Docket: A-421-15
Citation:
2016 FCA 203
|
CORAM:
|
NADON J.A.
DAWSON J.A.
WEBB J.A.
|
|
BETWEEN:
|
|
HER MAJESTY THE
QUEEN
IN RIGHT OF
CANADA
|
|
(THE
COMMISSIONER OF THE CORRECTIONAL SERVICE OF CANADA, THE WARDEN OF KENT
INSTITUTION and THE WARDEN OF MISSION INSTITUTION)
|
|
Appellant
|
|
and
|
|
JEFFREY G.
EWERT
|
|
Respondent
|
REASONS
FOR JUDGMENT
DAWSON J.A.
[1]
The Correctional Service of Canada employs
certain psychological tests, referred to as assessment tools or actuarial
tests, to assess the risk of criminal recidivism and to assess psychopathy in
inmates. The respondent in this Court, Jeffrey Ewert, commenced an action in
the Federal Court in which he alleged that the assessment tools are unreliable
when administered to Aboriginal inmates such as himself and that, in the
result, their use violated rights protected by section 7 and 15 of the Canadian
Charter of Rights and Freedoms. At trial, he sought injunctive and
declaratory relief.
[2]
For reasons cited as 2015 FC 1093, a judge of
the Federal Court found that the use of the assessment tools in respect of
Aboriginal inmates was contrary to subsections 4(g) and 24(1) of the Corrections
and Conditional Release Act, S.C. 1992, c. 20 (Act) and also violated
section 7 of the Charter in a manner that could not be justified. The
Federal Court found it unnecessary to consider the application of section 15 of
the Charter noting that “[t]he facts in this
case are not sufficiently developed to usefully engage in the nuanced analysis
called for in s 15” (reasons, at paragraph 109).
[3]
In consequence, the Court stated its intent “to issue a final order enjoining the use of the assessment
tools in respect of the Plaintiff and other Aboriginal inmates until, at
minimum, the Defendant conducts or has conducted a study that confirms the
reliability of those tools in respect to adult Aboriginal offenders”
(reasons, at paragraph 114). Meanwhile, the Federal Court issued an interim
order that:
i.
enjoined the defendant from using results
obtained from the use of five specified assessment tools in respect of Mr.
Ewert; and,
ii.
required the parties to “file
written briefs in respect of the type, methodology and other relevant matters
in respect of the study to be conducted to assess the reliability of these
psychological tests in respect to adult Aboriginal offenders”.
[4]
This is an appeal from the interim order of the
Federal Court. For the reasons that follow, I conclude that the Federal Court
erred in law in finding that Mr. Ewert had established both a breach of the Act
and a violation of section 7 of the Charter. It follows that I would
allow the appeal and dismiss Mr. Ewert’s action. In the circumstances, I would
not award costs.
I.
The Decision of the Federal Court
[5]
After setting out the factual background and
describing the five assessment tools at issue, the Federal Court described the
expert evidence tendered by Dr. Hart for Mr. Ewert and Dr. Rice for the
defendant. The Court preferred the evidence of Dr. Hart, finding Dr. Rice’s
evidence to be of little assistance (reasons, at paragraphs 26, 47 and 53). Dr.
Hart’s evidence was summarized; the Federal Court described his salient
testimony as follows:
i.
the assessment tools are more likely than not to
be “cross culturally variant” (reasons, at
paragraph 28);
ii.
given the pronounced differences between
Aboriginal and non-Aboriginal groups, Dr. Hart would not apply scores derived
from the assessment tools to Aboriginal persons (reasons, paragraph 31);
iii.
in Dr. Hart’s view, the better approach is a
structured clinical assessment of an Aboriginal offender, which would include
some consideration of the information derived from the assessment tools, seen
in the totality of the circumstances known about the offender (reasons at
paragraph 32); and,
iv.
there are three methods to establish that an
assessment tool is free from cross cultural bias. None of the methods had been
applied to the assessment tools (reasons, at paragraphs 34-36).
[6]
The Federal Court accepted what it characterized
to be Dr. Hart’s evidence that the assessment tools “are
not good predictors of recidivism in Aboriginals … [t]hey suffer from cultural
bias” (reasons, at paragraph 53). It followed that the “expert evidence establishes that the test scores, in and of
themselves, ought not to be relied upon” (reasons, paragraph 56).
[7]
The Federal Court went on to find that
psychologists and the Correctional Service rely on the test scores produced by
the assessment tools (reasons, at paragraph 58). The use of the scores
adversely impacted institutional decisions that affected Mr. Ewert’s
eligibility for parole, his security classification and his ability to be
granted escorted temporary absences (reasons, at paragraphs 60-65, and 75).
[8]
The Federal Court then moved to its legal
analysis, finding that:
i.
while Mr. Ewert “framed
his case principally as a Charter breach”, more properly the case
involved a breach of statutory duty (reasons, at paragraph 76-77);
ii.
by “relying upon
questionable tests and in failing to ensure that the tests are reliable”
the Correctional Service breached subsection 24(1) of the Act which required it
“to take ‘all reasonable steps’ to ensure that information
about [Mr.] Ewert … is accurate, up-to-date and as complete as possible”
(reasons, paragraph 81);
iii.
it was not necessary on this issue for Mr. Ewert
to establish definitively that the tests are biased. It was sufficient “if he raises a reasonable challenge to their reliability
pursuant to the aforementioned statutory requirements” (reasons, at
paragraph 82);
iv.
contrary to Mr. Ewert’s submission, the
Correctional Service did not owe an “overarching
fiduciary duty to him” (reasons, at paragraph 86);
v.
with respect to section 7 of the Charter,
Mr. Ewert’s liberty interests were engaged because the assessment tools were
used to restrict or deprive his liberty by making parole “virtually impossible”, by setting his security
classification at a high level and by negatively affecting requests for
escorted temporary absences (reasons, paragraphs 88-89);
vi.
Mr. Ewert’s security of the person interest was
engaged because he had been labelled psychopathic (reasons, at paragraph 92);
vii.
there was no evidence that the scores and the
conclusions that flowed from the use of the assessment tools predicted
recidivism in Aboriginal offenders as accurately or reliably as they did for
non-Aboriginal offenders (reasons, at paragraph 99);
viii.
the “continued use of
the assessment tools is overbroad of the purpose and objective of the
legislation and of [the Correctional Service’s] decision making
responsibilities” (reasons, at paragraph 102);
ix.
using “infirmed [sic] factors
upon which to make decisions that affect s 7 rights is clearly arbitrary”
(reasons, paragraph 105);
x.
it was unnecessary to engage in a section 15
analysis (reasons, paragraph 109);
xi.
the use of the assessment tools was not
justified under section 1 of the Charter (reasons, at paragraph 112);
and,
xii.
an interim order would issue enjoining the
defendant from using the results of the assessment tools in regards to Mr. Ewert
and a remedies hearing order would issue to address the best and fairest manner
of implementing a final order (reasons, at paragraphs 115 and 116).
II.
The Issues
[9]
In my view, the following issues are raised on
this appeal:
1.
Did the Federal Court err in finding that the
defendant breached the statutory obligation imposed by subsection 24(1) of the
Act?
2.
Did the Federal Court err in finding that the
defendant violated Mr. Ewert’s liberty and security interests contrary to
section 7 of the Charter?
III.
Standard of Review
[10]
I agree with the parties that the standard of
review to be applied to the decision of the Federal Court is that articulated in
Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 S.C.R. 235. Thus, questions
of law are to be reviewed on the standard of correctness. Findings of fact,
inferences of fact and findings of mixed fact and law may only be reversed if
the trial judge has made a palpable and overriding error, or erred in law on an
extricable question of law.
[11]
Having considered the standard of review, I now
turn to the application of that standard to the issues raised on this appeal.
IV.
Did the Federal Court err in finding that the
defendant breached the statutory obligation imposed by subsection 24(1) of the
Act?
[12]
Subsection 24(1) provides:
|
24 (1) The
Service shall take all reasonable steps to ensure that any information about
an offender that it uses is as accurate, up to date and complete as possible.
|
24 (1) Le Service
est tenu de veiller, dans la mesure du possible, à ce que les renseignements
qu’il utilise concernant les délinquants soient à jour, exacts et complets.
|
[13]
The appellant, which I will refer to as the Correctional
Service, advances two arguments with respect to this issue that on my analysis
of the case I need not address. First, the Correctional Service argues that the
Federal Court misconceived the purpose and scope of subsection 24(1) of the Act:
properly understood, the provision does not impose any duty on the Correctional
Service to conduct scientific research or investigate. Second, citing Holland
v. Saskatchewan, 2008 SCC 42, [2008] 2 S.C.R. 551, at paragraphs 7 to 9 and
11, the Correctional Service argues that the Federal Court erred by granting a
private law remedy for a breach of a public law.
[14]
The first argument is based on a limited body of
jurisprudence from the Federal Court that arose in an entirely different factual
context. The second argument overstates the holding of the Supreme Court in Holland.
Neither argument was considered by the Federal Court. In these circumstances, and
when on my analysis these arguments need not be addressed, I prefer not to deal
with these issues. This said, these reasons should not be seen to endorse the
Federal Court’s interpretation of subsection 24(1) of the Act.
[15]
My concern with the Federal Court’s analysis of
this issue arises from the legal test it applied when considering whether Mr.
Ewert met the burden upon him to establish his cause of action.
[16]
As noted above, the Federal Court found that Mr.
Ewert, the plaintiff, did not have to “establish
definitively” that the assessment tools were culturally biased. In light
of the statutory obligation on the Correctional Service to respect the special
needs of Aboriginal persons in its policies, programs and practices and the
statutory obligation on it to take all reasonable steps to ensure that any
information about an offender that the Correctional Service uses is as
accurate, up-to-date and complete as possible, it was sufficient that Mr. Ewert
raise “a reasonable challenge” to the
reliability of the assessment tools. Thus, for the Federal Court, the question
was whether the Correctional Service’s “lack of action
is sufficient to fulfil the legislated standard of all reasonable steps to
ensure accuracy, currency and completeness” of information (reasons, at
paragraph 82).
[17]
This was a question which the Federal Court
answered in the negative.
[18]
The Federal Court cited no authority for the
proposition that Mr. Ewert was not required “to
establish definitively” that the assessment tools were biased.
[19]
However, a plaintiff bears the burden of
establishing on a balance of probabilities the existence of each constituent element
of the cause of action asserted by the plaintiff.
[20]
In his third amended statement of claim, Mr.
Ewert alleged that the assessment tools “generate false
results and false conclusions when used on Aboriginal persons” (Third
Amended Statement of Claim, at paragraph 14). Thus, Mr. Ewert bore the burden
of establishing these facts on a balance of probabilities. The Federal Court
erred in law when it required that Mr. Ewert establish only a reasonable
challenge to the reliability of the results and conclusions generated by the
use of the assessment tools in respect of Aboriginal persons.
[21]
In the absence of evidence establishing, on a
balance of probabilities, that the assessment tools produce or are likely to
produce false results and conclusions, the Federal Court could not find that
the Correctional Service failed to take all reasonable steps to ensure that
information about Aboriginal offenders was as accurate as possible.
[22]
Given that the Federal Court did not address the
civil standard of proof, it is necessary for this Court to assess whether Mr.
Ewert met the burden upon him to demonstrate that the assessment tools
generated false results and conclusions when administered to Aboriginal
persons. In doing this, I accept the Federal Court’s characterization of Dr.
Rice’s evidence and will not rely upon it. Thus, the only expert evidence that
I will consider is that of Dr. Hart.
[23]
I begin with Dr. Hart’s expert report. There, he
opined that:
12. In the fields of clinical and
forensic psychology, it is generally recognized that the reliability and
validity of an assessment instrument may be biased by personal characteristics,
such as age, gender, and culture. For this reason, it is considered important
to systematically evaluate the extent to which measurements made using an
assessment instrument are susceptible to bias, especially when those
measurements are used to make high-stakes decisions.
…
14. In Canada, one of the most
important cultural characteristics that is likely to cause cross-cultural bias
is status as an Aboriginal person. Status as an Aboriginal person is a general
or higher-order characteristic that encompasses numerous specific
characteristics, including such things as language, religion or spirituality,
self-concept, and fundamental social norms and attitudes.
15. It is common practice in CSC to
use instruments including the VRAG, SORAG, STATIC-99, VRS, and VRS-SO to assess
violence risk in adult male offenders, including those who are Aboriginal
people; and to use instruments including the PCL-R to assess psychopathic
personality disorder in adult male offenders, including those who are
Aboriginal people.
16. There is a large body of research
evaluating and supporting the reliability or validity of assessment instruments
such as the VRAG, SORAG, STATIC-99, VRS, VRS-SO, and PCL-R in heterogeneous
groups of offenders in CSC.
17. There is no research evaluating
the cross-cultural bias in reliability or validity of assessment instruments
such as the VRAG, SORAG, STATIC-99, VRS, VRS-SO, and PCL-R due to status
as an Aboriginal person using scientifically adequate procedures.
18. It cannot be safely assumed
that research findings derived from a heterogeneous group apply equally to its
constituent elements; this would be making the logical error of division.
What holds true for a group does not necessarily hold true for individuals or
sub-groups within that group.
(emphasis added)
[24]
I pause to note that the absence of research
evaluating the cross-cultural bias in reliability or validity of the assessment
tools does not establish on a balance of probabilities that the assessment
tools generate, or are more likely than not to generate, false results and
conclusions when administered to Aboriginal persons.
[25]
Dr. Hart gave the following evidence on his
direct examination before the Federal Court:
Q Now, I want to focus now on the
potential for bias in respect of the use of these actuarial tests on or for
aboriginal persons in Canada. Can you speak generally about the circumstances
in which a person considering applying, or using an actuarial test, should be
concerned about cultural bias?
A Certainly. Test-users are always
supposed to be concerned about cultural bias. Test developers are always
supposed to be concerned about the possibility of culture bias. We know that
culture bias can have either subtle or profound effects on the way that a test
is used and the reliability and validity of those test scores.
…
Q In respect of aboriginal people,
could you speak to the - - your paragraph 14 in which you write,
One of the most important cultural
characteristics that is likely to promote cross-cultural bias is status as an
aboriginal person.
I wonder if you can unpack that a little bit
and explain why status as an aboriginal person is likely to cause cross-cultural
bias in actuarial data.
A Certainly. I think it’s generally
recognized and accepted both within the field of psychology but also outside,
that there are major differences in life circumstances and the way that people
live their lives, as a function of whether one is aboriginal versus
non-aboriginal. We know that there is differences in history, in language, in
child-rearing practices, in beliefs and attitudes and norms and these
differences are so important that we make sure that we recognize them
officially, and in fact we’re required to do so as a matter of law in some
cases. For example, with respect to sentencing in the Criminal Code, or
that we know that when we’re doing research on things like crime and
delinquency, that we have massively different rates of some risk factors in
aboriginal communities. For example, if we looked at things like history of
victimization or the prevalence of substance use, we might find very different
rates in aboriginal versus non-aboriginal communities. Even things like
employment and income, or housing, and we can see massive differences between
aboriginal and non-aboriginal communities with respect to some of these risk
factors.
But also in the rates of perpetration of
violence. We know that aboriginal communities are affected more by problems
like violence than our non-aboriginal communities. And therefore, it’s sort
of a no-brainer to figure out that there is a good strong possibility that
status as an aboriginal person may affect the way that a violence risk
assessment test works. If you’re going to develop it with people who are
non-aboriginal it may not work the same in people who are aboriginal.
But I would say the same thing holds true
for gender, of course. Men and women have very different rates of perpetration
of violence. There is different sex roles, or expectations about behaviours as
a function of gender, different life opportunities. So it’s again kind of a
no-brainer, that if we develop a test with men, it may not work the same with
women. So, we have to assume that there may be problems.
And this is so obvious and clear that, for
example, in some of the tests of risk, the assessment of risk for recidivism
that are used even in the Correctional Service of Canada, they have gone back
over the years and after developing them, tested out whether they worked the
same way in aboriginal and non-aboriginal people, and they did not. So some of
the tests they’ve said they can’t use with aboriginal people, because it’s
clear that they don’t work the same way, that they’re not as effective and they
don’t yield good, meaningful information.
Q You mentioned strong
possibility, but I’m wondering in addition to the possibility of cultural bias,
is there a probabilistic relationship between the extent of cultural
differences and the likelihood of cultural bias when an actuarial test is used?
A Yes. I’m not going to be able to
give you a kind of a number probability, but if I can frame it in language,
natural language, I’d say that as somebody who works in the field of risk
assessment and works with offenders and so forth, and has worked with both
aboriginal and non-aboriginal offenders, any professional who works in this
field is alive to the distinct possibility that there is bias, and furthermore
that we would assume that there is some from [sic] of bias. Now, it may
be relatively small and it may be tolerable. But it could actually be large and
it may be intolerable. The only question is really how big it is, and how big
that bias is, and what impact it has.
But I would say that, you know, my own
professional opinion would be, it would be more likely than not that there is
some kind of bias. It’s just a matter of how big and what impact it has.
(emphasis added)
[26]
Read fairly, this evidence falls short of
establishing that the assessment tools generate, or are more likely than not to
generate, false results and conclusions when administered to Aboriginal
persons. This is because Dr. Hart acknowledges that cultural bias may have only
a subtle effect on the way a test is used and on the reliability and validity
of the resulting test scores - the bias might be relatively small and
tolerable. Notwithstanding this acknowledgement, Dr. Hart did not opine as to
the magnitude of the impact caused by any cultural bias. Nor was any other
evidence led on this point.
[27]
In the absence of evidence demonstrating that
cultural bias affected or is more likely than not to affect test usage or the
reliability and validity of the resulting test scores in a material way, Mr.
Ewert failed to establish on a balance of probabilities that the assessment
tools generate or are likely to generate false results and conclusions and
failed to establish on a balance of probabilities any breach of statutory duty
on the part of the Correctional Service.
[28]
This conclusion is consistent with paragraph 114
of the reasons of the Federal Court where it expressed its intent “to issue a final order enjoining the use of the assessment
tools in respect of the Plaintiff and other Aboriginal inmates until, at
minimum, the Defendant conducts or has conducted a study that confirms the
reliability of those tools in respect to adult Aboriginal offenders”.
Once the Federal Court concluded that it was uncertain whether the assessment
tools were reliable in respect of Aboriginal offenders, it ought to have
dismissed Mr. Ewert’s action. The Federal Court could not as a matter of law,
grant the remedy it did on the basis of an unequivocal evidentiary record.
V.
Did the Federal Court err in finding that the
defendant violated Mr. Ewert’s liberty and security interests contrary to
section 7 of the Charter?
[29]
A plaintiff pursuing a remedy under the Charter
must prove a violation of the Charter on a balance of probabilities (Canada
(Prime Minister) v. Khadr, 2010 SCC 3, [2010] 1 S.C.R. 44, at paragraph
21).
[30]
The Federal Court did not expressly consider
whether Mr. Ewert had established on a balance of probabilities that the scores
and conclusions generated by the assessment tools were inaccurate and
unreliable when the assessment tools were administered to Aboriginal persons.
Instead, it cited Dr. Hart’s testimony that there was no evidence that the
scores and conclusions predicted recidivism in Aboriginal offenders as
accurately or reliably as they do when administered to non-Aboriginal offenders
(reasons, at paragraph 99).
[31]
In my respectful view, by relying on the absence
of evidence proving the reliability of the assessment tools, the Federal Court
erred in law by failing to require Mr. Ewert to establish his claim on a
balance of probabilities. Moreover, as explained above, when the expert
evidence of Dr. Hart is read in its entirety it is, as a matter of law,
insufficient to establish on a balance of probabilities that the assessment
tools generate results that are inaccurate or unreliable in a material way.
[32]
It follows that the Federal Court erred by
failing to conclude that Mr. Ewert had failed to establish a violation of his
section 7 rights on a balance of probabilities. It further follows that I need
not consider the other errors in the section 7 analysis alleged by the
Correctional Service. It is sufficient to say that these reasons should not be
seen to endorse the Federal Court’s analysis of the liberty and security
interests said to be engaged.
VI.
Conclusion
[33]
Mr. Ewert argued that, notwithstanding the
Federal Court’s assessment that the evidence was not sufficiently developed to
allow for the analysis mandated by section 15 of the Charter, the remedy
granted by the Federal Court could be supported under section 15 of the Charter.
[34]
I disagree. I see no error in the Federal Court’s
characterization of the sufficiency of the evidence. Mr. Ewert failed to
establish that the use of the assessment tools generated false results and
conclusions when administered to Aboriginal persons. In consequence, his claim
under section 15 of the Charter could not succeed.
[35]
For these reasons, I would allow the appeal.
Pronouncing the judgment that the Federal Court ought to have pronounced, I
would dismiss Mr. Ewert’s action. In the circumstances, I would not award
costs.
“Eleanor R. Dawson”
“I agree.
Marc Nadon J.A.”
“I agree.
Wyman W. Webb J.A.”
Email this Content