Fighting Expertise with Expertise

Lisa Kerr, a brilliant colleague of mine at the JSD programme at NYU and soon-to-be professor at Queen’s, has recently published a fascinating article called “Contesting Expertise in Prison Law,” explaining the practical and normative importance of expertise and evidence in prisoners’ rights adjudication. I am no doubt biased, but I think it deserves to be read and thought about, both for its importance to its specific topic, and for what it can tell us about some much broader trends in Canadian law. This post is mostly a summary of the article. I will offer a critique, focusing on its relationship with these broader trends, in a separate one (hopefully tomorrow or this weekend).

Ms. Kerr’s argument, in a nutshell, is an almost Posnerian plea for judges to be attentive to facts and, in particular, to the information that various experts can provide about prisons, when they adjudicate constitutional claims brought by prisoners, as well as for lawyers to provide judges with such information. Armed with facts and expert opinion, Courts can and should stop being unduly deferential to prison administrations, which tend to cloak their imposition of unnecessarily harsh conditions on prisoners in claims of expertise. “Prisons,” Ms. Kerr writes, “do not need to be viewed as mysterious places by courts, nor as places where necessarily amateur outside intervention could trigger unknown dangers.” (74)

Looking at both Canada and the United States, Ms. Kerr traces the rise and decline of judicial willingness to intervene to protect the rights of prisoners. For a long time, she notes, inmates were effectively regarded as having lost all rights. Prisons were places where the ordinary law, including constitutional law, did not run. But beginning in the 1960s in the United States, and the 1970s in Canada, courts started intervening and imposing constitutional constraints on prison authorities. Yet the movement was reversed in the United States, with courts adopting increasingly deferential approaches to the claims by prison authorities that this or that repressive measure was necessary to ensure prison security or otherwise advance “legitimate penological goals.” (A welcome counterexample, too recent to have made it into Ms. Kerr’s article, is the U.S. Supreme Court’s unanimous decision this week in Holt v. Hobbs, rejecting the claims Arkansas Department of Corrections’ claim that allowing a devout Muslim inmate to wear, for religious reasons, a half-inch beard would be too dangerous.) In Canada, although one might have expected the coming into force of the Charter to spur the courts to accept prisoner claims, their “hands-off” instincts have proven remarkably resilient. Both lower courts and, on occasion, the Supreme Court have been disinclined to look into the issues of sentence administration, and preferred to treat prisons as separate universes properly subject to their own rules.

Yet there have been hopeful signs, Ms. Kerr notes. Sauvé v. Canada (Chief Electoral Officer), 2002 SCC 68, [2002] 3 S.C.R. 519), which held that prisoners could not be disenfranchised, was one case where their constitutional rights were granted the same significance as those of other citizens. And, more recently, challenges supported by sophisticated evidentiary records have succeeded, at trial level, in thwarting the prison administrators’ repressive instincts. The presence in the record of social science evidence can make it clear ― and indeed force prison authorities to acknowledge ― that the impugned decisions were made without any real justification or even thought going into the process, and that alternative, less repressive, courses of action are available. Probably the most important theoretical point Ms. Kerr makes is that

[t]he question of evidence has, in fact, been the critical dimension for claimants who experience chronic marginalization and popular resentment. This is at least partially because the evidentiary record is the means by which counsel can insist that constitutional adjudication not mirror conjecture and stereotyping from the wider culture. (76; emphasis mine.)

The practical takeaway follows immediately, and inexorably:

Counsel for prisoner claimants should continue to focus on the issue of expert evidence, notwithstanding the difficulties of doing so, and should be aware that there is an extraordinary range of expertise and literature that could bear upon future Charter claims.(76)

It is a bit ironic, of course: if you want to contest expertise (that of prison authorities), you have to counter it with more expertise (that of social scientists or even, for that matter, that “former prison administrators, and administrators from other jurisdictions” (74)). And since the government is, as Ms. Kerr acknowledges, itself well-positioned to collect and present expert evidence, it could potentially play that game too. But the end point is not tto stick it to the government ― it’s to make prisons a more human place (and, by way of consequence, those who eventually come out of them better, or at least less-worse, people!). If the government starts winning cases on the basis that its policies are actually in keeping with at least a reasonable take on the best expertise that exists in the realm of prison administration, rather than simply because courts reflexively defer to the administrators, Ms. Kerr will have succeeded, and we all will be better off.

The Elephant in the Conference Room

In my post discussing the Supreme Court’s recent decision in R. v. Fearon, 2014 SCC 77, which held that (some) warrantless searches of cell phones incident to arrest were constitutional, I “wonder[ed] whether Justice Cromwell was swayed by his conclusion ― irrelevant and apparently unsupported ― that the robbery of which Mr. Fearon is accused is ‘a crime that has become depressingly routine.’” [5] I’m not alone in having found Justice Cromwell’s comments strange. Unlike me, however, Addison Cameron-Huff has done the right thing and fact-checked them.

What Mr. Cameron-Huff has found is that armed robberies have become much less, not more frequent in the last 15-20 years. He points, for instance, to Statistics Canada finding that between 1999 and 2008, the incidence of armed robberies went down by more than 25%, and that of armed robberies involving a firearm, by more than 35%. I guess it’s possible that this trend was reversed in the last few years, but that seems not likely at all: the RCMP points out that one particular type of armed robbery, namely “[b]ank holdups have declined even more dramatically — 46 per cent — across the country from a total of 1,098 robberies in 2000 to 591 in 2012,” while Statistics Canada reports that the overall robbery rate has gone down by 17% between 2002 and 2012.

Mr. Cameron-Huff also notes that he couldn’t “find any reference in the factums filed by the parties to an increasing crime rate. Justice Cromwell seems to have taken judicial notice of this (incorrect) statistic.” Less politely, one might suggest that Justice Cromwell took that “statistic” out of thin air. To be sure Justice Cromwell’s words might be (very charitably) read as suggesting that armed robberies are now considered not to be a big deal, rather than that they are more common. Yet not only would that reading be a stretch, but that assertion too would be unsupported by anything.

That a Supreme Court opinion would make such unsupported assertions ― as best we can tell, falsifiable, and actually false, unsupported assertions ― is remarkable. One thought here is that, for all the problems with social science evidence that I have been blogging about here and elsewhere, it is even worse when courts assert “facts” on which such evidence bears without looking at that evidence at all. Relatedly, it might be bad when judges google, but even worse when they make assertions that a fairly simple search would have proven wrong.

It is also remarkable, I think, that Justice Karakatsanis’s dissenting opinion ― which I otherwise quite like ― does not challenge Justice Cromwell on this point at all. Did Justice Karakatsanis think that Justice Cromwell’s assertion is not important? Did she not want to make her dissent more confrontational than it needed to be? Or did Justice Karakatsanis, as well as Justices Abella and Lebel, who dissented with her, not realize that Justice Cromwell was wrong? Did they not even suspect that he might be, enough to ask a clerk to check? That last possibility, suggesting that the Supreme Court is at serious risk of not noticing the elephant of falling crime rates in its conference room when it discusses future criminal law cases, is deeply worrying.

More on the Empirical Turn

My latest post for the CBA National Magazine’s blog has now been published. Continuing on the topic of the “empirical turn” in Charter litigation exemplified by decisions such as Canada (Attorney General) v. Bedford, 2013 SCC 72, [2013] 3 S.C.R. 1101, I argue that the courts’ reliance on social science evidence can be dangerous because the science in question is often affected by ideological bias.

I have recently blogged about the inextricable mixing of advocacy and social scientific expertise manifested in the evidence submitted in Trinity Western University v. Nova Scotia Barristers’ Society, 2014 NSSC 395. But there, at least, the advocacy was manifest. The problem which I now discuss is more subtle, and therefore more dangerous. As I explain in my post, even social science that does not look like advocacy, to its authors or even to outside observers, can be tainted by political or ideological bias.

This is not just speculation or conjecture ― my post is largely based on a paper by a group of social psychologists, including the celebrated Jonathan Haidt. (If you want to see him in action, watch this recent talk at NYU; it’s some of prof. Haidt’s comments, mostly in the Q&A, that set me off on this topic.) Of course, you may think that there is something ironic about using a paper written by a bunch of social scientists to attack the use of social science evidence. But, as I have already explained, I am not saying that we should give up on social science evidence altogether. Rather, we should be aware of the dangers that it brings, and adjust our procedures to deal with them as best we can. Both lawyers and judges, I argue, have an important role to play in this regard.

Inside the Empirical Turn

A judge’s take on dealing with the social science evidence involved in Charter litigation.

I have written about the doubts that I and others have about the increased reliance of Canadian courts on social science evidence in rights litigation. A recent decision of the Supreme Court of Nova Scotia, Trinity Western University v. Nova Scotia Barristers’ Society, 2014 NSSC 395, seems to me to suggest that to some judges, at least, this “empirical turn,” as it’s been described, does not come quite naturally.

The University has applied for judicial review of the Barristers’ Society decision to require it to abandon the “covenant” it imposes on its students, which among other things prevents them from engaging in sex outside a heterosexual marriage, before recognizing its law school’s degrees. The application has not yet been heard on the merits. In the meantime, the University has applied to strike parts of affidavits submitted by the Society’s witnesses, on the basis that they contained “material that amounts to a submission, plea or argument. The concern would be that the arguments might be accorded more weight or deference by virtue of their being sworn and in a report of an expert. Expert opinion and argument can be confused.” [23]

As Justice Campbell explains, these affidavits, despite their ostensible form, “are essentially reports intended to provide expert opinion,” [3] one of them on “the history of discrimination against gays, lesbians and other sexual minorities, the effect of exclusion on sexual minorities and the role that the state has taken in combatting that discrimination” [3] and the other on “the effects of the [University’s] covenant on sexual minorities” and “the impact on recruiting sexual minorities if the [Society] were to recognize degrees from [the University’s] proposed law school.” [4]

Justice Campbell’s explanation of the differences between this evidence and that which courts are used to is worth quoting at length:

[5]     These affidavits, as evidence, occupy a grey area. Affidavits are intended to be statements of facts. They do not contain opinions about evidence or speculation about what evidence could be. But expert reports are intended to convey an opinion based on assumed facts using specialized knowledge, training or experience. These are both affidavits and expert reports. Strictly applying the rules that pertain to affidavits would be impractical.

[6]     Reading through the affidavits gives a sense of just how different they are from the usual, sometimes rather dry, recitation of facts found in an affidavit. They clearly contain information that is not within the personal knowledge of the affiant. Of course they contain opinions. They also include statements that are intended to persuade.

[7]     They are also a departure from what might normally be expected in an expert report. It’s a bit hard at times to differentiate between conclusions based on expertise, assertions of facts, and advocacy.  The reports contain statements about historical events and trends that are supported by their assertion with no intent that they be subject to formal proof.

[8]     The language used is at times powerful. It is at times impenetrably academic. And, at other times takes the form of advocacy. One is not left in any doubt at all about whether these experts have a definite view about what the outcome of the matter should be. … That isn’t what an expert opinion has traditionally sounded like. It isn’t what affidavits have traditionally sounded like either.

[10]   Technically, a report of an expert would contain a statement of assumed facts that would have to be proven in court. The expert would then offer an opinion or a theory relevant to the case, using training, knowledge and expertise of a kind not possessed by the judge. That opinion would be based on those assumed and to be proven facts. That opinion would be testable or verifiable.

[11]   Charter litigation has changed that.

[12]   Social science evidence is critical in making decisions on the interpretation of the Charter especially when those matters involve public policy. Reports from experts in those areas convey a kind of information that may be quite different. The way in which those reports are used by courts can also be quite different. Adjudicative facts are those that are proven by evidence and relate directly to the subject matter of the proceeding. Judges can take judicial notice of facts that are widely known and beyond dispute. Social science evidence or legislative facts fall between those more traditional categories.

[14]   The Supreme Court has noted that when social and legislative facts are put before a trial judge he or she has to “evaluate and weigh that evidence in order to arrive at the conclusions of fact necessary to decide the case”. The trial judge can be involved in a time consuming exercise of reviewing material and reconciling differences between the experts, studies and research results.  Social and legislative facts are also intertwined with adjudicative facts. The judge has to analyze and give appropriate weight to the evidence in its different forms.

[15]   Social science evidence is presented through expert witnesses. The assessment of that evidence relies heavily on the trial judge. Social science evidence will necessarily be different from the reports of experts in the natural sciences. In those areas, where there are controlled experiments, testable predictions, and quantifiable data an expert opinion will usually involve the application of rigorous testing methods. There is no hierarchy of sciences and social sciences are no less rigorous than natural sciences. That statement itself could be the subject further debate requiring expert opinion.  Suffice it to say, social sciences and natural science are just different.

[16]   A trial judge has to consider the reliability of the particular social science evidence. If the report goes beyond the scope of the expertise of the writer, that has to be considered. If it is tendentious, or argumentative, or unbalanced those are considerations as well. Social science evidence that is supported by empirical research may be given different consideration from that which is based on anecdotal observation.

To dispose of the University’s motion, Justice Campbell applied a two-part test developed by Justice Doherty, of the Court of Appeal for Ontario, in R. v. Abbey, 2009 ONCA 624, for the admissibility of such evidence. As he described it,

[17]   …  First, the party who puts forward the evidence has to show the existence of some preconditions to the admissibility of expert evidence. The proposed witness must actually be an expert. The opinion has to relate to a subject matter that is properly the subject of expert evidence.  The opinion has to be logically relevant to a material issue. Finally, the opinion must not “run afoul of any exclusionary rule apart entirely from the expert opinion rule”.  …

[18]   Second, if the trial judge decides that the expert evidence meets the preconditions of admissibility, the issue is whether it’s sufficiently beneficial to the trial process to warrant its admission. That involves a discretionary cost-benefit analysis.  The “costs” in that analysis are “a consumption of time, prejudice, and confusion.” The benefits are found in the potential value of the evidence to actually prove something of significance. Whether the expert evidence does indeed have that potential requires a consideration of the reliability of the evidence. Reliability can be assessed by considering the subject matter of the evidence, the methodology used by the expert, the expert’s expertise and the extent to which the expert is shown to be impartial and objective.

Reviewing the affidavits in detail, Justice Campbell rejected most of the University’s concerns. However, he observed that one of them contained “parts … that amount to commentary on the current state of Canadian society, politics and public opinion,” [40] and the other “advocacy on the very matter before the court … openly and unapologetically argumentative.” [59] Nevertheless, he refused to strike any part of the affidavits, holding that “[t]he elements of argument contained in both involve the kind of prejudice that can be minimized by acknowledging them for what they are.” [60] Justice Campbell added that it was better to let the merits judge see the “report” ― he regularly referred to the purported affidavits in that way ― in its entirety, in effect suggesting, it seems to me, that these argumentative parts might cause the judge to discount the weight to be given to the affidavits’ conclusion.

This slightly subversive conclusion is probably right. Yet it also suggests that my and others’ worries about the “empirical turn” are not groundless. As I explained here, I do not believe we should renounce any and all resort to social science evidence. But as Justice Campbell makes clear, this evidence is quite different from what judges, especially trial judges, are  used to. It almost inevitably contains advocacy, backed by an expert’s aura of credibility on topics judges are likely not be well-versed in. Judges need to pay attention to research methods and weigh the reliability of the studies put before them, but how many of them are trained to do so? Arguably, it would be useful to have some guidelines on dealing with such evidence laid down by appellate courts, but since the Supreme Court held, in Canada (Attorney General) v. Bedford, 2013 SCC 72, [2013] 3 S.C.R. 1101, that they must defer to the trial judges’ findings, they are not really in a position to do so. In short, as I have already suggested, the “empirical turn” could well take us off course.

More Dead than Ever

While the Supreme Court is getting ready for the oral argument in Québec’s challenge to the abolition of the long-gun registry by the federal government (set for October 8), a different challenge to the constitutionality of the Ending the Long-Gun Registry Act was dismissed by Ontario’s Superior Court of Justice earlier this month in Barbra Schlifer Commemorative Clinic v. Canada, 2014 ONSC 5140. Justice Morgan held that, contrary to the Clinic’s claims, the abolition of the gun registry did not infringe the right of women either to the security of the person or to equality, and thus did not contravene sections 7 or 15 of the Charter.

Justice Morgan began his s. 7 analysis by discussing the “state action problem” with the Clinic’s argument. The Clinic claimed that the abolition of the gun registry infringed s. 7 because it increased the risk that women would suffer gun violence, in particular from their domestic partners. In its view, its argument was similar to those that prevailed in Canada (Attorney General) v PHS Community Services Society, 2011 SCC 44, [2011] 3 SCR 134 (the Insite case) and in Canada (Attorney General) v Bedford, 2013 SCC 72, [2013] 3 SCR 1101, where the Supreme Court found, respectively, that a denial of an exemption from drug laws to a safe-injection clinic and the prostitution-related provisions of the Criminal Code infringed s. 7 by increasing the risks to which the clinic’s users and sex workers were subject. But the fact a government policy fails to reduce a risk to which a person is exposed as much as possible is not enough, Justice Morgan found:

[c]ontemporary society is permeated by risk, including the risk of violent crime and injurious use of firearms, but unless that risk is a creation of state intervention it is not within government’s constitutional responsibilities. (Par. 25; emphasis mine.)

There is, furthermore, a crucial difference between the alleged increase of risk as a result of the abolition of the long-gun registry and the facts of the Insite case and Bedford, in that in those cases the government had prevented people from taking steps, on their own, to minimize the risks they were subject to, whereas the gun registry is a government-created risk mitigation scheme (if indeed it is that). The upshot of the Insite case and Bedford is that the government cannot prevent people from reducing the risks they run, not that it has a duty to do it by itself. Indeed, Justice Morgan pointed out,

[t]he Supreme Court indicated in Bedford, at para 88, that while the state cannot make prostitution even riskier than it is, the claimants [in that case] were “not asking the government to put into place measures making prostitution safe” (par. 34).

When Parliament itself creates a risk-reduction scheme, it can expand or limit it as it in its wisdom sees fit. So long as it doesn’t interfere with people’s autonomous risk-reduction endeavours, it does not act unconstitutionally.

The Clinic’s position also suffered, according to Justice Morgan, from a “baseline problem.” As he saw it, the Clinic

use[d] the 1995 Firearms Act [which had created the long-gun registry] as a baseline against which to measure the [new] licensing and registration system … , and finds the latter lacking. In the Applicant’s characterization, while the regulatory regime that existed from 1995 until … 2012 “provided an essential element in the effective protection for the s. 7 rights of women”, the “[Ending the Long Gun Registry] Act’s purpose appears to relate primarily to political aims”. (Par. 41)

But a statutory scheme cannot be made into a constitutional standard in this way. The gun registry’s creation was no more “politically neutral” than its abolition (par. 45). It was a choice of policy and politics, and it can be reversed or revised.

Justice Morgan then turned to the evidence which the Clinic adduced in support of its s. 7 claim, and found it insufficient. That evidence seems mainly to have consisted of statistics showing that gun violence declined in Canada while the long-gun registry was in operation. But, Justice Morgan said, gun violence had been declining before the registry was set up, and the early indications are that gun violence against women has not increased after it was abolished. In fact, as the government argued, there is a long-term trend towards the reduction of violent crime, so that what the Clinic showed was correlation, but not causation. Furthermore, police officers ― including those who served as the Clinic’s expert witnesses ― suggest that the effectiveness of gun registration (as opposed to licensing and background checks, which the abolition of the registry does not affect) is debated and, at best, limited. Thus even if an increase of risk resulting from the abolition of the long-gun registry were a constitutionally cognizable harm, the Clinic, in Justice Morgan’s view, had not demonstrated that the abolition had that effect.

Justice Morgan found that similar evidentiary problems undermined the Clinic’s claim that the abolition of the long-gun registry infringed the Charter’s equality guarantee becaused it would disproportionately affect women would be the victims of domestic gun violence. Again, there is evidence that gun both violence and domestic violence are decreasing. But given the multiplicity of causes likely to be involved, it is very difficult to isolate the gun registry’s role, if any, in these processes. The point, Justice Morgan said, is “not that there is no place for statistical data, but rather that statistical data can rarely, if ever, do the work of proof on its own” (par. 100). Nor was there anything to show that whatever effects the abolition of the gun registry might have on women were discriminatory in the sense of stereotyping or perpetuating past disadvantage. The overall statutory scheme for regulating firearms still tries to reduce gun violence, including in particular gun violence against women. Changing the mix of criminal and regulatory elements in that scheme was not, Justice Morgan said, discrimination.

I think this is the right decision. Justice Morgan’s point about the “state action problem” and the “baseline problem” are well taken, and his concerns about the insufficiency of the evidence to show that the gun registry’s abolition would have the disastrous consequences its supporters expect are justified.

I don’t know whether the Clinic intends to appeal his decision, but its chances of success would be very low indeed, given the Supreme Court’s insistence, in Bedford, that a trial judge’s assessment of social science evidence in constitutional cases is entitled to as much deference on appeal as any other form of fact-finding. Even if Justice Morgan were wrong in has approach to the issue of state action (and I don’t think that he is), his conclusions regarding the evidence would still stand.

For better or worse, this is the brave new world of Charter litigation ― a world in which cases live and die depending on the lawyers’ ability to assemble an evidentiary record and to convince a trial judge of this record of that record’s persuasiveness. And as I wrote here in discussing the potential dangers of this approach, “[t]he government, as the best funded and most powerful interest group of them all, is more likely to have the resources to put together a solid record than those who challenge it.” So here the government was able to show that the general trends in crime reduction, and its other (purportedly) crime-preventing policies meant that it was not possible to establish the effects of the gun registry’s abolition or continued existence with any degree of certainty.

This is not to say that I regret the outcome of this particular case, though others will find it regrettable. I will, however, note the irony of the government’s reliance on statistics showing a long-term general decrease in crime to defend this one element of its criminal-law legislative programme while implementing other, “tough-on-crime,” policies with a total disregard for these statistics. I hope that the courts which will consider the constitutionality of these measures will take good note. The gun registry is more dead than ever ― though the Supreme Court might still resurrect it in Québec. But in killing it, the government might just have inflicted some collateral damage on its other policies. No regrets from me there, either.

The Empirical Turn

In a post on the National Magazine’s blog, Kerri Froc discusses (among other things) what she refers to as “the empirical turn in Charter jurisprudence” ― the  tendency of Courts to decide Charter cases on the basis of social science evidence instead of “expounding on the nature of human values embodied by rights.” Perhaps most recent Supreme Court decision exemplifying this trend is the one in Canada (Attorney General) v. Bedford, 2013 SCC 72, [2013] 3 SCR 1101, which invalidated the provisions of the Criminal Code relative to prostitution. I had some thoughts of own on this topic here, so it is interesting to compare notes.

If I understand her correctly, Ms. Froc worries that increased reliance on social science evidence, in tandem with a narrow and technical approach to the application of the Charter, may lead to “democratic deficits” in rights adjudication. In particular, these twin trends, and the corresponding judicial retreat from more “abstract forms of philosophic reasoning,” cause “Charter analyses to become more divorced from the actual realities represented in the case,” while leaving the courts with excessive discretion in their treatment of the evidence, which they can use to either enable or stifle rights claims. Although Ms. Froc questions whether a more philosophical approach to the Charter would produce “more consistent and democratic outcomes,” she seems to suggest that it would at least be preferable from a process standpoint.

My own doubts about the direction in which the “empirical turn” is leading us concern both process and outcomes. Starting with the latter, my worry is specifically that deferring to findings of social science-based “legislative facts” by trial judges, as Bedford requires appellate courts to do, will lead to more mistakes than either more principle-based judicial decision-making or even a fact-based adjudication that allowed plenary review of such findings of fact on appeal. This is in part because trial judges may lack the skills to engage with social science evidence, and in part because, since trial judges sit alone rather than in panels, mistakes that they will make in assessing such evidence will not be caught before they issue their decisions. Besides, the government, whose resources are much larger than those of any civil-society organization, will be at an advantage in the evidentiary food-fights required by the “empirical turn.” We not are noticing this yet, because the current government is incredibly evidence-averse, and often legislates without bother to collect any facts at all, but that may change with the bill it is preparing in response to the Bedford judgment. Furthermore, the requirement to put together voluminous and complex social-scientific records to support Charter claims can not only lead adjudication astray, but slow down and maybe even prevent the filing of meritorious Charter claims.  As Sonia Lawrence put it, “[i]t’s a long road to justice this way around, folks,” given the time it takes to put together such records (and the money necessary to do so).

As for my process-related concerns, they might, like Ms. Froc’s, also be described as having to do with democracy, at least in a broad sense. One such concern is with accountability. Shifting the weight of rights adjudication from principle-based reasoning by appellate courts to fact-finding at trial risks making it less visible, especially to the media which tend not to devote nearly as much attention to the decisions of trial courts as to those of the Supreme Court. As I wrote in the post linked to above,

under the Supreme Court’s approach in Bedford, by the time a case comes up for appeal, and eventually reaches the highest ― and most visible ― court, it may well already be effectively decided. The Supreme Court can then avoid responsibility for controversial decisions, saying that “the trial judge made us do it.” (Indeed, I wonder whether this was not the real attraction of this approach to the Supreme Court in deciding Bedford.) Call it judicial leading from behind.

My second “democratic” concern with the “empirical” turn is that the costs and difficulty of assembling a trial record required by this approach will make Charter litigation the province of well-funded organizations and interest groups, leaving individuals without such support unable to pursue their claims.

The irony of it all is that the “empirical turn” in rights adjudication is probably motivated by judicial concerns over the “democratic deficit” of judicial decision-making in the realm of social policy and/or morality. As I wrote earlier, it seems likely enough that

the Supreme Court prefers to rely heavily on fact-finding ― preferably on fact-finding by others ― in order to avoid responsibility for controversial decisions, or at least, to put the point less harshly, in order to make such decisions appear more inevitable and hence less questionable and more legitimate.

Yet if Ms. Froc or I are right about the problems we see with the “empirical turn,” this strategy risks backfiring and making judicial review appear deeply problematic, perhaps even “undemocratic” ― to an even greater extent than it necessarily, and rightly, is. The point, as I argued here, is not that courts should ignore social science evidence altogether. In some cases, an “empirical turn” can be a useful contribution, perhaps even a correction, to Charter adjudication. But courts should be aware of its dangers, and should try to mitigate instead of compounding them. If that makes them appear more assertive and powerful, they need to take responsibility. Strong judicial review is not for the faint of heart.

Lies, Damned Lies, and Judicial Review

As the federal government considers its response to the Supreme Court’s ruling in Canada (Attorney General) v. Bedford, 2013 SCC 72, which invalidated the prostitution-related provisions of the Criminal Code, one can be forgiven for wondering whether its response will be guided by facts and research, or by ideology. Unfortunately, as a depressing but important guest post by Maggie McNeill on the Washington Post’s “The Watch” blog (run by Radley Balko, a civil libertarian) shows, when it comes to sex work, facts and ideology are often inextricably linked.

Ms. McNeill writes that

many of those who represent themselves as sex work researchers don’t even try to get good data. They simply present their opinions as fact, occasionally bolstered by pseudo-studies designed to produce pre-determined results. Well-known and easily-contacted sex workers are rarely consulted. There’s no peer review. And when sex workers are consulted at all, they’re recruited from jails and substance abuse programs, resulting in a sample skewed heavily toward the desperate, the disadvantaged and the marginalized.

Much of what passes for research on sex work, says Ms. McNeill, is produced not by impartial researchers, but by interest groups campaigning for the prohibition of any and all forms of prostitution, who often try to present them as invariably involving sex trafficking and exploitation of children, knowing that this is the most effective way to produce the response they desire from the public, the media, and legislators. Ms. McNeill also points out that the very fact that prostitution is criminalized in most of the world makes reliable statistics hard to come by (although even for countries where much sex work is legal, biased “researchers” choose to focus on its illegal side). Yet the media, which shape the perceptions of the issue among the general public and legislators, swallows dubious or outright distorted figures whole, without questioning the methodologies by which they are arrived at. (The Economist’s recent article based on a “study” which, as Ms. McNeill points out, involved fewer than 40 sex workers, drawn from one small segment of the profession, is a sad example of this trend, from a supposedly intellectual publication.)

The reason I am writing about this, apart from the obvious importance for all of us as citizens of being wise to what is going on, is that the litigation which will almost inevitably follow Parliament’s response to Bedford (especially if, as rumour has it, this response will include a criminalization of the purchase of sex) will heavily rely on studies of and statistics about sex work. What is more, under the Supreme Court’s holding, in Bedford, that the findings of “legislative fact” by a trial judge are entitled to full deference appellate courts, the way these studies and statistics will be handled at first instance is likely to determine the outcome of the new constitutional challenge, just as they determined that of Bedford itself.

As I wrote in my comment on what Bedford meant for the future of Charter litigation, this is a worrying perspective:

Trial judges ― most of whom are former litigators, without any sort of systematic training that would make them suitable for assessing social science evidence from which “legislative facts” are drawn ― are not chosen in the expectation that they will exercise determinative influence on the outcome of key constitutional cases. Many trial judges are capable of undertaking this responsibility. But many, with the best will in the world, are not. Of course, this may be true of appellate judges too, although presumably more of them are chosen for their expected capacity to deal with important, challenging cases. More importantly regardless of initial qualifications, appellate judges are more likely to develop an expertise relevant for exercising such functions, because there are relatively few of them, so that each one is much more likely to come across complex constitutional cases than a trial judge, of whom there are relatively many, so that each one may face significant constitutional cases only once or twice in his or her career. Furthermore, appellate judges do not sit alone. This means that the odds that an individual judge’s error will be caught and corrected before the court’s decision is issued are higher. Shifting power in constitutional cases from appellate to trial courts may thus lead to more errors in the dispositions of such cases.

The federal government will go into any new litigation armed with a record made up of studies and statistics about sex work ― in all likelihood, studies and statistics of the sort Ms. McNeill describes. A lot will depend on whether the lawyers who will challenge it will be up to the task of exposing the flaws of such a record. But, ultimately, the outcome of the case is likely to depend on the ability of just one trial judge to understand and critically assess this evidence. With all due respect to trial judges, I find this disquieting.