Understanding the Factors at Play in the Sender-Receiver Dynamic During Telepathy Ganzfeld: A Meta-Analysis 1

: Objective . To use meta-analysis to explore five previously uninvestigated factors related to the sender-receiver dynamic in the telepathy ganzfeld. The five factors of interest are: a) did the receiver see the sender’s room prior to the session?; b) could the sender hear the receiver during the mentation period?; c) could the sender hear the receiver during the judging period?; d) was the sender explicitly told to be silent?; and e) did the experimenter assist in the review section of the session? Method : Telepathy ganzfeld studies conducted post Joint Communiqué , with one session per day and the receivers rating the targets, were chosen. Two mixed-effects models were fit: 1) using the study hit rates as the binomial mean; and 2) using the study hit rates as a proportion. Both models have the five factors as binary moderators. Results : Both the binomial mean and proportion models suggest a significant effect of the moderators overall and two factors individually: 1) the sender being able to hear the receiver during the mentation period; and 2) a review period after the mentation period. Permutation tests for both models also show significant effects of the moderators and the two factors. Conclusion : The sender being able to hear the receiver’s mentation appears to increase overall study success, while the review period decreases overall study success.


Highlights
• There was a significant overall effect of the moderators on study success (hit rate) • The sender being able to hear the mentation period was associated with a significant increase in study success • A review period after the sending period was associated with a significant decrease in study succes 1 Address correspondence to: Abby Pooley, Department of Psychology, School of Philosophy, Psychology and Language Sciences, 7 George Square, University of Edinburgh, EH8 9JZ, United Kingdom, apooley@ed.ac.uk.
Preliminary results were presented at the Society for Psychical Research's 2021 Annual Conference.The analysis discussed at the conference only contained one analysis that used the hit rate as the outcome variable and based upon the ratings by the first author, before inter-rater disagreements could be resolved.The current report details the final methods, analyses and results.The authors would like to thank Umberto Noé for assisting with the standard deviation formula for the Binomial approximation distribution, and rater Stacey Bruce.
Caroline Watt thanks the Bial Foundation for Grant 190/18 to support her ganzfeld work.abundant evidence that the ganzfeld creates a psi-conducive environment, we need to be more systematic in investigating which elements of the ganzfeld procedure are important.Therefore, the focus of this meta-analysis is to explore how five previously uninvestigated ganzfeld telepathy study design features may influence study outcome.

Factors of Interest
Factor 1: Do Receivers See the Sender's Room?
Psi-conduciveness is often mentioned in the ganzfeld literature, with some stating that creating a warm and pleasant atmosphere creates a more psi-conducive session (Dalton, 1997;Milton, 1997).However, more information is required about the detailed protocol of each study, especially concerning experimenter-participant interactions, which are quite extensive in the ganzfeld due to the one-to-one nature of the testing, and the duration of each session (Smith & Savva, 2004).Ganzfeld researchers often mention rapport building chats between experimenter and participants, however it is not known whether having the receiver see the sender's room before the session is a key aspect of the study.Being told there is a sender in a different room may be unnerving to a new participant, especially given the length and intensity of the study.Furthermore, perhaps any emotional or social connection between the sender and receiver will be stronger if the receiver has been introduced to the sender's environment and the sender, in turn, is aware that the receiver knows where they are.Thus, the rationale for assessing this factor is to understand whether scoring is higher when the receiver has seen the sender's room before the session commences.

Factor 2: Do Senders Hear the Receiver's Mentation Live?
It takes some effort to set up a one-way audio connection from the ganzfeld receiver to the sender.Researchers often make this effort thereby enabling the sender to hear the receiver describe out loud their feelings, impressions and sensations during the 2023, Vol. 3, No. 1, pp. 42-77 the impression period, a part of the session typically referred to as the review period.After the impression (and review) period, the receiver views a random selection of decoy video/image clips, along with the target (the clip the sender was aiming to communicate).While the receiver and experimenter remain unaware of the identity of the actual target, the receiver ranks the similarity of their impressions with the presented targets.If the highest rated target is the same as the target that the sender was viewing the session is registered as a hit.
Parapsychologists have been reporting significant results in ganzfeld studies since the 1970s, however there has been little systematic investigation of which aspects of the experimental set-up are associated with elevated hit-rates.Ganzfeld design features such as target type have been analyed, with dynamic targets producing larger study effect sizes (Honorton et al., 1990), though this observation was not confirmed in Milton and Wiseman's (1999) analysis of "new generation" ganzfeld studies.Honorton (1977) reported that successful sessions have on average 37 minutes of ganzfeld exposure.Bem and colleagues (2001) found that more standard studies obtained higher hit-rates, although there is little consensus on the definition of the standard ganzfeld (Milton, 1999;Schmeidler & Edge, 1999).
Most attention has been paid to the role of the sender, with studies hoping to shed light on the sender's influence and whether it is instrumental (inherent to the communication process) or peripheral (pertaining to psychological or motivational factors).Honorton's (1995) meta-analysis of the ganzfeld literature reported that studies using senders perform better than those without.However, later studies designed directly to compare sender and no sender conditions generally report no significant difference between conditions (e.g.Morris et al., 1995;Roe & Holt, 2005;Roe et al., 2003).Other potentially important aspects of the ganzfeld set-up, especially around the senderreceiver dynamic, remain unexplored.As Cardeña (2020)  been used in study protocols throughout the ganzfeld literature (e.g., Berger & Honorton, 1986;Honorton et al., 1990).Thus, assessing the prevalence of this factor may help us to evaluate whether subtle sensory leakage (such as vibrations) may potentially influence study outcome, even if acoustic shielding in these studies is assumed to be adequate (see Wiseman et al., 1994).

Factor 5: Mentation Review
It is common for ganzfeld studies to have review periods (after the sending and before the judging period) in which the experimenter reviews the mentation notes and allows the receiver to elaborate or clarify their mentation (Kanthamani & Broughton, 1994;Roe et al., 2004;Watt et al., 2020).The review period may assist participants in processing their experiences, remembering their mentation and in making connections between their mentation and the targets that they may have otherwise not noticed (Wooffitt, 2003).However, there has not yet been a systematic review of the importance, or otherwise, of the mentation review.

Objective
This meta-analysis is exploratory as there has been no previous systematic review of the above aspects of the ganzfeld study procedure.The research questions originate from a pragmatic motivation: to provide evidence to guide the design of future ganzfeld telepathy studies.Hence, there are no expectations from the analysis.Nonetheless, the null hypothesis is that there will be no effect of the moderators (five factors) on study success (hits significantly greater than chance).Ganzfeld telepathy studies published between January 1988 and September 2021 are included, to assess studies conducted with the potential benefit of the methodological guidelines from the Joint Communiqué (Hyman & Honorton, 1986). 2023, Vol. 3, No. 1, pp. 42-77 sending period.The stated justification for this design feature is to allow the sender to mentally reinforce the correct images and impressions to the target (Dalton, 1997) and to add an air of excitement and active involvement for the participants (Parker et al., 1997).
If the sender's influence is peripheral via motivational factors (as suggested by Honorton, 1995), then the sender being able to hear the receiver during the sending period may be the most important aspect of a telepathy design because by hearing the receiver's mentation the sender should, theoretically, be able to reinforce receivers at times when they seem to be describing the target.Further, not every ganzfeld study features this audio channel (or perhaps does not clearly report this aspect of procedure), so it is also of interest to establish how common this practice is.If the sender can hear the receiver produce their mentation during the sending period and reinforce impressions linked to the target (Factor 2), the same logic extends to the judging period -the time when the receiver decides which target clip is most like their experiences.As the receiver (and/or the experimenter) are reviewing the mentation report and making decisions about the ratings, the sender being able to hear this should, in theory, be able to mentally reinforce the target.Nevertheless, to date there has been no analysis of the impact of the sender hearing the judging period.

Factor 4: Sender Told to be Silent
This factor is primarily assessed from a target security concern: even if senders are physically distanced from the other experimenters and receiver and the target is shielded, explicitly telling the sender to be silent provides an extra layer of security.
Although a minor aspect of the ganzfeld procedure, if this factor is significantly related to study outcome then it suggests that previous studies may have been susceptible to sensory leakage.Instructions for the sender to "silently communicate" the target have

Information Sources
We first extracted all telepathy studies from Tressoldi's ganzfeld database, accessible at the Society for Psychical Research's open-access data website Psi Open Data (Tressoldi, 2019), which contains all studies conducted since 1974 to 2018, used for the two recent meta-analyses of the ganzfeld literature (Storm & Tressoldi, 2020;Storm et al., 2010).
To check for studies produced since 2018, a literature search on Google Scholar was conducted, using the search terms "ganzfeld," "telepathy," and "study," using the Boolean connector "&" in the title and abstract fields for the years 2018 to 2021.Inspection of reference lists of included papers was also used as a part of the search strategy to ensure all relevant studies were included.The literature search resulted in one addition, a telepathy study published in 2020 (Cardeña & Marcusson-Clavertz, 2020).ALP contacted the author of the 2020 paper to establish if the multiple sessions had been performed on a single day by the participants -which they had not.

Study Selection
The study selection procedure is outlined in Appendix Figure C1.Studies were excluded based on these criteria: • Duplicated: for example, if a published paper was produced from a conference proceeding, the conference proceeding was removed from the database.Published papers have more detailed and full analyses and usually all planned sessions are complete.
• Studies using external judges: the factors Hear and Hear judging asses if the sender hearing the receiver during the two periods (outlined above) influences study outcome.Thus, using external judges would not help this assessment, especially for the Hear judging factor (see Hyman, 1995).
• Multiple sessions a day: if the studies had a repeated participants design which ran multiple sessions on the same day, they were removed.Participants contributing to multiple sessions violates the independence assumption of most statistical hypothesis tests.Likewise, there is literature noting a decline effect due to fatigue in experimenters (Broughton & Alexander, 1997;Parker et al., 1998;Wezelman & Bierman, 1997) 2023, Vol. 3, No. 1, pp. 42-77 The independent variables are the five factors (detailed above) rated by two raters (details in Appendix Figure C2).The dependent variable for both meta-analytic models is the study hit rate (percentage of hits).The first author created two models, the first treating the study hit rate as a mean, following the approximated binomial distribution.
The intention was to use the z-scores for each study, which are approximated from the binomial distribution, but the standard deviations could not be computed.The model effect size metric is defined in Appendix Formula A6.The second, supplementary, model uses the study hit rate as proportion and uses the effect size metric defined in Appendix Formula A8.Homogeneity analysis was automatically calculated by the model function, which calculated the I 2 value (See Appendix Formula A3).The I 2 statistic describes the variation across studies due to heterogeneity, rather than chance and is a simple description of the inconsistency of studies' results (Higgins & Thompson, 2002;Higgins et al., 2003).Analyses were conducted with RStudio Workbench Version 1.4.1717-3© 2009-2021 RStudio, PBC, and models built with the metafor package (Viechtbauer, 2010).
A protocol was not pre-registered for this project due to its exploratory nature.The data set used for analysis, analysis code, and information are all publicly and freely available at https://github.com/yeloopa/telepathyMA.git

Method Inclusion and Exclusion Criteria
We include all ganzfeld telepathy studies using visual targets and human participants from January 1988 to September 2021.All studies had a measure of hit rate (%) as well as session binomial z-score as their outcomes, and a four-option design with one target and 3 decoys (thus resulting in a 25% mean chance expectation; MCE).Study hit rate is a percentage calculated as the overall number of hits obtained across the study sessions (see Appendix Formula A2).The associated z-scores are the related binomial distribution z-ratio for situations of the general "k out of n" type with the formula as defined in Appendix Formula A1.

P A G E 5 0 P A G E 5 1
Journal of Anomalous Experience and Cognition (JAEX)

• Number of participants • Number of trials
Studies were then organized according to the five factors of interest: • Did the participant (receiver) see the sender's room before the session?(See) • Could the sender hear the receiver produce their mentation?(Hear) • Could the sender hear the receiver during the judging period?(Hear judging) • Was the sender explicitly told to be silent?(Silent) • Did the experimenter review the receiver's mentation notes with the receiver, after the sending period?(Review) Each factor is rated on a binary scale (0 for no presence, 1 for presence of factor).
See Appendix Figure C2 for the instructions given to the raters.For the five factors, the first author first assessed each paper and provided ratings based on the instructions.A second rater then did the same, following the same instructions.However, because of health issues with the second rater, it was not possible to arrange a meeting to resolve the discrepancies in ratings and those ratings were disregarded, so CW was recruited as a third rater.Discrepancies in ratings between Raters 1 (ALP) and 3 (CW) were resolved in a meeting (see Table 1 for the inter-rater reliability scores).Due to the ambiguity of what constitutes a "review period," ALP and CW agreed in the rating meeting that if the study report explicitly stated a review occurred, then it was rated a 1.If a review period was not explicitly stated in the paper, it was up to the rater to decide if a review stage could be inferred: hence the rating instructions included the opportunity for the receiver to add, alter and/or discuss their mentation with the experimenter (Appendix Figure C2).• Multiple trials per session: like the point above, only one trial per session in the current study design.
• Stimuli material: Studies which used non-visual targets were removed.Visual targets are considered standard (specifically dynamic targets; Bem et al., 2001).
• Multiple or no senders: Some studies included designs that involved 0, 1, or 2 senders.In these reports, it was unclear if the analysis combined all of these trials into one analysis, or the different sender options had low sample sizes (< 10).
• Low sample size: studies with samples of 10 or less were removed because of potential bias in results stemming from sampling error.
The first author conducted an initial analysis presented at the Society for Psychical Research's 2021 Conference (Pooley, 2021) using the data set composed of the studies rated by her using the study hit rate as the outcome measure.She stated then that this was not the final analysis, with corrections and changes still to be made, such as resolving the rater discrepancies and analysis with z-score as an outcome.After the conference and study selection was assumed to be final, numerous reports in the Journal of Scientific Exploration reported the serious fraudulent actions and widespread plagiarism conducted by Alejandro Parra (Braude, 2021;Cardeña, 2021;Nahm, 2021).
Because of the seriousness of the accusations and evidence collected, we deemed it best to remove all the Parra studies from the dataset (a total of 5 data points).Likewise, there was duplicate reporting of some results in the Gothenburg study series (Parker & Westerlund, 1998;Parker et al., 1997).Given the removals and corrections to the database, the final data set is 41 studies.

Data Extraction and Coding
As the data are primarily sourced from a freely available ganzfeld database (Tressoldi, 2019), the number of variables of interest for experiments were reduced to: • Study author(s) and year (and series number if multiple series per study) • Study hit rate (%; see Appendix Formula A2) • Study z-score (see Appendix Formula A1)

P A G E 5 2 P A G E 5 3
Journal of Anomalous Experience and Cognition (JAEX) number of trials are required to calculate the appropriate effect size.We provide the formulae for the model in Appendix Formulas A7 and A8.Due to the similarity between the primary and secondary models, we report the primary analysis in full in the Results section with the model prior to the removal of outliers.The secondary model results can be found in full in Appendix A1-A4.

Methods of Synthesis
Because of the exploratory nature of the study and lack of previous relevant research, a meta-regression for both models was created.Meta-regression not only provides a summary of the selected studies but also evaluates how the five potential moderators may influence study outcome.Study results are not weighted.

Publication Bias and Selective Reporting
Funnel plots for each analysis were created (a funnel plot is a useful visual aid to assess publication (and other) potential bias in the database).Because of the inclusion of moderators in the models, we could not perform the trim-and-fill method to assess publication bias.However, multiple reviews of the ganzfeld literature have reported no suggestion of publication bias problem (Baptista & Derakhshani, 2014;Cardeña, 2018;Storm et al., 2010).

Inter-Rater Reliability
For the sake of transparency, Table 1 presents the results of the initial ratings between Raters 1 and 3.These disagreements were resolved before the final analyses reported below (also see Appendix Figure C1).

Primary Analysis: Binomial Test with Mean Number of Hits
For the primary analysis we intended to use the z-score as the outcome measure.
However, we noted that a recent pre-publication of a ganzfeld meta-analysis has been criticized during open peer review for the effect size calculation for the study z-scores as not being scientifically sound (see Tressoldi & Storm, 2021).Thus, for the current analysis, ALP performed the Binomial test using the mean number of hits rather than the total number of hits, using a random-effects model with study hit rate treated as a mean.The five factors (all binary) of interest were added as moderators in the model, thus resulting in a mixed-effects model.The final model is as follows: First, the number of trials, study hit rate and the associated standard deviation of the binomial distribution for treating the hit rates as a mean were calculated and entered into the escalc function, which resulted in the observed effect sizes and sampling variances in order to fit the meta-analytic model (see Appendix Formula A5 and A6, respectively).We estimated the heterogeneity of the effect sizes by fitting the model with a restricted maximum-likelihood (REML), which is better when working with smaller samples (Viechtbauer, 2005), and using the Knapp-Hartung adjustment (Knapp & Hartung, 2003).

Secondary Analysis: Proportion of Hits
For this model, we used the proportion of hits as the outcome measure.As the meta-analysis used aggregate scores that provide data about individual groups in respect to a dichotomous dependent variable (hit or miss), the number of events and

Table 1
Unweighted Kappa Scores and Agreement Between Raters 1 and 3

Descriptive Statistics
Appendix Table C3 details the studies included in the models and the respective measures.A total of 41 studies (or series reported as part of a wider study) were conducted by 17 different lead authors who reported their results in a total of 23 articles.
All 41 studies used a four-choice design therefore the mean chance expectation is 25%.A total of 1,496 participants contributed to 1,624 ganzfeld telepathy sessions.The average hit rate across the studies is 32% (SD = 10.40%) with a skewness of 0.72.An exact binomial test is reported to assess if the mean hit rate in the data set is significantly greater than chance.There are 520 hits in 1,624 trials, resulting in a significant difference from chance at the 5% significance level: Binomial Exact p <.001, one-tailed.The mean z-score is 0.91 (SD = 1.37), the sum of z-scores is 37.20 and Stouffer's Z is 5.81, although, as discussed earlier, the z scores were not used in the models reported.

Results of Binomial Mean Model (Model 1)
The model was first fit with all 41 studies.The model output is reported in Table 2.The test for moderators is non-significant at the 5% level F(5, 35) = 1.89, p = .12.However, the coefficient Review is significant suggesting that the presence of a review session in a ganzfeld telepathy study decreases the average study success by 12%, when all the other study features are set to 0. See  The estimated amount of residual heterogeneity for the model is very low (τ 2 <.001, SE = 0.001), unaccounted variability is also low (I 2 = 2.8%) and total amount of heterogeneity accounted for by the model is very high (R 2 = 92%).Model funnel and forest plots are shown in Figures 3 and 4, respectively.
Because of the inclusion of moderators in the model, the trim-and-fill method could not be performed (the trim-and-fill method was applied to the binomial mean model without the moderators, using the original dataset and the dataset that had influential studies removed.Trim-and-fill analyses did not estimate missing studies).However, as shown in Figure 3, there is little evidence to suggest publication bias in the data set.
Furthermore, a mixed-effects meta-regression model Egger's regression test for funnel plot for asymmetry was performed and was non-significant t(31) = 0.17, p = .87.

Binomial Mean Model with Review Factor Removed (Model 1.2)
Because of the high incidence rate of the Review factor (only 3 studies were rated to have no identifiable review period), the Binomial final mean model (Model 1.1) was performed with the Review factor removed to assess if the model results changed with this factor removed.The model formula is the same as the primary analysis, just with this With the three influential cases removed, the test for moderators is significant at the 5% level F(5,32) = 3.78, p = .01.As in Model 1, Review is significant and Hear now is also significant, as shown in Table 3.On average, study success rate is increased by 7% when the sender can hear the mentation period, when all other factors are set to 0. However, the addition of a review period in a ganzfeld telepathy study decreases the average study success by 10% when all other factors are set to 0. A permutation test (5000 iterations) was performed and confirmed the findings, with the test for moderators significant F(5,32) = 3.78, p = .016.The factors Hear (Factor 2) and Review (Factor 5) were significant, as shown in allowed the sender to hear the receiver's mentation (Factor 2) were associated with an increase in study hit rate by approximately 7%.This suggests that the sender hearing the receiver may motivate them to actively reinforce the target during the sending period and keep the receiver "on track."An alternative interpretation arises from work by Fox (2004), who found senders are susceptible to wandering and boredom, suggesting that perhaps an audio link is sufficient reinforcement to keep the sender engaged and motivated throughout the session.
In contrast, the inclusion of a mentation review period (Factor 5) was associated with a significant decrease in study hit rate by 10%.This supports Dalton's observations her that the experimenter-receiver interaction during the review period decreases study success (Dalton, 1997).However, this finding is based upon only three studies (after influential cases were removed) that the raters judged did not have an identifiable review period.Nonetheless, perhaps the review period with the experimenter and receiver discussing correspondences between the mentation and clips does not

Figure 2
Funnel Plot for Model 1.1

Discussion
In this study, we used a meta-analysis to explore how five potentially important aspects of ganzfeld telepathy study procedure might be associated with study outcome., and found two factors that had significant impacts on study outcome.First, studies that more clearly defined review periods where the receiver had to explicitly elaborate their mentation and experiences during the sending period.This a limitation with the rating instructions due to the broad criteria for the Review factor.
One weakness of the current analysis is that the studies are not rated or weighted in terms of their study quality.For the current report, only studies produced years (1988 onwards) after the Joint Communiqué (Hyman & Honorton, 1986) were included, in an attempt to exclude earlier studies that did not have the benefit of Hyman and Honorton's methodological recommendations.Nonetheless, studies with better security protocols and clearer method sections could have been given more weight in our meta-analysis.
Even are still some valuable findings from our analysis.First, this metaanalytic review shows that the results found in both models (Model 1.1 and Model 2) were confirmed in the permutation tests of the model coefficients.The omnibus test for the moderators shows that there was a significant effect of the factors, with the Hear and Review factors significantly affecting study outcome.Second, even when using different outcome and effect size measures than previous meta-analyses (Storm & Tressoldi, 2020;Storm et al., 2010), the results previously reported in the literature still stand: the noise-reducing ganzfeld protocol significantly produces hit rates greater than MCE.This is not surprising as the studies we used were primarily extracted from the same database; however, there is no need to account for selected participants in the current analysis as heterogeneity is low in both models, unlike what Storm and colleagues (2010) found.Third, although our study is limited in its generalizability, it has provided a new angle to look at the telepathy ganzfeld literature and can perhaps aid in providing an evidence-based procedure for future ganzfeld telepathy studies.The vast majority of reported ganzfeld telepathy studies have a review period, even though it appears from our analysis to be detrimental to study outcome.Likewise, the difficulty we had in coding some study reports should encourage future researchers to provide more detail when ganzfeld telepathy literature (Wiseman et al., 1994), but this factor is an indirect measure of sensory leakage.
One caveat to our finding that the mentation review is associated with lower hitrates, was that the rating instructions did not distinguish between a review period which was stipulated in the procedure in order for receivers to clarify or elaborate upon their mentation, but a simple opportunity for the receiver to add, comment upon, or discuss their mentation period.This is in contrast to a scenario where the receiver moves immediately from the sending/mentation period to the judging period.Hence, there is scope for future research to look more closely at the researcher-receiver relationship and to investigate how interactions during the session may influence study outcome.Likewise, the inter-rater reliability (kappa) scores are noticeably varied: Factor 5 (Review) was significant in all models, but this factor had the lowest kappa values.This illustrates a limitation with kappa values: kappa is not reliable for rare observations and low values of kappa may not necessarily reflect low agreement overall (Viera & Garrett, 2005), and given that there were four studies rated as not having a review period, the kappa value is unsurprisingly low for this factor.During the meeting to resolve disagreements, both raters discussed their ratings.Some discrepancies were merely mistakes whereas others were different interpretations.For example, for the Hear judging factor, it was decided during the rating meeting that if (after the receiver had logged their ratings) the sender was summoned via the one-way audio link then this implied that the sender could hear the whole judging procedure (given they could already hear the mentation period).
Hence, the ambiguity of the phrasing used by the authors in their study methods allows for different interpretations of the study designs and we recommend future ganzfeld researchers take care to give a comprehensive description of possibly important aspects of study procedure.Similarly, with the Review factor, the rater clarified during their meeting that if there was an opportunity for the receiver to add, comment, or discuss their mentation notes with the researcher then this was rated as a review period, as were the Journal of Anomalous Experience and Cognition (JAEX)

Figure B4
Funnel Plot for Model 2

Appendix B Table B1
Model 2: Proportion of Hits Summary with same 3 Outliers Removed as Model 1.1

Figure D2
Rating Instructions For each of the papers your task is to assess if they have certain characteristics present.If the characteristic is present then give it a 1, if not then a 0. Give a 1 if the characteristics are explicitly stated.
There are 5 characteristics to assess: 1. Did the receiver see the sender's room?a.
Some papers may say that both participants were shown the whole operation.

2.
Did the sender hear the receiver produce their mentation (verbal report)?

3.
Did the sender hear the receiver during the judging procedure?

4.
Was the sender explicitly told to be a.shouting/loud noises from sender's room would abort the session.

5.
Did the experimenter review/allow additions to the mentation notes with the receiver, after the sending period?a.
Some papers say they review the notes with the participant and allow for additions/changes.Use your judgement and common sense to assess if these characteristics were present, some will require re-reading and thinking but the main question is, "Is this clearly stated?Would I able to run the exact same procedure given the detail in this paper?"

Important Notes:
-Some shorter papers refer to other, already published papers and their procedures.Unless the authors state there were specific deviations from the previous design, you can give them the same ratings.
-Also, be careful for footnotes and procedural information outside of the 'Methods' section.It's worth skim reading all sections.Information about the study design may also be in the 'Participants'/'Procedure'/'Lab set-up'/'Design' parts of the paper (depending on how the paper is formatted).
-Some papers have multiple studies in them -you will be given a list of the studies of interest.However, this might require you to distinguish any differences in the procedures between the series, so it may take some deeper reading.

Factor 3 :
Do Senders Hear the Receiver During the Judging Period?
2023, Vol. 3, No. 1, pp. 42-77 2023, Vol. 3, No. 1, pp. 42-77 Journal of Anomalous Experience and Cognition (JAEX) 003.As shown in Figure1, the funnel plot suggests that there is publication bias in the dataset.A mixed-effects metaregression model Egger's regression test for funnel plot asymmetry was performed and was significant t(34) = 2.15, p = .04.Further assessment of the model revealed there were influential studies in the dataset.First, Honorton 302(Honorton et al., 1990) was removed due to having the highest standardized residual value of 3.41, where standardized residuals between -2 and 2 are commonly used as acceptable limits.The model with Honorton 302 removed still flags an influential case,Goulding et al. (2004)  with a standardized residual value of -2.7.With the removal of Honorton study 302 and Goulding et al. (2004), the model checks flag another influential study exceeding the limit: Broughton and Alexander (1997) FT2 has a standardized residual value of -2.7.The model was run once again and checked for influential values and returned no more influential cases, thus we detail the final model below.2023,Vol. 3, No. 1, pp. 42-77 Figure 1Funnel Plot for Model 1 Before Removal of Influential Studies provide clarity but rather introduces ambiguity and could allow the experimenter to direct the receiver away from the target.Alternatively, perhaps the review period introduces confusion about who (experimenter or receiver) is making the final judgement.In support of this latter interpretation, Model 1.2 with the Review factor (Factor 5) removed resulted in Hear (Factor 2) becoming more significant than in Models 1.1 and 2 and Hear judging (Factor 3) becoming significant (as shown in Appendix B).Our analysis revealed two factors to be clearly unrelated to study outcome.Factor 1 (See) was not significant in any model nor near significance at any point, suggesting that knowledge of the sender's location may not have a major influence over telepathy study success.Likewise, the non-significance of Factor 4 (Silent) may suggest that sensory leakage is less of a concern than has been suggested for some of the early2023, Vol. 3, No. 1, pp. 42-77    factor removed.Given the similarity between the Binomial mean model (Model 1.1) and Proportion of hits model (Model 2), only the Binomial mean model was performed.Full results are reported in Appendix B5, B6, and B7.

Figure B3 ForestP
Figure B3Forest Plot for Model 2 with same 3 Outliers Removed as Model 1.1 argues, although there is Journal of Anomalous Experience and Cognition (JAEX)

Table 2
Model 1: Hit Rate as Binomial Mean Summary Prior to Influential Studies RemovedThe unaccounted variability in the model is moderate (I 2 = 43%) and the QE test for residual heterogeneity is significant QE(35) = 62.38, p = .

Table 3
Model 1.1: Hit Rate as Binomial Mean Summary Output

Table 4 .
The forest plot for the final model (Model 1.1) is shown in Appendix Figure A1.
P A G E 5 8 P A G E 5 9Journal of Anomalous Experience and Cognition (JAEX)