A Rebuttal of Lau

Abstract

Peter Lau raises an appropriate concern for inferences drawn from a comparison between groups in a natural experiment; but in his zeal to overturn our conclusion about the existence of a community-college effect, he fails to consider the range of evidence on which this conclusion rests. We remain convinced that, his assertions to the contrary, our analysis constitutes strong evidence for such an effect and evidence, moreover, in a theoretically interesting site-the open-admissions program. We will deal with Lau's major point before taking up some of the other issues his comment raises. Lau suggests that even though we made great efforts to construct equivalent groups of seniorand community-college students, the equivalence falls apart because the community-college students may have been reacting to the negative treatment of placement in schools that were not their first choices.' We do not dispute that this issue is worthy of consideration, and Lau seems correct in assuming that it is difficult, if not impossible, to circumvent it either through the research design or suitable control variables. But these facts hardly mean that we must surrender our central conclusion without further ado, because here, as in many similar situations, one can distinguish between competing hypotheses by teasing out the full implications of the evidence. Such implications are not hard to find in our article. For example, students' academic life chances varied according to the particular community college at which they were placed, and in fact the students placed at some of the community colleges had life chances equal to those placed at some senior colleges. Such findings are hard to reconcile with the notion of reaction to an unfavorable placement, but even more troublesome for this notion is that the worst life chances were found for students at a community college that had long been publicly viewed as a good transfer route into the fouryear schools (Alba and Lavin, 1981:236). The most compelling evidence on behalf of our conclusion lies in one of the article's major findings-namely, the temporal unfolding of the divergent careers of the two groups of students. If reactions to initial placement played a key role in accounting for academic outcomes, then one would expect differences between the two groups to manifest themselves early in the freshman year, because this is when students placed in the two-year schools should feel most discouraged. A further reasonable expectation is that the community-college students who survive this initial period of discouragement should begin to approach their senior-college peers in terms of academic progress in second and subsequent years as it becomes gradually clear that they can attain their goals. If, on the other hand, the ultimate differences we observed are due to effects of community-college attendance, then one would expect the differences to emerge more slowly; it should take time for the community-college students to become deflected from their original goals. It is precisely the latter pattern that we found, not the former. In fact, community-college students did at least as well as their senior-college peers during the first two years; for example, with high school background controlled, they were significantly more likely to return for the second year (fully 80 percent of them did so). It was only after the second year that differences tavorable to the senior-college group began to emerge. To our mind, this constitutes powerful evidence against Lau's thesis that ultimate differences are due to the devastating effects of rejection. In considering the larger concerns Lau raises, it may be helpful to recapitulate the starting point for the article we published in Sociology of Education (Alba and Lavin, 1981), especially since his comment ignores some of the important features of our own (and others') research. The article marks a culmination point. It was preceded by a lengthy set of analyses, published in our book (Lavin, Alba, and Silberstein, 1981), in which we examined in depth the academic careers of four-year and two-year college students and in which we did precisely what Lau claims ought to be doneI For clarity's sake, it should be noted that applications to CUNY permit the ranked choice of up to six colleges. Since all of the students in our two control groups had academic records that made their admission to a senior college questionable, presumably most or all of them would have included some community colleges among their choices. Thus, the great majority of the students placed at a community college would have been placed at a school among their choices.