Can Big Science Be Too Big? | Hacker News Reader

Can Big Science Be Too Big?

nytimes.com

23 pointsundefined112 comments

> larger research teams did the work of consolidating the ideas and solidifying the evidence

As someone who works in science, I hate that this was just kind of a side note in this article.

If you want to know how to create a replication crisis, just value "novelty" and "disruption" way more than correctness and replicability.

The scientific approach is the gold standard for validating knowledge. An individual makes observations, formulates a hypothesis, and experimentally tests the hypothesis with controls. I believe the reason why science is so important to humans as a species is because it allows individuals to overcome their own mental and psychological biases about how they believe the universe should work.

Now consider this quote from the article: "Psychologists have found that people working in larger groups tend to ... become less receptive to ideas from outside."

When humans work together in large groups, new problems arise. I think new frameworks are needed to set the gold-standard for how humans interact in groups to prevent group biases and group think from strangling innovation and new ideas.

I think we all know from experience that working in a big organisation can be stifling.

What I find interesting about it is it's the larger organisations that encourage specialisation. You would think that would increase innovation but it seems to do the opposite. YouTube, GitHub, Whats App, DeepMind, LinkedIn, these companies weren't created by big business, they were bought by them.

I've generally viewed large organizations at the best as being better at refining, expanding on, and scaling ideas more than creating brand new ideas on their own. After a certain size, companies tend to favor working on things that are 'proven' and tested as their priority shifts from exploring new ideas to maintaining and iterating on their primary products.

Have Psychologists really proven they use the scientific method? Aren't they one of the worst offenders when it comes to replication of results?

I don't know if they're the worst, but psychology certainly is one of the first fields that pops into my head when we talk about hard to replicate results. I find it incredibly ironic that the results that they may or may not have found are now used as an argument in favour of changing actual science.

Psychologists are stretching the scientific method to its limits. They are forced to do it due to nature of the target of their research. Scientific method was developed mainly by a physicists who works with a less complex systems. It was adopted then to a chemistry and even biology. But psychology needs more adoptation.

So the answer is yes. They are using scientific method and they are constantly inventing new scientific tools. Physicists tell us that philosophy is dead[1], but it is just because philosophy is dead in physics: physics research now is a settled technology, while psychology needs more powerful methods which do not exist yet. So psychology is searching for this methods and philosophy now lives in the domain of psychology, not in physics.

The funny conicidence: when I searched ddg for reference of Hawking using "philosiphy is dead" as a search term, the first search result led me to news about Hawking, the second to some philosopher who didn't agree, the third one led to Psychology Today, where cognitive scientist argued against Hawking.[2] Though I see his argument as overcomplicated shot in the blue: it is pointless to agrue that physics and philosophy have something in common in the present times.

[1] https://www.telegraph.co.uk/technology/google/8520033/Stephe... [2] https://www.psychologytoday.com/intl/blog/hot-thought/201011...

The fact that they know there's an issue with replication suggests they're using the scientific method. Their experimental design might need work, but they're working within a scientific framework.

Title is misleading, that is not what the study really suggests. That study also doesn't do a good job at discriminating between causality and consequence. Nonetheless, they are careful in their conclusion and suggests that it is more an organic way of doing things. Smaller teams are more focus and will work with a lot less, and the PI will try to get famous, so new stuff it is. While bigger team are going to carry own, try to get it to the next level. It's lot more incremental, and that is the way it should

Some of the largest studies ever have been performed recently, using very large particle colliders or very large-sample, multi-site clinical trials. These scaled-up versions of smaller previous experiments show the big money and big stakes in science, and also the pushed drive towards publication as a metric for scientific success.

Can we trust psychologists to tell us anything empirical anymore?

Plus, any article about big science that doesn’t mention CERN is quite literally pissing in the wind.

Article tl;dr: Consensus stifles innovation. This is not new knowledge. But it’s the organization’s culture that decides which is more important: See CERN, Bell labs, Xerox-PARC, HP, etc

Reply on news.ycombinator.com