Hacker News Reader

Comment by cryoshon | Hacker News Reader

This is trash science.

Nobody else has been able to replicate their "results".

Seems like they are running experiments and collecting data to test their hypothesis. Doesn't that count as science? Or do you call it trash science simply because the hypothesis seems unreasonable?

It's a good first step, however unless the experiments are repeatable/independent, it's not particularly trustworthy.

Consider that they may be falsifying or misinterpreting data. The only way to be sure they aren't -- or pretty sure, anyway -- is to have independent researchers be able to follow the same repeatable process in order to replicate the results.

I don't know enough about this case to say whether they have a good repeatable experimental process -- it's possible that the website just smells so strongly of woo that nobody has even tried to replicate their results in the first place. But without that independent verification, the proposed results seem questionable.

Here's a good post about this [0], including this gem:

> Schiltz is a psi believer [... but Wiseman] is a psi skeptic [...]

> The idea was to plan an experiment together, with both of them agreeing on every single tiny detail. They would then go to a laboratory and set it up, again both keeping close eyes on one another. Finally, they would conduct the experiment in a series of different batches. [...]

> The results? Schlitz’s trials found strong evidence of psychic powers, Wiseman’s trials found no evidence whatsoever.

Running experiments and collecting data is necessary, but not sufficient. Bias and other failures are extremely hard to avoid.

Saying "it's not science" is (derogatory) shorthand for the assertion that their results are due to failures in their process, as opposed to being reflective of reality.

0: http://slatestarcodex.com/2014/04/28/the-control-group-is-ou...

I'm just going to leave this here:

http://en.wikipedia.org/wiki/The_New_Inquisition

Agreed. Sites not loading for me, but I've looked into this nonsense from time to time over the last 15 years.

A couple of issues I remember:

1. They only publish results for positive events. They only use pseudorandom data, not sampling other time frames of their existing data

2. Their random source showed consistent bias (I believe it was something like 51% chance of 1 vs 49% of 0. This is after Von Neumann whitening which means they have some additional biases in their random sampling

1. They also publish for negative events. Look at the really low Z-scores: http://noosphere.princeton.edu/results.html Also you can get all their data (downloading it day by day).

2. Possibly, but unavoidable to get rid of a systematic error which would pollute everything else. Details are here: http://noosphere.princeton.edu/gcpdata.html

1. They publish events that don't have high scores, but they don't publish non-events that do have high scores.

2. I was actually wrong on the VN whitening, they use XOR, either 10 for some devices, or a 560 bit mask for others(!). This is just silly.

  the XOR is used to exclude an important class of potential spurious effects — biases that might arise from temperature changes, component aging, etc.

It's not excluding spurious effects, it's hiding it.

Inclined to agree but lack of replication does not of itself point to trash science. There may well be other good reasons for so-defining this project.

Reply on news.ycombinator.com