<p>Mollie's recollection tallies with what I know of science admissions at Harvard and MIT. The traffic between the two goes both ways.</p>
<p>Let me speculate about the proportion of graduates from top research universities going into top Ph.D. programs as opposed to LAc graduates.
This does not mean that LACs do not produce a large number of Ph.D.s, but my argument is limited to top Ph.D. programs.</p>
<p>When I was in grad school, I don't recall there was much concern about how long it took for students to finish their Ph.D.s I do recall, however, that there were several categories of admissible students based on merit, However, when it came to admitting them, some would be receiving full funding and some would bring in outside money--an outside fellowship or their parents' funding. This led to rather unevenly qualified cohorts of graduate students.
I don't quite know when top Ph.D. programs began to admit Ph.D. students with full funding, severely limiting their intake compared to before, but also making for better qualified cohorts (Princeton, if I recollect, had always provided full funding). At the same time, universities began to put pressure on students to finish in specific periods of time, rather than dragging things out.
This had led to a far more cautious approach to admission that favors graduates of "known" institutions; in certain fields, it also advantages students at research universities who have access to a wider range of courses. It may not matter in fields such as American history to have gone to a LAC or to a research university, but it matters in fields such as East Asian or Middle Eastern of South Asian studies whether an applicant has taken more than introductory courses and has already studied the relevant language(s).It matters as well in fields such as math, where graduate students are encouraged to take their generals in their first year of graduate school.
This cautious approach favors schools such as HYPSM which are all mid-sized research schools and top state universities such as UCLA, Berkeley Michigan.
This is probably why Mollie's cohort does not include as many LAC graduates.</p>
<p>Now, I'm not arguing that graduates from LACs do not get into top Ph.D. programs. It depends of course on the LAC, but also on the profs' recommendations and the fit between the applicant and program. Sometimes, as Mini found, a LAC has a better program than an Ivy. But as a very general rule, mid-sized research university give their students access to a greater range of advanced courses. This has to factor into the admission process.</p>
<p>I definitely agree with you, marite, although I just want to add the caveat that the information for my cohort is only one year of one PhD program, so it's data, but it's only slightly better than anecdotes.</p>
<p>Marite:</p>
<p>This is consistent with what the study on baccalaureate origins of PhD students at top programs in the sciences. With the exception of Swarthmore and to a lesser extent Williams, graduates from LACs were not as present in the top PhD programs in the STEM fields as graduates of elite private universities such as HYPSM. </p>
<p>Again for departments that rely heavily on outside grants (generally the case in the sciences and engineering) to fund their programs, it makes a lot of sense to recruit PhD students that are already familiar with the research conducted in the lab and can be productive right away. In the increasingly competitive environment for research grants, an already trained cadre of incoming graduate students is a major asset.</p>