Andrew Coyne: Political parties — not Statistics Canada — are the real bad guys of privacy invasion

Valid critique of the double standard:

For the past week, question period has been dominated by accusations from Conservative MPs that a government agency has been spying on Canadians — improperly gathering sensitive personal information, it is suggested, on behalf of the ruling party.

That shadowy cabal? You guessed it: Statistics Canada.

The Conservatives have invested much effort in recent years attempting to persuade Canadians that StatCan is their enemy: witness the campaign against the long-form census. The current hysteria was kicked off by a letter from the agency requesting Canada’s banks make available to it personal financial data from 500,000 of their customers.

The program is not secret: the agency briefed reporters on it a month ago. Neither does it apply only to banks. StatCan is reaching out to a range of public and private organizations, hoping to tap the databases they maintain. The reason? People aren’t filling out the surveys the agency has traditionally used to keep track of consumer purchases and the like in anything like the numbers they used to: the data is increasingly unreliable. Without access to “administrative data” to replace it, the agency would be stumbling in the dark.

Privacy concerns are worth taking seriously, of course. Canadians would be right to worry if StatCan were proposing to set up personal files in their name, or to combine bits of data collected from a variety of sources into individual profiles. Needless to say, that is not what the agency is proposing. And while data security is increasingly a concern, StatCan’s record in this regard is unblemished.

Indeed, of all the organizations that now monitor, collect and compile your personal data, StatCan would seem among the least threatening. Your cellphone provider, to take one example, not only keeps tabs on who you called at what hour and for how long, but where you were at the time — in fact, where you are at all times. In the wrong hands, that sort of detailed personal information could be used to manipulate, intimidate and defraud. Whereas the broad aggregates StatCan extracts from it are essential to good public policy.

And of all the wrong hands it is possible to imagine, among the wrongest are those of the political parties — the same parties that are so quick to mount the privacy soapbox when it comes to other organizations. It’s StatCan this week, but it was the big social media companies before and it will be somebody else next – everyone, that is, but the parties themselves. Yet the scale of what the parties are up to, and the potential for abuse — no, the reality of abuse — is far greater than anything StatCan might propose.

All of the parties keep detailed personal files on literally millions of voters. Unlike last year’s scandal over Cambridge Analytica’s use, on behalf of its political clients, of information illegally scraped off of Facebook users’ pages, the data here is acquired legally, which is to say the law has been written in such a way as to allow it.

For example, the parties all have guaranteed access to Elections Canada’s voter lists, though there is no obvious reason why they should. Combined with data purchased from private market-research companies and their own proprietary data collected from interviews with individual voters, the parties are able to assemble quite fantastically “granular” profiles of the voters they are trying to reach, with which not merely to anticipate their responses to events but to shape them, via the sort of highly customized, micro-targeted messages that modern media make possible.

All of which would be objectionable enough — there is, again, no need for any of it, and much reason to object to all of it — if it were subject to even the barest regulatory safeguards. But while government agencies like StatCan are covered by the Privacy Act and private companies come under the Personal Information Protection and Electronic Documents Act (PIPEDA), the parties have taken care to exempt themselves from federal privacy laws.

And, what is more, they seem determined to keep it that way. Federal and provincial privacy commissioners have called for bringing the parties within the law; so has the head of Elections Canada; so, too, has an all-party committee of MPs. Yet Bill C-76, the package of changes to the election laws currently before the House, makes no requirement of parties other than that they should publish their privacy policies on their websites, with no guarantee they will even abide by their own standards, let alone the kind they impose on others.

Asked to justify this, Liberal spokespeople burble on about the need to “engage” voters. “Understanding the interests and the priorities of Canadians,” Liberal adviser Michael Fenrick told the Commons access to information, privacy and ethics committee last week, “helps us to speak to the issues that matter most to them and in turn mobilizes democratic participation in our country.” Those hot-button fund-raising emails and Facebook ads that cater to your worst fears? That’s what he’s talking about, behind all the high-falutin’ language.

A Conservative official said much the same, adding that of course his party was willing to live with whatever Parliament decides, which is how an opposition party traditionally hides behind the government’s skirts. Only the NDP and Greens have publicly expressed support for bringing the parties under the privacy laws — though since neither is likely to be in a position to put this into effect, this, too, seems awfully convenient.

We’ve been this way before. The parties thoughtfully exempted their own solicitations from the do-not-call rules that apply to other telemarketers. Third-party advocacy groups are subject to much tighter election spending limits than those the parties apply to themselves. Corporate advertisers must conform to truth in advertising laws; not so the parties.

And now privacy. It’s tempting to say there’s one law for the parties and another for everyone else but, in this case, there isn’t any.

Source: Andrew Coyne: Political parties — not Statistics Canada — are the real bad guys of privacy invasion

There is nothing Orwellian about collecting accurate, real-time data: Barrie McKenna

Good commentary. Given the Conservatives legacy in downgrading the Census to the less accurate National Household Survey in 2011, their record on these kinds of issues is suspect.

And, as McKenna notes, “Worse than Big Brother is Blind Brother:”

To hear Conservatives spin it, Statistics Canada’s plan to gather the banking and spending records of hundreds of thousands of Canadians is akin to “Big Brother on steroids” and an “Orwellian intrusion into the lives of Canadians.”

The truth isn’t nearly as sinister. Rest assured, the government is not plotting a massive surveillance campaign to find out what you ate for lunch or your monthly mortgage payment.

Guess what? Ottawa already has your social insurance number – because it gave it to you. And it has your tax returns.

The government does, however, need better data to provide a complete and accurate portrait of Canada’s economy and society, in real time. As part of a “modernization” of its operations, Statscan wants banks, cellphone companies, retailers and other companies to share more of the so-called big data they have, and leverage them for the collective public good.

As Canada’s chief statistician Anil Arora put it: “Traditional statistics gathering methods are no longer sufficient to accurately measure Canada’s economy and social changes.”

Yes, some of the information Statscan wants to gather is personal. But all personal identifiers, including names, addresses and social insurance numbers, would be removed before any of it is compiled and released to the public. That’s what the agency already does routinely with census data, the monthly household survey and vast amounts of competitively sensitive corporate information.

Statscan has been peeking into our lives for a long time. Unfortunately, response rates from the agency’s traditional surveys have been falling, leaving it with often suspect and outdated data to feed into its key reports. The agency says getting access to financial transactions is vital to producing a timely, accurate picture of the economy.

As it should, Statscan is working closely with the federal Privacy Commissioner Daniel Therrien to ensure personal data are not put at risk, or shared publicly. It’s up to Mr. Therrien, who last week launched an inquiry into Statscan’s big data pilot project, to set the rules, and then let the agency do its job.

Statscan is hardly unique. Statistics agencies around the world are similarly leveraging big data for public policy purposes. And that’s unambiguously a good thing, according to University of British Columbia economist Kevin Milligan.

“This research is vital to forming good government policy and providing good economic information to the private sector,” Mr. Milligan says. “Statistics Canada should and does work with the privacy commissioner to balance the good that comes from research to the potential challenges to privacy.”

It’s ludicrous to suggest Ottawa is spying on Canadians. What Statscan is doing is tapping into what the private sector already knows about all of us, and aggregating it for public consumption.

If you’re seriously concerned about letting others see your financial records, shopping habits and internet surfing behaviour, well, that horse left the barn a long time ago.

Just think for a minute what companies such as Toronto-Dominion Bank, Bell, Facebook, Google, Amazon or the operator of the Highway 407 toll road already know about what you did today, or in the past month. Stitch it all together, and it’s your life in bits and bytes.

Canadians should be more concerned that there are adequate controls over what these companies are doing with your data. Perhaps Canada’s big banks are resisting giving your data to Statscan because they are more interested in exploiting it themselves.

The more ominous privacy threat may not be Statscan. The greater risk may lie with the major private-sector collectors of big data, many of which are foreign owned and store it all far beyond the reach of the government. And they often operate with far weaker privacy constraints than government agencies.

Governments already know plenty about you. There are census data, passport photos and records, tax filings, municipal property records, health records, driving offences and court records. No reasonable person would suggest this is somehow part of a nefarious Big Brother spying plot.

The agency’s data-collection pilot is not the problem. It is part of the solution. For years, Statscan’s ability to do its job was eroded by steady budget cuts. The current Liberal government reinstated some that funding in this year’s budget, with an additional $41-million over five years to improve the agency’s ability to do its job.

Worse than collecting more data is having a data deficit. Governments, and businesses, risk making major mistakes without accurate, real-time data.

Worse than Big Brother is Blind Brother.

Source: There is nothing Orwellian about collecting accurate, real-time data: Barrie McKenna

StatsCan – Police-reported hate crime in Canada, 2016 – Analytical Note

Given the government’s diversity and inclusion focus, StatsCan has vastly improved the analytical note on hate crimes, including age and gender data on those accused. While personally I prefer longer-term comparisons rather than year-to-year as in the above charts, this analysis is nevertheless helpful (StatsCan summary below):

  • In 2016, police reported 1,409 criminal incidents in Canada that were motivated by hate, an increase of 3% or 47 more incidents than reported the previous year. Accounting for the population, this amounted to a rate of 3.9 hate crimes per 100,000 Canadians in 2016.
  • The increase in the total number of incidents was largely attributable to an increase in police-reported hate crimes motivated by hatred of a sexual orientation (+35 incidents) or of a race or ethnicity (+25 incidents). Hate crimes accounted for less than 0.1% of the nearly 1.9 million police-reported crimes in 2016 (excluding traffic offences).
  • Police-reported hate crimes targeting sexual orientation rose 25% in 2016 to 176 incidents, compared with 141 incidents in 2015. These incidents accounted for 13% of hate crimes reported in 2016 and 11% of hate crimes reported in 2015.
  • Between 2015 and 2016, the number of police-reported crimes motivated by hatred of a race or ethnicity increased 4% (from 641 to 666). In all, 48% of all police-reported hate crimes in 2016 were motivated by hatred of a race or ethnicity. Much of this increase was a result of more hate crimes targeting South Asians (+24 incidents) and Arabs and West Asians (+20 incidents). Despite posting a decrease in 2016, crimes targeting Black populations remained one of the most common types of hate crimes (15% of all hate crimes).
  • Overall, 33% of hate crimes reported in 2016 were motivated by hatred of religion. Compared with 2015, the number of hate crimes motivated by religion decreased 2% in 2016 (from 469 in 2015 to 460 in 2016). Police-reported crimes motivated by hate against the Jewish population rose from 178 incidents in 2015 to 221 incidents in 2016 (+24%). In contrast, the number of crimes targeting the Catholic population fell from 55 to 27 incidents. Similarly, crimes targeting the Muslim population decreased 13% (from 159 incidents in 2015 to 139 incidents in 2016).
  • The provinces of Quebec and British Columbia, and more specifically Vancouver (+30 incidents), Québec (+29 incidents), and Montréal (+25 incidents), were the census metropolitan areas where hate crimes increased the most in 2016. The increases in Montréal and Québec are associated with a rise in hate crimes targeting the Jewish, Arab and West Asian, and gay and lesbian populations. The increase in Vancouver was primarily explained by a rise in hate crimes against the East Asian, Southeast Asian and South Asian populations.
  • Based on data from police services that reported characteristics of hate crimes, 43% of police-reported hate crimes in 2016 were violent offences. Violent offences included, for example, assault, uttering threats and criminal harassment. Overall, the number of violent hate crimes rose 16% from the previous year (from 487 to 563 violent incidents), driven by increases in common assault, criminal harassment and uttering threats.
    Crimes motivated by hatred of a sexual orientation continued to be among the most violent hate crimes. In 2016, 71% of these types of police-reported hate crimes were violent, compared with 45% of crimes motivated by hatred of a race or ethnicity and 27% of hate crimes targeting a religion.
  • Non-violent offences made up 57% of police-reported hate crimes in 2016. Mischief, which includes vandalism and graffiti, was the most commonly reported offence among police-reported hate crimes and accounted for 41% of all hate crime incidents in 2016. Between 2015 and 2016, the total number of non-violent hate crime incidents fell 6%. In 2016, 73% of crimes targeting religion were non-violent. This proportion was 55% for non-violent crimes motivated by hatred of a race or ethnicity. Conversely, hate crimes motivated by sexual orientation were less often non-violent (29%).

via Police-reported hate crime in Canada, 2016

Shree Paradkar: Census vastly undercounts Indigenous population in Toronto, study says

One of the harder groups for StatsCan to count despite their ongoing efforts, with this alternative study being instructive in terms of the possible gap:

For decades, Indigenous communities have said their numbers are far higher than reported by government agencies.

Not so, according to officialdom.

“Always our studies in the past have been critiqued or undermined as not having a scientifically sound approach,” says Sara Wolfe, founding partner of Seventh Generation Midwives Toronto, which caters to Indigenous mothers and babies.

“Or there’s been concerns about bias or questioning of the relevance… of the study that’s been done.”

The tables were turned recently.

The census released in October pegged Toronto’s Indigenous population at 23,065, up from the 19,270 census estimate in 2011.

Not so, says a new study that confirms what Indigenous people have been saying all along.

The study by researchers from York University and St. Michael’s Hospital, in collaboration with Indigenous agencies, was published in the British Medical Journal Open.

It says the census — that gold standard in population counting — vastly underestimated the Indigenous population in Toronto. The study’s most conservative assumption places it between 45,000 and 73,000 people, or two to four times the 2011 census estimate.

This finding has major implications, particularly in funding for health care and community services.

Statistics Canada is receptive to the study. The agency’s chief priority is accuracy and precision, said Marc Hamel, director-general of the census program.

When there are reports of discrepancies, “we review all the processes we have internally. We also try to work with these groups to better understand the way the study was conducted,” he said. “We always have to be careful when we compare results from different studies because different methodologies are being used, different concepts.”

Lead scientist Janet Smylie, from St. Michael’s Hospital, and lead author of the study Michael Rotondi, a York University professor, employed a statistical method called respondent-driven sampling, which leveraged the inherently strong social networking of Indigenous populations.

.

Specifically, 20 people called “seeds” completed the survey and were given five uniquely coded coupons. They gave these to other Aboriginal people who then filled the survey, and those people gave out coupons to others in their social networks, and so on.It allowed Indigenous community members to recruit each other for the study which then reached a large sample of more than 900 adults.

“This helps better track Indigenous community members who might be homeless and otherwise unstably housed,” says Rotondi.

They partnered with Wolfe’s midwifery clinic, which led a multi-agency collaboration to plan the questions, recruit trained Indigenous interviewers and disseminate the survey that took more than an hour to complete.

The census, on the other hand, uses the concept of usual residence and is based on private dwellings.

“It doesn’t measure, for example, where people would be temporarily residing for whatever purpose, whether it be work, school or receive certain types of services,” says Hamel.

“The census is never perfect, like any study. We know we have unaccounted populations. We have measures to identify and account and to make adjustments to the population estimation programs that are used by the government to make decisions.”

The survey included a question of whether or not the respondents had completed the 2011 census.

“Even under a conservative model we were able to say only about 19 per cent (of individuals) had even completed the census,” Rotondi says.

“One of the big reasons is people don’t trust governments, long forms and mandatory surveys,” says Wolfe, who is Ojibwe from Brunswick House First Nation and was the community lead for the study.

“We might be afraid to tell someone on the phone that says they’re from the government that we’re Indigenous,” says Smylie, who is Métis. “We might purposely not want to participate. We might be opting out because we feel socially excluded or frustrated with the government. Or, it’s not on our priority list ’cause we’re too busy trying to get enough groceries on our shelf and we’re running around and didn’t even know the census was happening. Or (we’re) renting a room somewhere or couch surfing.”

These were some of the barriers the respondent-driven sampling broke down.

The impact of this study will be tremendous and long-term, the researchers say.

“It doesn’t mean that just because there are more Indigenous people everyone’s going to have to pay more taxes. It could mean if we’re counting properly (and allocating correctly) we’re paying less taxes,” said Smylie.

“This is irrefutable evidence,” said Wolfe. “There’s no way you can say the population is not this big any more.”

This is relevant because, “Indigenous people are not getting asked for input and consulted on the decisions being made… because there’s a presumption that we are not a significant or substantial portion of the population,” says Wolfe.

Why does it matter if the people accessing care are Indigenous as long as they have access to it? Two reasons: to counter ongoing racism and to redress intergenerational trauma produced by historic wrongs.

In a report titled First Peoples, Second Class Treatment, Smylie says she wrote that if you’re a First Nation person living in the province of Alberta having a heart attack, “you’re less likely to get a picture of your heart, called a coronary angiogram, and more likely to die just because you’re First Nation. It doesn’t matter if you live in the city or a rural area or if you’re rich or poor.”

Residential schools, the last of which closed 20 years ago, left Indigenous people with a painful legacy. Abuses that are only just being seriously documented have left a community history of complex trauma.

“That might be something you’d need specialized services and responses,” said Smylie. “We also know that some Indigenous people benefit greatly from access to traditional healing and traditional counseling and a revitalization of Indigenous culture.”

Says Wolfe: “Everyone needs to make a concerted effort to work together to close these (health) gaps so we can have as good a chance as everyone else in society to reach our full potential.”

Source: Shree Paradkar: Census vastly undercounts Indigenous population in Toronto, study says

Ex-StatsCan chiefs make last-ditch appeal to fix ‘egregious flaw’ in stats agency governance bill

Tend to agree with Smith and Fellegi:

A pair of former chief statisticians made a last-ditch plea to Senators last week to fix what one said was an “egregious flaw” that “fundamentally undermines” a government bill’s aim to give Statistics Canada more independence.

In an appearance in front of the Senate’s Social Affairs, Science, and Technology Committee Nov. 30, Wayne Smith and Ivan Fellegi restated the case they made in front of the House of Commons committee that studied Bill C-36: that the changes to the Statistics Act that purport to give the chief statistician more independence, should come with a more-stringent hiring process.

“I’m very conscious of the fact this is the 23rd hour in terms of this legislation,” said Mr. Smith, who served as the chief statistician from 2010 until he resigned in September 2016 in protest over the role Shared Services Canada plays in handling Statistics Canada’s information technology.

Mr. Smith called a lack of specificity in the bill over how the chief statistician is selected the “one egregious flaw in the legislation that fundamentally undermines the achievement of its objective.” The bill would change the chief statistician’s term to a fixed five years served under good behaviour, instead of the current term that lets them stay on as long as the government wants them to.

Both Mr. Smith and Mr. Fellegi—who served as the country’s chief statistician from 1985 to 2008—are calling for the creation of a three-person non-partisan selection committee to create a shortlist of candidates for the governor-in-council appointment (which would ultimately be decided by cabinet), as well as for the bill to clearly define what requirements a chief statistician should have.

“The new proposed bill gives a great deal more authority on professional issues to the chief statistician,” Mr. Fellegi said. “That makes it that much more important that he or she should be properly qualified. And an appropriately composed search committee should have that task.”

But Liberal Senator Jane Cordy (Nova Scotia), a member of the Social Affairs Committee and the bill’s sponsor in the Senate, told The Hill Times she doesn’t think the bill needs changing.

Sen. Cordy, who said she was first approached about sponsoring the bill in the spring by the government’s representative in the Senate Peter Harder (Ottawa, Ont.), pointed to comments made by Independent Senator Tony Dean (Ontario) during the Nov. 30 committee meeting that selection committees don’t eliminate potential bias.

“I’m a fan of search committees,” Sen. Dean told the former chief statisticians, but added he didn’t think the bill necessarily required changing. Instead, the recommendation could be rolled into an observation for the minister to consider.

There is precedent for choosing the chief statistician by committee, according to Mr. Fellegi, who said the appointment of his predecessor, Martin Wilk, was conducted that way. Mr. Wilk’s appointment came after a period in the late 1970s when Statistics Canada didn’t have the stellar reputation it enjoys today, Mr. Fellegi said, and needed a “transformative” leader.

“We found a transformative chief statistician who wouldn’t have applied because he was vice-president of AT&T, being paid probably five times as much, at least, as the offer from the government of Canada,” Mr. Fellegi said, adding that the search committee “basically appealed to his conscience” to have him return to Canada from the United States to take the job from 1980 to 1985.

Conservative Senator Linda Frum (Ontario), the official opposition critic for the bill in the Senate, said during her second-reading speech on Oct. 5 that the chief statistician should also be subject to approval by both houses of Parliament.

“If this government wants to demonstrate its sincere desire for a more arm’s-length relationship between the agency and the government of the day, it should support such an amendment, that parliamentary approval must be required before appointing a new chief statistician,” Sen. Frum said.

On Dec. 1, Sen. Frum’s office told The Hill Times that following the start of the committee’s study of the legislation, she still saw this as a shortcoming, but that there were no proposed amendments yet formalized.

via Ex-StatsCan chiefs make last-ditch appeal to fix ‘egregious flaw’ in stats agency governance bill – The Hill Times – The Hill Times

Erreur dans le recensement linguistique: Statistique Canada s’explique

Not an easy time before parliamentarians:

Statistique Canada avait «détecté certains changements» dans les données sur la langue à l’étape de la validation, mais «n’a pas, à ce moment-là, capté» qu’il aurait fallu procéder à une révision avant de diffuser les données linguistiques qui ont provoqué un tollé au Québec.

«Je sais ce qui s’est produit. Mais comment on a manqué cette erreur-là, c’est cette partie que je ne sais pas encore», a lâché devant les députés du comité permanent sur les langues officielles Marc Hamel, directeur général du programme du recensement.

L’agence fédérale avait déjà fait son mea culpa en août dernier, expliquant que l’erreur avait été causée par le logiciel de compilation de données. Celui-ci a inversé les réponses dans des formulaires en français d’environ 61 000 personnes, dont environ 57 000 au Québec.

La bourde avait eu pour conséquence de surestimer la croissance de l’anglais dans la province et dans certaines de ses régions, tant pour la langue maternelle que pour la langue parlée à la maison, ce qui avait inquiété politiciens et défenseurs de la langue française.

«Ce n’est pas le système qui n’a pas détecté (l’erreur). Ce sont les gens qui ont testé le système qui n’ont pas détecté que le système ne lisait pas le questionnaire de façon conforme», a spécifié Marc Hamel aux élus.

Le député conservateur Alupa Clarke lui a demandé si des têtes allaient rouler chez Statistique Canada, déplorant que «de plus en plus, aujourd’hui, on vit dans une société où on ne met jamais au banc des accusés les responsables».

«Dans un cas comme celui-là, on ne parle pas des individus, on parle des processus. Si à chaque fois que quelqu’un faisait une erreur, il était congédié, on en congédierait peut-être plusieurs. Les erreurs sont rares», lui a répondu M. Hamel.

«On a fait les correctifs appropriés pour éviter que ce genre de situation comme ça se reproduise encore. Est-ce que je peux vous dire aujourd’hui que dans les 100 prochaines années, ça n’arrivera pas encore? Absolument pas. L’erreur est humaine», a-t-il ajouté.

Au haut fonctionnaire, qui s’est défendu de «prêcher par nonchalance», Alupe Clarke a suggéré d’envoyer une «lettre diplomate» aux 5000 employés de l’agence pour leur dire de faire gaffe à l’avenir, établissant un parallèle avec son expérience dans les Forces armées.

«Moi, j’ai fait l’armée, puis nous, ça ne niaise pas, là. Il y a une discipline (…) puis quand on fait la guerre, ça marche», a-t-il lâché.

Un peu plus tôt, son collègue néo-démocrate François Choquette s’était étonné que l’agence ait diffusé les données linguistiques alors que certaines d’entre elles, en particulier dans certaines villes à forte majorité francophone, étaient clairement suspectes.

«Attendez que je comprenne comme il faut: 164 pour cent d’augmentation de la population anglophone à Rimouski, 115 à Saguenay, 110 à Drummondville. Vous avez eu ces chiffres-là, qui n’étaient pas normaux, et vous avez quand même décidé de les sortir?», a-t-il questionné.

Le directeur adjoint de la division de la statistique sociale, Jean-Pierre Corbeil, a répondu que ce n’était «pas aussi simple» et qu’il «fallait être prudent quand on faisait des comparaisons historiques», surtout compte tenu des changements survenus sous les conservateurs en 2011.

Ces données contenues dans la livraison initiale de données du 2 août dernier étaient passées sous le radar jusqu’à ce que le président de l’Association d’études canadiennes, Jack Jedwab, lève un drapeau rouge après avoir passé les chiffres au peigne fin.

Les données revues et corrigées publiées quelques jours après ont confirmé que le français avait effectivement perdu du terrain au Québec, mais moins qu’annoncé initialement, et que l’anglais n’avait pas progressé, mais plutôt reculé, dans la province.

En présentant les nouveaux chiffres, l’agence fédérale avait fait acte de contrition et reconnu que cette erreur était d’autant plus regrettable qu’elle concernait un enjeu fort délicat au Québec.

«Nous sommes très conscients de l’aspect très sensible de cette question, de ces enjeux, et Statistique Canada va corriger le tir, simplement», affirmait Jean-Pierre Corbeil, directeur adjoint de la division de la statistique sociale et autochtone, qui était aussi au comité, mardi.

Source: Erreur dans le recensement linguistique: Statistique Canada s’explique | Mélanie Marquis | National

The Daily — Study: International Students, Immigration and Earnings Growth

Important study showing the importance of pre-landing work experience to earnings:

International students are increasingly regarded as an important group of young and well-educated individuals from which to select permanent residents. In December 2015 there were 353,000 international students with a valid study permit in Canada, up from 84,000 in December 1995. Of the international students admitted to Canada in the early 2000s, 25% became permanent residents over the 10 years that followed. Of these, nearly one-half applied as principal applicants in the economic class.

A small number of studies from Australia, Canada and the United States suggest that the earnings advantage that former international students have over other economic immigrants may be either small or non-existent. This suggests that pre-landing study experience in a destination country such as Canada may not in and of itself improve immigrants’ labour market outcomes over university degrees acquired abroad. Policy-makers and researchers are thus shifting their attention to the complementary role played by other factors, such as pre-landing work experience. A study released today by Statistics Canada offers new evidence on this issue.

The study examines the earnings trajectories of three groups of university graduates: international students who obtained a university degree in Canada and then became landed immigrants (i.e. Canadian-educated immigrants); individuals who had a university degree from abroad at the time they immigrated to Canada (i.e. foreign-educated immigrants); and university graduates born in Canada. The earnings trajectories of these groups were examined over 6 years for the cohort of individuals aged 25 to 34 in 2006, and over 20 years for the cohort of individuals aged 25 to 34 in 1991.

Among the 2006 cohort of male Canadian-educated immigrants, average annual earnings one year after landing were 48% lower than those of Canadian-born graduates. This gap narrowed to 34% six years after landing. Among female Canadian-educated immigrants, the earnings gap vis-à-vis Canadian-born graduates was 39% one year after landing and 32% six years after landing.

Most of these earnings gaps were accounted for by differences in the work histories of immigrant and Canadian-born graduates. Prior to becoming landed immigrants, 12% of male Canadian-educated immigrants had no work experience in Canada and 40% had prior work experience with annual earnings under $20,000. Among male Canadian-born graduates, virtually all had prior work experience and almost 90% had prior work experience with annual earnings of $20,000 and over. These patterns were broadly similar among women.

When group differences in prior Canadian work experience were taken into account, the earnings gap between Canadian-educated immigrants and Canadian-born graduates in the 2006 cohort disappeared among both men and women. Likewise, prior work experience accounted for much of the earnings gap observed among the 1991 cohort.

Canadian-educated immigrants had higher post-immigration earnings than foreign-educated immigrants, but prior work experience once again played an important role. Five years after landing, male Canadian-educated immigrants with no pre-landing work experience had annual earnings 20% below those of male foreign-educated immigrants. Among women, the shortfall was 7%. This takes into account a broad range of socio-demographic and source country characteristics. Canadian-educated immigrants who accumulated pre-landing work experience fared far better relative to their foreign-educated counterparts.

Canadian-educated immigrants with three years of pre-landing work experience that paid less than $20,000 had annual earnings five years after landing that were similar to, or higher than, their foreign-educated counterparts. Those with three years of pre-landing work experience that paid $20,000 to $50,000 had annual earnings five years after landing that were 42% to 61% higher. For the approximately 10% of Canadian-educated immigrants who had three years of pre-landing work experience that paid more than $50,000, their earnings five years after landing were more than double those of foreign-educated immigrants. These differences in earnings were larger among the 2006 cohort than the 1991 cohort.

These results suggest that pre-landing Canadian work experience and earnings play an increasing role in differentiating the post-immigration labour market outcomes of university-educated immigrants.

Source: The Daily — Study: International Students, Immigration and Earnings Growth

An increasingly diverse linguistic landscape: Highlights from the 2016 Census

Excerpts from the StatsCan summary:

Immigrant languages show strong growth

“Immigrant languages” refer to languages (other than English and French – the national official languages) whose existence in Canada is originally due to immigration after English and French colonization. This expression excludes Aboriginal languages and sign languages, in addition to English and French.

The first results from the 2016 Census, released on February 8, 2017, showed once more that international migration is the key driver of population growth in Canada. As such, Canada’s linguistic landscape is constantly changing. In the 2016 Census, over 7.7 million people reported an immigrant mother tongue (alone or with other languages). This corresponds to 22.3% of the Canadian population.

Over 7.3 million people reported speaking an immigrant language at home. The main immigrant languages spoken at home by Canadians in 2016 were Mandarin (641,100 people), Cantonese (594,705 people), Punjabi (568,375 people), Spanish (553,495 people), Tagalog (Pilipino) (525,375 people) and Arabic (514,200 people). Proportionally speaking, the number of people who speak each of these languages individually represents between 1.4% and 1.9% of the Canadian population.

Some languages saw significant growth from 2011 to 2016. Among the languages spoken by at least 100,000 people, Tagalog (Pilipino) (+35.0%), Arabic (+30.0%), Persian (Farsi) (+26.7%), Hindi (+26.1%) and Urdu (+25.0%) experienced the largest increases. The number of people who spoke a Chinese language at home rose 16.8% from 2011 to 2016 (see note to readers).

Chart 2  Chart 2: Variation between 2011 and 2016 in the population who reported speaking certain immigrant languages, Canada
Variation between 2011 and 2016 in the population who reported speaking certain immigrant languages, Canada

Chart 2: Variation between 2011 and 2016 in the population who reported speaking certain immigrant languages, Canada

Conversely, some European languages were reported by fewer people as the language spoken at home, led by Italian (-10.9%), Polish (-5.5%), German (-3.3%) and Greek (-2.3%).

These trends reflect the changes that Canada has undergone in terms of the geographic origin of its immigrants. The number of people who speak languages from countries that are recent sources of immigration, primarily Asian countries, is on the rise. Meanwhile, the number of people who speak certain European languages—which reflect older waves of immigration—is declining.

Immigrant languages are more commonly spoken in Canada’s large census metropolitan areas (CMAs). The infographic Immigrant Languages in Canada gives a general overview of the main immigrant languages spoken in the different regions of Canada.

The population with an immigrant mother tongue is increasing across Canada

The population with an immigrant mother tongue rose in every region of Canada. In absolute numbers, Ontario (+352,745 people) and Western Canada (+414,260 people) saw the largest growth from 2011 to 2016.

In relative terms, the Atlantic provinces (+33.2%) and the territories (+27.6%) saw the largest increase in the population with an immigrant mother tongue, despite accounting for only 1.2% of this population in 2016. In 2011, these two regions accounted for 1.0% of this population.

The population with an immigrant mother tongue is largely concentrated in large CMAs, with nearly two-thirds living in the CMAs of Toronto (35.3%), Vancouver (14.1%) and Montréal (13.0%). These proportions were down slightly from 2011, when they were 36.3%, 14.3% and 13.3% respectively.

In relative terms, the population with an immigrant mother tongue experienced more rapid growth in the CMAs of Edmonton (+31.1%), Calgary (+28.0%) and, to a lesser extent, Ottawa-Gatineau (+15.5%). This growth was 10.3% in Toronto, 10.6% in Montréal and 11.5% in Vancouver.

The document Linguistic diversity and multilingualism in Canadian homes presents the main immigrant mother tongues in the large CMAs in Canada. In Montréal and Ottawa-Gatineau, Arabic was the main immigrant mother tongue. In Calgary and Edmonton, the three most common immigrant mother tongues were, in order, Tagalog, Punjabi and Cantonese. In Toronto and Vancouver, they were Cantonese, Mandarin and Punjabi.

Cree languages are the Aboriginal languages most commonly spoken at home

The 2016 Census of Population provides data on close to 70 Aboriginal languages.

Cree languages were the Aboriginal languages most often reported as the language spoken at home in Canada (83,985 people) in 2016. Inuktitut was spoken by 39,025 people, while 21,800 people spoke Ojibway, 13,855 people spoke Oji-Cree, 11,780 people spoke Dene and 10,960 people spoke Montagnais (Innu).

Overall, the number of people who speak an Aboriginal language at home (228,770 people) is higher than the number of people who have an Aboriginal mother tongue (213,230 people). This difference, particularly significant among youths aged 0 to 14, shows the growing acquisition of an Aboriginal language as second language. In this age group, 44,000 people have an Aboriginal language as a mother tongue, while 55,970 people speak an Aboriginal language at least on a regular basis at home.

More detailed analysis highlighting the richness and diversity of Aboriginal languages will be made available with the release of 2016 Census data on Aboriginal Peoples on October 25, 2017.

English and French are pathways of integration into Canadian societyLinguistic diversity is also measured by the growth of multilingualism in Canadian homes. Multiple languages in the homes of Canadians of all origins are becoming more common.

The proportion of the Canadian population who speak more than one language at home rose from 17.5% in 2011 to 19.4% in 2016. There were also more multiple responses to the question on mother tongue, with the proportion of people who reported more than one mother tongue rising from 1.9% in 2011 to 2.4% in 2016.

Multilingualism primarily occurs when official languages are used increasingly with an immigrant language. For more information, see Linguistic diversity and multilingualism in Canadian homes.

For example, 69.9% of people with an “other” mother tongue (reported alone) spoke English or French at home in 2016—mostly in combination with the mother tongue.

Similarly, 69.8% of people who spoke an “other language” at home (regardless of mother tongue) did so in combination with at least one of the two official languages.

Overall, 98.1% of Canadians reported that they were able to hold a conversation in at least one official language in 2016, and 93.4% spoke English or French at home at least on a regular basis.

Strong growth of Arabic in the Atlantic provinces

There was an increase in immigrant languages as mother tongue and as a language spoken at home in the Atlantic provinces. Arabic in particular saw strong growth from 2011 to 2016 and was the main immigrant language spoken at home in three Atlantic provinces.

The only exception was Prince Edward Island, where Mandarin was the main immigrant language spoken at home (2,290 people).

Montagnais (Innu) (1,505 people), an Aboriginal language, was the other language most frequently reported spoken at home in Newfoundland and Labrador.

Decline in the use of French at home in Quebec

The various linguistic indicators show an increase in other languages and in English, and a decline in French in Quebec.

Arabic was the most common immigrant language spoken at home (213,055 people) in 2016 in Quebec, up 23.7% from 2011.

English as a mother tongue rose from 9.0% in 2011 to 9.6% in 2016, and as a language spoken at home it increased from 18.3% in 2011 to 19.8% in 2016.

French saw a decline as a mother tongue (from 79.7% in 2011 to 78.4% in 2016) and as a language spoken at home (from 87.0% in 2011 to 86.4% in 2016).

Nearly half of Canadians with an immigrant mother tongue lived in Ontario in 2016

Ontario accounted for nearly half (49.5%) of Canadians whose mother tongue or language spoken at home was an immigrant language in 2016, down slightly from 2011 (50.9% for mother tongue and 51.2% for language spoken at home).

Immigrant languages spoken at home rose significantly in Ontario from 2011 to 2016, led by Arabic (+30.5%), Persian (Farsi) (+24.0%), Urdu (+21.3%), Tagalog (Pilipino) (+19.3%), Chinese languages (+17.4%) and Punjabi (+14.5%).

Asian languages see strong growth in the western provinces

Tagalog (Pilipino) is the main immigrant language spoken at home in the Prairie provinces. From 2011 to 2016, Tagalog (Pilipino) increased 123.1% in Saskatchewan, 68.3% in Alberta and 42.3% in Manitoba.

In numbers, Punjabi was the main immigrant language spoken at home in British Columbia (222,720 people) in 2016, up 10.9% from 2011, followed closely by Mandarin (202,625 people) and Cantonese (200,280 people).

Strong growth for Tagalog in the territories

The number of people who reported speaking Tagalog (Pilipino) rose sharply in Yukon (+105.4%), the Northwest Territories (+58.8%) and Nunavut (+54.5%). The main “other” languages spoken at home were Dogrib (Tlicho) in the Northwest Territories (2,005 people) and Inuktitut in Nunavut (25,405 people). From 2011 to 2016, the number of people speaking Inuktitut in Nunavut rose 12.1%.

All French linguistic indicators increased in the three Canadian territories.

Source: The Daily — An increasingly diverse linguistic landscape: Highlights from the 2016 Census

Politicians can’t let another year of hate crimes pass without action – Macleans.ca

Not really much insights in this column on the latest StatsCan hate crimes report, nor any particular startling or new recommendations. No real clarity of what government’s acting “forcefully” would entail beyond the Ontario government’s strategy and its emphasis on wider collection of race-based data to inform policy and programs.

The longer-term view shows no clear overall trend: a decline 2009-2011, an increase 2013-15. And no recognition that the recent increase may also reflect a greater willingness to report hate crimes as well as an actual increase.

While any hate crime or equivalent is an abomination, are the numbers really so high compared to the population? How do they compare to other countries?

When racial and religious groups insist discrimination is a hindrance to their success and well-being in Canada, governments must act forcefully to remove this barrier to demonstrate that mistreating someone based on their race or religion is unacceptable in contemporary Canadian society. This display of solidarity from politicians may act as a deterrent to future hate crimes and finally bring down the stubbornly high incidents of hate crimes towards Blacks and Jews, as well as the spike against Muslim Canadians.

Source: Politicians can’t let another year of hate crimes pass without action – Macleans.ca

A closer look at the rise in hate crimes in Canada

Good to see wide coverage of the latest hate crimes report. Of interest are the comments of NCCM on the increase in the number of hate crimes against Canadian Muslims (Muslim group urges Ottawa to speed up release of hate-crime statistics):

The National Council of Canadian Muslims connected the anti-Muslim bias to a backlash over two terror attacks in Paris in 2015. But the group also singled out Conservative Party election campaigning under former Prime Minister Stephen Harper.

“The Canadian Muslim community bore the brunt of sinister political rhetoric surrounding the federal election, which painted Muslims as terrorists or terrorist sympathizers, as well as being anti-woman,” council vice-chairman Khalid Elgazzar said at a press conference on Parliament Hill.

In an interview, Mr. Elgazzar referred to Conservative pitches in favour of “snitch lines” for so-called barbaric cultural practices, as well as a ban on face veils at citizenship ceremonies.

“Words matter and those words had an impact,” he said. “There was an immediate uptick in terms of incidents of hate being reported to us.”

The Statscan data indicate that hate crimes targeting Muslims in Canada rose to 159 incidents, a 61-per-cent spike over 2014. Jewish people remain the most targeted religious minority in Canada, though reported anti-Semitic incidents declined in 2015 over the previous year, the federal agency said.

Meanwhile, the percentage of women targeted by violent hate crimes increased because of a hike in the number of victimized women in the Jewish and Muslim communities. Over all, the sharpest rise in hate crimes was in Alberta, where officials have already noted an increase in total crime due to the province’s economic downturn.

Still, the true picture of hate in Canada is probably darker than the numbers released on Tuesday suggest. Statscan said the figures “likely undercounts” the real extent of hate crime in Canada because not all crimes are reported to police.

The two-year lag in releasing the figures is problematic at a time when Muslims feel the effects of turmoil linked to global radicalization, the presence of far-right groups in the West and the anti-Muslim rhetoric adopted by U.S. President Donald Trump.

Mr. Elgazzar’s organization has received an influx of complaints about anti-Muslim incidents this year, but they won’t be reflected by Statscan until 2019, he said. The data released on Tuesday are already two years old.

“You can’t build a case without evidence, and the evidence we have is stale,” he said. “It’s 2017 and I’ll tell you we’re having a pretty rough year. But we’re only going to hear about it in 2019.”

I suspect that international news events were a more important factor than the previous government’s playing identity politics (no excuse). Another possible factor, hard if not impossible to measure, is the degree to which Canadian Muslims are more willing to report hate crimes to the police, which has been an issue in the past. Higher numbers may reflect in part better Muslim-police relations.

In terms of timelines required to produce these reports, it would be nice, and should be possible, to have a one-year time lag rather than 18 months as at present, while ensuring the necessary data integrity and consistency.

One of the better overviews, with the relevant charts (just comparing the past two years compared to my eight year comparison The Daily — Police-reported hate crimes, 2015 (with annual 2008-15 data)):

The number of hate crimes in Canada jumped five per cent in 2015 from the year before, according to a Statistic Canada report released Tuesday.

The report looked at a variety of hate-crime statistics—from crime motivations and violations to the demographics of victims and the accused.

In total, 1,362 hate-crimes were reported across the country that year. To put that in perspective, there were nearly two million criminal incidents reported to police in the same year.

An increase in hate-crimes based on religion and race

Two major factor explain the increase—an uptick in religiously-based and race-based hate crimes. Nearly 50 per cent of all hate crimes reported in Canada in 2015 were motivated by hatred of race or ethnicity.

The largest increase in religiously-based hate crimes was against Muslims (an increase of 61 per cent to 159 incidents) and Catholics (a 57 per cent increase to 55 incidents). Jewish people faced the highest level of religiously motivated hate crimes (178 incidents) despite seeing a 16 per cent drop over the two years.

Hate crimes targeting Blacks were still the highest of all racially or ethnically motivated crimes in 2015 (224 incidents), though that was down slightly from the year before.

Hate crimes targeting sexual orientation fell by nine per cent between 2014 and 2015.

Violent hate crimes also increased

Violent hate crimes increased 15 per cent from 2014 to 2015, accounting for more than two-thirds all police-reported hate crimes. The most common types of violent hate-based crimes were assaults, which jumped13 per cent from the year before, and uttering threats, up 22 per cent.

Most victims younger than 35 years old

Nearly 60 per cent of hate crime victims in 2015 were younger than 35 years old, according to the report—a similar percentage as in 2014.

When it comes to victims of hate crimes motivated by religion, however, victims were younger than the year before—people under 35 accounted for nearly 60 per cent of victims in 2015, up from around two-thirds the year before.

FINAL---Characteristics-of-hate-crime-victims,-Canada,-2015-(%)

People accused of religious hate crimes are most likely to be under 18 years old

In more than 22 per cent of religious hate crime incidents, young people aged 12 to 17 years old were the perpetrators. Meanwhile people under the age of 24 were responsible for slightly more than half of hate crimes that targeted sexual orientation.

FINAL---age-distribution-of-persons-accused-of-hate-crimes-nationally,-2015--ungrouped

In its report, StatsCan suggested that the actual number of hate crimes could be considerably higher than what it found. It estimated that in two thirds of cases of hate crime, victims don’t file complaints with police. The agency also cautioned that the reporting rates can also vary by the targeted population—for example, some demographic groups might be more willing to report than others.

Source: A closer look at the rise in hate crimes in Canada – Macleans.ca