[text only] [privacy]
  about PSI


the PSI team




press centre


our research


new articles


PSI news  







Investigation methodology


This section briefly details our methodology, for a full discussion download a sample investigation report.  Investigations are compared to conventional scientific research.

The section explains more about what PSI does on investigations, and more importantly why we do what we do. Italicised words carry definitions to their right.  PSI methodology is split into three sections, which are outlined below.

Ethical issues associated with investigations are also considered.


 PSI conducts investigations in the context of three research objectives:
1. To develop a case study of a particular property
2. To conduct qualitative, naturalistic research over the course of the investigation
3. To perform quantitative field research

Investigations as a case study

Investigations as naturalistic research

   Reliability: Human testimony

   Reliability: Human psychology

   Reliability: Recording equipment

   Internal validity

   Construct validity

   Ecological validity

Quantitative field research

   Reliability and validity

Field Experiments (opens in separate window)

Analysis and conclusions

Ethical Issues

Investigations as a case study
One research objective of conducting a paranormal investigation is to build up an overall picture of a property. The purpose of this objective is to build a catalogue of allegedly ‘haunted properties’ and provide a rich and detailed report for the benefit of owners of the properties.
As such, the report brings together several strands of research:
■ A historical report of the venue
■ A catalogue of anomalous phenomena reported over time
■ Witness reports of ‘anomalous phenomena’ derived from interviews with witnesses
■ An account of any phenomena recorded over the course of the investigation
■ An account of any information established through non-scientific spiritual methods over the course of the investigation.

Case studies are not randomly selected and are either self-selective, by venue owners approaching PSI or venues are purposefully chosen because of accounts of anomalous phenomena.
Case study reports do not impose limitations or constraint by requiring all phenomena to be operationally defined. There is no attempt to manipulate variables and no attempt to systematically determine how variables interrelate.

No analysis or critical assessments are made of case studies, and no conclusions are drawn from them as to whether reported phenomena are valid or accurate, and are not used to generalise findings.

Back to top


Investigations as naturalistic research
Investigation settings are slightly artificial in nature. Operationally an investigation consists of a group of trained participants being used to self-report, in non-operational terms, any phenomena witnessed. Individuals act as participants as they are split into groups, spending forty-five minute sessions in different locations, being asked to report what is experienced into voice recording equipment. Such accounts are subsequently transcribed.
Attempts are made to record by various means any phenomena reported for reasons of verifiability (see below). In addition, attempts are made to record phenomena independently of participants, following the tenacious paranormal theories that phenomena can include unexplained voices when white noise is recorded, and that phenomena can include spontaneous PK. Thus sound recorders are set up to record against static radio or television channels, and camcorders record ‘trigger’ objects to test the PK theory.
Further tenacious paranormal theories are occasionally tested during investigations. Methodologies for these are noted in the relevant investigation reports.
Participants also act as ‘investigators’ by engaging in quantitative field research, which will be described later.
Such naturalistic research presents a large number of threats to reliability and validity. Attempts are made to minimalise such threats, but it is not possible to eliminate all threats:

Reliability: Human testimony
Human participants’ testimony is flawed, representing a problem of reliability:
■ Participants thus carry photographic equipment to attempt to capture phenomena witnessed. However, photographic equipment can also be flawed, as discussed later.
■ Self-reports are made instantly, as opposed to reliance on retrospective reporting, a method associated with even more problems of reliability.

Reliability: Human psychology
Empirical phenomena reported by participants can be psychologically caused, presenting further problems of reliability:
■ Participants are open to ‘suggestion’. Efforts are made to minimise the level of suggestion. For example groups of participants are not permitted to discuss with other groups anything witnessed during sessions, to avoid the experiences of one group affecting the reporting of another group. It is more difficult, however, to limit the level of suggestion within a group, as participants can be influenced by what others are experiencing. These problems cause treating witness reports as ‘reliable’ to be flawed, making analysis difficult.
■ Participants can be primed by expectation. Participants are not permitted to research the history of venues in advance of investigations, so as not to be influenced by the knowledge of past events, owners or previous phenomena reported. Similarly, in the selection of venues, efforts are made to avoid venues previously ‘studied’ by television programs focusing on the paranormal. Any evidence reported by investigation leaders who have researched or dealt with venue owners in advance of the investigation is treated as suspect during analysis. Finally, participants are deemed to be ‘primed’ by having participated in paranormal investigations before.
■ Empirical reports can be the result of hallucination, including group hallucination and imagination. It is noted when phenomena are witnessed by more than one participant, in the attempt to promote inter-observer reliability, but the problem of group hallucination remains. Efforts are made to control conventional factors that could cause affected perception, such as a prohibition of use of alcohol or mind-altering drugs on the day of investigations. Similarly, any evidence reported after, say, 0200 can be viewed as especially unreliable, due to the potential of cause by sleep deprivation. The problem of witnesses reporting events that may have not taken place is possibly the greatest problem of reliability, calling all evidence into question that was not verified by photographic equipment.
■ Sampling errors can effect phenomena recorded. The individuals own experiences and mindset can affect what is reported. For example studies have shown that those who self define as ‘sceptics’ have less ability to report phenomena when it is witnessed, whereas those who self-define as ‘believers’ have been shown to report phenomena where none is in evidence.

Reliability: Recording equipment
Reliability of photographic equipment is a threat to verification of recorded empirical phenomena. For example digital cameras may appear to record anomalies where it is in fact distorting light, dust, moisture or reflections. Where possible these variables should be controlled, but otherwise analysts should have an awareness of the location and any potential factors which could interfere with photographic equipment. Further, use of digital cameras are backed up with use of 35mm cameras and camcorders, which are held to be more reliable. Finally, all locations that are to be the settings of sessions are photographed on two occasions before the investigation to act as a baseline.

Internal validity
Problems of internal validity can be identified and attempts made at control. In a natural largely uncontrolled environment there can be problems associated with outside events, which can be confused by participants as anomalous phenomena. All possible outside events must be recorded and verified with other groups or by venue owners before any inferences are drawn. For example all potential phenomena recorded are noted with the time of the event, and all teams have synchronised time pieces. This allows events reported outside of the session setting to be verified by other groups. Venue owners are also interviewed between sessions concerning any possible outside events. Similarly, analysts should have an appreciable knowledge of the venue, to be able to analyse any structural factors. However, it is not possible to eliminate all outside events, so even corroborated reports must be treated scepitcally.

Construct validity
■ One identifiable threat to construct validity is bias, including demand characteristics, social desirability and experimenter bias. Again, it is nearly impossible to guarantee elimination of all such biases. Trained, experienced participants who are not reinforced for reporting phenomena help to overcome these biases, but it is difficult to fully rule them out.
■ A further threat to construct validity is that alternative theories can be found to explain phenomena. Any explanation found should be deemed to be the most ‘likely’ explanation. Therefore at the stage of analysis, evidence should be ruled out if there is a more likely explanation for a reported phenomena, other than it being ‘unexplained’.

Ecological validity
An arguable strength of paranormal investigations is the relative ecological validity of the research. That is investigations are conducted in ‘natural settings’, in properties actually reputed to associated with anomalous phenomena. There are, however, significant threats to the ecological validity of investigations and to the nature of ‘naturalistic’ research specifically. Investigations are forced environments with use of monitoring equipment. Efforts are made to minimalise the intrusiveness of such operations, but ‘investigations’ can never fully represent ecologically valid natural conditions.

Following naturalistic research ‘investigations’ all evidence is analysed against the weaknesses in reliability and validity as outlined above.
Analysis tends to be a process of elimination of all evidence, and the occasional drawing of tentative conclusions that recorded phenomena may be, as yet, unexplained.

Back to top


Quantitative field research
Such research contends to test the tenacious paranormal theory that there is a correlational – not in this case causal – relationship between phenomena reported by participants and and fluctuations in the environment recorded using equipment.
Tenacious theories expound that there is a correlation between phenomena and fluctuations in: temperature, air pressure, humidity, presence of ionic particles and in the electromagnetic field.
Thus baseline readings are recorded on two occasions in advance of the investigation, once immediately before. Participants record readings every twenty minutes throughout sessions to establish a baseline against which readings recorded when phenomena are reported can be contrasted.
There are numerous problems of reliability and validity associated with such correlational research:

Reliability and validity
■ Human error can compromise the reliability of equipment readings. Whilst participants are well trained in the use of equipment and are supervised at all times, human error cannot be discounted.
■ Equipment error can further compromise reliability. All equipment is duplicated and readings compared with each other before investigations begin, but equipment error can never be fully ruled out.
■ Problems of internal and construct validity can compromise equipment readings. That is, outside a laboratory setting it is difficult to establish a direct link between equipment readings and reported phenomena: extraneous factors can account for fluctuations. Thus, investigation leaders must attempt to control or at least have knowledge of the causes of such extraneous factors, so elimination can take place. For example natural or artificial temperature fluctuations should be controlled or established. Further, external factors that cause electromagnetic field fluctuations should be controlled (for example turning off electrical equipment and mobile phones) or established (for example establishing the location of power lines).

During the stage of analysis, correlation between reported phenomena and environmental fluctuations can be highlighted. However there is need to establish if the frequency or intensity of the fluctuations constitutes a significant correlation, that could not have occurred naturally.
The frequency of such correlation can be too few within one session or investigation to attempt to establish a correlation, and may be more appropriate to meta-analyse across investigations.

Back to top


Analysis and conclusions
As noted previously case study findings are not suitable for either analysis or conclusion.
Evidence recorded through naturalistic and field research should be analysed in respect of the methodological weaknesses in reliability and validity as outlined above.

In practice, the majority of empirically recorded phenomena are ruled out under the tests of reliability and validity. However, occasionally evidence withstands scrutiny and can be declared ‘unexplained’. It should be noted that on the rare occasions that evidence is not explainable it should not be described as paranormal and will almost certainly not ‘prove’ the existence of the paranormal. Firstly, by definition spontaneous anomalous phenomena tend not to be repeatable, so proof cannot be established. Secondly, outside of a laboratory setting, it cannot be positively established that the phenomena are not caused by more mundane extraneous factors.

If a significant relationship between phenomena reported and environmental fluctuations is established it should be emphasised that this is a correlation, not a causal relationship. That is, without the ability to manipulate the independent variable, is it not possible to conclude that such readings are caused by the paranormal. For example, studies have shown correlation between fluctuations in the electromagnetic field and the reporting of phenomena, but argued that electromagnetic fluctuations caused perception of phenomena, as opposed to electromagnetic fluctuations being caused by phenomena that had been reported.

Finally and conversely, it should be noted that an investigation revealing no evidence of spontaneous anomalous phenomena does not prove that a venue is not ‘haunted’, at best it proves that there was no phenomena reported on that investigation.

Back to top



Ethical issues

Ethical conduct is of great importance when conducting investigations, especially as human participants are involved and individual’s properties are the setting.
A number of investigation groups operate with unethical practices. PSI offers assurances of its strong ethical standards, and presents the basis tenets of its ethics, below:

Doing no harm. This relates to not putting participants at risk of undue physical or emotional harm during an investigation, and being insured against liability. Participants are chosen via a rigorous selection and training procedure that ensures they are not of a nervous disposition. PSI carries out safety assessments in conjunction with venues. PSI mediums offer spiritual protection for participants who self-define as ‘believers’. PSI insists each investigator completes and carries with them a full medical form.
Consent. All individuals participate with informed consent and no individuals under the age of 18 are not permitted on PSI investigations.
Deception: All participants are fully briefed, trained and aware of all research objectives.
Freedom to withdraw. All participants are free to withdraw at any time. Further, venue owners have the right to halt any investigations at any time.
Confidentiality. All evidence and information gathered on investigations is held in confidence within the organisation unless agreed otherwise, this applies to work with the press. For these purposes the organisation is defined as constrained to trained team members, and specialists associated with PSI to whom confidentiality is occasionally extended.
Debriefing. All participants are fully debriefed and unload experiences at the end of investigations, to ensure they leave the in the same psychological state in which they arrived. PSI has follow up procedures should participants need to discuss any issues after any investigations.
Reporting research. PSI has a duty to take due care any reporting of results, especially to the media. PSI take the responsibility as being seen as ‘representing’ the field very seriously. PSI never sensationalises evidence or vies for attention. PSI never makes unscientific claims, including concluding that a property is ‘haunted’, or not ‘haunted’.

Treatment of venues and owners. All venues and owners are treated with respect at all times. PSI always shares results with venues, abides by their wishes and continues contact for as long as venues and owners need. Ethical responsibility extends to not conducting investigations in any way which might cause a venue to feel any 'presence' they feel has been exacerbated.  PSI never performs 'clearances', but can refer venue owners to suitable individuals or groups should they request that information.  PSI has special ethical procedures for dealing with private homes and families.


Back to top

Free membership of PSI





 About PSI
 Our approach
 Outside Links
Latest PSI investigation

 Contact PSI
 Text only


Methodology: A set of rules and practices used in an experiment, investigation or enquiry.
Case Study: Typically qualitative field research involving the in-depth study of a single environment or group.
Spontaneous anomalous phenomena: Instant, extraordinary events that run contrary to the common order; typically not repeatable.
Naturalistic research: Qualitative research that tends to be conducted within a participant’s or environments natural or normal setting.
Tenacious paranormal theories: Unscientific theories about the nature of the paranormal. Beliefs that are held without scientific basis, which self-sustain as superstitions without scientific proof. Examples of TPT include the theory that orbs and mist are stages of ghost manifestation. An investigator's objective should be to prove or disprove TPT, not to automatically accept them as true.
Spontaneous PK: Typically instant and unrepeatable psychokinesis. In this case the movement of objects by individuals or spirits by the power of the mind.
Reliability: Accuracy of evidence gathering tools, or consistency of results over time.
Empirical: Evidence gained through observation or experimentation.
Primed: Informed beforehand. Information can be held unconsciously and retrieved with the participant not necessarily remembering that he or she had the knowledge previously.
Hallucination: The perceptions of things not present, including somatic (external), visual, tactile (felt), olfactory (smelt) and auditory (heard).  Apprehension, expectation and imagination combined with unusual stimuli and the tendency of the brain to try to reorder ambiguous information. Also ‘group hallucination’ shared by more than one individual.
Inter-observer reliability: The concept that empirical evidence is more accurate if observed by two or more participants simultaneously.
Baseline: The level of behaviour or conditions without intervention.
Internal validity: A measure of whether an identified action caused an observed response, or whether it was caused by something else.
Validity: A measure of whether a study or investigation accurately measures what it intends to measure.
Construct validity: A measure of how accurately a construct is measured.
Demand characteristics: Where a participant perceives the demands of the experimenter and adjusts his or her behaviour accordingly.
Social desirability: Behaving in a way in which a participant believes will result in social approval from a group or person.
Experimenter bias: Methodological flaws resulting from the desires of an experimenter to produce a certain outcome or set of results.
Ecological validity: A measure of whether or not one set of results can be generalised to another environment.
Quantitative research: Gathering information objectively with pre-arranged constraints, typically used for statistical analysis.
Correlation: Typically the statistical strength of relationship between two or more variables. A correlation is not a causal relationship, i.e. it does not determine whether one variable caused the effect in the other.
Extraneous factors: Unplanned variables that can interfere with results.
Meta-analysis: A statistical test drawing evidence from a number of different experiments.
Repeatability: The idea that in order to be ‘proven’ an effect must be repeated many times.
Significant: Objective and statistical tests are often used to determine whether evidence is ‘significant’. E.g. whether an environmental fluctuation is ‘significantly’ at variance with the baseline.
Independent variable: Where questioning if ‘A caused B’, ‘B’ is the dependent variable and ‘A’ is the independent variable: the variable an experimenter manipulates to determine a causal relationship.