Does grade deflation really exist at Berkeley?

[sakky]

<p>

Finally, another indication that a high upper division GPA itself is tantamount to research potential is that many/most upper-division engineering classes at Berkeley have you write a paper that comprises the bulk of your grade. This paper is in journal format and is graded by a research professor.

</p>

<p>Now I’m afraid I have to emphatically disagree. Many/most upper-division Berkeley engineering courses requiring writing a paper in journal format? Exactly which ones? I know I never had one, nor can I think of a single engineer who actually did. Note, to be clear, a few such courses may indeed exist, but they seem to be clearly in the majority. Heck, I would submit that the overwhelming majority of Berkeley engineering graduates don’t even know what ‘journal format’ even is. </p>

<p>

After spending a year in a lab, I think any reasonable undergrad should be able to write an A-level paper. Remember, this is engineering, not pure mathematics. So you can, and actually often have, reasonably smart students who have gotten very lucky with collected data. Often, in an established lab, undergrads will just copy the lines that graduate students and postdocs take. Thus, research potential, as evaluated at the undergrad level is “soft” because its often a function of:</p>

<p>a) mentoring strength. Most undergrads I know of just follow a grad student around and do the grunt work. Because of their grunt work they often get a publication.
b) which lab the undergrad chooses to join
c) the element of luck in research, which is always there.

</p>

<p>And it is for the above reason that I continue to maintain however soft you may think research output may be, GPA is even softer. Sure, publication success is highly luck-driven. Nobody is arguing otherwise. On the other hand, as I said, GPA has little if anything to do with research success. Nobody in academia will ever care about your GPA. Like I said, high GPA is a function of your understanding of the material (and ability to beat the curve). It says practically nothing about whether you can generate new material. </p>

<p>

By the way, students are accepted at top programs without any publications, and often with very little research experience.

</p>

<p>But now it seems as if you’ve contradicted yourself. You said before that any ‘reasonable’ undergrad could obtain an A-level pub in just one year. That is an assertion to which I must emphatically disagree, having known many ‘more-than-reasonable’ students who have worked in labs for years yet lack even a single pub, let alone an A-level. But let’s assume it’s true anyway. Now you’re contending that students are accepted to top programs without any pubs (to which I agree). But how does that square with your assertion above. Why don’t these applicants have any A-level pubs? Are they not ‘reasonable’ students? If so, why are these programs admitting ‘unreasonable’ students? </p>

<p>The far more parsimonious explanation is that most applicants to top academic programs have no impressive pub record whatsoever. Even after admitting the small minority students who do have such a record, the programs still have numerous slots to fill, and the only criteria they can then use is GPA, which while clearly soft, is better than nothing at all. </p>

<p>

For what its worth a lab technician who graduated Berkeley 3 years ago and has been doing fulltime postbac research with 2 first-author publications in A-level journals was only accepted to the same school he had done the postbac research in due to grades. If he had better grades, he would have been in anywhere.

</p>

<p>But now you’ve implicitly changed the terms of debate. Implicit within any discussion of a pub record is the amount of time you spent to obtain those pubs. For example, it may be highly impressive for a graduating engineering PhD student on the academic job market to have 10 A-level pubs. That’s far less impressive for an associate professor up for tenure review. The latter is supposed to have far more, perhaps by an order of magnitude, by virtue of how much longer has had been in the game. </p>

<p>Similarly, a lab tech with 3 years of full-time research experience, frankly, should have more than 2 A-levels. However, an undergraduate having even a single A-level is a notable achievement, for as you noted, the vast majority of undergrads have zero. I struggle to think of even a single engineer who actually had an A-level pub by the time he graduated. </p>

<p>

However, a poor GPA will not get a candidate into a top program, no matter the strength of his research, unless he back-doors/networks his way in by meeting a professor at a conference and having that professor push for him.

</p>

<p>But that’s a major caveat you just laid out right there, and indeed, that is how I suspect that most grad admissions are performed. Whoever you had been coauthoring with is clearly going to be strongly pushing for your admission to his program, if for no other reason, so that he can continue his collaboration with you. If one faculty member strongly endorses one candidate, it is hard for the rest of the adcom - who is comprised mostly of faculty themselves -to deny him.</p>