Discovery Lab Answers

The premise, first

“Strongest evidence base” is a comparative judgment, and ESP-Nexus takes no position on whether any psi effect is real — so what follows is a description of which lines of research have accumulated the most systematic, meta-analytic attention in the sources retrieved here, not a verdict that any of them is established.

Two caveats that shape everything below:

• This is a small, non-random slice of the literature. The sources retrieved for this question cover free-response ESP meta-analysis, after-death-communication survey research, and a methodological commentary on replication. That is not a measured survey of the field, and the share of the published psi literature ESP-Nexus holds on this comparative question has not been measured here. Treat the picture below as what these few sources show, not as a ranking of the field.
• The metrics are not comparable. A standardized effect size from free-response ESP studies, an incidence/prevalence percentage from bereavement surveys, and a commentary on meta-analytic method are different kinds of quantity. They cannot be pooled into a single “strongest” number.

What the retrieved sources actually show

Free-response ESP (ganzfeld, remote viewing, dream, precognition)

The most quantitatively developed line in these sources is the free-response anomalous-cognition literature. Storm and Tressoldi (2020), reviewing free-response studies from 2009–2018, report mean effect sizes broken out by modality: telepathy mean ES = .074, clairvoyance mean ES = .038, and precognition mean ES = .079 ^[2]. Precognition produced the numerically largest figure, but they state the difference among the three modalities was not statistically significant (F(2, 22) = 1.74, p = .200) — so the data in that paper do not support ranking one psi modality above another ^[2].

This body of work is the one with a decades-long meta-analytic tradition — Storm and Tressoldi explicitly frame their analysis as a ten-year update to a “noise-reduction model” line running back through earlier Storm et al. work and the Bem & Honorton (1994) ganzfeld results they cite ^[2].

After-death communication (ADC)

The other quantitatively substantial source is a different kind of evidence. Streit-Horn, Holden, and Smith (2022) systematically reviewed 35 studies of spontaneous after-death-communication experiences, ranking them by methodological quality (mean quality score 31.51, range 23.25 to 41.25) ^[1]. Their headline finding is about incidence, not an ESP effect size: within a few months of a loved one’s death about three-quarters of bereaved people report an ADC, dropping to about half after a year ^[1]. That is a prevalence pattern in spontaneous reports — a fundamentally different measure from a lab effect size, and not something that can be placed on the same scale as the Storm–Tressoldi figures.

The methodological caution

Morris (1991), commenting in Statistical Science, is not a study but a methodological critique bearing directly on the “strongest evidence” question — he raises how effect size is estimated in meta-analysis, noting that a measure like Cohen’s h can badly underestimate effect size for very low-probability events ^[3]. This is the kind of internal-to-the-field caution that any claim about which phenomenon has the “strongest” meta-analytic support has to sit alongside.

Bottom line

Within the sources retrieved here, the free-response ESP literature (ganzfeld and related) is the line with the most developed meta-analytic infrastructure and small, replicated-across-updates effect sizes ^[2]; after-death-communication research offers large incidence figures but on an entirely different metric ^[1]; and Morris (1991) is a reminder that how those effect sizes are computed is itself contested ^[3]. No single “strongest phenomenon” ranking follows from these three, and no significant difference among ESP modalities appears in the one paper that tested it ^[2].

For the site’s own treatment of how consistently these effects reproduce, see Replication in Psi Research: Behavioral Science, and the topic pages under Phenomena.