Re-Investigating Un-Haunted Houses
Eight couples who had never experienced any ‘haunting’ activity in their houses and had no reason to expect they would experience ‘ghosts’ were asked by the author to keep a diary for one calendar month from 17th October – 17th November 2012 in which they recorded unusual experiences. 62.5% of the participant couples recruited completed the task and submitted the diaries for analysis. Of the five participant couples who submitted diaries, four reported at least some phenomena that met the criteria, and one couple reported no unusual activity at all. The study was a larger scale replication of Houran and Lange (1996). My findings are compared with those of the original study which featured only one couple.
UPDATE: Within 20 minutes of the first draft of this paper going live on my blog I was contacted by one of the missing research participants and was able to locate the couples data which had been submitted at the time but by Facebook message rather than e-mail. I have therefore revised the figures to take in to account the new material. It had no impact on the overall findings, fortunately.
Introduction: Houran & Lange’s 1996 Study
James Houran and Rense Lange have been the authors of a number of innovative studies in parapsychology. In this 1996 paper they were exploring if hauntings and in particular it would seem poltergeist cases were explicable in terms of a self-reinforcing-psychological contagion hypothesis. In essence the idea is that once one notices unusual anomalies in one’s home, and has ones attention drawn to it, more such anomalies are noticed. The paper is often cited (for example Wiseman 2011; Wiseman 2011b) as it provides an elegant psychological explanation for purported “hauntings”. The original paper is based on the experiences of one couple, mature students, who were requested by the researchers to keep a 30 day diary of unusual events in their home which was in no way believed to be ‘haunted’ before the study began. The small number of participants (one diary) troubled me: it seems dangerous to draw too many conclusions form a single innovative pilot study like this, and I could find no replications, yet the paper is repeatedly cited by sceptics without mention of this limitation. I therefore decided to replicate the study, on a larger scale.
The basic idea behind Houran and Lange’s paper appears simple. Imagine one day you come home and find your books are symmetrically stacked in the living room. You don’t recall doing it, and your housemate is never so neat! Later on, an egg starts to fry on your kitchen worktop, and then you hear an odd voice say “Zuul”. Your attention may now well be extremely focussed upon the weird things happening in your house – you probably approach the fridge with trepidation – and when the cat knocks over a flower vase later and the hot water system causes knocking in the pipes, you are only to quick to jump to the “ghost did it” conclusion. In short, ghosts are by this hypothesis merely a narrative we create to explain little mysteries (anomalies) in our daily lives. When my door keys go missing, I search and search and eventually find them on the shelf where I thought I had looked first, I may be more willing to blame a spook than my poor perception.
In the same year I suggested something similar (Romer; 1996); what did not occur to me was that such observations of purportedly paranormal phenomena would eventually die out. Houran and Lange argue this based upon
“the assumptions that (1) the environment provides a stable supply of events that can be interpreted as paranormal and (2) the probability of noticing an additional anomaly is directly proportional to the number of anomalies already noticed as well as the number of remaining potential anomalies. Under these assumptions, it can be shown that the cumulative frequency distribution of perceived anomalous events should follow the familiar logistic curve.” (Houran and Lange 1996: my emphasis)
So you notice something odd going on; you start to look for it as your attention becomes focussed on the “ghost” – that much seems straightforward. However I am slightly confused by the “number of remaining possible anomalies.” This implies there are a limited number of such events in the environment, and eventually you will reach a point where you have observed most potential anomalies, causing the number of new experiences observed to tail off. I am puzzled as to how Houran and Lange came to this conclusion. If perceptual mistakes give rise to some anomalies, and others are simple misunderstanding of mundane events, I see no reason for them to “run out” as suggested. What limits the “number of remaining anomalies”? Yet this is an important aspect of the paper, even if not explained within it. As has been pointed out, poltergeist type events usually run out if steam in a fairly short period – the “logistic curve” Houran and Lange hypothesise would explain this within their psychological explanation. Here is the graph of the cumulative experiences that were reported by one couple in their diary study. As we can see it neatly fits the predicted logistic curve.
Yet without understanding why the potential anomalies in any given environment (house) are limited in the time period, it is hard for me to understand. Why they predict the classic logistic curve above. I would have predicted an exponential rise in cumulative frequency: the problem is that while this neatly represents reports of ‘actual’ poltergeist cases, which trail off over time, I can’t see why it should be suggested in the first place. What limits the potential anomalies?
Replicating The 1996 Paper: A New Diary Study
Given the fact that people citing the original 1996 paper have at times drawn rather strong conclusions from this single diary study with only one couple involved, I decided to attempt a replication. I intended to recruit ten couples as participants, though that proved impossible. I wanted to see if the couples reported similar experiences to those in the original study, and if the puzzling logistic curve was borne out in the new data.
Recruitment was via volunteers through the authors Facebook account. 8 couples living in the UK volunteered to keep the diaries, and then again two days before the end of the study. Five couples mailed me completed diaries. No reason was given by the other three couples for failure to complete (though simple forgetfulness is one possibility). The couples were all aged between 30 and 50 years, though I did not ask for precise ages, marital status, or other personal information. Two of the couples have strong interests in the paranormal, and two in religion. This was not intentional selection, nor even a feature of the couples who initially responded to my request for participants, but it may be a reason why they stuck with the study till the end.
Of the four couples who submitted diaries, one had experienced nothing unusual which met the criteria at all in the time period, and a second had a relevant experience while staying in a place other than their own home (discussed briefly later.) So from the original eight couples, five participated and three had experiences that met the criteria. The fact three did not is in itself of interest. They were certainly aware of the study – Couple C reported several events which met the criteria, but which occurred while they were away from their home, hence were excluded. Couple D reported that no such experiences occurred in the time frame, though one partner had experienced anomalous experiences in the past. Couple E had one very striking experience.
Those who expressed an interest in participating were sent the following instructions (along with some introductory text and contact numbers for myself. No one called during the study). The instructions, and the 8 categories were based on those employed in the 1996 study – the categories they employed derived from an Lange paper on ‘Contextual mediation of perceptions in hauntings and poltergeist-like experiences’. (Lange et al. 1996) I attempted to replicate as faithfully as possible the original research. Here are the instructions I emailed out to the prospective participants.
“For the next month, until November 17th, please pay particular attention to any unusual occurrences in your residence. These occurrences may be emotional feelings, physical sensations, or environmental events in your residence. Please keep detailed and accurate notes, even if you know or believe to know what caused the occurrences to happen. I will need the gender and age of adult occupants, and who had each experience noting. If you have children please do not discuss this with them. I have no desire to upset children! The types of unusual experiences I am interested include but are not limited to
* Visual – seeing things not there
* Audio – hearing stuff with no known cause *
Tactile – the feeling of being touched with no obvious reason
* Olfactory – strange smells
* Sensed “presences”
* Intense emotion for no apparent cause beyond that you might normally experience
* Object movements with no apparent cause
* erratic function of equipment.
At the end of the month I would like you to send me the file with your notes. Obviously the experiment requires the full consent and participation of your house mates. I’m asking for volunteers on my Facebook because I want people who I can trust and know. My final report will be anonymized to prevent personal details being shared, and will credit you by name if you wish in the credits. You can end participation at any time.
You can always contact me if necessary on (numbers removed). This is a very important piece of research and I’ll be hugely grateful if you can assist.”
The Phenomena Reported
Only two couples (labelled A and B for ease of reference) provided phenomena that occurred in their own homes. Couple C reported phenomena that occurred to their car, and a phenomena that met the criteria but occurred while she was working elsewhere overnight in the period in question, and while of considerable interest this had to be excluded as not occurring in their own home from this study: however it was still of great interest. Couple D reported no phenomena. Couple A reported 19 events, couple B 10, Couple E 1 – compared with the 1996 couple where in the slightly shorter period of 30 days (as opposed to 32 days in this study) 22 events were reported.
In this study the five couples reported an average of 6 experiences that met the criteria and were in their homes, but of course 50% of the participants reported none – so the actual figures are 0,0, 1, 10,19. Only on the 7th November did three events occur to the same couple on the same day: No more than 3 events are reported on any given day. Halloween (October 31st) gave us only one event – which rather knocks traditional beliefs in this respect!
The nature of the phenomena can be classified by the eight categories used in the original study. There was however a new category that emerged strongly. “Sense presences” were inferred by both couples by the behaviour of there cats seemingly staring at things not there and behaving unusually. Given that this is not a “sensed presence” by a human percipient, but certainly can be seen as building towards the narrative of a psychologically induced haunting, I included these in a new 9th category (which might be called Unusual Pet Behaviour in any replication). The single human “sensed presence” was of a deceased cat, sensed by the owner on November 3rd, and appears in the Sensed presences category as the percipient was human. A visual experience reported was also of a cat where no real cat was; this was from the other couple.
|Phenomena||Couple A||Couple B Couple E||Total (Percentage)|
|Visual||0||2 0||2 (6.6%)|
|Auditory||5||1 0||6 (20%)|
|Tactile||2||0 0||2 (6.6%)|
|Olfactory||1||0 1||1 (3.3%)|
|Sensed “presences”||1||0 0||1 (3.3%)|
|Intense emotion||0||2 0||2 (6.6%)|
|Object movements||0||8 0||8 (26.6%)|
|Equipment Erratic||1||4 0||5 (16.6.2%)|
|Cat Behaviour||1||2 0||3 (10%)|
One of the issues when tabulating the data was what to call an “experience”. For example, on one experience a cat was heard to jump on the sofa, and the black tail briefly glimpsed out of the corner of the eye – and no cat was there. (A very mundane common hallucination, any cat owner must be used to). As the two events followed each other in quick succession, I recorded them as 2 events – auditory and visual. However for a strange noise heard coming from a bookcase one night, I recorded it as one experience, despite it recurring a few minutes later. Such subjective judgements are unavoidable in dealing with diary studies.
So as we can see “un-haunted” houses can appear surprisingly haunted once we pay attention to the anomalies, just as the 1996 paper said, and as I argued in my (also) 1996 piece a cumulative narrative can be composed from non-associated and presumably non-paranormal occurrences. (We will return to this seemingly solid conclusion later however.) What is also clear is that while there are commonalities the specifics of our two haunts vary considerably, with Couple B reporting object movements and classic poltergeist “small object displacement” or “jottle” effects while Couple A report significantly more strange noises and auditory experiences. So we appear to have a general confirmation of expectancy/priming effects and focussing awareness leading to the development of a ‘ghostly’ narrative – though it is important to note neither couple actually reported their experiences in those terms, and both were aware that the experiment led to them paying attention to the anomalies obviously. Just to confound matter further Couple B included with their diary a query as to whether I was familiar with Houran and Lange (1996), the paper that I was attempting to replicate. While I trust them obviously this could colour their dairy, as they were clearly aware of the hypothesis I was testing. In this day and age finding “naive” subjects for any experiment is increasingly difficult while meeting the needs of informed consent!
The Logistic Curve
So what of Houran and Lange’s hypothesis that the experiences would follow a logistic curve? Let us firstly remind ourselves of what this looks like in the original study.
As I currently lack the software to plot the logistic curve all I can note is this looks more like a straight line distribution to me: it levels off , but if we just plot the experiences the effect of the curve is far from apparent. I see less evidence of the purported “running out of anomalies” effect, and given the tedium of keeping up a diary study, it is just as possible the whole logistic curve tells us more about the enthusiasm of research subjects for participation in a project than the nature of hauntings.
Let us move to Couple B. Here are there results, presented the same way. Firstly graphed as in the original paper.
Again, despite the levelling off in the middle, there is no resemblance to the logistic curve. I am fairly sure that if tested the relationship between the observed values and the expected values would be non-significant. Just to be consistent, here are just the cumulative experiences depicted.
Again we see as I hypothesised a fairly straight line progression. The evidence does not seem to support a logistic curve, and hence does not support a “running out of anomalies” factor. There is no apparent reason why in 32 days the effect should tail off – which is an important criticism of the idea that it explains why poltergeist events are short-lived and episodic, if the psychological hypothesis theory is correct. Let us finally combine all three couples results (with single experience of couple E included) and examine them.
The Logistic Curve is nowhere to be seen. Our couples did not “run out of anomalies” – they continued to find new odd occurrences to remark upon. The very nature of a diary study where the research participants may strain to find things to comment upon to “do their homework” and feel they are justifying their participation may lead to this result, but then one would have expected it to show up in the original study.
Comparing the Experiences
The original paper gives relatively little information about the actual phenomena reported. Equipment behaving erratically was the most common experience, with 16 of the 22 reported events, followed by 5 counts of object movement and one subjective experience. So in the 1996 study the phenomena classes described were far more limited than in this replication. Furthermore it is surprising to read in Houran and Lange (1996)
“Further, in agreement with the focussing effect described by Roll (1977), three out of the five objects which were found to have moved were the same, and all of the erratic functioning involved the same piece of equipment.” (emphasis mine)
If I had the same piece of equipment malfunction 16 times, I would suspect that there was something broken with it, not spooks. 72.7% of the phenomena reported in the original study were malfunctions of this one piece of equipment, the nature of which is not specified. I find this quite incredible. The pattern does not repeat in this replication – all object moved were unique, and Couple B’s 4 cases of erratically behaving machinery only involved two the same, both involving the lounge lamp, several days apart. There is no evidence to support the kind of effect seen in Roll’s poltergeist cases as cited in the new study.
UPDATE: re-reading Richard Wiseman’s Skeptical Inquirer piece gives additional information cited as from the paper, but not contained within the paper.
“Reporting the results in the paper “Diary of Events in a Thoroughly Unhaunted House,” he noted that the couple reported an amazing twenty-two weird events, including the inexplicable malfunctioning of their telephone, their name being muttered by a ghostly presence, and the strange movement of a souvenir voodoo mask along a shelf.” (Wiseman 2011b)
I am not sure what Richard’s source is, presumably the author’s themselves: however of the ‘amazing’ 22 experiences 16 (72.7%) involved the telephone malfunctioning.
Nonetheless the replication provides greater diversity and similarly impressive numbers in some of the 5 diaries. While the original study found a significant case for a ‘focal person’ as often found in poltergeist-like cases, who witnessed 16 of the phenomena while by themselves (72.7%) no such effect is apparent in the replication. It is impossible based on the ambiguity of the records regarding who exactly was present or first discovered an object had moved to tabulate exactly, but the experiences are generally framed from the author’s perspective (in both cases a female) but seem to have involved and been witnessed by their partners (both male) on several occasions, and in some instances the males was the percipient. Again, an effect found by Houran and Lange and common to the case history of poltergeists does not appear in this replication.
The greater diversity of experiences reported seems to me to strengthen the case for a psychological contagion effect, but it is important to note that a) the participants did not come to the conclusion they were being haunted and b) for those familiar with the Census of Hallucination (1894) research, I do not think any of the experiences reported would meet the exclusion criteria used there: object movements were not included in that study. To compare these experiences with say the witnesses at Enfield (Playfair 1980) or Cardiff (Fontana, 1991) or Andover (Colvin 2008) appears unreasonable. These experiences may well lead some people to believe their house is haunted, but with the possible exception of the object movements (none of which were witnessed moving, and for 60% of which the participants offered likely mundane causes) none of them are likely to cause resort to paranormal explanations.
So What Have We Learned?
The replication has provided significantly stronger evidence for the psychological contagion case than the original paper does, in that it shows that a wider range of “paranormal-type” experiences can occur in everyday life, with the potential to be misinterpreted and develop in to a ghost story narrative. Yet we must note several things.
Firstly, the phenomena involved would not I fear withstand an objective external investigator. The participants themselves repeatedly “explain away” the phenomena – after all, as in the original study, they were instructed to report such things even if “even if you know or believe to know what caused the occurrences to happen.” Secondly, the study may simply show the priming effect of participating in the experiment.There is no reason to think the participants would have thought very much if at all about what occurred, let alone ascribed it to spooks, if they had not been participating in the diary study. It is important to note that 40% of those who responded, and quite possibly the other three participants who did not submit diaries, experienced no notable phenomena. If the three who had expressed willingness to participate but never got back to me had noticed anything similar occurring, you might have expected them to respond.
Yet I have no doubt that life is full of tiny anomalies: during the day it has taken me to write up this replication my partner has texted to say she had her sat nav come on while lying on her bedroom floor and make her jump by telling her to “turn right”; I myself thought I saw Cuddles my black cat sitting on top of a cupboard, but on looking again he was not there, and was still sleeping in my bedroom when I returned to the computer. Neither of us have jumped to the conclusion we are haunted: but I can see how it could well happen, and I think the psychological contagion hypothesis requires much more study, and am thankful to Houran and Lange for their pioneering and important work. Houran and Lange (1996) wrote
‘This resulting cumulative frequency distribution of event times closely follows a logistic curve… thereby providing strong support for our hypothesis that perceptions of anomalous events are an artefact of attentional contagion. This finding implies that explanations of anomalous events need not invoke such untestable notions as “discarnate agents” or “recurrent spontaneous psycho-kinesis”.’
This study found no evidence for the logistic curve – and the author is still confused as to why it was invoked, as it appears to be difficult to justify as a hypothesis. While the replication was relatively small scale, it was of course still larger in scope than the original study, and leads to the question as to why no one appears to have attempted to replicate it in the intervening sixteen years given the elegance and simplicity of the research design. Widely cited, and fascinating in its implications, the Houran and Lange study opens up new vistas for research in to people’s interpretation of ambiguous stimuli, but one must question whether it really demonstrates all that some sceptical proponents have made out.
Chris Jensen Romer, January 2013
Note: I would to acknowledge the kind assistance of Tom Ruffles of the SPR in helping me locate articles used in writing this piece. Participant Bryan Saunders has kindly agreed to be waive his identity, and I would like to thank him and Barbara for their faithfully maintaining their diary throughout the month and all their help. It is always pleasing to have some non-anonymous participants, as it it lowers the potential for fraud (I did not make up the results, but you don’t know that). :) I would also like to thank the SPR for their research grant support of my ongoing research.
Colvin, B (2008) The Andover Case: A responsive poltergeist, Journal of the Society for Psychical Research, 72, p. 1-20. Fontana, D (1991) A responsive poltergeist: A case from South Wales, Journal of the Society for Psychical Research, 57, pp. 385-402.
Houran, J. and Lange, R. (1996), Diary of events in a thoroughly unhaunted house, Perceptual & Motor Skills, 82, 499-502
Lange R, Houran, J, Harte T.M. & Havens R.A. (1996) Conceptual mediation of perception in hauntings and poltergeist -like experiences, Perceptual & Motor Skills, 82, 755-762
Lange, R., and J. Houran. 1997. Context-induced paranormal experiences: Support for Houran and Lange’s model of haunting phenomena. Perceptual and Motor Skills 84: 1455–58.
Playfair, Guy Lyon, (1980) This House Is Haunted: the Investigation of the Enfield Poltergeist, Stein & Day, London.
Roll, W.G (1977) Poltergeists in B.B. Wolman (ed) Handbook of Parapsychology, Jefferson, NC; McFarland p.382-413
Romer, C. (1996) The Poverty of Theory: Some Notes on the Investigation of Spontaneous Cases, Journal of the Society for Psychical Research, 61, 161-163
Sidgwick, Eleanor; Johnson, Alice; and others (1894). Report on the Census of Hallucinations, Proceedings of the Society for Psychical Research, 10.
Wiseman R, (2011), Paranormality, Macmillan, London.
Wiseman R, (2011) The Haunted Brain in Skeptical Inquirer 35.5 (available online at http://www.csicop.org/si/show/the_haunted_brain/)