Affecting behavioral changes in others: psychokinesis or expectations?

[Ninshub]

This short NPR piece talks about the classic Bob Rosenthal experiment and follow-up research.

Can You Affect Another Person's Behavior With Your Thoughts?

That's interesting as far as it goes. However it's not at all what I was expecting from the thread title. The experiment described contains, throughout, a whole lot of ordinary communication going on, from the words and language, to body language, as well as all sorts of inner mental states resulting from these communications. It's possible that there could be a real effect, but the experimental set-up was designed to confound and avoid finding such effects with any certainty.

What I had in mind was something closer to Sheldrake's experiments on 'the sense of being stared at', where there is no ordinary communication whatsoever. I've mentioned before, anecdotally and not as proof, that I've noticed strange effects when observing people on CCTV. Here, Sheldrake has also said that private investigators, store detectives and animal trackers also have to take care so as to avoid alerting the observed person or animal, even when watching CCTV. In the latter case again, it is not ordinary senses, such as inadvertently snapping a twig, but the mental state and intention which is the issue.

That's interesting. The video gives the impression that this was a huge effect - "the smart rats did almost twice as well as the dumb rats" - which would perhaps make it unlikely to be a psi effect. I had a look at the paper by Rosenthal and Fode (1963) to see if it calculated an effect size:
https://onlinelibrary.wiley.com/doi/epdf...3830080302

I couldn't see a figure for effect size, but looking at the t values in the first table of results, based on the values for individual days the effect size looks to be roughly 0.5. (I don't see that it's valid to combine the data for all the days and treat them as several hundred independent trials - as the authors did in the bottom line of the table - because each experimenter was working with a fixed group of rats over the whole period of the experiment.)
[Image: Rosenthal.jpg]
So that is quite a bit bigger than the effect sizes reported for psi, but perhaps not quite as big as the video would suggest.

Incidentally, I can't see any figures in the paper showing that the smart rats did almost twice as well as the dumb rats. From that table, the rats who did almost twice as well as the dumb ones were those in the column labelled "Asst." Those were rats experimented on by a research assistant who knew about the experiment, including that all the rats were just randomly chosen.  But as there was only one such experimenter, it doesn't look as though that difference was statistically significant (unless the results from all the days are combined, which I don't think they should be).

[Edit: Actually, I think the difference between "Asst." and "Dull" is significant, but not the difference between "Asst." and "Bright".]