PSI Experiments (Phase 2)

(Frozen copy of this post located here: https://pastebin.com/fmsS6WBA)

I'm preparing to perform "phase 2" of my psi experiments in the coming months.

For those unfamiliar with it, you can find the details of my previous experiment here. (Summary of the results, in graph form, seen here.)

Phase 2 will be quite different than the first. There are two major differences planned:

Difference 1) I am planning to use a "divide and conquer" approach, to isolate what may have been causing the deviation found in my previous experiment (recap: peak rarity by chance of 1 in 32,258, while the control series remained within the expected range the entire time, having a peak rarity by chance of only 1 in 5.96). Maybe it was "actual psi", maybe it was just a mistake in my code, some sort of interaction leakage, etc. -- either way, I want to confirm/understand it better.

So what I'm going to do is come up with all sorts of different variants of the experiment, and perform a "mini-session" of each variant, within each day's session. This will let me get "incremental results" from those (possibly dozens) of variants each day, which will hopefully let me narrow in on the possible cause more quickly.

Some potential "variant" ideas: (as deviating from the original protocol)
A) Perform the experiment without touching your mouse at all. (The software only used the mouse to change the location of an on-screen "intention-focusing" marker, with the simulation code being separate. The RNG-sample modulation layer [details in linked thread] also made mouse-directed interference very unlikely. That said, it's always good to add further confirmation.)
B) Perform the experiment from another location, only viewing it through remote desktop. (mitigating the chance of "temperature-based" interference, from proximity)
C) Perform the experiment with pseudo-random data rather than data from the hardware random-number-generator.
D) Perform the experiment with your eyes closed. (Only visualizing your own imagination of the session state; again, the point is not that this is expected to work, but that we want to try all sorts of crazy variations, to get some sort of handle on what factors might be involved. Think of all these pieces as "partial-result shards", from which inferences can be made later, during analysis.)
E) Perform the experiment, except set your intention to make the results go exactly opposite of what you normally would.
F) Try variations on the code that selects numbers from the hardware RNG, for example discarding every other number, using PRNG to select from each second's buffer of numbers, etc.
G) Try variations on the way the session state is displayed in the UI, while the session is running. (there are many variations possible here)
H) Perform the experiment with and without music playing. (for whether the sound waves, or one's enjoyment level, could have an impact somehow)

Anyway, the idea is that I'll be doing a short bit of experimentation with each of these variants each day. And then at the end of each day, I'll load up the statistics page, and get a historical view of how each of the variants is "performing". If any of those variant lines starts greatly rising above, or greatly falling below, the rest of them, then it might provide a clue as to what the "cause" of the earlier anomaly may have been. (I will then, of course, attempt to "confirm" the deviation of that line, through extensive additional testing.)

Difference 2) This one may confuse people: The phase 2 experiments will be done without sharing of the detailed results.

"Why would you do this!? Isn't the whole point of these experiments to improve society's understanding of the subject? Why would you be concealing data!? This is the opposite of what you've spoken of previously!"

I'll tell you exactly why. Well actually, there are two major reasons:
1) To allow freeform, expansive experimentation with dozens of variations, without the concern of selective reporting when I (according to plan) do not publicly share the numerous, overlapping, multi-faceted results. In other words, I am pre-committing myself to not sharing the results; this lets me quickly iterate on ideas, without those numerous mini-experiments creating potential for selective-reporting/data-dredging/p-hacking. Those are real concerns, but it's also the go-to excuse people use for experiment results they "find implausible" (ie. "these results are surely just the best selection from a large set of previous attempts, and they're presenting it without telling us that, skewing the true significance"). This approach of committing to a pre-defined, minimal "Q/A set" for the results, solves both concerns. (My next post will contain the text specifying exactly what the set of questions will be, with each one having a pre-defined set of answer options, with future me required to select one and only one answer for each. Additional details are included in that text, explaining the rationale for the selected questions and answers, as well as protocols for handling various "gray areas" that might come up.)
2) I think it's possible that the strength of the psi effect for a given experiment, depends partially on what the experimenter's conditions/intentions are for sharing it. I'm aware that this will sound ridiculous, perhaps even delusional, to some; nonetheless, I consider psi "more likely than not" given the full set of data, and that it has evaded detailed understanding so far suggests that, if it is real, there are some additional factors we've been missing which need to be accounted for. So I want to try a set of experiments in complete "isolation mode", to see if that has some sort of an impact. (When there's something you don't understand, you don't keep staying within your existing assumptions, you expand them even to things that seem very unlikely -- if for no other reason than to loosen up your thought process to further possibilities. For example, when you've lost your keys/wallet, and you've looked every place you can reasonably imagine it to be placed, that often means you need to expand your search/thinking even to those "ridiculous" location possibilities. Many times you'll find it was placed in an odd place for a very specific reason, but you just couldn't think of it at the time, because those peculiar reasons were only active for that brief window of time.)

Even given those reasons, I can see how it would be annoying as a reader to not be able to see the raw results. But not to worry! After this "phase 2" isolated experimentation period, I plan to take what I learn from it, to construct the protocol for "phase 3" -- which is where I will do everything I can to make the experiment as fast and streamlined as it can be, for other people to either:
1) Perform the experiment themselves, once I've made the software public. (regarding hardware, it may be possible to make use of my existing HRNG over the internet; if not, the HRNG device is commercially available, and I can maybe help subsidize the costs for serious inquirers)
2) If it appears that the protocol is robust to (the hypothesized) intention/reporting effects, then I'll work to set up a nice online interface, letting users review all data obtained from the experiment, from users around the world. (well, for those who authorize the sharing anyway; though for that, we'd need some way to prevent selective reporting, perhaps by requiring users to select their setting before doing any experimentation)

Anyway, with that background explained, my following post (well, after one empty post, reserved for updates) will provide the draft of the "Q/A set", that I will eventually use (after I've completed my phase 2 experiments) to share (the predefined portion of) its results.

(Post reserved for future updates.)

(Frozen copy of this post located here: https://pastebin.com/ieBLu61T)

As described in my first post, the below is the draft of the full list of questions (and answers) which I will eventually be using to summarize the "phase 2" period's results. (I will make a post soon declaring its start/activation [at which point the protocol will be finalized], and it will continue until a corresponding "end of phase" post is made sometime in the future.)

Official question/answer set
==========
* How many "substantial learnings" occurred? (ie. observations or theory-advancements occurring during the experiment, which you consider significant steps forward in understanding the situation)
Options: [number]

* For the learnings that occurred, roughly what percentage of them were the result of "obvious lines of inquiry", ie. things that the majority of people attempting a similar experiment would, on their own, think of testing? [the answer may not exactly match a ratio of the count in question 1, due to some instances being "in-between" obvious and non-obvious]
Options: [percentage]

* Was there anything distinct, and very unexpected, that happened? (ie. beyond the isolation-effect you theorized, and a general psi effect, regardless of strength) [it is expected that this will be a no, but is included on the off-chance of some highly bizarre/specific sequence seen in the data]
Options: Yes, No, Unclear/Other

* For whatever patterns in the data seemed most plausible to be a psi effect, were those patterns roughly homogeneous, or rather, multi-stage? (ie. appearing to have distinct shifts in operation at certain points)
Options: Single-Stage, Multi-Stage, Unclear/Other

* Overall, how "interesting" were the results? Express this in terms of how many hours/days you'd be willing to spend isolated (in an empty-room/void, though comfortably and with external time paused), to unlock the ability to share it freely. [think your answer through carefully; it should match with what you'd actually assent to]
Options: [number of hours/days; max answer: "30+ days"]

* For the most interesting results obtained, what likelihood do you give that it would be replicable by a random other, if they sincerely attempted to follow the same protocol? (with isolation, same mindset, equivalent variants and duration, etc.) [this may be difficult to guess prior to any sort of external feedback, and not possible to directly confirm [due to externalization of the protocol being a gradual process, with possible unavoidable changes along the way], but do your best to guess anyway]
Options: [percentage]

Notes
==========
* The sharing restrictions are only for "in system" results, ie. the standard data that is recorded by the experiment software. "Out of system" observations are fine to share. For example, if the HRNG device starts glowing and emitting sparks that fries all the electronics in my bedroom (!), this would qualify as an "out of system" observation, which I would not be restricted from reporting. So rest assured, my lack of reporting such things does indeed imply that they did not happen. ^_^

* The intention of the experimentation period is to not share the in-system results with anyone, for the rest of my time on earth (for reasons explained in the main post above). That said, there is a part of me somewhat hesitant about making oaths of this sort, as they could theoretically lead to disruptive situations I didn't anticipate, making me somehow regret it. Because of this hesitance/abundance-of-caution, I offer the following "safety hatch" to my future self: If there is some aspect of the data or situation, such that the restriction of sharing causes distinct real-world negatives (excluding my regular motivations, eg. the general desire to advance understanding), then you are permitted to share the full data with up to three people (but no more). Furthermore, you must obtain a clear (and, as far as you can tell, sincere) statement from such people, stating that they intend to not subsequently share the data (or even "revealing derivations" of it) with a wider audience. This is a generic sort of safety hatch, which I intend to use for any future "oaths", to ease my overly-cautious personality, while also having minimal potential impact on the experiment itself.

Responses to future questions/concerns
==========
> What options are available for questions readers may have about the experiment, that were not answered in these original explanation posts?

Well, it's not a problem to ask them, even after the experiment has started; I just might not be able to answer in detail. However, a good portion of questions will still be able to be answered, because they deal purely with clarification on the "intent" I had prior to the experiment's start. Those questions are generally fine, because they don't reveal information about the results of the experiments. Occasional questions of that sort may be left unanswered though, if they are of a type that I'm concerned may lead to my answer being "tainted" by knowledge of the actual results (or highly specific details of methods utilized, as exact replication is intended for a later phase); but I think this would be rare, unless a person specifically designed it for that.

> Why are the questions/answers included left so vague? For example, why do you not include specific questions about the highest rarity-by-chance value obtained, the number of variations tried, etc?

I can see why this would seem frustrating or suspicious to readers, perhaps even seeming as though I'm intentionally trying to "induce mystery" around the experiment artificially. However, the main reason I don't include questions of that sort (with answers highly informative about the results and/or methods) is because for this "phase 2" experiment, one of the main things I'm trying to test, is on whether there is some sort of "intention/reporting effect" for psi experiments.

That psi has eluded detailed understanding for so long makes me think that, if it is real, we need to question deeper assumptions about how to perform experimentation on it -- and one of these assumptions (which I'm trying to test here) is the [generally left implicit] one that "Whatever effect is being studied will not be impacted by the reporting intentions of the researcher, and/or the societal impact likely to result from the experiment" (while far-fetched, there are some reasons I find this worth questioning -- some from inside the field and some from out). And I think a step in checking this assumption, involves preventing "overly informative windows" into the results of the experiment, through means of potential significance-signaling/worldview-disrupting question-and-answer sets.

Apologies! But I expect that I will be able to reduce these restrictions in future experiments, as I get a better handle on how much (if at all) they may be a factor.

> Taking a step back, does the overall premise of this "phase 2" experimentation not seem a bit delusional to you? Like, do you have no regard for the possibility that you've walked yourself into this odd malformation of standard protocol (which values full transparency) on account of your "psi effect" ultimately being illusory, rather than just "evasive" as you call it?

First of all, let me assure you that I value transparency very highly. I think it is tremendously valuable for readers to have complete information about an experiment, empowering them to perform replication attempts, look for errors the researcher may have made, etc. It goes against my regular nature/aspirations to be performing an experiment of this type, with results that will never get to be shared in detail with the wider world. I'm experimenting where I am only because: a) it avoids the concern of selective reporting with relation to the many variations tried [by pre-committing to the minimal sharing defined above], b) it's a response to the observation that the field of parapsychology has "stalled" somewhat for the last 50 years -- with some remarkable successes, some new developments (such as the Radin double-slit experiment), and *some* level of replicability (eg. as seen in the meta-studies of Ganzfield trials), but still failing to produce the "100% replicable experiment" that has long been sought. Rather than forming another "open-air experiment", I want to look into an area that I haven't seen much attention placed, relating to variations on the design process and intent for sharing the results.

Secondly, I think there is, in fact, a fair chance that the skeptics are right, and the "evasiveness" of psi is instead an artifact of the effect being illusory, and the striking results obtained at various points being due to a variety of traditional explanations that merely combined in hard to unravel ways. That said, there are many experiment results I've seen which just do not seem to fit into any of the traditional explanations (even after looking at them closely, and hearing out the skeptics' arguments), and so while it's possible a *combination* of those factors may ultimately explain it, the situation is far from solved/confirmed in my eyes. And overall, I think the evidence does tilt toward there being an actual effect, even after accounting for the lower "base plausibility" of the concept (though I also think this "base plausibility" is higher than traditionally assumed, due to most people not taking the time to research/construct theories which elegantly -- or at least semi-elegantly, at this point -- explain a large portion of the results [including observations outside of the field] in a consistent way).

In any case, I think it's worth at least testing the ideas. It's an example of the "low probability, high reward" concept. For me, the reward of finding a piece of the puzzle for better understanding psi, would be tremendously rewarding; so I consider it worth trying, even if realistically I will most likely not find anything earth-shattering (since, probably, someone has tried something like this before; that said, I haven't read of any such explicit tests yet, with the follow up intent of gradually reducing the restrictions to "find the point of psi diminishing", anyhow). Besides, I enjoy the process of experimentation inherently; so even if it ends up a dead end, I would not view it as a total waste, as even a negative result is interesting to confirm, and I will have had some fun in the process.

Time of activation
==========
I will leave the description above up for a few days, on the chance that someone has specific thoughts, questions, or criticisms to add. If you have some, please share! -- as in a few days I will make a post declaring the "official start" of phase 2, and I won't be able to make changes to the protocol after that point. (as explained, this is mostly done to avoid [understandable] concerns of selective/skewed reporting, from the potential changing of the information set to be shared, after having obtained partial results)

(Frozen copy of this post located here: https://pastebin.com/BanGfGWM)

The activation of phase 2 was delayed for a few weeks, as I decided to finish writing the code for the new version of the software first.

Well, the software's finally ready! Thus, in a few minutes, I will "officially start" the phase 2 experimentation period, at which point the "protocol" for it will be locked.

However, before I do that, I'm adding one final "addendum" to the protocol (as allowed for by its not-yet-active status). The addendum has less to do with "changes" to the protocol, as it does with "clarifications", relating to further questions/ambiguities that I anticipate may eventually come up.

Here is the addendum, in question/answer form:
==========
> Is it okay to repeat only a certain portion of the Q/A set, to a given audience?

Yes. (though no manipulative communication, eg. Morse code or the like)

> It it okay to style parts of the Q/A set (eg. adding bolding)?

Yes. (though generally I'd avoid this; and again, no manipulative communication using it)

> Is it okay to report only parts of the Q/A set, in later contexts?

Yes. (though I'd almost always provide a link to the full Q/A set -- with exceptions for very limited space, it already being shared, etc.)

> Is it okay to make statements that compare future results against those within phase 2?

Only to an extent; it's okay to make generic statements about results differing from those previously obtained or the like, but you cannot give details about what specific variants were the most promising, number-based approximations of the difference amount, etc.

Now why would I allow this? Partly it's because I know it would be super annoying if I could not even express frustration at results not matching ones found previously. But the more important reason is that it's unavoidable that information of this sort would leak out, through body language, etc. Thus it's better to make the boundary between allowed and disallowed expression explicit, rather than having the continuous mental tension of knowing the results don't match, but not knowing how much can be said (as defined by the protocol -- which, to be clear, includes this addendum, since it's being written and made public prior to phase 2's activation).

> Is it okay to share the answer to questions in the Q/A set prior to phase 2's completion?

For the most part, no. However, in the interests of being able to share *some* information, I looked through the list, and question #1 appears fine to share preliminary results of. (since it only gives the "number of learnings", not information on the details of the variants, strength of the results, etc.)

Thus, I hereby give my future self permission to share the preliminary answer to question 1 as desired, even prior to phase 2's completion. (as explained earlier, this protocol change/clarification is only possible because phase 2 hasn't started yet)

> What degree of expression is acceptable when the subject comes up privately? (ie. among family/friends)

Something similar to the above: It's okay to express high-level emotions on the experiment, but it's not okay to provide quantifiable estimates of things, nor to confirm/deny specific details that would not be answerable from your pre-experiment thoughts.

Examples of things that are okay:
1) Generic expressions of excitement, surprise, etc., in response to something you found in your experiment. (the contents of which you can't share, of course)
2) Responding to someone asking about general things, like how much time you're spending on the experiment recently -- for which responses like the following are okay: "Quite a lot. I had a new idea last night that I'll be starting this evening after coding it."

Examples of things that are not okay:
1) Giving a numerical estimate of the ratio of hours spent each week currently, compared to a concrete point in time (eg. the start, the peak).
2) Responding to a suggestion someone has, saying that you've already tried it out and confirmed/rejected it within the experiment.

> What degree of expression is acceptable after the experiment is done, and people have questions about the design, eg. for attempts at replication?

My general inclination for this, is just to facilitate "conceptual replication", by distilling the solid elements from phase 2 into a new design (with some additional ideas and features) that anyone can attempt. This avoids a couple issues (mentioned below), as well as potentially helping avoid the "decline effect" that seems to occur sometimes with exact replications (presumably because of less novelty/excitement, but there are some other theories for it as well).

That said, there may still be questions people have regarding the design, if those attempted phase 3 designs do not yield similar results.

In that case, I suppose it would be best to share a closer copy of the codebase used in phase 2, in the interests of those trying to perform replications. So I give my future self permission for this.

However:
1) Place care on what schemes/variants are shared; some of these may be overly specific / formed in response to data obtained during the experiment. It is fine, at least, to share the list of schemes/variants that were present when the experiment was started; and ones that were created after that are fine as long as they're not overly specific / revealing of experiment subresults (for those that are too specific, you can perhaps give a high-level summary of them).
2) Place care on details within the codebase that may also be too specific / tied to the experiment's results -- for example, code comments (referencing real-world feedback) explaining why a code change was made, the conditional code run for those variants, etc. Like the above, it is fine to share the codebase as it existed when the experiment started; but changes to the code after that depends on how tied to the results they were.

I understand if this point is confusing to people; it may seem like I'm being overly restrictive in the sharing of things constructed mid-experiment. However, I'm trying to be cautious, because one of the main points of this phase 2 experiment is to test the idea that the "social impact"/"remote interactions" of an experiment plays a role in how successful the results are. And every bit of information shared that exposes details of the experiment's results/data, weakens the theorized "isolation" that I'm trying to produce around it. Yet, I'm too greedy to completely isolate the results (ie. never tell anyone about it at all) -- so I go for a compromise, where I share some information, but "not too much". And being cautious, it's possible I err too much on that side.

In any case, I'm hoping this restrictive period will only be temporary: that I'll find an approach that works well regardless of the isolation level (or at least, one that can quantify the effects of isolation levels, and thus find a compromise that allows for the results to still be attempted by many others, and thus accumulate a database that others can analyze).

Until that point (where I do my best to make the experiment easily runnable by anyone else), I choose to err on the side of caution/gratuitous-isolation.
==========

And that's it! The above is the third and final section of the protocol for my phase 2 psi experiments.

Now comes the time to "activate" it; this is done by the short statement below. (yes, this is surely viewed as over-dramatic by some; but it's meaningful to me ^_^)

Oath content
==========
I (Stephen Wicklund) hereby declare my commitment (effective at the point of posting on the DreamViews forum) to follow the "limited sharing" protocol described above (more precisely, the meaning behind it -- which I did my best to put into writing), for the purpose of creating a context in which -- perhaps -- further understanding of the apparent mind-matter interaction effects we call "psi" can be obtained.
==========

Time to get to work, and figure this thing out. :)