Proposals to "staple a green card" to every science, technology, engineering, and mathematics graduate degree earned in the United States appear to imply that all degree-earning international students are equally valuable in producing high-quality research. But, as anyone who has ever stood before a classroom or worked in a lab knows, the individuals enrolled in a given program, regardless of their nationality, vary widely in academic ability and creativity. A wise policy would therefore consider the quality of international students in making visa decisions, write Eric T. Stuen of the University of Idaho, Ahmed Mushfiq Mobarak of Yale University, and Keith E. Maskus of the University of Colorado, Boulder, in an article  published in The Economic Journal in December 2012.
Using statistical means to examine the contribution of graduate students to scientific productivity, the authors conclude that "both foreign and domestic doctoral students contribute significantly and positively to the production of publications and citations at US academic departments. … Our analysis thus documents a key benefit of high-skilled immigration in the US, which relies on innovation for growth and helps to explain the vociferous complaints by US academic departments about visa restrictions on foreign graduate students. However, our estimates also indicate that the marginal effects of foreign and American students are statistically comparable … . Universities would have equally strong reasons to complain about any policy distortion that might keep Americans away from graduate study in S&E fields."
"[T]he quality of entering foreign doctoral students matters greatly for research productivity," they continue. "[I]f an important objective is to expand the research capacity of American universities … a key criterion for issuing a visa could be indicators of student quality." Their analysis suggests that "from the perspective of US science education and innovation policy, visa restrictions for foreign students should not be applied uniformly or on the basis of financial means; they ought to account for student-quality differences."
You can read the full study here .