FORENSIC PSYCHOLOGY: HOW TO PROVE THE TRUTH BY SHOWING THAT THE UNTRUTH IS FALSE

Judges; police interviews; evidence; statement verification; the scientific method; an experiment you can try

By our guest author, Arjen Akkersdijk.

Imagine you are a judge and have to pass judgement in a criminal case. In this case, a woman accuses her husband of a serious crime, but he vehemently denies the accusation. There is no physical evidence. It is therefore a case of testimony against testimony. Of course, you know that only one spouse is telling the truth, but you cannot say for sure who. Both testimonies seem equally credible to you. You are faced with an ethical dilemma: either you risk sending a possibly innocent person to prison, or you have to accept that the potential victim will not get justice and the offender will go unpunished. Difficult decision, isn't it? Being a judge sometimes sucks.

In this situation, wouldn't it be nice if there were a scientific method at hand that would enable you to tell who is lying? Of course, it would. In fact, the search for a reliable method of lie detection is something like the search for the Holy Grail of legal psychology. Numerous methods have been developed to give authorities an answer to the question of whether a testimony is true or false. The most popular example is perhaps the polygraph test, which is often colloquially (and misleadingly) referred to as a lie detector. Another well-known method is the analysis of non-verbal behaviour, such as so-called micro expressions (as depicted in the TV show Lie to Me starring Tim Roth). However, many of these methods - regardless of how well known they are - are either considered unscientific or frequently presented in a highly distorted way. But today, we would like to turn our attention to a perhaps less known, but scientifically sound and legally established psychological method of lie detection: Statement Validity Analysis (SVA).

SVA is one of the most common techniques for evaluating the validity of testimonies. First developed in the 1960s, it since has become a popular forensic practice especially in Europe. It is a complex methodology that combines semi-structured interviews, psychometric tests, systematic analysis techniques and clinical judgement.

A core component is the Criteria-Based Content Analysis (CBCA), which is the systematic search for linguistic cues associated with truthful statements. The underlying assumption is that ‘truthful, reality based accounts differ significantly and noticeably from unfounded, falsified, or distorted stories’[i]. Telling a sophisticated lie is a complex task. Anyone who has ever told a lie and been exposed certainly knows this; so most likely all of us. But what exactly makes lying so difficult? Just three simple examples for you: First, you have to take care that the lie itself is believable and realistic, and that it fits neatly into the framework of truthful events surrounding it. Secondly, if you tell that lie more than once, you have to make sure that it consistently stays more or less the same. And last but not least, you have to be prepared - especially if you give a false testimony in a criminal investigation - that questions will be asked that you did not prepare for, and which might force you to make up new lies on the spot. As a liar, you better be creative, a good actor and blessed with a good memory. In contrast, a person who is telling the truth doesn’t need to consciously take care of these things. Experience will usually allow him or her to effortlessly recall the event in question.

Time Saving Truth from Falsehood & Envy

Are you ready for a little experiment? You certainly cannot remember what you did on December 13^th 2018, can you? Perfect. Now, make up 10 random events that you could have experienced that particular day in chronological order. Maybe, you could have started the day with a visit to the gym, then had brunch with colleagues, then visited your mother in her elderly home, and so on. Write these 10 made-up events down on a piece of paper and carefully memorise them. Great! And now speak them out loud in chronological order without looking at the paper. That was quite easy, right? Most of you will have been able to do that. Now, speak out the same 10 events again, but this time in reverse order. What did that feel like? For most people this is way more difficult since they memorised the events in chronological order; and going against that order needs more cognitive effort. Real experience, on the other hand, makes you recall events regardless of chronology. If I asked you what you actually did yesterday, you certainly can - without any preparation - tell me your whole day in reverse chronological order with relative ease. Did you try? Do you believe me now that lying is cognitively difficult?

Considering the difficulty of complex lying, we can now better understand the basic assumption of CBCA: Truthful and false testimonies often show certain linguistic differences. CBCA is based on the analysis of a total of 19 features (so-called truth criteria), such as the quantity of details, logical consistency, contextual embedding, mentioning of unusual details, spontaneous corrections, and self-deprecation. The significance of one feature alone is very low, but it is assumed that the combination of different truth criteria is a reliable indicator of the truthfulness of a statement. ‘The absence of [such] criteria, on the other hand, does not imply that a statement is fabricated’, since the ‘low quality [of the testimony] could also be the result of a person’s lack of motivation to make a detailed statement’[ii].This is why CBCA has to be embedded into the holistic procedure of SVA, in which it is complemented by the psychological assessment of other personal and situational factors.

Regardless of its concrete methods, SVA shows a formal peculiarity - at least in Germany - that may seem somewhat strange, unnecessarily complex, counterintuitive, or even unfair to some readers.

Following a series of judicial scandals in which SVA played an infamous role, the German High Court laid down mandatory quality standards for SVA in 1999. One of these standards is the methodological rule that SVA experts must start their assessment from the basic hypothesis that the witness is not telling the truth; the so-called null-hypothesis. The task of the SVA expert then is to critically examine this null-hypothesis by means of SVA. You may wonder now why this procedure is necessary. Perhaps you are also wondering why one cannot simply assume as the null-hypothesis that the witness is telling the truth and then test this hypothesis instead. Wouldn't that somehow be fairer than accusing the witness of being untruthful by default?

Well, you need to know that there are compelling logical reasons to assume that scientific hypotheses can never be verified. This generally applies to all empirical sciences. What is logically possible, however, is the falsification of hypotheses. So if I want to investigate whether a person is telling the truth or not, I must not simply search for confirming information that supports the truth hypothesis. Instead, I have to critically examine the falsehood hypothesis; and if I succeed in falsifying it with certainty, then its counter-hypothesis that the testimony is true must itself be the truth. If you haven’t yet fully understood the above explanations on scientific logic, then rest assured: It is complicated, and probably worth its own essay. Let us know in the comments if you want to know more about it.

So is SVA the Holy Grail of legal psychology? Unfortunately, it is not. Like any other psychological method, it has limitations. One major methodological problem is that of ecological validity; a topic that has been discussed on this substack recently. Many studies on SVA in general and CBCA in particular are based on laboratory studies. Only in a laboratory setting can scientists control which testimonies used to test SVA are true or false. Using testimonies from real-life criminal cases, on the other hand, is complicated, as it is often uncertain whether a given real-life testimony is actually true or false; which is the reason why SVA was developed in the first place. Another big problem is that of false memories; a topic which has also recently been discussed on this substack. While CBCA might be a useful tool for distinguishing between true statements and deliberate lies, it is not well suited for separating true testimonies from those based on false memories, since the latter frequently share many characteristics with truthful statements, such as richness of detail and contextual embedding. This is a problem that fundamentally affects all existing and future methods of lie detection. How can you prove the falsity of a testimony if it actually feels like the description of a real experience to the witness making that statement?

These and other problems of SVA ultimately lead us deep into the realm of philosophy. What is truth anyway? And will we ever be able to reliably discover it? These are questions that probably cannot be answered. But what we already know (or at least hopefully agree on) is that detecting lies is a very complex task, and that any method that promises a simple, quick and safe solution to this complex problem should be met with suspicion. As we have seen, SVA is neither simple nor fast, and does not always provide a clear answer. Perhaps that is why it is less well known and why it is not yet the subject of any popular TV show. But perhaps that is also what makes SVA a scientific method: its limitations.

That is all for today. I hope you enjoyed this brief excursion into the world of lie detection. By the way: Please don't forget to hit all the blue buttons on this page. Hit them hard. The owner of this substack really deserves a coffee break. Do you have any idea how much work it is to supervise a German Master's student? He certainly knows now.

---

[i] Vrij, A. (2005) Criteria-Based Content Analysis. A Qualitative Review of the First 37 Studies. Psychology, Public Policy, and Law 11(1): 3-41.

[ii] Oberlader, V., Naefgen, C., Koppehele-Gossel, J. et al. (2016) Validity of content-based techniques to distinguish true and fabricated statements: A meta-analysis. Law and Human Behavior 40(4): 440-457.