A peer-reviewed paper published in early 2026 contains a citation that none of its authors wrote. None of them remember seeing it. None can trace it to any stage of drafting, editing, or peer review. It was not generated by an AI tool. It simply appeared inside a finished scientific document, was flagged by a colleague, and has never been explained. That single untraced reference triggered a formal investigation into what happens to researchers when they spend serious time writing about ghosts, hauntings, and poltergeists, and what that investigation found is that the phantom citation was not an isolated incident. It was the beginning of a pattern.
Findings published in the Journal of Scientific Exploration in March 2026 formally measured, for the first time, a phenomenon paranormal investigators have reported informally for decades. A team of five researchers surveyed 167 people who had written about ghostly episodes in some professional or semi-professional capacity, covering everything from academic case files and conference papers to documentary scripts and magazine features on haunted locations. The survey listed 15 specific categories of disruption. Every single one was confirmed by at least somebody in the group. Not one category scored zero.
The experiences split into two types. Internal ones: vivid and thematically related dreams, sudden emotional shifts while writing, a felt sense of time distorting, unexplained physical sensations like chills or pressure, and a feeling that something unseen was present in the room. External ones: files going missing without explanation, audio recordings corrupting or producing distorted playback, devices malfunctioning without cause, completed work vanishing despite being saved, and text appearing altered in documents without anyone touching it. Both types were reported across the full group. Internal experiences were confirmed roughly 1.6 times more often than external ones, but the external events were confirmed too, at rates between 7% and 21% of 167 people.
What happened next in the analysis is where this stops being a collection of anecdotes. When the researchers applied a statistical method called Rasch scaling, which determines whether a set of reported experiences belongs to a single underlying phenomenon or whether they are independent and unrelated, 14 of the 15 items aligned on one coherent dimension. In plain terms: these experiences are not random. They belong together. They scale. The internal consistency of the instrument came in at 0.87, a score that sits in the range researchers classify as excellent. Something structured is happening to people who write about the paranormal, and it follows a specific order.
That order runs in one direction without exception. Vivid thematic dreams sit at the easiest, most common end, confirmed by 30% of participants. Meaningful coincidences during writing sessions sit just above, at 26%. Emotional pull toward or revulsion from the material came in at 25%. Moving up the scale, the experiences become rarer and more physically concrete: audio distortions in recordings, time warping, unexplained bodily sensations, the felt presence of something watching. At the top, confirmed by only 7% of the group, sits full erratic computer failure with no technical explanation. The progression runs from inside the researcher outward into the physical environment, and the statistical model treats this progression as a single coherent phenomenon, not coincidence stacked on coincidence.
The hierarchy held stable across every demographic split the researchers tested. It did not shift between men and women. It did not shift between older and younger participants. It did not shift between the seven credentialed professional scientists in the group and the 160 amateur investigators. One item, audio distortion in recordings, appeared more frequently among women than men at equivalent overall experience levels, a gap of 1.65 logits on the measurement scale. Everything else held its position regardless of who was reporting it. That consistency across completely different types of people, with completely different relationships to paranormal belief, is what the researchers call a trickster chain: a structured sequence of escalating disruptions that behaves the same way no matter who it happens to.
The seven professional scientists in the sample did not report nothing. They reported missing or misplaced files at 29% and disappeared or altered text at 29%, rates comparable to the broader group. What they reported less of was the internal, subjective end of the scale. The amateur investigators covered the full range, with marked dreams at 31%, device interruptions at 18%, emotional agitation at 19%, and equipment malfunctions at 14%. Certain categories, including lighting and phone disruptions and odd experiences reported by people nearby, were almost exclusively confirmed by the amateur group. The professionals clustered toward the concrete, hard-to-explain end while amateurs spread across everything.
American journalist Arthur Myers documented his own version of this pattern across multiple published books in the 1980s and 1990s. He wrote in one introduction that every time he worked on a paranormal project, notes disappeared and reappeared, film rolls vanished from plain sight never to return, and audio recordings of interviews produced missing or reversed voices. He also wrote that when he finished a paranormal book and moved on to other subjects, everything stopped. Myers recorded this decades before any scientific framework existed for measuring or categorising such experiences. The 2026 survey places missing notes and audio distortions in the middle range of its hierarchy, precisely where Myers placed them in his own accounts, written a generation earlier with no knowledge of what would eventually be measured.
The hitchhiker effect sits directly inside this framework. This is the pattern in which anomalous experiences connected to a specific investigation location follow researchers back to their homes and continue there. The most extensively documented case is Skinwalker Ranch in Utah, where military and scientific observers documented orbs, poltergeist-type disturbances, and physical anomalies appearing in their own residences after returning from the site. Some of those witnesses reported effects that persisted and escalated over months. The trickster chain measured in the 2026 survey does not require a physical location to trigger it. It attaches to the act of writing, to sustained focused immersion in paranormal material, and it follows the same escalation pattern whether the researcher is a credentialed scientist at a university or an amateur ghost hunter preparing a documentary script.
One participant, a credentialed researcher writing in July 2025, submitted an account alongside their survey responses. During or shortly after completing a paranormal project, their personal author copy of a book they had referenced vanished from the precise location where they had placed it. It has not been found in over a year. The phantom citation in the paper that launched this entire investigation, a fully formed reference to a real journal and real authors, attached to a statistic that was accurate but sourced to entirely the wrong publication, remains untraced as of the date of publication in March 2026. The journal issued a formal correction. The correction identified six additional untraced citation errors in the same paper. None of the authors can account for any of them.
The Trickster-Like Experiences Inventory produced by this research scores on a standardised scale with a mean of 50 and a standard deviation of 10. It is freely available to other researchers. The median raw score among the 167 participants was 1, meaning most people reported zero or one anomalous event. A small subset reported multiple confirmed disruptions across several categories, running the full length of the hierarchy from vivid dreams through to physical device failure. The tool does not yet have the validation history to classify those high-end cases with certainty. It has enough internal structure to confirm they are not random, that they belong to the same underlying phenomenon as the lower-level experiences, and that the phenomenon scales consistently across the full population of people who write seriously about the paranormal.
The researchers are actively designing follow-up protocols involving diary-based monitoring and psychophysiological measurement to test whether the hierarchy holds under controlled conditions. The phantom citation in the paper that started this remains without a confirmed origin. The author whose book disappeared has not found it.
SOURCE:
Houran, J., Lange, R., Massullo, B., O’Keeffe, C., & Schumacher, D. (2026). Trickster-like experiences while documenting the paranormal: Rasch analysis of an initial survey. Journal of Scientific Exploration, 40(1), 52–74. https://doi.org/10.31275/20263837
The idea of a phantom citation leads me to wonder: If what we call “ghosts” are actually just residual energy from past/future people, then these citations could be some kind of an artifact from that encounter. It could explain the time warping effect, too. Because if the trickster effect increases the more the researchers investigate, it’s very possible (in my opinion) that researching the paranormal is something similar to “messing with time” on that they’re observing something from a different time/place (ie. Ghosts)