[Excerpt (notes omitted) from Bruce A. Hake and David L. Banks, “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 Immigration Bulletin 403-420 (March 1, 2005). Copyright © 2005 by Bruce A. Hake and David L. Banks. All rights reserved. Bruce A. Hake is a lawyer in Maryland. His practice concentrates on J-1 exceptional hardship waiver cases. See www.hake.com/pc. David L. Banks is Professor of Statistics, Duke University. First-person references in the text are to Mr. Hake.]


[First 11 pages omitted.]

Genesis of the Idea

I first started trying to invent a quantitative tool for the assessment of hardship in immigration cases in 1992, 12 years ago. The idea has finally matured.

I never accept a hardship waiver case unless it meets certain criteria: (1) I personally believe it involves serious hardship; (2) there is a very good chance that it is approvable under established law and practice; (3) I am free to take the case at the time; and (4) there is no special reason to decline, such as a conflict of interest. Applying these criteria, I turn down the majority of people who want to hire me.

Of those criteria, the most important is the estimation of approvability. For years my rule of thumb was not to accept a case unless I felt it had at least a 75 percent chance of success. Because I have so much experience in this narrow area, my own gut prediction about success is probably about as accurate as could be found anywhere. But it always bothered me that I did not have a stronger empirical basis.

Before I completed this article, in thinking generally over my work and about the thickets of reasoning in the case law, I had thought that a comprehensive listing of relevant hardship factors would need to be very complex. In April 2002, however, I had an inspiration one night and discovered to my delight that the opposite is true. In fact, one can put the entire structure of pertinent hardship waiver factors on a single page.

Original Version

The original 2002 version of the Hake Hardship Scale was based on a systematic analysis of the major hardship factors in my last 50 successful Form I-612 hardship cases. I discovered through this empirical analysis that every single important hardship factor fell within just six hardship categories, with no loose ends. Altogether, one needs just 10 categories for a complete and practical analysis of hardship waiver cases:

A. Three categories for persons involved in the case–

1. U.S. citizen spouse or child?

2. LPR spouse or child?

3. Third persons facing very serious hardships?

B. One category for the public interest–

4. Significant public interest factors?

C. And just six categories for specific hardship grounds–

5. Medical hardships to spouse or child?

6. Psychological hardships to spouse or child?

7. Career or educational disruptions to spouse or child?

8. Very serious financial hardships?

9. Sociocultural hardships upon relocation to the home country?

10. Significant risk of physical harm due to political violence?

Usefulness of a Limited Number of Categories

It can take great effort to establish the gravity or probability of different kinds of harms. For instance, it may take many documents to prove that a family faces a risk of physical harm from political violence that is so serious that it must be given weight. Moreover, there may be much overlap between related categories. For instance, certain extreme sociocultural factors (such as the ongoing genocide being inflicted on the Shia religious minority in Pakistan) may cause (1) a significant risk of physical harm due to political violence, (2) serious psychological hardships, (3) serious medical hardships, and (4) profound career disruption, and they may defeat J-1 program and policy goals by making futile any effort by the applicant to employ his U.S. training in the home country. Nonetheless, it is useful to realize that the fact development can be channeled into such a small number of main categories in every single case.

The Hardship-Minimizing Travel Alternative

A major reason why people lose at J-1 hardship waiver applications if they try on their own or with an inexperienced lawyer is due to ignorance of the USCIS “two-step” analysis, which is the central dogma of J-1 hardship waiver law. One has to prove that the U.S. qualifying relatives would face hardships in the home country if the entire family were to relocate together, but usually this is easy and in any event it is not sufficient. The core of the case is to show that the U.S. qualifying relatives would face comparably exceptional hardships if the family were to adopt the travel alternative that minimizes hardship to the qualifying relatives, which typically involves several of the family members staying in the United States. (If the applicant’s spouse is not a U.S. citizen or LPR, then the analysis is somewhat simplified, but one still needs to prove that the children could not stay in the United States for two years.)

Therefore, in using the Hake Hardship Scale to assess the approvability of a case, the primary focus is on whether or not the situation scores a sufficient number of points under the hardship-minimizing travel alternative. In many cases (including most of the ones involving a U.S. citizen spouse), the hardships would be even more serious if the entire family relocated for two years or more to the applicant’s home country. But the key is the hardship-minimizing travel alternative.

In practice, it is often difficult to maintain clarity about distinctions under the “two-step” analysis, especially where there are more than two hypothetical travel alternatives. I make a point to discover the most likely alternative that the family really would follow if forced to choose (one has to find out; real answers are all over the map), and I tend to emphasize this reality, while giving less attention to the merely hypothetical alternatives.

Scoring the Various Hardships

Once one has identified the major hardship factors in a case, one needs a way to score them to make an assessment as to whether the case is likely to be approved. The scoring perhaps may give a modicum of credit to the lawyer’s belief about what the law “should” be, but to be useful in practice it should be based almost entirely on an objective and accurate reflection of the government’s action in real cases.

It is crucial to emphasize that one must be extremely skeptical and conservative in assigning point totals for categories that permit a range. Only the most clearly serious facts justify the higher numbers, and only when those facts can be supported by authoritative evidence. For instance, the mere fact that one can articulate some kind of “medical hardship” does not necessarily get you even one point.

Details of the Scoring System

After much thought, one night of inspiration when years of fuzzy thinking seemed to snap into clarity, and then two years of working with the scale, I propose that the factors should be weighted (scored) as follows:

1. U.S. citizen spouse or child? Five points for a U.S. citizen spouse and/or five points for a U.S. citizen child. One point for each additional U.S. citizen child. If I were the adjudicator, I would give five points for each U.S. citizen child, but in practice the government does not decide like that. If that were the actual rule of decision, a hardship finding would be made by the USCIS in every case involving two or three U.S. citizen children, and that is plainly not the reality (notwithstanding the authority cited above regarding the dependence of the probability of approval on the number of citizens involved). If a spouse or child obtained U.S. citizenship through naturalization, subtract one-half point. Under the law, all U.S. citizens are equally deserving of protection from their government. In practice, however, the government gives somewhat less weight to the suffering of naturalized citizens. In AAO decisions, for example, during terse summaries of the material facts, the AAO nearly always goes out of its way to mention that a spouse was naturalized where that is the case.

2. LPR spouse or child? Four points for an LPR spouse or child. One point for each additional LPR child. The scoring here is based on the bedrock principle of American immigration law, which has been consistently affirmed by the Supreme Court in many cases for over 100 years, that aliens’s rights increase over time as their ties to the community increase. The fundamental American legal principle is equality before the law and morally all persons are equal, so I’m uncomfortable to give less weight to the suffering of a green card holder than to a citizen. Nonetheless, in practice the government gives less weight. Indeed, as noted above, it sometimes seems to give less weight to naturalized citizens, and such discrimination sometimes appears to reflect ethnic biases.

3. Third persons facing very serious hardships? One to five points (per person). These situations are unusual and very fact-specific. In the great majority of cases one could not assign any points in this category. Even in cases involving significant suffering to third persons, such as extended family members, one usually could not accurately assign more than one or maybe two points. Nonetheless, in a few of my victories the only significant hardship has been extremely serious hardship to a third person, such as a grandparent of a qualifying relative dying of cancer. Such rare cases may merit four or five points.

4. Significant public interest factors? One to three points. The statute requires that all J-1 waiver approvals must be grounded on a finding by the Attorney General that it is in the public interest. All cases involve some degree of public interest in view of the ties of the applicant’s family to the community. One or rarely two or three points should be assigned here in unusual cases where there is some special, strong public interest factor. My favorite example is a client who was asked to serve on a special project to develop an anthrax vaccine during the time of the anthrax terrorist attacks in 2001. My preference would be to permit higher scores in this category in certain cases, but my impression is that the government typically will not do so.

5. Medical hardships to spouse or child? One to six points (per person as appropriate). This is the big enchilada. One has to be very skeptical and honest in assessing the evidence. An assignment of five or six points requires a definite life-and-death risk. If there are several qualifying relatives with medical hardships, one adds up the points for each. Note that the State Department’s Waiver Review Division routinely sends claims of medical hardship to a separate bureau for an opinion on whether the medical condition can be treated in the home country, so it is crucial to provide documentation from credible medical authorities in the home country.

6. Psychological hardships to spouse or child? One to five points (per person as appropriate). Again, one has to be very skeptical and honest in assessing the evidence. An assignment of four or five points requires an extremely serious risk of catastrophic mental breakdown or suffering that would be unconscionable to inflict. In practice, it is difficult to prove to the satisfaction of the government that a psychological hardship is exceptional. This topic alone could support an entire article. In brief, I rarely use psychiatric letters unless there is a pre-existing, substantial history of clinical psychiatric illness. Exceptional cases may require a report from a forensic psychologist. In the rare cases where there is no apparent outward hardship, but there is in fact very serious and unusual inward hardship (based, for instance, on past trauma such as torture in Bosnia or suicide of a brother), I have had success relying on legal authorities insisting that the government must look at the individual’s actual circumstances, with analogies to the “thin-skinned plaintiff” rule in tort law (a tortfeasor is ordinarily liable for all the plaintiff’s injuries, even if most persons would not have suffered injury from the same act). Reports from treating mental health professionals are often of little use in proving psychological hardship, but they are useful to prove the fact of treatment. In practice, the USCIS and the State Department are often more reluctant to tear asunder the bond of an existing, prior relationship with a mental health professional than with a spouse or child.

7. Career or educational disruptions to the spouse (or, in theory, child)? One to two points. This factor has strong support in the case law. There must be real proof of disruption.

8. Very serious financial hardships? One point. Only rare cases get even one point for this usually disparaged factor. In the 50 cases I analyzed for the first version of this article, only 13 identified this as a major hardship. My rule of thumb is whether there is a real risk that children may go without essential needs or that a mortgage could not be paid.

9. Sociocultural hardships upon relocation to the home country? One point. This factor includes things like mistreatment of women in Muslim societies, language problems, educational deficiencies, and the like. There is quite a bit of discussion of factors like this in the case law, and some lawyers give great emphasis to this category. I personally assigned just one point in this category in only seven out of the 50 cases I analyzed for the first version of this article. Some lawyers will differ with me, but I just do not think this category is compelling or effective. When “sociocultural” hardships are sufficiently extreme to be counted on the hardship scale, I think it is usually better to treat them as psychological or occasionally medical hardships, or in terms of the risk of physical harm due to political violence.

10. Significant risk of physical harm due to political or sectarian violence? One to three points. No matter how white-hot the danger, such risk is always inherently attenuated. If the applicant has been specifically singled out for harm, a better option may be to file on Form I-612 for a waiver based on risk of persecution. My preference would be to sometimes permit up to five points in this category. In practice, the government usually does not give that much weight. I assigned the full three points in this category to only 10 of the 50 cases I analyzed for the first version of this article. Proof of the danger in this category can require extensive documentation, organized into numerous subcategories. In Pakistan, for example, an applicant’s family may face significant risks (1) due to the danger of kidnapping, (2) due to their religious affiliation, (3) due to their American ties, (4) due to past political affiliations, and so forth. Nonetheless, this all falls under one core category, where the key concept is risk of physical harm due to political (or sectarian) violence.

11. Adverse factors. U.S. immigration law has many kinds of applications for relief where the government performs a balancing process, weighing positive factors (“equities”) against adverse factors. In my latest work on the Hake Hardship Scale, I have started to use an additional column to record adverse factors, which cause a reduction in the total points scored for a case. As noted above, I deduct one-half point for a naturalized spouse. I deduct one point for each specific problem, of the kinds likely to be articulated by the AAO as negative factors. Examples include absence of documentary evidence for specific points, recency of marriage of a J-1 exchange visitor to an American, and so forth. In addition, I deduct FIVE points if the J-1’s program was supported by U.S. government funding.

Enough to Win

What does one do with those scores? It might appear that an exceptional hardship finding should require just 10 points, but in practice one needs 11 or more. A score significantly above 11 should be approved quickly and smoothly. A case scoring less than 10 points is not even in the ballpark and should not be accepted by a lawyer.

In my view, a case involving a U.S. citizen child and a U.S. citizen spouse (10 points), and nothing more, should always be enough. Clearly the government does not agree. One also needs at least one substantial articulable hardship in one of the six hardship categories. Therefore, one needs at least 11 points.

My final hypothesis is that a winnable case requires (1) at least 11 points, plus (2) at least one clearly exceptional and provable hardship (or, one might say, 11 points and a good story), plus (3) if the J-1 program was funded by the U.S. government, then substantial special additional factors must exist, such as spectacular levels of hardship to qualifying relatives or spectacularly high-level political help.

What about a case involving one U.S. citizen child with a very serious medical hardship, and no other hardships? If it really is a very serious medical problem, that application will almost always be approved. That is why I have the medical hardship category weighted up to six, because five for the child plus five for a serious medical hardship would only total 10, not enough compared to the previous example involving a citizen spouse and child. Five for the child plus six for the very serious medical condition would total 11, which is sufficient. If one clearly scores 11 points, one does not need to go extensively into all the other hardships that may exist for the family.

In reviewing cases to compile the scoring ranges for the Hake Hardship Scale, I posed many such comparisons. I tried my best to give accurate numbers. So, for instance, childhood asthma without a history of hospitalization might get a 1 or maybe a 2, but not more. Scoring each case as accurately as possible, in the preliminary version of this article I found that all 50 approved cases did in fact score 11 or more. Moreover, the range of scores that significantly exceed 11 accurately reflects my subjective impression of the seriousness of the cases, and in most of them the government’s response time was appropriate. The highest score on the list was 27, in a case where a permanent resident spouse, a wonderful woman, died in childbirth giving birth to the applicant’s fifth child. On my emphatic urging, that case was approved by the State Department’s Waiver Review Division 40 minutes after its arrival from the USCIS service center.

Usefulness of Case Law

I have always tried in each case to identify and emphasize a small set of “main hardship factors.” My main tools are intuition and empathy. Over the years I have been as confused as anybody by the case law in this area. As discussed above, I have a low opinion of most of the case law in this area. But knowledge of the case law does prevent certain mistakes. For example, the average man on the street, faced with the prospect of being forcibly separated from his wife and children for two years or more, would probably regard the pains of spousal separation, and the emotional and developmental hardships of parent-child separation, as the dominant hardships. The USCIS, however, has always followed the cruel rule that such hardships do not count in the Form I-612 context, because they are imagined to be normal. Similarly, people are often greatly traumatized by things like worries over the decreasing chance of having children that could result from a two-year interruption of infertility treatments. But this argument has nearly always fallen on deaf ears. There is no point beating these dead horses. Aware of the case law and actual administrative practice, one must simply state the truth about these kinds of hardships (for these kinds of real suffering must be treated with dignity), but these factors must not be emphasized. Instead, one focuses on the factors that will “work,” ever mindful of the need for absolute fidelity to the truth of the family’s situation.

Latest Supporting Data

As noted above, the first version of the Hake Hardship Scale was based on an analysis of my 50 previous approved J-1 hardship waiver cases. The first published article included a spreadsheet summarizing the hardship factors in those 50 cases. This analysis was useful. Among other things, it confirmed that 11 points did appear to be the accurate breakpoint between likely success and likely failure. None of the 50 successful cases scored less than 11 points, and the overall range was from 11 to 27.

But there are problems with that data set. First, one needs to review comparable numbers of approvals and denials in order to speak with scientific authority. Second, since all 50 cases were prepared by me, it is impossible to know whether the result is biased by my personal reputation or skills. For instance, it is conceivable that other lawyers, not knowing my manner of presenting a case, might not also consistently win cases that score 11 or 12 points. Third, this data set is inescapably biased toward my own impressions as to what are the most important factors in hardship waiver cases. It is conceivable that other lawyers might win cases by looking at very different kinds of factors that I tend to ignore.

Accordingly, over the past two years I have thought about ways to base the Hake Hardship Scale on better data. I could not rely on my own cases. First of all, I did not have enough comparable denials, because I’ve had only 19 denials in 11 years, and 12 of those came during the anomalous year of 1999. Further, even if I had sufficient denials to compare, I could not avoid the other possible biases caused by using only my own cases.

The best alternative data set I have been able to assemble consists of 140 decisions of the AAO spanning the years from 1985 to 2002. Of those 140 decisions, 85 are denials and 55 are approvals. All are decisions on the merits on Form I-612 hardship waiver applications.

I coded all 140 decisions using a 30-line coding sheet. The data were analyzed in several ways by statisticians at the Institute of Statistics and Decision Sciences at Duke University. This article focuses on the results from fitting a logistic regression model to the data.

This data set is far from perfect. First, it only includes cases that were initially denied by the INS. Therefore, it does not include cases initially approved by the INS. This significantly skews the varieties of major hardship factors identified in the decisions. I am certain from my own experience that the most significant kind of hardship is serious medical problems of a spouse or child. Such cases are often approved comparatively quickly, and they almost never require an appeal to the AAO. Therefore, this most important kind of hardship factor rarely appears in the 140 AAO decisions I analyzed. My initial data set of 50 of my own cases contains a more accurate distribution of the most important hardship factors.

Second, this AAO data set is silent on the question whether the cases were ultimately approved after favorable recommendation of the USIA or State Department. It is certain that many of the 55 cases approved at the AAO level had to be subsequently denied by the INS after negative recommendations from USIA or DOS, especially those cases that involved U.S. government funding of the exchange program. My initial data set of 50 of my own cases was also superior in this respect, since it only included cases that were finally approved after favorable recommendation of both the INS and the USIA or State Department.

Third, many of the 85 denials were of very poor quality. Some involved waiver applications that never should have been filed in the first place, because there were no qualifying relatives and thus the government did not have statutory power to approve. Many were filed in ignorance of the “two-step” rule described above and thus actually did not identify even one exceptional hardship under the hardship-minimizing travel alternative. Such cases have little value in assessing the significance of different kinds of hardship claims.

Nonetheless, this AAO data set is probably the best and most neutral data available on the question of which factors are important in Form I-612 hardship waiver cases.

Using an established technique called logistic regression, the statistical analysis found that the Hake Hardship Scale is a significant predictor of approving or denying cases. . . . The interpretation is that the results are highly significant. If the Hake Hardship Scale were not related to the probability of success, then there are only approximately two chances in a million of obtaining a result that supported its value so strongly as does this data. This is so notwithstanding the potential problems with the AAO data set identified above. Results in all cases may very well match the Hake Hardship Scale even more closely.

In short, the structure of the Hake Hardship Scale, and its assignment of weights to different factors, is highly accurate from a statistical perspective, at least insofar as it predicts the results in the analyzed cases.

[statistical chart omitted]

Of the 140 cases analyzed, 16 scored exactly 11 points. Of those, 15 were approved, representing a 94 percent success rate. That is an even higher success rate than indicated by the chart above.

Of the 140 cases analyzed, 55 scored 11 or more points. Of those, 50 were approved, representing a 91 percent success rate.

Moreover, of the 140 cases analyzed, 129 turned out to have results that are consistent with the predictions of the Hake Hardship Scale. Only 11 decisions had contrary results. In particular, six cases were actually granted for which the Hake Hardship Scale would have predicted denials. Those cases scored 9, 10, 10, 10, 9, and 10 points, respectively.

Five cases were actually denied for which the Hake Hardship Scale would have predicted approvals. Those cases scored 15, 13, 12.5, 11.5, and 11 points, respectively.

Of these 11 “outlier” decisions, all are close in score to the predictions of the Hake Hardship Scale, except only for the denials that scored 15, 13, and 12.5 points. In my opinion, those three cases are clearly anomalous and wrongly decided.

As described above, there are three main components to a score on the Hake Hardship Scale: the total for the persons involved, a possible score for special public interest factors, and a total for the exceptional hardships involved in the hardship-minimizing travel alternative. As noted above, statistical analysis showed that the total score on the Hake Hardship Score is highly significant statistically. An additional statistical analysis that regressed the logit(P[granted]) over the three main components showed that the total for persons, as well as the total for hardships, are also highly significant statistically. In other words, not only is the Hake Hardship Scale’s overall score highly statistically significant, but the balance among the two main components of the score also is statistically accurate.

The score for public interest, in this data set, was not found to be statistically significant. However, this factor should not be overlooked. Of the 140 decisions, a special public interest factor or factors was recognized by the AAO in only 10 decisions. Of these, 6 cases were denied and 4 were approved. Of the 4 that were approved, 3 of them needed the public interest points in order to reach 11 points on the Hake Hardship Scale. Sometimes this one detail means the difference between victory and defeat. This fact, incidentally, is also supported by many of the published precedent decisions in this area.

Unfortunately, the AAO data set did not permit much in the way of intelligent nuancing of the originally proposed Hake Hardship Scale, because the range of hardships represented in these decisions turned out not to be particularly representative of what one observes in practice. Additional statistical analysis found that only three specific hardship categories–psychological hardship to a spouse, psychological hardship to a child, and financial hardship to the qualifying relatives–were statistically significant. Moreover, the data set was too small to permit refinement of the range of scores that are permitted within particular categories on the hardship scale. Nonetheless, the statistical analysis showed very clearly that the overall system is highly accurate at predicting success.

Role of U.S. government funding. Of the 140 AAO cases that were analyzed, 72 (or approximately 50 percent) involved U.S. government funding of the exchange program, according to the AAO. Some of these specifications of U.S. government funding, especially from the earlier cases, may be inaccurate, but the overall total is probably fairly close to accurate. Of the 72 that were said to involve U.S. government funding, 53 were denied and only 19 were granted. Note further that of these 19 that were successful at the AAO level, many or most were probably ultimately denied after USIA or State Department review, because it is the latter agencies that are primarily concerned with the so-called “program and policy” considerations, unlike the Immigration Service, which is primarily concerned with the existence or not of exceptional hardship to qualifying relatives. These totals clearly confirm the almost overwhelming problems posed by U.S. government funding. My estimate that the existence of U.S. government funding should be assigned an adverse weight of five points appears to be remarkably accurate, at least in terms of predicting the chance of a case being recommended for approval by the Immigration Service. It is unknown whether that number accurately reflects the final results after State Department review.

[Addendum omitted.]