I prepared this material for a presentation this week at the Northwest Regional Conference for…
Psychological evaluations are routinely submitted in support of applications for “extreme hardship” waivers of inadmissibility. Yet despite widespread use, practitioners must rely on their own experience and anecdote to gauge whether such an evaluation will be worth its substantial expense to a client. Expense aside, practitioners are often skeptical about whether such evaluations will help their client prove up an application. Jointly authored by an immigration attorney, clinical psychologist and law professor, article uses statistical analyses of waiver decisions to answer some preliminary questions about whether psychological evaluations help waiver applicants demonstrate extreme hardship.
This content was originally published as: Greg McLawsen, Julia McLawsen and Prof. Kevin Ruser, Demonstrating Psychological Hardship; A Statistical Study of Psychology Evaluations in Hardship Waivers of Inadmissibility, 16 BENDER’S IMMIGR. BULL. 10 (JAN. 1, 2011).
Psychological evaluations are routinely submitted in support of applications for “extreme hardship” waivers of inadmissibility. Yet despite widespread use, practitioners must rely on their own experience and anecdote to gauge whether such an evaluation will be worth its substantial expense to a client. Expense aside, practitioners are often skeptical about whether such evaluations will help their client prove up an application. This article uses statistical analyses of waiver decisions to answer some preliminary questions about whether psychological evaluations help waiver applicants demonstrate extreme hardship.
I. Hardship waivers
There are many excellent guides to hardship waivers of inadmissibility, so we give only a summary orientation here. Veterans may skip to the next section. Because it is important for a psychologist who authors a hardship evaluation to have a clear understanding of the legal standard for waiver applications, the following discussion may also be helpful in that regard. Note that all unpublished cases discussed in this article are available either from LexisNexis or on the website of the Administrative Appeals Office (AAO).
In short, certain grounds of inadmissibility are subject to waiver upon a showing that a non-citizen’s qualifying family member would face “extreme hardship” if the noncitizen were denied a waiver of inadmissibility. The non-citizen’s own hardship is statutorily irrelevant. Rather, the non-citizen must demonstrate that waiver denial would cause extreme hardship to a U.S. citizen or lawful permanent resident (LPR) family member. For inadmissibility due to fraud or unlawful presence, only a spouse or parent meets the definition of qualifying relative, though for criminal inadmissibility a child of the noncitizen also qualifies. In addition to proving that a qualifying relative would experience extreme hardship, the waiver applicant must also merit a favorable exercise of discretion. In the absence of negative equities, this final requirement adds little to the applicant’s burden, since the relevant criteria are essentially identical to those used to judge extreme hardship.
For waivers of inadmissibility there is no statutory definition of extreme hardship, nor is one contained in the federal regulations. Indeed, the Board of Immigration Appeals (BIA) has announced time and again that extreme hardship has no “fixed and inflexible meaning.” While this may be candid, it is hardly helpful for those attempting to demonstrate statutory eligibility. (Imagine a jury receiving the following instruction: “there are no fixed and inflexible requirements for contract formation.”)
Although there is a limited body of case law discussing hardship waivers of inadmissibility, a multitude of immigration benefits turn on a determination of “extreme” hardship. The contexts most familiar to practitioners – and from which the largest body of relevant case law derives – are pre-IIRAIRA suspension of deportation (“extreme hardship”), and cancelation of removal for non-LPRs (“exceptional and extremely unusual hardship”). While courts caution against the cross-application of standards, it is widely accepted that the term “extreme hardship” constitutes a single legal standard throughout its use in the Immigration and Nationality Act (INA). Bruce Hake has demonstrated this convincingly in prior Bulletins. “Exceptional and extremely unusual hardship,” as required for non-LPR cancelation of removal, is a qualitatively identical but heightened standard as compared to “extreme hardship.” Similarly, the AAO has stated that “exceptional hardship,” as required for waivers of the two-year foreign residency requirement under INA § 212(e), is a qualitatively identical but less stringent standard as compared to “extreme hardship.”
In order to demonstrate extreme hardship for purpose of a waiver application, the non-citizen must prove the following:
That denial of a waiver application would (1) cause (2) a qualifying relative to (3) experience a magnitude of hardship (4) that substantially exceeds the magnitude of hardship an average individual would be expected to suffer if their ken was denied such a waiver.
The causation requirement is straightforward and follows from the plain text of the INA, which asks whether “denial of admission would result in extreme hardship” to the qualifying relative. Thus in all waiver applications the adjudicator’s determination is counterfactual at its core: what will likely happen to the qualifying relative in the event the waiver is denied? By default, there will be two possibilities for the adjudicator to consider. Since the inadmissible noncitizen will be residing outside the U.S. – that is what inadmissibility entails – the qualifying family members will have to choose whether (1) to remain in the U.S. and thus be separated from the inadmissible family member, or (2) relocate to the foreign country with the noncitizen. A failure to demonstrate hardship on both contingencies is clear grounds for denial. If one of the two options (staying or leaving) would constitute a hardship, adjudicators will presume that a family member is able to take, and will take, the option not constituting a hardship. Yet applicants can overcome this presumption by demonstrating that only one outcome is factually likely.
The heavy lifting in waiver applications comes in bolstering the claim that a qualifying relative would face an adequate magnitude of hardship. Case law identifies the well-known list of hardship factors, which are best understood as describing the categories of personal adversity encompassed in the legal concept of extreme hardship. Though formulations vary slightly, the list is something like the following: the length of the individual’s presence in the U.S.; the individual’s age; family ties to the U.S. and outside the country; the conditions in the country to which the individual may relocate; the financial impact of departure from this country; significant conditions of health, particularly when tied to an unavailability of suitable medical care in the country to which the alien will return; and the possibility of other means of adjustment of status or future entry into this country.
Finally, hardship qualifies as extreme by virtue of exceeding some level of adversity described as the normal consequence of inadmissibility. Yet appellate decisions are somewhat vague as to what exactly the normal consequences of inadmissibility are. Perhaps most importantly, it is unclear whether the “normal consequences” are an empirical baseline (i.e., the types of hardship that in fact generally result from inadmissibility) or a normative one (i.e., types of hardship ascribed to be an acceptable consequence of inadmissibility). By way of possible comparison, an individual may be liable at tort law for failing to exercise the degree of caution of a reasonable person under like circumstances. The reasonable man, however, is normative rather than empirical baseline for conduct. The standard reflects “a standard of conduct demanded by the community,” rather than an observation about how individuals are in fact likely to act in particular circumstances. Would we be able to characterize the “normal consequences” of inadmissibility if we could accurately describe the forms of hardship that in fact follow from deportation, somehow accounting for frequency and intensity of those hardships? Or, do “normal consequences” – like the conduct of the reasonable person – represent the quantum of hardship imputed as the normatively acceptable consequence of inadmissibility? This question is thus far unsettled.
II. Psychological hardship
Though not always set forth on the canonical lists of hardship factors, it is settled that “emotional and psychological impact” must be considered in regards to a hardship waiver application. Likewise, an adjudicator must consider any psychological evaluation offered in support of claimed hardship.
Severe mental illness, standing alone, is sometimes enough to prove extreme hardship. For instance, the AAO has held that relocation to Brazil would be an extreme hardship for an individual who was medicated for Schizo-affective disorder, had been hospitalized after threatening suicide, and was able to continue work only given the “presence, support and encouragement” of his wife. Yet it is probably safe to say that the cases will be few and far between in which psychological hardship, standing alone, will demonstrate a sufficient quantum of hardship to meet the waiver standard.
In the case of children, issues such as Attention Deficit Disorder may directly impact a child’s ability to adjust to a new environment abroad. In one case, for example, a child presented a “compelling example of hardship” where his psychologist opined that Attention Deficit Hyperactivity Disorder rendered the child “incapable” of learning the foreign language. As ever though, it must be apparent the child’s individual difficulties raise the hardship of relocation above the norm. Despite the de jure presumption that a qualifying relative may either remain in the U.S. or relocate to the foreign county, adjudicators more readily presume that children will follow an inadmissible parent, perhaps especially in the case of an inadmissible noncitizen mother.
By far the most prevalent psychological issues raised in waiver applications relate to emotional hardship. In this regard, the AAO routinely cites Matter of Pilch for the proposition that “emotional hardship caused by severing family and community ties is a common result of deportation.” In Matter of Pilch, two qualifying relative children had been present in the U.S. 11 and nine years respectively. They resided in the country with three children, though their eldest son and extended family members lived in their native Poland. Amongst other arguments, the Board rejected the contention the male respondent’s “emotional hardship” rose to the statutory standard of extreme hardship. It is worth emphasizing, however, that Matter of Pilch is perhaps most notable for what it is not – an example of particularly extreme psychological circumstances.
Matter of Pilch notwithstanding, applicants sometimes succeed in proving that their emotional hardship will rise above the level normally “caused by severing family and community ties.” At the extreme end, a multitude of waiver appeals have been sustained where the qualifying relative demonstrated serious issues of emotional distress, and produced expert testimony that waiver denial had the potential to lead to suicide. For example, the AAO sustained a waiver appeal where a psychologist testified the applicant suffered from depression, and that loss of access to medical care and anti-depressant medication could cause the relative to commit suicide. But emotional hardship sometimes forms the core of successful waiver applications on far less severe facts. Consider the following cases in which the AAO concluded the non-citizen had established extreme hardship:
- Based on an evaluation by a clinical social worker, the subject would continue to suffer from a “high level of anxiety” without the companionship and care of her husband. The subject had been in abusive relationships in the past, and the loss of her husband’s support would constitute an extreme hardship.
- Considering a letter from a psychologist concerning spouse’s “mental state,” there was a “serious question” as to whether the spouse would be able to continue performing her job duties if separated from the applicant.
- A “detailed” psychological evaluation documented the spouse had been physically and sexually abused as a child, that the applicant had been a source of stability since they were 15-years-old, and that it was “entirely possible” separation would be devastating to the wife.
- Both the applicant and qualifying relative suffered from depression, and waiver denial would cause extreme hardship to the relative in part due to his concern over the mental health of the applicant.
As Laurel Scott as observed, the small number of precedent decisions on waivers of inadmissibility are of somewhat limited use to practitioners. Relying on individual unpublished decisions is even more problematic. We undertook this study to explore whether statistical analyses of a relatively larger number of cases would reveal trends that could help practitioners make a more informed decision about recruiting the expertise of a mental health professional.
For purpose of the present article, we were interested in characterizing when expert psychological testimony appeared to influence the outcome of waiver adjudications. After reviewing published and unpublished waiver cases, and for the reasons below, we decided to explore the following two hypotheses:
- Expert testimony predicts success on appeal, but only if there is an ongoing therapeutic relationship with the clinician.
- Success on appeal is not related to whether or not a qualifying relative received a psychiatric diagnosis.
The first hypothesis was driven by the AAO’s frequent admonition that psychological evaluations have little probative value where the evaluator has met with the subject only one or several times. Decisions often draw attention to the number of times an evaluator met with a subject, and voice skepticism that evaluations can be probative if based on as few as one to three sessions with a subject. To the extent the AAO believes accurate psychological evaluations require many hours of time with a subject, we would expect greater weight given to evaluations submitted by providers with a standing relationship with a subject. Alternatively or additionally, emphasis on an ongoing therapeutic may provide the AAO with “objective” evidence of the alleged psychological distress; that a subject has sought out mental health treatment lends additional credibility to the claimed distress. (In a similar vein, the AAO cites to past use of psychotropic medication to bolster the credibility of the current allegations of distress.)  We therefore anticipated the mere presence of testimony from a mental health professional would not increase the likelihood of successful appeal outcome in the absence of an ongoing relationship between the clinician and subject; conversely, we anticipated that an applicant would be more likely to prevail if she did have such an ongoing therapeutic relationship.
Our second hypothesis – essentially, that psychiatric diagnoses make no difference – reflects two considerations. First, we suspected Matter of Pilch has been given broad application, to the extent it is tantamount to the proposition that almost any magnitude of emotional distress is a normal consequence of deportation. On the supposition that most psychiatric diagnoses would be for Major Depressive Disorder, or other mood-related disorders, we anticipated diagnoses would have little impact on appeal outcome. Second, we suspected the AAO does not view psychiatric diagnoses as the sort of credible and “objective” evidence that can bolster a claim of psychological hardship. Just as we anticipated the AAO would credit reports of clinicians only where they had an ongoing therapeutic relationship with the subject, we anticipated the AAO would discount mood disorder diagnoses as too easily manipulated.
Driving each of our hypotheses was a theory about the AAO’s overriding concerns about psychological testimony in general. Decisions by the AAO suggest an alarm that attaching too much weight to expert psychological testimony could open the flood gates to otherwise non-approvable waiver applications. In other words, that psychological experts take cases with facts that would otherwise fall short of extreme hardship and apply scientific language to exaggerate the merits. If so, it would be no surprise the AAO frequently cites – as it in fact does – to the “objective” underpinnings of an evaluation, rather than the expert’s “subjective” analysis of a subject. In discussing evaluations it ultimately considered probative, the AAO cites to physical manifestations of the purported mental condition (e.g., weight loss, night sweats), the use of clinical assessment instruments (e.g., Minnesota Multiphasic Personality Inventory, Beck Depression Inventory). Similarly, the AAO has found testimony persuasive that discussed the interplay of demonstrable physical ailments, such as hyperthyroidism, and psychological distress.
A. Source of data
This article studied decisions by the AAO. The AAO is part of Citizenship and Immigration Services (CIS) with appellate authority to review adjudications of benefit applications. Though not specified by statute, the AAO reviews benefit applications under a de novo standard of review. The AAO makes available a portion of its decisions publically available, both online and at its brick and mortar library. These decisions, however, have no legal precedential value unless formally adopted by the Secretary of Homeland Security with concurrence by the Attorney General, an option which is rarely utilized.
For this article we were interested in AAO decisions pertaining to appeals of Form I-601, Waiver of Inadmissibility. From the CIS website we acquired 2,035 decisions issued during calendar years 2006, 2007 and 2008. We limited our initial sample to recent years to minimize the chance our results would be inadvertently influenced by a change in law or policy during the period of time from which the sample was drawn.
Note we were primarily interested in how expert testimony influences the factual/legal conclusion of whether an applicant had met the extreme hardship, rather than on the ultimate appeal outcome. The rationale, of course, is that expert testimony is proffered as a tool to meet the legal hardship standard, not to overcoming factors cutting against an exercise of discretion. Regardless, the distinction is academic since applicants who established extreme hardship on appeal almost always prevailed.
There are important limitations to this study based on our choice to rely on AAO decisions. First, since adjudications are appealed only if unsuccessful at the first tier, appellate decisions fail to capture information about those applications in which immigration practitioners are probably most interested: the ones that are immediately approved! Yet it may be reasonable to presume that the factors tending to predict success on appeal are the same factors tending to predict success before the first adjudicator. A second limitation is that we know only as much about the applications described here as the AAO decision tells us. In this study we did not know what was contained in the applications themselves, only what was reported by the AAO in the course of explaining its holding.
We created data coding procedures to ensure maximum confidence in the reliability and validity of our results. These terms of art are well known to social scientists, but perhaps less so within the legal community. A procedure is reliable to the extent the extent it renders the same results when performed at different times and/or by different individuals; a procedure is valid to the extent it accurately reflects the phenomenon it purports to describe. Thus it was important that our coding methodology would render the same results as between multiple coders in the current study, and as to any future researchers who wished to replicate or extend this research. Coding rules are documented in a comprehensive manual available from the first author.
Cases were coded by the first author along with two research assistants. The first author is an attorney and the two research assistants were second year law students at a large Midwestern university. Research assistants were given a selection of readings to provide them with background on the law governing waivers of inadmissibility along with further training from the first author.
We conducted a preliminary coding of decisions to identify opinion discussing testimony by a mental health professional. Preliminary coding consisted of the following three variables:
|Hardship||The AAO’s finding on whether waiver denial would result in extreme hardship to one or more qualifying relatives|
|Psychological hardship||Whether the decision addresses an assertion by the applicant that waiver denial would result in psychological hardship to any individual|
|Testimony of a mental health professional||Whether the decision states the record includes testimony of a mental health professional|
Coders were provided with randomized samples of decisions from our pool of 2,035 cases and coded randomized blocks of decisions until we had identified an adequate number of cases involving a mental health professional. Ultimately, 657 cases received preliminary coding, which allowed us to identify 91 cases (35.1%) involving testimony by a mental health professional.
C. Inter coder agreement
After achieving sufficient inter coder agreement through practice coding, the three coders began independently coding cases to yield data for the current study. To ensure satisfactory inter coder agreement was maintained, coders overlapped on approximately 10% of the cases. Coders did not know which cases would be used to compute inter coder agreement. Data from those overlapping cases allowed us to measure the degree of consensus between coders for each variable by computing intra-class correlation coefficients using one-way fixed effects models.
Overall, variables were coded with impressive levels of inter coder agreement. As most results reflected perfect concordance between coders, only exceptions are reported. All other variables, save one, were coded with moderate inter-coder agreement. Intra-class correlation coefficients between 0.4 and 0.74 are considered moderate; those .75 and above are considered satisfactory. See Table 1. One variable, which recorded the number of sessions the second qualifying relative had with the expert witness, was coded with poor inter coder agreement.
The 91 cases involving testimony of a mental health professional received coding for the following variables:
|Adjudication details||Location of adjudication; representation by counsel; date of adjudication; statutory basis for application; appeal outcome.|
|Personal descriptors||Citizenship of applicant, immigration status of qualifying relative, gender of applicant and qualifying relative, relation of applicant to qualifying relative.|
|Treatment history||Whether the subject has ever or is currently participating in therapy or taking psychotropic medication.|
|Functional impairments||Whether the subject contends that his mental impairment has ever manifested in a physical symptom.|
|Identity of expert||Whether the expert had an ongoing relationship with the subject; duration of relationship with the subject; the label used to characterize the expert’s role; the label used to characterize the expert’s testimony.|
|Expert’s testimony||Whether the expert posited a formal psychological diagnosis, and which one(s); reliance on formal psychological assessment instruments; whether the expert recognized a relationship between a mental health issue and physical symptom; whether the expert recommended future treatment, and what sort.|
V. Results and discussion
a. General descriptive statistics
Of the 664 decisions we coded, applicants succeeded in establishing extreme hardship in 9.0% of cases. In 56% of all cases the AAO found the applicant failed to demonstrate extreme hardship, and in 35% of cases the AAO did not reach the question, usually due to a procedural irregularity of the case. See Figure 1 (“Did application establish extreme hardship?”). In cases reaching the question of extreme hardship, about one quarter noted the applicant had made some contention regarding psychological hardship.
The remaining statistics reported in this Section derive from a sub-sample of 259 decisions that included expert testimony from a mental health professional. Most applications (67%) sought waivers of inadmissibility due to fraud or misrepresentation. A small portion (28%) concerned criminal grounds of inadmissibility. A small percentage (6.0%) involved multiple grounds of inadmissibility. See Table 2 (“Grounds of Inadmissibility Sought Waived”). Applicant nationality was heavily skewed to Mexican citizenship (33%). In fact, Mexico was the only country represented in more than five percent of the sample.  We found no correlation between the citizenship of an applicant and success on appeal.
Ninety percent of applications were appealed to the AAO from CIS offices within the U.S. The remaining 10% were appealed from State Department offices abroad. We found no relationship between office location and success on appeal.
An overwhelming majority (87%) of applicants were represented by legal counsel on appeal. Notably, this is over twice the portion of noncitizens who secure representation in immigration court proceedings. We found representation by counsel correlated significantly with success on appeal. As encouraging a statement as that may be for advocates, it is important to note this study was not designed to identify other factors that might account for that trend.
We found no support for the proposition that an application is more likely to succeed when a qualifying relative is a U.S. citizen rather than LPR. Commentators – and one outspoken BIA member – have suggested adjudicators give more weight to the hardship of native-born U.S. citizen than to naturalized one, and more weight to the hardship of a U.S. citizens than to an LPR. We coded decisions for whether the qualifying relative was a native citizen, naturalized citizen, citizen not identified as naturalized or native (“citizen not-specified”), or LPR. Native citizens and citizens not-specified both prevailed more frequently (57%; 51%) than LPRs (47%), but LPRs prevailed more frequently than naturalized citizens (32%). See Table 3 (“Immigration status of qualifying relative”). Yet none of the relative rates of success between these groups were found to differ, statistically-speaking. Because of the relatively small group sizes used in these statistical comparisons, these results should be interpreted with caution.
Without controlling for other factors, the number of qualifying relatives did not significantly correlate with appeal outcome. Of 259 cases, most applications (n = 189; 73%) were based on only a single qualifying relative. The remaining applications involved between two and eight qualifying relatives. See Table 4 (“Number of qualifying relatives and appeal outcome”). Observe, in Table 4, however, that as the number of qualifying relatives increased, the number of sustained and denied appeals converged. In other words the data suggest a trend (albeit statistically non-significant) that a larger number of qualifying relatives is more frequently associated with a successful appeal. Moreover, once we controlled for representation by legal counsel and the presence of mental health testimony, the number of qualifying relatives did significantly relate to appeal outcome. These findings lend some support to the intuitive and widely-held belief that an applicant is more likely to succeed with more, rather than fewer, qualifying relatives.
Hypothesis I: Expert testimony predicts success on appeal, but only if there is an ongoing therapeutic relationship with the clinician.
We were surprised to observe – contrary to our hypothesis – the presence of testimony from a mental health professional was in fact associated with favorable appeal outcomes, without regard to whether the professional had an ongoing relationship with the subject.
We found noncitizens were significantly more likely to prevail in appeals that included testimony from a mental health professional than in cases without such testimony. When an appeal included testimony from a mental health professional, applicants prevailed about half of the time; when such testimony was not included, applicants prevailed in only one out of three cases. See Table 5 (“Mental Health Professional Testimony and Appeal Outcome”).
Contrary to our hypothesis, we discovered no significant relationship between appeal outcome and factors showing an ongoing therapeutic relationship with a psychologist. We tested a binary logistic regression model consisting of the following three variables, which were identified as indicators an ongoing relationship between a psychologist and subject: whether the decision expressly stated that the subject was currently receiving therapy; whether the psychologist was characterized as a therapist versus evaluator; and whether the testimony was characterized as a psychological evaluation or letter from a therapist. None of these variables, individually, significantly contributed to the model or shared a significant and unique relationship with appeal outcome. Nor did the overall model (i.e., the ability of all independent variables acting collectively to predict appeal outcome) significantly predict appeal outcome.
Similarly, appeals were no more likely to be successful in cases where the psychologist had opined the subject should enroll in some mode of treatment, whether participating in therapy, taking psychotropic medication, or both. Just nine (18.0%) of the 50 sustained appeals included an evaluation that recommended further treatment, and six (14.3%) of the 42 denied appeals included an evaluation that made such recommendations.
Although we observed no increased success in cases where the testifying expert had a therapeutic relationship with a subject, success on appeal was significantly associated with whether the subject had participated in mental health treatment at some point in the past. Analysis via binary logistic regression revealed that appeals were more likely to be sustained if the qualifying relative had at some point participated in therapy, without regard to whether the individual was currently in therapy with the testifying expert. For coding purposes, therapy was defined broadly to encompass any form of treatment by a mental health professional aimed at addressing a psychological issue. Of the 32 cases in which a subject had at some point participated in therapy, 24 appeals (75.0%) were sustained while 8 (25.0%) were denied. In contrast, of the 59 appeals in which an applicant had never participated in therapy, 25 (42.4%) were sustained, and 34 (57.6%) were denied. The difference in proportion of successful appeals was statistically significant.
We were surprised to observe the statistically significant relationship between the presence of a psychological expert testimony and successful appeal outcome. Practitioners should not be too quick to assume, however, that bolstering a waiver application with a psychological evaluation will necessarily have an impact on the adjudication. Our results may say less about the efficacy of a psychological evaluation than about the types of applicants who include such evidence in their waiver packet. Such individuals, as a class may have access to more resources overall, including better legal counsel. Also, applications with psychological evaluations, as a class, may tend to have other factual elements that favor successful outcomes. Applicants and legal counsel who obtain psychological evaluations may simply do a more thorough job with the application overall. Similarly, insofar as hardship waivers are frequently denied for inadequate evidence, the presence of any additional documentary evidence may tend to correlate with improved outcome.
Our results did not support the proposition that more weight is given to testimony given by professional with an ongoing therapeutic relationship with a subject. Yet our results suggest that an applicant is more likely to prevail when the qualifying relative claiming psychological hardship has been involved in therapy at some point in the past.
Focusing on a qualifying relative’s history of formal mental health care may be a questionable reason to discredit otherwise meritorious mental hardship claims. As a class, relatives of noncitizens may be less likely than the U.S. population as a whole to have access to mental health care. For instance, Hispanics who are U.S. citizens or LPRs are far more likely than the U.S. population as a whole to lack health insurance. Moreover, insofar as qualifying relatives themselves are from minority cultural background, as a class they may be less inclined to seek western-style mental health care. If adjudicators view formal therapy as a litmus test for serious psychological distress, this may systematically undervalue claims by individuals who cannot, or chose not to participate in western medicine.
Finally, although our results suggest the AAO may be influenced by psychological testimony, it appears adjudicators may disagree with the testimony far more frequently than in other legal contexts where expert psychology receives greater deference. As reported above, applicants prevailed on appeal in about half of cases where the application was supported by psychology testimony, but in only one of three cases where such testimony was lacking. Assuming the psychology testimony always offered direct support for a finding of extreme hardship, this suggests the AAO disagreed with the psychologist about half of the time. By contrast, there are much higher rates of agreement between judges and psychology evaluators in criminal law contexts. For example, when assessing a criminal defendant’s psychological competency to stand trial, judges tend to agree with the opinions of evaluators in over 90% of cases. Although there are any number of potentially legitimate reasons for these strikingly different rates of agreement, it is at least not obvious why there should be such a discrepancy.
Hypothesis II: Success on appeal is not related to whether or not a qualifying relative received a psychiatric diagnosis.
We found support for the proposition that a waiver applicant is no more likely to prevail on appeal where an application is supported by a psychological testimony containing a psychiatric diagnosis than where the application is supported by psychological testimony but no psychiatric diagnosis is given. In our data set, appeals were just as likely to result in a successful outcome where an evaluation contained a formal diagnosis as were appeals supported by evaluations that lacked diagnoses.
Of the 43 cases in which a mental health professional’s testimony mentioned a subject’s psychiatric diagnosis, 24 appeals (55.8%) were sustained and 19 (44.2%) appeals were denied. Similarly, of the 48 cases in which a mental health professional’s testimony did not mention a subject’s psychiatric diagnosis, 25 appeals (52.1%) were sustained, and 23 appeals were denied (47.9%). Of the 73 cases that lacked a mental health professional’s testimony, irrespective of diagnostic content, just 23 appeals (31.1%) were sustained while 50 appeals (67.7%) were denied.
Of the 165 cases where psychological hardship was asserted, 50 individuals (30.3%) received at least one psychiatric diagnosis. The majority of individuals who asserted psychological hardship did not receive a diagnosis (n = 115, 69.7%). Of those who did receive a diagnosis, approximately half (52%) were for depression. Table 6 (“Psychiatric Diagnoses and Appeal Outcome”) displays diagnostic frequencies, along with the relationship between diagnosis and appeal outcome. Note that because we report occurrences of diagnoses, rather than the frequency of individuals’ diagnoses, counts sum to more than 50 and percentages sum to more than 100%. That is, when an individual received multiple diagnoses (e.g., Depression and Posttraumatic Stress Disorder), we increased our counts by one for both the “Depression” category, and the “Posttraumatic Stress Disorder” category.
Our study suggests the AAO may not rely on psychiatric diagnoses in weighing the input of a mental health professional. It is unclear, however, why the AAO would tend to devalue diagnoses. Under Matter of Pilch, an adjudicator should take into consideration any credible evidence that a qualifying relative will experience a degree of emotional distress that exceeds the normal consequences of a family member’s deportation. On its face, a psychiatric diagnosis would seem to be per se relevant to this question if the symptoms of the disorder would be triggered by the events flowing from waiver denial. It is worth quoting at length the definition of “mental disorder” as used in the DSM-IV:
In DSM-IV, each of the mental disorders is conceptualized as a clinically significant behavioral or psychological syndrome or pattern that occurs in an individual and that is associated with present distress (e.g., a painful symptom) or disability (i.e., impairment in one or more important areas of functioning) or with a significantly increased risk of suffering death, pain, disability, or an important loss of freedom. In addition, this syndrome or pattern must not be merely an expectable and culturally sanctioned response to a particular event… Whatever its original causes, it must currently be considered a manifestation of a behavioral, psychological, or biological dysfunction in the individual.
A psychiatric diagnosis necessarily communicates that an individual is understood to have a dysfunctional, rather than expectable response to a particular event. Where a psychological expert explains a family member’s deportation would trigger an abnormally dysfunctional response in a qualifying relative (e.g., in the form of a diagnosable mental abnormality), that opinion is directly relevant under Matter of Pilch. That a psychiatric diagnosis is relevant to the determination of extreme hardship certainly does not imply the diagnosis is conclusive on the inquiry. The fact deportation would trigger an unexpected “syndrome or pattern” does not entail those consequences rise to the level of extreme hardship.
We also explored whether an evaluation was more likely to be associated with a successful appeal outcome where the evaluation relied on formal assessment instruments. Those instruments, such as the Minnesota Multiphasic Personality Inventory-2 (MMPI-2), assess personality constructs and psychological functioning. Because they represent an “objective” means of supporting an evaluation, we wondered if the AAO would be more likely to credit an evaluation that incorporated such instruments. Our appraisal, however, was hampered by small numbers. Out of the 50 appeals that yielded successful outcomes, just 5 (10%) included testimony based on formal assessment instruments, while the remaining 45 (90%) did not. Of the 42 appeals that were denied, only three (7.1%) included testimony based on formal assessment instruments, while the remaining 39 (92.9%) did not. These numbers fail to show assessment instruments are associated with favorable appeal outcomes, but no conclusions should be drawn due to the small numbers involved.
We did observe increased rates of success when evaluations mentioned somatization – a physical manifestation of a mental health issue. Twenty-one sustained appeals (42%) included expert testimony that mentioned symptoms of somatization, while just eight (19%) denied appeals included evaluations with this information. This finding supports the contention the AAO assesses the credibility of claimed mental hardship by looking towards relatively objective manifestations of the claimed distress. Although a subject could presumably exaggerate claims of stomach pain, the AAO may be more likely to credit a report that cites such somatization than one that does not.
Our data suggest a psychological evaluation may sometimes be an important component of a strong hardship waiver application. The basic conclusions of our statistical analyses are:
- Applications with expert psychological testimony were more frequently associated with favorable appeal outcomes than applications without psychologist evaluations. The relationship to favorable appeal outcome was no more pronounced where the testimony was by a psychologist with an ongoing therapeutic relationship with the subject. Yet applications were significantly more likely to be sustained where a qualifying relative had some history of participating in mental health treatment, regardless of whether it was with the psychologist who offered testimony.
- Whether or not a psychologist evaluation contained a formal psychiatric diagnosis had no relationship to success on appeal. On the other hand, success on appeal was more likely where an evaluation suggested a physical manifestation of the subject’s mental health condition.
Perhaps the single most important take-home from this article is that advocates should not be discouraged from referring a client for a psychological evaluation merely because the client is not already seeing a clinician, or because there is too little time to for the client to participate in many therapy sessions before submitting the waiver applications. Despite the oft-repeated admonition that an evaluation based on only one or several sessions is of limited evidentiary value, such evaluations are associated with favorable appeal outcomes.
|Variable and Coder Agreement|
|Number of Qualifying Relatives||ICC = .558, p < .001|
|Immigration Status of Subject||ICC = .656, p < .001|
|Was Subject #2 ever in therapy||ICC = .610, p < .001|
|Did Subject #2 experience somatization?||ICC = .483, p = .006|
|Did Subject #2 receive a diagnosis?||ICC = .531, p < .001|
|Did the expert opine about Subject #2’s somatization?||ICC = .610, p < .001|
|Relationship of Subject #2 to applicant||ICC = .501, p = .005|
|Was expert testimony relating to Subject #2 characterized as an evaluation or letter?||ICC = .868, p < .001|
|Number of therapy sessions with Subject #2||ICC = .368, p = .033|
|Was subject #2 recommended to get treatment||ICC = .657, p < .001|
|Was Subject #2 recommended medication?||ICC = .632, p < .001|
|Ground of Inadmissibility Sought Waived|
|Unlawful presence & misrepresentation||7||3%|
|Criminal & misrepresentation||7||3%|
|Immigration Status of Qualifying Relative|
|Sustained : Denied||Percent sustained|
|Lawful Permanent Residents (n = 30)||14 : 16||46.67%|
|Native citizens (n = 14)||8 : 6||57.14%|
(n = 31)
|10 : 21||32.26%|
|Citizens, status not specified (n = 81)||41 : 40||50.62%|
|Number of Qualifying Relatives and Appeal Outcome|
|No. of QRs||1||2||3||4||5||6||8|
|Mental Health Professional Testimony and Appeal Outcome|
|Psychiatric Diagnosis and Appeal Outcome|
|Posttraumatic Stress Disorder (PTSD)||6||12.0||4||66.7||2||33.3|
|Adjustment Disorder, with Mixed Anxiety and Depressed Mood||6||12.0||5||83.3||1||16.7|
|ADD or ADHD||4||8.0||4||100||0||0|
|Adjustment Disorder, with Depressed Mood||2||4.0||1||50.0||1||50.0|
|Adjustment Disorder, with Anxiety||1||2.0||0||0||1||100|
|Autism Spectrum Disorder||1||2.0||1||100||0||0|
|Dependent Personality Disorder||1||2.0||1||100||0||0|
 For an overview of application procedures see the recently added section to Charles Gordon et al., Immigration Law and Procedure, § 63.12 (2004), researched by the first author of this article. For introductory advice on compiling successful applications see Laurel Scott, An Overview of I-601 Waivers and Extreme Hardship (Jan. 11, 2010), in State Bar of Texas, 2010 Advanced Immigration Law Course (Feb. 25, 2010), available at http://www.scottimmigration.net/I601Memo.pdf (last visited Sep. 4, 2010) [hereinafter Scott (2010)]. See also Laurel Scott, Evidence for an Extreme Hardship Waiver of Inadmissibility: Boldly Going Where no Case Law has Gone Before, 09-1 Immigr. Briefings 1 (2009) [hereinafter Scott (2009)]; Gerald Seipp, Waivers of Inadmissibility – From Basic Principles to Advanced Practice Considerations, Part I, 03-09 Immigr. Briefings 1 (Aug. 2003) [hereinafter Seipp (Part I)]; Gerald P. Seipp, Waivers of Inadmissibility – From Basic Principles to Advanced Practice Considerations Part II, 9 Immigr. Briefings 1 (Sep. 2003) [hereinafter Seipp (Part II)]; Bruce A. Hake, The Hake Hardship Scale: A Quantitative System for Assessment of Hardship in Immigration Cases Based on a Statistical Analysis of AAO Decisions, 10 Bender’s Immigr. Bulletin 403 (Mar. 1, 2005) [hereinafter Hake (2005)].
 Decisions available from LexisNexis may be accessed through the Board of Immigration Appeals (BIA) & AAU Non-Precedent Decisions database. The decisions are best located searching by alien number, but note the numbers are only sometimes delineated with dashes (e.g., Axx-xxxx-xx). For decisions without alien numbers search by full date (e.g., January 30, 2009).
 For decisions available on the AAO’s website, go to http://www.uscis.gov/portal/site/uscis, click on “Laws” on the header, then click on “Administrative Decisions” on the left-hand bar. Decisions discussed in this article may be found in the “H2” folder on that page. Because decisions are listed by date, and multiple decisions are frequently issued in a single day, citations in this article include the file name in parentheses. See, e.g., Matter of [name redacted] (AAO Oct. 27, 2006) (Kansas City) (unpublished decision), available at www.uscis.gov (Oct272006_02H2212.pdf) (last visited Oct. 4, 2010).
 INA § 212(h)(1)(B), 8 U.S.C. § 1182(h)(1)(B) (certain criminal grounds); INA § 212(i), 8 U.S.C. § 1182(i) (immigration fraud and misrepresentation); INA § 212(a)(9)(B)(v), 8 U.S.C. § 1182(a)(9)(B)(v) (unlawful presence).
 The term “qualifying relative” is used ubiquitously in case law and commentary, but does not appear in the immigration statutes. See Immigration and Nationality Act of 1952, Pub. L. No. 82–414, § 101(a)(15), 66 Stat. 163 (codified as amended at 8 U.S.C. §§ 1101(a)(15)) (hereinafter INA).
 See Seipp (Part II), supra note 5, text accompanying n. 27 (“In the real world of decision-making the analytical concepts, involving existence of extreme hardship versus the ultimate issue of whether a favorable exercise of discretion is warranted, are often ignored”).
 But see 8 C.F.R. § 1240.58(b) (setting forth hardship standards for suspension of deportation commenced prior to April 1, 1997).
 See, e.g., Matter of Monreal-Aguinaga, 23 I. & N. Dec. 56, 59 (BIA 2001); Matter of Cervantes-Gonzalez, 22 I. & N. Dec. 560, 565 (BIA 1999).
 Illegal Immigration Reform and Immigrant Responsibility Act of 1996, Pub. L. 104-208, 110 Stat. 3009-546.
 INA § 244(b), 8 U.S.C. § 1254(b) (1995).
 INA § 240A(b)(1), 8 U.S.C. § 1229b(b)(1).
 See, e.g., Matter of Cervantes-Gonzalez, 22 I. & N. Dec. at 565 (cautioning against the “cross application” of different types of relief, but stating case law addressing extreme hardship in different statutory contexts would be “helpful”).
 Compare USCIS, Immigrant Waivers; Procedures for Adjudication of Form I-601 for Overseas Adjudication Officers, at 50 (Apr. 28, 2009), available at www.uscis.gov (last visited Oct. 18, 2010) (“The “extreme hardship” standard is always the same, whether it is used for unlawful presence purposes or whether it is applied to an applicant seeking the waiver of a criminal ground”), with Seipp (Part II), supra note 5, below text accompanying n. 78 (noting informal reports that waivers of unlawful presence are held to a less stringent standard than for criminal grounds and fraud).
 Bruce A. Hake, Hardship Standards, 7 Bender’s Immigr. Bull. 59 (Jan. 15, 2002) [hereinafter Hake (2002)]; Hake (2005), supra note 5; Bruce A. Hake, Hardship Waivers for J-1 Physicians, 94-2 Immigr. Briefings 1, text accompanying n. 99-155 (Feb. 1994).
 Cf. Hake (2002), supra note 19 (arguing hardship standards are “essentially identical with the one exception of the ‘exceptional and extremely unusual’ standard for suspension/removal”).
 See, e.g., Matter of [name redacted], at 3 (AAO Feb. 06, 2007) (Nebraska Service Center) (unpublished decision) (copy on file with appellant’s attorney, Bruce Hake and with the first author of this article) (“The AAO notes that suspension of deportation cases require a finding of extreme hardship, whereas I-612 waivers require the lesser standard of exceptional hardship. Therefore, [suspension of deportation cases] are not precedent…”). The authors thank Bruce Hake for drawing our attention to this case.
 See citations supra note 7.
 See Scott, supra note 5, at text below n. 4 (“Submitting only evidence to prove the extreme hardship if the qualifying relative moves abroad is going halfway and one should expect such a waiver to be denied”).
 See, e.g., Salcido-Salcido v. INS, 138 F.3d 1292, 1293 (9th Cir. 1998) (BIA abused discretion in denying suspension of deportation, where testimony given as offer of proof showed applicant would be separated from her child).
 While we refer to the legally sufficient “magnitude” of hardship, Bruce Hake has argued that sufficient hardship may be either exceptional, “unusual in terms of probability of occurrence,” or extreme, “unusual in terms of gravity of harm.” Hake (2002), supra note 19, at text accompanying n. 40. For purposes of psychological hardship as discussed in this article, those two concepts may converge. As Hake suggests, exceptional hardships may be those having an idiosyncratically severe impact on a particular individual, due to his/her vulnerabilities. See id., at text below n. 4 (analogizing tort law’s eggshell plaintiff). From a psychological standpoint, the inquiry necessarily turns on hardship as experienced by the subject.
 Matter of Pilch, 21 I. & N. Dec. 627, 630 (BIA 1996) (citing Matter of Anderson, 16 I. & N. Dec. 596 (BIA 1978)).
 See, e.g., Matter of Cerventes-Gonzales, 22 I. & N. Dec. (BIA 1999) (holding § 212(i) waiver applicant failed to hardship “over and above the normal economic and social disruptions involved in the deportation of a family member”); Matter of O-J-O-, 21 I. & N. Dec. 381, 387 (BIA 1996) (holding respondent met requirements for suspension of deportation because the case presented hardships “over and above the normal economic and social disruptions involved in deportation”)
 Restatement (Second) of Torts, § 283 cmt. b (1965).
 See Matter of O-J-O-, 21 I. & N. Dec. at 385 (quoting Santana-Figueroa v. INS, 644 F.2d 1354, 1357 (9th Cir. 1981)).
 Ravancho v. INS, 658 F.2d 169, 176 (3d Cir. 1981) (psychological report must be considered in connection with all other relevant factors); Batoon v. INS, 707 F.2d 399, 402 (9th Cir. 1983) (BIA erred in denying motion to reopen application for suspension of deportation; Board ignored psychiatrist’s testimony opinion deportation would result in long-term psychiatric hospitalization), after remand, Gonzalez-Batoon v. INS, 767 F.2d 1302 (9th Cir. 1985), rehearing en banc, 791 F.2d 681 (9th Cir. 1986).
 Matter of Duncan, A79-643-536, Lexis: BIA and AAU Non-Precedent Decisions (AAO May 16, 2007) (San Francisco) (unpublished decision).
 Watkins v. INS, 63 F.3d 844, 849 (9th Cir. 1995) (holding BIA erred in failing to reopen suspension of deportation application). See also Matter of Maria Isabel, A79-559-641, Lexis: BIA and AAU Non-Precedent Decisions (BIA June 30, 2006) (unpublished decision) (granting cancellation of removal; child had ADHD requiring medication and ongoing psychiatric evaluation).
 See Perez v. INS, 96 F.3d 390, 392 (9th Cir. 1996) (denying suspension of deportation; report by school psychologist failed to explain how child’s “adjustment difficulties” would exceed the norm).
 Cf. Jennifer Lidsley, Comment, All Relevant Factors: Gender in the Analysis of Exceptional and Extremely Unusual Hardship, 19 Wis. Women’s L.J. 337, 348-53 (2004) (arguing the BIA more readily presumes children will relocate with a mother than with a father).
 In our sample, of all cases where any issue relating to psychological hardship was discussed, 95% of those cases expressly discussed emotional hardship in particular.
 21 I. & N. Dec. 627 (BIA 1996).
 Waiver appeals were sustained in each of the following cases: Matter of [name redacted] (AAO June 18, 2007) (Miami) (unpublished decision), available at www.uscis.gov (Jun182007_01H2212.pdf) (last visited Oct. 5, 2010) (relative had threatened suicide if applicant were removed); Matter of [name redacted] (AAO May 11, 2007) (Los Angeles) (unpublished decision), available at www.uscis.gov (May112007_01H2212.pdf) (last visited Oct. 5, 2010) (subject had extensive history of drug abuse, arrests and hospitalization, and psychoanalyst opined loss of applicant’s support could cause “life threatening” relapse); Matter of [name redacted] (AAO May 23, 2009) (New Delhi) (unpublished decision), available at www.uscis.gov (May232008_01H2212.pdf) (last visited Oct. 5, 2010) (spouse’s depression had worsened to suicidal in absence of applicant); Matter of [name redacted] (AAO Dec. 21, 2007) (Los Angeles) (unpublished decision), available at www.uscis.gov (Dec212007_01H2212.pdf) (last visited Oct. 5, 2010) (therapist of seven years testified stress of applicant’s removal could cause client to contemplate suicide); Matter of [name redacted] (AAO June 7, 2006) (Los Angeles) (unpublished decision), available at www.uscis.gov (Jun072006_02H2212.pdf) (last visited Oct. 5, 2010) (it was “very possible” applicant’s removal would lead to full-fledged clinical depression and “suicidal ideation,” which symptoms applicant’s presence was needed to combat).
 Matter of [name redacted] (AAO Oct. 27, 2006) (Kansas City) (unpublished decision), available at www.uscis.gov (Oct272006_02H2212.pdf) (last visited Oct. 5, 2010).
 Matter of [name redacted] (AAO Oct. 27, 2006) (Chicago) (unpublished decision), available at www.uscis.gov (Oct272006_01H2212.pdf) (last visited Oct. 5, 2010).
 Matter of [name redacted] (AAO Oct. 20, 2006) (San Francisco) (unpublished decision), available at www.uscis.gov (Oct202006_01H2212.pdf) (last visited Oct. 4, 2010).
 Matter of [name redacted] (AAO May 16, 2006) (Los Angeles) (unpublished decision), available at www.uscis.gov (MAY162006_01H2212.pdf) (last visited Oct. 4, 2010).
 Matter of [name redacted] (AAO Sep. 24, 2007) (San Francisco) (unpublished decision), available at www.uscis.gov (Sep242007_07H2212.pdf) (last visited Oct. 4, 2010).
 See, e.g., Scott (2009), supra note 5 (noting BIA decisions give “[v]ery little guidance on evidence” for waiver applications).
 In order to preserve the special import of the term “evaluation,” we use “testimony” to refer generically to the written statement offered by a mental health professional in support of an immigration application. We do not mean to incorporate the special legal meanings of testimony as it is used in the INA. See, e.g., INA § 101(f)(6), 8 U.S.C. § 1101(f)(6) (false testimony is a per se bar to establishing good moral character).
 See, e.g., Matter of [name redacted], Lexis: BIA and AAU Non-Precedent Decisions (AAO Jan. 9, 2008) (Athens) (unpublished decision) (noting “the psychological evaluation of the applicant’s spouse’s emotional state is based on a single interview and would ordinarily be viewed as having limited evidentiary value,” but evaluation was supported by adequate documentary evidence to merit consideration).
 Matter of [name not provided] (AAO Oct. 27, 2006) (Kansas City) (unpublished decision), available at www.uscis.gov (Oct272006_02H2212.pdf) (last visited Oct. 4, 2010) (sustaining appeal, and citing to qualifying relative’s prior use of antidepressant medications).
 See, e.g., Matter of [name not provided], 2009 Immig. Rptr. LEXIS 86, at *8-9 (AAO May 1, 2009) (unpublished decision) (granting waiver under INA § 212(a)(9)(B)(i)(II); psychologist’s report diagnosed depression, along with erratic sleep and loss of 35 pounds, and found separation was “tak[ing] a toll”).
 Matter of [name redacted] (AAO May 24, 2006) (Cleveland) (unpublished decision), available at www.uscis.gov (May242006_02H2212.pdf) (last visited Oct. 4, 2010).
 Matter of [name redacted] (AAO June 8, 2006) (Los Angeles) (unpublished decision), available at www.uscis.gov (Jun082006_05H2212.pdf) (last visited Oct. 4, 2010) (evaluator met with subject only twice but assessment instrument were consistent with severe depression and paranoia).
 Matter of [name redacted] (AAO Mar. 13, 2008) (San Francisco) (unpublished decision), available at www.uscis.gov (Mar132008_01H2212.pdf) (last visited Oct. 4, 2010) (evaluation based on single interview was nonetheless credible because of its reliance on clinical testing instruments).
 See, e.g., Matter of [name redacted] (AAO Sep. 11, 2008) (California Service Center) (unpublished decision), available at www.uscis.gov (Sep222008_03H2212.pdf) (last visited Oct. 4, 2010) (sustaining appeal where evaluator opined hyperthyroidism could exacerbate depression).
 See Dor v. INS, 891 F.2d 997, 1002 n. 9 (2d Cir. 1989); Spencer Enterprises Inc. v. U.S., 229 F.Supp.2d 1025, 1043 (E.D. Cal. 2001), aff’d 345 F.3d 683 (9th Cir. 2003).
 8 C.F.R. § 103.9(b).
 8 C.F.R. § 103.3(c). No codified standards exist for when an AAO decision is formally adopted, though the Office has said that decisions are adopted when they present a “novel issue of law or fact” or where the applicable regulations would benefit from clarification. Memorandum from Prakash Khatri, CIS Ombudsman, to Robert Divine, Acting Deputy Director, CIS, Recommendation from the CIS Ombudsman to the Director, USCIS (Dec. 6 2005), available at http://www.dhs.gov/xlibrary/assets/CISOmbudsman_RR_20_Administrative_Appeals_12-07-05.pdf (last visited Sep. 4, 2010) [hereinafter Khatri Memo].
 See Khatri Memo, supra note 52, 3 (noting no decisions were adopted in the period between August 1998 and December 2005). Since 2005 CIS has also designated a small number of AAO decisions as “USCIS Adopted Decisions,” to be relied upon as policy guidance within CIS. Id.
 Of 102 cases in which the applicant demonstrated “extreme hardship,” only one was denied, all others were sustained.
 Cf. Lee Epstein and Gary King, Exchange: Empirical Research and the Goals of Legal Scholarship, 69 U. Chi. L. Rev. 1, 83-97 (2002) (discussing and distinguishing reliability and validity).
 Scott (2009), supra note 5; Seipp (2003 – Part I), supra note 5; Seipp (2003 – Part II), supra note 5; and Hake (2005), supra note 5.
 The following procedure was used to draw a randomized sample from our initial pool of decisions. Each appeals decision was initially downloaded as an Adobe .pdf file and renamed in chronological order with numerical file names from 1-2,035. Using a Microsoft Excel spreadsheet, we assigned a random numerical value (0 to 1) to each digit between 1 and 2,035. Decisions were then coded based on the random numerical value (i.e., beginning with 0.000160642, which corresponded to file name 635).
 What constituted an “adequate number” of cases was determined via a priori statistical power analysis. Power analysis guides decisions about what sample size will likely be necessary to enable accurate and reliable statistical tests, as well as the likelihood that pre-determined statistical methods will detect relevant patterns within the data.
 Cf. Joseph L. Fleiss, Design and Analysis of Clinical Experiments (John Wiley & Sons 1986).
 Assertion of psychological hardship (n = 106; 25%); no assertion of psychological hardship (n = 321; 75%).
 For administrative reasons, we examined only the 14 countries whose citizens are most frequently placed in removal. See Executive Office of Immigration, FY 2009 Statistical Year Book, at p. E2, Table 5 (Mar. 2010), available at http://www.justice.gov/eoir/statspub/fy09syb.pdf (last visited Oct. 4, 2010) [hereinafter FY 2009 Statistical Year Book]. Such coding captured 57% of cases (n = 148). Other countries were coded by continent only.
 X2 = 14.835; p = .786; F(20, 234) = .723, p = .801, ES = .24. The same was true when countries were grouped by continent. X2 = 3.469, p = .748.
 This was true both when comparing domestic offices to foreign offices grouped by continent (X2 = 6.195, p = .288), and when comparing all domestic offices to all foreign offices (X2 = 3.135, p = .077).
 See FY 2009 Statistical Yearbook, supra note 63, at p. G1, Fig. 9 (reporting that 39% of respondents were represented in immigration court proceedings in fiscal year 2009).
 r = .212, p < .01.
 For instance, attorneys may undertake appellate representation in cases that would be more likely to prevail regardless of representation.
 Hake (2005), supra note 5, below text accompanying n. 15; Matter of Cervantes-Gonzales, 22 I. & N. Dec. 560, 586 (BIA 1999) (Rosenberg, Board Member, dissenting) (arguing the majority discounted the importance of hardship to the respondent’s naturalized U.S. citizen wife).
 When decisions involved more than one qualifying relative, we coded information for each individual. Our discussion in this paragraph, however, focused on only one qualifying relative for each case (the first relative discussed individually in the AAO decision).
 The number of qualifying relatives did not significantly correlate with appeal outcome. r = -.114, p = .069, n = 358.
 r= -.153, p = .017.
 X2 = 8.208, p < .01; r = .224, p < .001. When looking at this relationship, we excluded cases in which the applicant made no assertion whatsoever concerning psychological hardship to a qualifying relative, since our research question pertains to whether mental health professional testimony bolsters assertions of mental hardship.
 A binary logistic regression is a statistical tool that tests whether a group of independent variables collectively are associated with a dichotomous outcome (e.g., appeal approved/denied), or dependent variable. The independent variables measure the degree of statistical relationship between the group of independent variables and each value of the outcome variable. Moreover, binary logistic regressions identify how each independent variable uniquely contributes to the overall predictive model.
 χ2 = 8.235, df = 4, Nagelkerke’s R2 = .148, p = .083.
 χ2 = 072, p = .511.
 This binary logistic regression included the same three variables as the previous model, along with a fourth variable representing whether an applicant had ever participated in therapy. This time, the overall model significantly predicted appeal outcome, χ2 = 12.938, df = 5, Nagelkerke’s R2 = .225, p = .024. Although no variables contributed significantly to the overall model, the only variable that approached significance (β = 2.681, p = .067) was the added variable, which represented whether the applicant had ever participated in therapy.
 The remaining cases lacked sufficient information to code one way or the other.
 χ2 = 8.887, p = .004.
 Gretchen Livingston, Hispanics, Health Insurance and Health Care Access, at 2 (Pew Research Center, 2008), available at http://pewhispanic.org/files/reports/113.pdf (last visited Sep. 4, 2010).
 See DHHS, U.S. Public Health Services, Mental Health: Culture, Race, and Ethnicity—A Supplement to Mental Health: A Report of the Surgeon General, at 42 (2001), available at http://mentalhealth.samhsa.gov/cre/default.asp (last visited Sep. 4, 2010) (concluding that cultures of racial and ethnic minorities relates to the types of mental health services they use, and that cultural misunderstandings with provides may be a barrier to appropriate care).
 See supra, text accompanying note 78.
 We did not code for whether the psychology testimony stated an opinion on the ultimate legal issue of whether the subject would experience extreme hardship. Nonetheless, it is a fair inference that an applicant would include psychology testimony in a packet only if it either expressly endorsed a finding of extreme hardship or implicitly endorsed that finding.
 Due process requires a criminal defendant have the present ability to understand the nature of the proceedings against him and assist in his own defense. See Dusky v. United States, 362 U.S. 402 (1960) (per curiam).
 Melissa L. Cox & Patricia A. Zapf, An Investigation of Discrepancies Between Mental Health Professionals and the Courts in Decisions About Competency, 28 Law & Psychol. Rev. 109 (2004) (finding agreement rate of 99.6%); Jennifer L. Skeem at al., Logic and Reliability of Evaluations of Competence to Stand Trial, 22 Law & Hum. Behav. 519, 520 (1998) (citing agreement rate of over 90%). See also Ivan Kruth, Juvenile Competence to Stand Trial: A historical and Empirical Analysis of a Juvenile Forensic Evaluation Service, 5 Int’l J. Forensic Mental Health 109, 117 (2006), available at http://www.iafmhs.org/files/v52Kruh.pdf (last visited Sep. 16, 2010) (finding 95% agreement in assessing juvenile competence to proceed).
 For instance, we are comparing apples and oranges to an extent. The rate of agreement we observe on appeal may be lower (or potentially higher) than the rate of agreement at the first-tier adjudication.
 X2 = .127, p = .442.
 American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders: DSM-IV-TR, xxxi (4th ed. text rev. 2000) (emphasis added).
 X2 = .264, p = .447.
 X2 = 5.905, p = .015.
 One application was based on unlawful presence, a handwritten designation on its cover page shows it was erroneously filed in the folder from which our sample was drawn. Another application sought waiver of both a criminal ground of inadmissibility and inadmissibility due to HIV-positive status.
 Attention Deficit Disorder or Attention Deficit Hyperactivity Disorder.