestions:

-Blame Stats Can for bad advice (it doesn't have to be true, it just has to be plausible), or

-Drop a question or two from the long form and declare it to no longer be excessively intrusive

[Jon]

No matter how much you want that census data, it still isn't appropriate for a civil, modern government to insist upon such a thing.

The next (right) step here is to refer a case to the ICC against those politicians and bureaucrats who mandated and enforced that mandate. These kinds of human rights violations have got to stop.

[Patrick]

Jon: are you being serious? It's precisely the sort of thing a civil modern government MUST do to be able to manage the welfare state.

Here's my suggestion: Canada is a really big country, and most of it is fairly empty. Let's carve off a chunk of it somewhere and declare it to be Libertarianastan. All the libertarians and anti-society morons can go live anonymously in their huts without gov't interference, taxes, medical care, roads, or schools. Heck, we could even broadcast it as a reality TV series and make millions. My guess is it wouldn't take more than a few years of watching people revert back to the stone age to cure us of the idiotic and childish delusion that we can live with any semblance of civility without gov't.

[Wendy]

Other face saving option:

Put it to a vote in the house, ensuring that enough opposition members will vote the change down. Then tories can blame other parties and still look like the "anti-state" old reformist party to that base (who likely otherwise feels a bit betrayed by the government spending growth in recent years).

[David]

Simon is bang-on. I've done briefing notes for poli sci classes and for co-op terms in government, and you list possible options and give a recommended option. It's quite conceivable that Stats-Can gave the option of making the long form voluntary as an option, but recommended something else.

When I read Clement's comment, I wondered what the higher-ups at StatsCan thought about it because I'd be surprised if they endorsed making the long form voluntary or said it was "good enough."

[Mike Moffatt]

Fraser Institute just made an annoucement on the Census. Wonder if Prof. Gordon pushed them into it?

[Adam P]

Patrick@12:28pm

Very well said, thanks. I enjoyed reading that.

Anyone who benefits from the services that government provides, meaning everyone living in the society, is obligated ot help the government provide those services as needed. If that means giving a bit of personal information so be it.

[Kevin Milligan]

I don't know much about the ins and outs of how decisions are made in government, but here is an interesting opinion of a Conservative who does know how things work.

A quote:
"It would be interesting to know where the idea originated. Was it the PMO Policy shop or at the ministerial level or at the departmental level? Who was the driving force behind this initiative?

Usually if an idea originates on the political side, staff check with departmental officials to get their comments. It is part of the checks and balances in policy making. Both the political side and the departmental side are not immune from coming up with bad ideas and one checks the other usually resulting in a better proposal or on occasion no proposal going forward. Was this done this time? Did the department support the decision or did they originate it?"

[StatCan Insider]

As a Statistics Canada insider, I can assure everyone with 100% certainty that no one at Statistics Canada would suggest making the long form voluntary. When I first heard of this possibility from Census staff (way back in January), everyone was shocked at the mere suggestion. Furthermore, on the first workday back after the news appeared in the gazette, Munir Sheikh (The Chief Statistician) assembled everyone for a meeting on how to deal with the media. It quickly became obvious to us why the meeting was taking place at that time. He confirmed our suspicion by then telling us to not talk to the media about the changes to the census long form. Only Rosemary Bender or Peter Morrison can do so. Never before have we been told not to discuss a particular topic with the media. The reason for the new measure is clear: Statistics Canada is 100% opposed to making the long form voluntary.

[Andrew F]

Thanks for having the courage to speak out, Insider. It's a shame that is a sign of bravery in the public service these days.

[Mike Moffatt]

I just find it funny how many of the Fraser Institute folks work in publicly funded Universities. They have a problem with the government collecting census data, but they have no problem drawing a paycheque from an quasi-government institution funded by tax dollars? (which, I'm told by libertarians, are taken by force!)

[Stephen Gordon]

What Andrew F said, StatCan Insider. Thanks.

[Scott Wolfe]

To StatsCan Insider: Thank you, thank you, thank you. The term 'patriotism' has been so sullied over the past decade that I hesitate to use the word. But, without going off on a tangent, I believe that your willingness to share your experience, and to bring these details to light, is what I would characterize as Canadian patriotism. Thank you for sharing what you know, and providing an example for others. I hope others will follow suit.

I, for one, will stand firm behind any and all of you on "the inside" who take a stand...and I suspect that waves and waves of Canadians will do likewise.

[Another public servant]

[quote]

“They gave me options and we chose one of those options,”

If I were the Chief Statistician (Munir Sheikh) and the government asked me to give options for eliminating the compulsory form, I would do so....it's my job.

[/quote]

The former chief statistician, Ivan Fellegi, told the Canadian Press that if he had been asked to provide options instead of a mandatory long-form census, he would have resigned rather than do so.

The job of civil servants is to speak truth to power, and yes occasionally suffer the consequences. It is not to Dr. Sheikh's credit that he shirked this larger responsibility.

[CampbellGord]

Ivan Fellegi was on CBC's The Current on Wednesday and gave an incredibly reasoned and detailed answer as to why this is the wrong move. He also said that the government might as well cancel the census entirely rather than make compliance voluntary. http://www.cbc.ca/thecurrent/2010/07/july-14-2010.html

[westslope]

Never thought I'd see the day that Canadian newspapers wrote op-eds about stratified sampling.

Nicely done Stephen! Good on Fellegi.

[Geoff NoNick]

I have no doubt the Minister was given the option to eliminate the long-form census, having drafted TB submissions and ministerial briefing notes in the past. It would have looked like this:

Option (a): Give us $200M for our Cadillac option.
Option (b): Give us only $50M for our reasonable compromise option - for a limited time only!
Option (c): Eliminate our sacred cow (unthinkable to technocrats).

Guess what? When you submit "options" like this, occasionally the minister will exercise his discretion instead of dancing to your puppeteering. It's called "democracy".

[Kevin Milligan]

Except, Geoff, that the option he chose was

(d) Eliminate our sacred cow and replace it with something costing much more (the proposed household survey) that does less.

If you believe this story, then (c) was indeed on the menu and Statcan preferred it to (d).

That's called "professional competence." Hand in hand with enlightened democratic oversight and hey we'd have good governance here.

[Jon]

"Jon: are you being serious? It's precisely the sort of thing a civil modern government MUST do to be able to manage the welfare state."

Are you? P A R O D Y. The claim that sky will fall because of this is utter nonsense. Its really rather outrageous.

To get your reality check, the US Census Bureau has pleaded with Congress and the Courts for years to switch to a sampling based approach. Why? Because surveying everyone is a waste of time and money and neglects illegal populations who usually successfully evade being counted.

Unfortunately, the courts determined that a census rather than a sample was mandated by the constitution.

Private businesses seem to manage demand forecasts just fine without resorting to enumerating everyone.

[Patrick]

Jon: Absolutely I am serious. But I wasn't asking a rhetorical question. I couldn't tell if you were being sarcastic or not ... I see now that you weren't. More's the pity.

I have no idea what your point is. What does the US census have to do with the Canadian mandatory long form, which by the way *is* a sampling. And the validity of the sampling is the crux of the matter at hand.

[Jon]

As you should be aware sampling does not require compulsion of any kind to be effective, again the mindset that you get the right answer by mandatory collection alone is wrong per se.

You'll never get responses from the underground population, so you have to add them in. ALWAYS. You need to go into it with that mindset and calibrate the results accordingly.

Not dress it up by claiming you achieve that end by "compelling" responses...

[Kevin Milligan]

"As you should be aware sampling does not require compulsion of any kind to be effective." It does require representativeness of sample. Compulsion is one way to get there. What is your suggestion for an alternative way to get representativeness?

[StatCan Insider]

Jon:

I'm going to try to be as polite as possible, but you are quite confused. Sampling does not work if people are allowed to decide whether to answer or not. Here's my evidence (you provide yours):

1. 25% of potential respondents do not respond to Statistics Canada surveys when asked.
2. 0% of potential respondents do not respond when the survey is mandatory.
3. Comparing the two, we KNOW (from Kevin Milligan's work with Statistics Canada, posted yesterday by Kevin) that low-income people are far less likely to answer voluntary surveys. In fact, incomes of the low-income population are twice as high in a vulntary survey (this is because the very low-income population simply does not respond to a voluntary survey).

That's my evidence. Where's yours? (just saying something doesn't count - all of what I have said can be backed up by detailed references, which I will provide if you insist).

Besides, have we heard from any competent expert in Statistics who has said that sampling is fine if response is voluntary (I know that we've heard from a ton of them suggesting the contrary).

[Kevin Milligan]

" The claim that sky will fall because of this is utter nonsense."

I submit that your proposed test for public policies is a poor one. There are many bad policies that will not make the sky fall.

No long form WILL make both our private and public sectors more inefficient. More wasteful. If we can get the good that comes out of the census by requiring 30 minutes of each citizen's time, once every 25 years, that is not too much to ask.