Deception detection is an interdisciplinary field attracting researchers from psychology, criminology, computer science, and economics. Automated deception detection presents unique challenges compared to traditional polygraph tests, but also offers novel economic applications. In this spirit, we propose an approach combining deep learning with discriminative models for deception detection. Therefore, we train CNNs for the facial modalities of gaze, head pose, and facial expressions, allowing us to compute facial cues. Due to the very limited availability of training data for deception, we utilize early fusion on the CNN outputs to perform deception classification. We evaluate our approach on five datasets, including four well-known publicly available datasets and a new economically motivated rolling dice experiment. Results reveal performance differences among modalities, with facial expressions outperforming gaze and head pose overall. Combining multiple modalities and feature selection consistently enhances detection performance. The observed variations in expressed features across datasets with different contexts affirm the importance of scenario-specific training data for effective deception detection, further indicating the influence of context on deceptive behavior. Cross-dataset experiments reinforce these findings. Notably, low-stake datasets, including the rolling dice Experiment, present more challenges for deception detection compared to the high-stake Real-Life trials dataset. Nevertheless, various evaluation measures show deception detection performance surpassing chance levels. Our proposed approach and comprehensive evaluation highlight the challenges and potential of automating deception detection from facial cues, offering promise for future research.

Experimental economic laboratories run many studies to test theoretical predictions with actual human behaviour, including public goods games. With this experiment, participants in a group have the option to invest money in a public account or to keep it. All the invested money is multiplied and then evenly distributed. This structure incentivizes free riding, resulting in contributions to the public goods declining over time. Face-to-face Communication (FFC) diminishes free riding and thus positively affects contribution behaviour, but the question of how has remained mostly unknown. In this paper, we investigate two communication channels, aiming to explain what promotes cooperation and discourages free riding. Firstly, the facial expressions of the group in the 3-minute FFC videos are automatically analysed to predict the group behaviour towards the end of the game. The proposed automatic facial expressions analysis approach uses a new group activity descriptor and utilises random forest classification. Secondly, the contents of FFC are investigated by categorising strategy-relevant topics and using meta-data. The results show that it is possible to predict whether the group will fully contribute to the end of the games based on facial expression data from three minutes of FFC, but deeper understanding requires a larger dataset. Facial expression analysis and content analysis found that FFC and talking until the very end had a significant, positive effect on the contributions.

Using a public goods laboratory experiment, this paper analyzes the extent to which face-to-face communication keeps its efficiency gains even after its removal. This is important as communication in real world is costly (e.g. time). If the effect of communication is long-lasting, the number of communication periods could be minimized. This paper provides evidence that there is a lasting positive effect on contributions even after communication was removed. Yet, after the removal, the contributions are lower and abate over time to the previous magnitude. This is referred to as the reverberation effect of communication. As we do not observe an effect of endogenizing communication, the strongest driver of the size of the contributions is the existence of communication or its reverberation. Eventually, the experiment provides evidence for a strong end-game effect after communication was removed, insinuating communication does not protect from the end-game behavior. In total, the results of the paper imply, that the effects of communication are not permanent but communication should be repeated. Simultaneously, results indicate no need for permanent communication. Since communication is conducted using video-conference tools, we present results from a machine learning based analysis of facial expressions to predict contribution behavior on group level.

In this paper, we investigate how technology has contributed to experimental economics in the past and illustrate how experimental economics can contribute to technological progress in the future. We argue that with machine learning (ML), a new technology is at hand, where for the first time experimental economics can contribute to enabling substantial improvement of technology. At the same time, ML opens up new questions for experimental research because it can generate previously impossible observations. To demonstrate this, we focus on algorithms trained to detect lies. Such algorithms are of high relevance for research in economics as they deal with the ability to retrieve otherwise private information. We deduce that most of the commonly applied data sets for the training of lie detection algorithms could be improved by applying the toolbox of experimental economics. To illustrate this, we replicate the “lies in disguise-experiment” by Fischbacher and Föllmi-Heusi with a modification regarding monitoring. The modified setup guarantees a certain level of privacy from the experimenter yet allows to record the subjects as they lie to the camera. Despite monitoring, our results indicate the same lying behavior as in the original experiment. Yet, our experiment allows an individual-level analysis of experimental data and the generation of a lie detection algorithm with an accuracy rate of 67%, which we present in this article.

MTVE is an open-source software tool (citeware) that can be applied in laboratory and online experiments to implement video communication. The tool enables researchers to gather video data from these experiments in a way that these videos can be later used for automatic analysis through machine learning techniques. The browser-based tool comes with an easy user interface and can be easily integrated into z-Tree, oTree (and other experimental or survey tools). It provides the experimenters control over several communication parameters (e.g., number of participants, resolution), produces high-quality video data, and circumvents the Cocktail Party Problem (i.e., the problem of separating speakers solely based on audio input) by producing separate files. Using some of the recommended Voice-to-Text AI, the experimenters can transcribe individual files. MTVE can merge these individual transcriptions into one conversation.