CIT Testing UK: The Concealed Information Test

What Is a CIT Test?

The short version: a memory test for one detail, read from the brain instead of from what the person says.

CIT stands for Concealed Information Test. You will also see it called the Guilty Knowledge Test, the name it was given when the psychologist David Lykken first described it in 1959. The idea has not changed since. Somebody who took part in an event carries details of it in memory. Somebody who did not take part does not. If you show both people the true detail mixed in with several plausible false ones, only the first person has any reason to react to the true one differently.

A CIT therefore never asks "did you do it?". It asks a much narrower question: "does this person's brain treat this item as familiar, when it treats the alternatives as ordinary?" The tested person is shown a short list on a screen, one item at a time. One item in the list is the real detail, called the probe. The others are made-up alternatives of the same kind, called irrelevants. One further item, the target, is there to give the person a simple job to do and to prove the recording is working.

In our version of the test the reaction we measure is the P300. This is a positive swing in the brain's electrical activity that appears roughly a third to half a second after a person sees something that is both rare and meaningful to them. It is recorded with EEG sensors resting on the scalp. The P300 was being used to detect concealed information in research laboratories by the late 1980s, and the method has been studied continuously since.

What a CIT Can and Cannot Show

Most disappointment with any form of credibility testing comes from asking it a question it was never able to answer.

What it can show

• That a specific item stands out for this person. If the probe draws a clearly larger response than five comparable alternatives, across dozens of showings, the person's brain is treating it as special.
• That a specific item does not stand out. If the probe behaves just like the alternatives while the target check shows the recording was sound, there is no sign the person recognises it.
• Agreement across several details. One recognised detail is suggestive. Three independent details, each recognised, is a much stronger pattern, because the chance of that happening by accident falls sharply with each one.

What it cannot show

• It cannot test a bare accusation. "Did you take the money?" has no probe. There is nothing to recognise in the question itself. A CIT needs a detail: the amount, the envelope, the drawer, the day.
• It cannot tell innocent knowledge from guilty knowledge. If the person has been told the detail, read it in a message, or heard it in an argument, they will recognise it whether or not they were involved.
• It cannot prove an act. Recognition is evidence about memory. What that memory means is a judgement for the people involved, taken with everything else they know.
• It is not a court instrument. We make no claim that a CIT result is admissible in any court or tribunal in the United Kingdom.

This is why the first conversation we have with anyone booking a CIT is about the details, not about the equipment. If there is no detail that only an involved person would know, a CIT is the wrong tool and we will say so. The page on who qualifies for a CIT test goes through that check step by step.

How the Test Works, in Brief

A fuller walk-through is on the how CIT testing works page. This is the outline.

1. The details are agreed in advance. Before the appointment we work with the person arranging the test to choose the details to be tested and to write the alternatives. The tested person is not told which item is the real one.
2. The cap is fitted. A soft EEG cap with four sensors is placed on the head, with two small clips on the ears. Contact is checked on screen until the traces are steady.
3. The items are shown. Words or short phrases flash up one at a time, each for about a third of a second, roughly one every two seconds. The person presses one key for the target and another key for everything else.
4. Each item is repeated. A single showing tells you nothing; the brain's response to one word is buried in background activity. Each item is shown twenty to forty times so the responses can be averaged.
5. The recording is analysed. The response to the probe is compared with the response to the irrelevants using a resampling test. The target is checked first: if the target did not produce a clear response, the block is marked unusable.
6. The examiner reports the result. The screen shows the tested person a neutral "test complete" message. The result is released by the examiner and written up afterwards.

Each detail is tested as its own block of five to ten minutes, with rests between. A typical appointment covers two to four details.

The Science Behind It

What the published research says, stated at the strength the research supports.

The P300 is one of the most studied signals in human neuroscience. It is not a lie signal. It appears whenever a person notices something that matters to them among things that do not: their own name in a list of names, a rare tone among common ones, the face of someone they know. The CIT borrows that property. The probe is rare, because it is one item among six or seven, and for a person who knows it, it is meaningful.

In 2014 a group led by Ewout Meijer published a meta-analysis in the journal Psychophysiology that pooled decades of CIT studies across four measures. It reported a mean effect size of 1.89 for the P300, against 1.55 for skin conductance, 1.11 for respiration and 0.89 for heart rate. In plain terms, all four measures separate people who know a detail from people who do not, and the P300 produced the largest separation. The same paper adds an important qualification: the P300 outperformed skin conductance only in studies that used personal items, such as a participant's own details, and not in mock-crime studies.

Laboratory studies of P300-based CITs commonly report that somewhere between 80 and 95 per cent of people who hold the tested information are detected. Those figures come from controlled experiments, usually with students, usually with details that were deliberately rehearsed. They are not results from our own testing. Field validation, meaning studies of real cases where the truth was later established, is limited for the P300 CIT, and anyone who quotes a single accuracy percentage for real-world use is going beyond the evidence.

Where the CIT is used

The largest field user of the CIT is the Japanese police service. A 2012 review in Frontiers in Psychology reported that around 100 trained examiners carry out about 5,000 CIT examinations a year in Japan, using bodily measures such as skin conductance, heart rate and breathing. Japan is the main example of the CIT being used routinely outside the laboratory, and the same review is frank about how field practice differs from laboratory studies.

Known weaknesses

Two are worth stating plainly. First, leaked information: if the real detail has reached an uninvolved person by any route, the test can no longer separate them from an involved person. Second, countermeasures: in 2004 Peter Rosenfeld and colleagues at Northwestern University showed that people who are taught a simple trick can reduce the difference between probe and irrelevants in the standard three-stimulus protocol, which is the protocol we use. Later protocols were designed to resist this. We watch reaction times and the target check for signs of it, but we do not claim the test cannot be beaten.

Who a CIT Suits

The test fits a case, not a person. These are the situations where it tends to fit.

• Workplace loss or leak. Something specific went missing or was disclosed, and the particulars have been kept within a small group: which cabinet, what sum, which file name, what the package looked like.
• A dispute over a specific event. Two accounts differ, and one of them, if true, involves knowledge of a place, an object or a sequence that the other person says they know nothing about.
• Relationship cases with a withheld detail. One partner has found something concrete, such as a first name, a venue or a town, and has not yet revealed it. The page on CIT testing for a suspected cheating partner explains why this is the hard requirement and how often it is met.
• People who want to show they do not know something. A person accused of involvement can volunteer for a CIT on details they say are new to them. A clean result on several withheld details supports that.

It tends not to fit where everything has already been said. If a couple have argued through every detail for months, or an employer has put the full allegation in writing to the employee, the details are no longer concealed from anyone.

Where the true detail is not known to anyone except, possibly, the tested person, there is a different form of the method. The searching CIT presents a set of candidate answers and looks for one that stands out. It is weaker than the standard test and we explain why on that page.

CIT and Polygraph Are Different Tests

People often arrive expecting a polygraph. The two answer different questions.

A polygraph examination usually uses the comparison question technique. The examinee is asked direct questions about the matter ("Did you take the money?") alongside comparison questions designed to be uncomfortable, and the examiner compares the body's arousal to each: breathing, sweating, blood pressure. The inference is that a deceptive person reacts more to the relevant questions.

A CIT asks no direct questions about guilt and does not rely on arousal. It compares the response to one item against responses to others of the same kind. An anxious person is anxious for every item on the list, so nervousness alone does not single out the probe.

Neither test is a substitute for the other. We do not conduct polygraph examinations. The full comparison, including what UK law says about polygraph use, is on the CIT test vs polygraph page.

What Happens on the Day

An appointment from the tested person's point of view.

The examiner arrives at the agreed place, which is either a home address or a private room we have arranged nearby. The first ten minutes are paperwork and explanation. The tested person reads and signs a consent form. Testing is voluntary; nobody can be tested without agreeing to it, and they may stop at any point.

The cap goes on next. It is a fabric cap with four small sensors and a lead to a matchbox-sized amplifier, which sends the signal to a laptop by Bluetooth. A little conductive gel under each sensor improves contact. There are no needles and nothing is passed into the body; the sensors only listen.

The examiner then explains the task using the first block's question. For example: "Words will appear one at a time. Press Y when you see the word Bristol. Press N for anything else. Answer every word." There is no trick in the instructions. The tested person does not need to say anything, and they are not asked whether any item is true.

Each block runs for several minutes with a rest halfway. Between blocks the person can move, drink water and blink freely. During a block they are asked to keep their eyes on the centre of the screen and their jaw relaxed, because eye and muscle movement produce electrical activity far larger than the brain signal being measured.

When the last block ends, the screen shows a neutral completion message and the cap is removed. The examiner reviews the recording quality before reporting anything. If a block is unusable because the contact drifted or the target response was missing, that is reported as unusable, not as a pass or a fail.

How Results Are Reported

Plain wording, with the working shown.

For each detail tested, the report states one of three things. Recognised means the probe's response exceeded the irrelevants' in at least 90 per cent of 1,000 resamples of the recording. Not distinguished means it did not. Unusable means the target check failed or too many showings were lost to movement, so no conclusion can be drawn from that block.

The report also gives the figures behind the wording: how many showings of each item were clean, the size and timing of the response to each item, how the tested person did on the key-press task, and the averaged waveform. We keep the complete unfiltered recording with the exact moment each item appeared, so the analysis can be repeated by someone else if the result is ever questioned.

We do not print a single accuracy percentage on a report, because no honest one exists for an individual test. What we can state is how consistently this recording separated the probe from the alternatives.

CIT Terms in Plain English

The vocabulary you will meet on these pages and in a report.

• CIT. Concealed Information Test. A test of whether a person recognises a specific detail.
• GKT. Guilty Knowledge Test. The older name for the same method.
• Probe. The true detail being tested, hidden among alternatives.
• Irrelevant. A false alternative of the same kind as the probe. A person with no knowledge should respond to it just as they respond to the probe.
• Target. An item the person is told to press a particular key for. It keeps them reading and proves the recording can detect a response.
• Block. One run of the test on one detail, lasting five to ten minutes.
• EEG. Electroencephalography. Recording the brain's electrical activity from sensors on the scalp.
• P300. A positive brain response beginning around 300 milliseconds after a rare, meaningful item.
• Pz. The sensor position over the parietal midline, where the P300 is usually largest. It is the channel we score.
• Epoch. The short stretch of recording following one showing of one item.
• Artefact. Electrical interference from blinks, muscle tension or movement. Epochs containing it are discarded.
• Bootstrap. The resampling procedure that tests how consistently the probe response exceeds the irrelevant response.
• Countermeasure. Anything a person does deliberately to distort the test.
• Known-answer CIT. The standard test, where the true detail is known in advance.
• Searching CIT. The form used when the true detail is unknown and the test looks for a stand-out among candidates.

Booking a CIT Test

What to have ready, and what happens after you book.

You can book online or ask a question first. Bookings are for anywhere in the United Kingdom; the examiner travels to you or to a private venue arranged near you.

After booking we will ask you three things. What happened, in your own words. What you know about it that the person to be tested has not been told. And who else knows those things. From your answers we will tell you honestly whether a CIT can be built for your case, and if it can, we will draft the item lists with you.

The person to be tested must agree to take part. It helps to tell them the truth about the format: that they will watch words on a screen, press a key, and will not be asked any accusing questions. Please do not tell them which details will be tested.