For Better or For Worse: Voodoo Science and the Immigration Marriage Fraud Amendments

“Instead of sham marriages, the only proven sham so far has been sham legislation.”

 -James A. Jones 1

Great power can be achieved through science. This power should be treated with care; while the benefits of legitimate scientific research are unquestionable, the damage caused by “junk science” or “nonscience” can be significant and lasting. Within the legal context, scientific research is of the utmost importance.

Throughout the legislative process, Congress relies on experts to describe, explain, and evaluate proposed policies through scientific research and statistical analysis; a process which helps Congress decide whether or not to legislate.2 This is essential, as the members of Congress are not typically experts in the fields in which they legislate. Due to this reliance, those testifying before Congress are in a position of substantial power and influence. The famously complicated field of immigration is no exception; for better or worse, the use of statistics has played a significant role in the development of our nation’s immigration laws.

While a presentation of legitimate scientific research findings is hopefully the norm when it comes to congressional testimony, the legislative history of the Immigration Marriage Fraud Amendments is a distressing example of nonscience being presented to and relied on by Congress in passing a law.

A brief discussion on the benefits of marriage within the immigration process is useful to understanding the context in which the Immigration Marriage Fraud Amendments were passed. Marriage to a United States Citizen (USC) or Legal Permanent Resident (LPR) offers many benefits for noncitizens seeking immigrate to the United States, as well as advantages in the realms of deportation and naturalization—including automatic exemption from, and discretionary waivers of, requirements or sanctions that may otherwise be imposed.3 In light of the significant visa backlogs that attend the U.S. immigration system, one of the most important and immediate benefits of marriage to a USC or LPR is its effect on the visa waiting period. Marriage to a USC confers “immediate relative” status on the immigrant, which automatically eliminates the visa waiting periods that would otherwise apply, with the only wait time being the time it takes for the visa’s processing.4 Marriage to a LPR gives “first preference” status to the immigrant, considerably reducing the waiting period.5 There is no question that this benefit might be sufficient to entice fraud; the waiting periods for certain categories of immigrants are significant, with some categories currently experiencing a wait period of over twenty years.6 Consequently, detecting fraudulent marriages has long been a priority for immigration officials.7 Marriage fraud is grounds for removal from the United States.8

Until 1986, the process for detecting a fraudulent marriage was to interview both spouses separately before approving the marriage petition.9 The officer would ask questions regarding the couple’s relationship— their courtship, marriage, day-to-day life, and other details—looking for discrepancies in their “stories.”10 Despite such procedural safeguards, the former Immigration and Naturalization Service (INS) still believed marriage fraud to be rampant.11

In 1985, Commissioner Alan Nelson of the INS testified before the Senate immigration subcommittee in support of legislation to combat this “rampant” marriage fraud.12 He announced that the INS had conducted a preliminary survey on marriage fraud, based on which “we believe as much as 30%, which is an extremely high figure, of the spouse relationships may be fraudulent.”13 During the hearing, Jules Coven, then-president of the American Immigration Lawyers Association, challenged the 30% estimate: “I would be extremely surprised to learn, if it could be shown statistically, that more than one or two percent of the ‘green cards’ issued annually on the basis of marriage involved fraud.”14 Nonetheless, the House Judiciary Committee recommended, and Congress passed, the Immigration Marriage Fraud Amendments of 1986 (IMFA).15

After the IMFA passed, the INS disclosed the basis for its findings: a 1983-84 INS study in which INS investigators reviewed selected spousal petitions.16 In conducting the study, the INS collected data in only three cities.17 Excluded from the sample were entire categories of marriage petitions in which officials believed fraud to be unlikely.18 The handpicked cases which remained were reviewed “in as much depth as possible given the [one to two months] for completion of the survey”; sometimes the parties were interviewed, sometimes they were not.19 The investigator was then asked whether he/she suspected fraud—not whether any fraud was actually found.20 A “suspicion” of fraud was reported in 30% of this already biased sample.21

In 1989, the further details regarding the truth about the study were uncovered when David Nachtsheim, the INS official in charge of the study, was deposed in conjunction with a case challenging the constitutionality of one IMFA provision.22 Nachtsheim testified that he and other top officials had known that the survey was “statistically invalid and lacked any probative value regarding the actual incidence of marriage fraud.”23  In fact, it was revealed that the study was actually completed to estimate the future workload of INS officials; it was never intended to measure the incidence of marriage fraud.24 Top INS officials ordered another study conducted, and prior to Commissioner Nelson’s congressional testimony, recommended the study not be cited as probative incidence of fraud.25 However, the Commissioner cited the figure anyway, and the IMFA became law.26 Incredibly, the INS never determined the exact number of cases of known fraud before Congress enacted the IMFA.27

The legislative history of the IMFA is a flagrant example of “junk” or “voodoo” science being used in the immigration context. There was a deliberate intent to mislead; the INS knew the study offered no evidence of marriage fraud, but its Commissioner cited it anyway. If top INS officials knew the study to be invalid, they should have and could have done more than simply “recommend” that the study not be cited; they should have prohibited it.

From a scientific standpoint, the problems with the research design of this study are numerous. First and most glaringly, it was not a random sample. In fact, the sample was totally biased, an example of the “sharp-shooter” fallacy in which shots are fired and the target is drawn around them. It was selective observation. The agency had at its disposal a sampling frame which included all marriage-based immigration petitions, yet it only focused on a select subset of the data that was already thought to support the agency’s hypothesis. Any data that did not “fit” was excluded from the sample. Because the sample was not random, it was not probability-based and therefore could not allow any generalization to the population of interest. Any concerns which could be raised about the efficiency of including the entire population within the sampling frame are totally unfounded; the agency must examine every petition for marriage fraud regardless of whether or not a study is being performed.

Beyond the research design flaws in the study itself, the other obvious issue here is the major break in the chain of reasoning: the conclusions presented to Congress were irrelevant to the purpose for which the study had been conducted. The study was intended to forecast future INS workloads—thus, even if the study was valid (which it was not, for the reasons discussed above), there would remain this blatant problem that the conclusions regarding the incidence of marriage fraud had nothing to do with the study’s purpose. By no terms is this good science.

Since the IMFA’s enactment, many of its controversial provisions have been criticized and constitutionally challenged. A few of the provisions have been struck down or amended, but most others have been upheld pursuant to Congress’ plenary power over immigration.28 Beyond whatever legitimate purpose the IMFA may or may not serve in combatting marriage fraud (which is beyond the scope of this discussion), perhaps its most important purpose is to serve as a reminder that junk science is not confined to the realm of the general public, or news reports, or the courtroom; it can and will permeate even our legislative process if permitted. When it does so, the effects are permanent and far-reaching; in the case of immigration, the effects reach not only U.S. citizens, but citizens from every corner of the globe.


1. James A. Jones, The Immigration Marriage Fraud Amendments: Sham Marriages or Sham Legislation?, 24 Fla. St. U. L. Rev. 679, 700 (1997).

2. See generally, Robert Park, Voodoo Science 96 (2009).

3. Stephen H. Legomsky & Cristina Rodriguez, Immigration and Refugee Law and Policy, 282 (Robert C. Clark et al. eds., 5th ed. 2009).

4. INA § 201(b)(A)(i).

5. INA § 203(a).

6. U.S. Dep’t of State, Visa Bulletin for March 2015, available at http://travel.state.gov/content/visas/english/law-and-policy/bulletin/2015/visa-bulletin-for-march-2015.html.

7. See Legomsky, supra, at 284.

8. INA § 212(a)(6)(C), (G).

9. Legomsky, supra, at 284.

10. Id.

11. Id.

12. Id. at 283-84.

13. Id. at 284.

14. Jones, supra, at 700.

15. Legomsky, supra, at 284.

16. Id.

17. Jones, supra, at 699.

18. Legomsky, supra, at 284.

19. Id.

20. Id.

21. Id.

22. Legomsky, supra, at 284.

23. Id.

24. Id.

25. Id.

26. See Id.

27. Jones, supra, at 699.

28. Legomsky, supra, at 285, 289.