Harvard Probe Finds Honesty Researcher Engaged in Scientific Misconduct

My personal field is irrelevant, since my work let me see across fields (for reference, most of my direct work as relevant to this discussion was in the social sciences, though -- that's the project which let me interact with large numbers of academics across institutions and fields).

The short answer is that what I saw was the level and means of cheating roughly correlates with the degree to which it is possible.

In a lot of engineering papers, if I publish a new material, mechanical structure, or electronic circuit, most people will be able to:

(1) tell how it works just by inspection

(2) if it's useful, try to use it in an engineered system

Cheating would be caught very quickly, at least for any interesting result, so it's much less common than in fields like social sciences.

Computer science is in between. There is a growing pile of b-llshit, since a lot of the work is squishy. Something like a user study can certainly be doctored, a data analysis can be p-hunted, and results can be overstated. On the other hand, something like the big-O of a new algorithm isn't something you can really cheat on.

There are also fields like [insert minority group] Studies or [insert domain of art / music / literature] Analysis, where it's not so much direct fabrication (since there is no data) as writing papers for an inner circle of mutual adoration. "Research" is mostly around stroking egos of people who make decision around hiring, promotions, publication, and funding. I'm not quite sure how to describe that, since it's not really cheating, but very little new knowledge is generated.

Footnote: What's interesting is MIT used to be super-honest in the nineties with computer science research at both AI Lab and LCS (which have since merged into CSAIL). The Media Lab really pioneered (by the standards of the time) b-llshit computer science research. It was a running joke at the Institute. In the decades which passed, the Media Lab stayed roughly where it was, while CSAIL's integrity fell, and now the Media Lab feels much more honest than CSAIL.

Footnote 2: This post should be viewed as much more anecdotal than the one before. I have a very high level of confidence making cross-cutting statements as in my original post. Doing field-by-field dives is just personal observations, so this post, I'm responding to your question, but please just take it for what it's worth. It's small n by field-cross-school.

What generation?

Academia was very honest for most of history. From WWII through roughly 1980-2000, there was almost exponential growth in the number of academic jobs available. At the time of WWII, about 5% of people had college degrees. Today, about 2/3 of people enter college.

That near-exponential growth meant there while each professor had a few grad students, and there were enough faculty openings for the students who were both talented and wanted to go into the academy since new schools were opening up.

The extreme competition came in when this growth stopped -- college enrolment can't go above 100% of the population. For every professor, there are still many graduate students, but a position at a research institution only opens up when that professor retires. The whole dynamic changed.

Culture doesn't change overnight either, so things remained pretty honest for a few years. The present level of wholesale academic fraud is a relatively new phenomenon.