1. Introduction
Users’ feedback is becoming more and more important in many different contexts of users’ interaction with computerised systems, such as in recommender systems, social network, e-democracy, quantified-self, affective computing and also in the IOT world. Most of these systems need users’ feedback for their proper operation (i.e. recommender systems, affective computing-based systems, reputation systems), or because it is linked to their inner nature (e-democracy and social network systems), or for adapting their behaviour according to a specific user’s behaviour while using an object in the real or virtual world (IOT and user-adaptive systems). Hence stimulating users to provide explicit feedback becomes an important challenge, especially as users are reluctant to provide it and using and relying on implicit feedback has its well-known limitations.
The explicit and/or implicit collection of user’s feedback (opinions, ratings, likes, physiological states, usage of virtual or tangible objects, etc.) is a central feature in the design of interactive, personalised systems, and their design may have an impact on the way the feedback is collected and interpreted.
2. Special issue papers contributions
The special issue examines factors that may influence users’ willingness to offer feedback, how feedback may be solicited and ways to encourage/convince users to provide it, as well as the impact of users’ feedback on the evaluation of system performances. More specifically, the accepted papers focus on several different aspects:
An application of digital nudging as a conceptual basis for interventions in social network privacy-related issues that is similar to persuasion but focuses on informed and consistent decision-making (Kroll and Stieglitz Citation2019)
An example of a novel modality of feedback elicitation by means of game adapted to real-time user conditions (Kihl, Park, and Park Citation2019)
Uncertainty in explicit user feedback, as well as its impact on the evaluation of system performances (Jasberg and Sizov Citation2019).
The first paper deals with persuasive elements on Facebook in relation to self-disclosure and privacy, as well as an investigation of the effect of such elements on trust, concern, perceived control and disclosure. In particular, the authors investigate on factors that influence self-disclosure in social network sites. After the identification of several persuasive elements in Facebook the authors performed a quantitative study, consisting of an online survey that was accomplished by 382 respondents. Some of the obtained results partially confirm literature evidences about influencing factors of self-disclosure and the role of privacy-related nudges, while some others provide new insights.
In particular, the two identified privacy-related nudges do not yield clear results. The results differ with respect to short- and long-term changes in perceptions. The authors have found indications that nudges may have a converse effect, meaning that reminders to change privacy settings trigger privacy concerns. The privacy-enhancing actions (taken by Facebook) do not have a long-term effect but may cause short-term changes, potentially leading to less interaction and information disclosure. When people are reminded of privacy issues, they seem to immediately feel a threat to their personal data and hence they perceive less control and a higher risk. According to Kroll and Stieglitz (Citation2019) more elaborated variations of privacy-related nudges (i.e. leveraging persuasion) may have a stronger effect even directly on self-disclosure for a longer period.
The main contribution of this paper in relation to the special issue theme is related to the implicit modalities for acquiring user feedback and the importance related to the privacy issue behind all the tracks that users leave in the (social) web.
The second paper of the special issue, (Kihl, Park, and Park Citation2019), is an example of a novel modality of feedback elicitation, exploiting a gamified approach enriched with real-time user’s physiological data provided by sensors. In this way, the game dynamically unfolds the story based on user response and she/he is stimulated to provide feedback, in this case to participate in the game.
In particular, the paper describes a system which supports children with Attention Deficit/Hyperactivity Disorder (ADHS) reading fairy tales by providing a gamified interactive reading experience where the reader acts as a helper of the character, and real-time feedback based on their concentration and behaviour monitored by sensors. While the child is reading the fairy tale, the platform continuously captures the attention and meditation values of the player through a brain–computer interface, and the motions of the player using a motion-sensing device. Then, the fairy tale narrative dynamically reacts (changing, for example, the text provided) to the players attention or behaviour, so the participant feels as if she/he is one of the characters in the game.
The approach has been tested both quantitatively and qualitatively, showing that it helps children with ADHD to improve their reading ability and attention span. The analysis of the sensors’ data shows improvements in the attention span and decreases in hyperactive behaviour over time. The analysis of the interview data supports the quantitative test results, in particular children participating in the reading were found to feel satisfaction from understanding the texts on their own and felt immersed in the game because of the feedback provided by the sensors.
Thus, the main contribution of this paper in relation to the special issue theme is related both to the use of (i) a gamified interaction modality; (ii) adapted to real-time user response in order to favour the user engagement with the system and thus feedback provision.
Starting from the well-grounded observation (see, e.g. Hill et al. Citation1995; Herlocker et al. Citation2004; Said et al. Citation2012) that repeated evaluations of the same item on the part of the same user lead to different results, the third paper, by Jasberg and Sizov (Citation2019), discusses the intrinsic unreliability of explicit user feedback, an issue that cannot be solved by simply collecting more data. Such variability is relevant to intelligent systems as it has a huge impact on the evaluation of their performance: in fact, it is unclear whether possible discrepancies between system predictions and their corresponding user evaluations should be ascribed to algorithmic flaws or to ‘human uncertainty’.
Borrowing from the fields of physics and metrology, the authors first investigate how uncertainty can be modelled and measured and then analyses the impact of uncertain data on prediction accuracy. Finally, they discuss the effectiveness of existing solutions and suggest that developing novel user interfaces for soliciting user feedback might help to decrease its intrinsic unreliability without causing distortions in user evaluations.
Considering the topic of this special issue, the paper by Jasberg and Sizov (Citation2019) provides different important contributions, in particular: (i) it suggests that systems that make use of explicit user feedback should always take into account its intrinsic uncertainty when interpreting it; (ii) it recommends that approaches and interfaces for soliciting user feedback should not primarily aim at quantity, but rather at reducing uncertainty.
3. Summary
All in all, the papers span a large variety of users’ feedback-related aspects, they demonstrate the subtle nature of persuasion/soliciting feedback and its potential negative impact, they present the potential contribution of sensor-based feedback that requires no active user involvement, and they present the uncertain nature of users' explicit feedback. It seems that with the advancement of technology, possibly relying on sensor-based measured feedback, the need to persuade users to provide feedback will diminish, but how to interpret this feedback may still be a challenge.
References
- Herlocker, Jonathan L., Joseph A. Konstan, Loren G. Terveen, and John T. Riedl. Jan. 2004. “Evaluating Collaborative Filtering Recommender Systems.” ACM Transactions on Information Systems 22 (1): 5–53. doi:10.1145/963770.963772.
- Hill, Will, Larry Stead, Mark Rosenstein, and George Furnas. 1995. “Recommending and Evaluating Choices in a Virtual Community of Use.” In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI ‘95). ACM Press/Addison-Wesley Publishing Co., New York, NY, USA, 194–201. doi:10.1145/223904.223929.
- Jasberg, Kevin, and Sergej Sizov. 2019. “Human Uncertainty in Explicit User Feedback and its Impact on the Comparative Evaluations of Accurate Prediction and Personalisation.” this issue (2019).
- Kihl, Taesuk, Kyungeun Park, and Seungie Park. 2019. “Fairy Tale Directed Game-Based Training System for Children with ADHD using BCI and Motion Sensing Technologies.” this issue (2019).
- Kroll, Tobias, and Stefan Stieglitz. 2019. “Digital Nudging and Privacy: Improving Decisions about Self-disclosure in Social Networks.” this issue (2019).
- Said, Alan, Brijnesh J. Jain, Sascha Narr, and Till Plumbaum. 2012. “Users and Noise: The Magic Barrier of Recommender Systems.” In User Modeling, Adaptation, and Personalization, edited by Judith Masthoff, Bamshad Mobasher, Michel C. Desmarais, and Roger Nkambou, 237–248. Berlin, Heidelberg: Springer Berlin Heidelberg.