Coffee Space

TL;DR

This article contains a few things that concern me about the current state of things in academia, concluding that he future will be tough going forwards.

Progressive Push

One thing that got me thinking about this was a recent article from Wired (which seems to repeated in other places - likely distributed via a news network). They argue the following (at least in response to specific papers):

Getting rid of harmful papers is a vital step toward reestablishing readers’ trust. Next, publishers should target articles that are flawed in other ways.

I really hate that people are jumping in joy about burning books, it’s always sad when some published material becomes unavailable, even if we never had any intention of experiencing it. For example, I’ve not watched the old episodes of Doctor Who, but it’s sad that actually this is partly an impossible task, as some episodes were lost. Even objectionable content deserves preservation. I’ve never read Mein Kampf and I have no intention to, but to hear of it’s loss tomorrow would be saddening, because it’s part of history. From influential publications all the way down to some random persons’ comedy Universe where they hunt a sasquatch, we should preserve it all the best we can.

Some basic questions I must ask:

1. If papers that are un-Scientific are getting through peer review, surely this points towards a problem in the peer review process. If the peer review process worked as expected and these reportedly racist papers have merit, then what business do they have in removing them? Either the peer review process of these publishers is broken or we are wrongly retro-actively attacking papers, pick one.
2. How harmful are a bunch of old papers really? Who is actually reading them? Who is seriously citing them for new pieces of work and getting their work published? I think it’s likely the impact of these papers is overrated. For example, the first paper mentioned in the Wired article, titled “Do pigmentation and the melanocortin system modulate aggression and sexuality in humans as they do in other animals?”, was “cited just nine times in eight years, according to Clarivate Analytics’ Web of Science”.
3. “publishers should target articles that are flawed in other ways” - if that isn’t ominous then I don’t know what is. What will be considered objectionable content tomorrow? Will even posing the question become reason enough to have your paper deleted? Bare in mind that this is during a time where certain studies are very unlikely to receive funding and even less likely to be published. There are questions academics refuse to touch, such as anything that reflects minorities in a negative way, anything that questions trans sexuality and generally anything that questions a progressive narrative.
4. The problem with terms like “racist”, “sexist”, “hate”, etc, is that their social definitions are changing. A young family member for example was punished at school just for factually stating that one of the other children in his class is black. The family member didn’t suggest this was a negative trait, merely one that this person in the class could be identified by (it made more sense in context). The goalposts are continuously being moved and it’s not clear where they will land. For example, just stating that this event occurred may land me in hot water in the future. Perhaps I’ll be tortured in room 101 until I reveal their identity, so that their social credit points can be retroactively negated.

Science was never supposed to be a place that was politically correct, Universities used to be a place where crazy ideas could be tested and debated, until eventually truth appears. We must ask the hard questions if we are ever going to find truth. Sometimes, we may not like the answers.

Bias & Reviewers

A friend recently complained of a paper they wrote being rejected, which I personally read some points around and was very impressed by. We’re talking about some very cool work, some stuff that could really influence some future work in the space. The problem is, their field is refusing to publish their work as it suggests that many preconceived ideas are possibly wrong.

This is ultimately a problem in academia, where people put their reputations on the line. I understand it and how people get into this space, but I certainly don’t like the fact that it exists. Sometimes these reviews are absolutely seething, for reasons that are really quite trivial to deflect.

Not every reviewer is equal. A reviewer once told me that most other reviewers are not very mathematical, so when they see lots of mathematics they think “wow”, when in fact it might be hot garbage. This reviewer said that when they see lots of equations this causes them to inspect the paper more closely, and many papers in their opinion were actually not robust and had glaring holes.

What the solution is I don’t know, but there is problems. Bad papers still slip through the net and good papers get rejected for trivialities, like a reviewer having a bad day.

Case-Studies

A bunch of times now I have seen people get their doctorates with relatively trivial PhDs based on case-studies. The subjects are suitably complex and I don’t doubt these are smart people, but ultimately they submit a thesis with zero supporting publications and it’s read by just two experts in the field. Typically they will interview between 10 & 200 people, so some questionable analysis and offer some ‘new’ process or diagram to ‘explain’ something, and apparently this is trivial enough to get the title of “PhD”. Usually the worst data is interviews, because these are exceptionally subjective.

Somebody I know did 40 interviews in a different Country and there were so many variables, I doubt anything useful came out of it. For example:

• When they conducted the interview (by themselves), they likely asked lead on questions. This person has a habit of asking suggestive questions, for example “do you think you should do X?”. Great technique for rallying support, awful for conducting a subjective study.
• They had to convert from one language to another using Google translate. It’s well understood that context can be easily lost (try to convert something to Russian and back for example).
• They then had to do “data analysis”, where they basically judged what they believed the person was saying. The PhD student of course has an extreme conflicting interest when doing this translation.
• They then drew conclusions from this analysis, which they pulled ‘statistical’ numbers from.
• During the defence, the reviewers only had time to read their thesis (just about), not enough time to read all the supporting documents, especially not time to perform translations and do “data analysis”. They just looked at the conclusions, which of course looked great.
• The last time I spoke to them, they said they had to drop a few outliers from the their case study of 40 people.

And wullah, they have a PhD, job done. I suspect nobody will ever read this thesis other than perhaps the conclusion section or copy their new diagram, and the content will never be challenged. Awesome.

Finally

It really is kind of scary and I am concerned about the future of academia. Between censorship of socially objectionable material, to biased reviewers, to potentially fundamentally broken PhDs, things don’t look great. I think these things will only get worse as time goes on.

Hopefully 2021 will be better, let’s see.