Social influence bias in ratings: A field experiment in the hospitality sector

We investigate the empirical phenomenon of rating bubbles, that is, the presence of a disproportionate number of extremely positive ratings in user-generated content websites. We test whether customers are influenced by prior ratings when evaluating their stay at a hotel through a field experiment that exogenously manipulates information disclosure. Results show the presence of (asymmetric) social influence bias (SIB): access to information on prior ratings that are above the average positively influences the consumers’ rating of the hotel. In contrast, information on ratings that are below the average does not affect reviewers. Furthermore, customers who have never been to the hotel before the intervention are more susceptible to prior ratings than customers who have repeatedly been to the hotel before. Finally, customers who are not used to writing online reviews are more prone to SIB than customers who frequently write online reviews. Our findings suggest that online rating systems should be adjusted to mitigate this bias, especially as these platforms become more relevant and widespread in the hospitality sector.

First-order visual interneurons distribute distinct contrast and luminance information across ON and OFF pathways to achieve stable behavior

The accurate processing of contrast is the basis for all visually guided behaviors. Visual scenes with rapidly changing illumination challenge contrast computation, because adaptation is not fast enough to compensate for such changes. Yet, human perception of contrast is stable even when the visual environment is quickly changing. The fruit fly Drosophila also shows nearly luminance invariant behavior for both ON and OFF stimuli. To achieve this, first-order interneurons L1, L2 and L3 all encode contrast and luminance differently, and distribute information across both ON and OFF contrast-selective pathways. Behavioral responses to both ON and OFF stimuli rely on a luminance-based correction provided by L1 and L3, wherein L1 supports contrast computation linearly, and L3 non-linearly amplifies dim stimuli. Therefore, L1, L2 and L3 are not distinct inputs to ON and OFF pathways but the lamina serves as a separate processing layer that distributes distinct luminance and contrast information across ON and OFF pathways to support behavioral performance in varying conditions.

Hybrid motion illusions as examples of perceptual conflict

Shapiro and Hedjar (2019) proposed a shift in the definition of illusion, from ‘differences between perception and reality’ to ‘conflicts between possible constructions of reality’. This paper builds on this idea by presenting a series of motion hybrid images that juxtapose fine scale contrast (high spatial frequency content) with coarse scale contrast-generated motion (low spatial frequency content). As is the case for static hybrid images, under normal viewing conditions the fine scale contrast determines the perception of motion hybrid images; however, if the motion hybrid image is blurred or viewed from a distance, the perception is determined by the coarse scale contrast. The fine scale contrast therefore masks the perception of motion (and sometimes depth) produced by the coarser scale contrast. Since the unblurred movies contain both fine and coarse scale contrast information, but the blurred movies contain only coarse scale contrast information, cells in the brain that respond to low spatial frequencies should respond equally to both blurred and unblurred movies. Since people undoubtedly differ in the optics of their eyes and most likely in the neural processes that resolve conflict across scales, the paper suggests that motion hybrid images illustrate trade-offs between spatial scales that are important for understanding individual differences in perceptions of the natural world.

Intervening in Anti-Immigrant Sentiments: The Causal Effects of Factual Information on Attitudes toward Immigration

If preferences on immigration policy respond to facts, widespread misinformation poses an obstacle to consensus. Does factual information about immigration indeed affect policy preferences? Are beliefs about immigration’s societal impact the mechanism through which factual information affects support for increased immigration? To address these questions, we conducted an original survey experiment, in which we presented a nationally representative sample of 2,049 Americans living in the United States with facts about immigrants’ English acquisition and immigrants’ impact on crime, jobs, and taxes—four domains with common misperceptions. Three of these factual domains (immigration’s impact on crime, jobs, and taxes) raise overall support for increased immigration. These facts also affect beliefs that are directly relevant to that information. Moreover, those beliefs mediate the effect of factual information on support for increased immigration. By contrast, information about English acquisition affects neither policy preferences nor beliefs about immigration’s impact. Facts can leverage social cognitions to change policy preferences.

Federated Disentangled Representation Learning for Unsupervised Brain Anomaly Detection

Recent advances in Deep Learning (DL) and the increased use of brain MRI have provided a great opportunity and interest in automated anomaly segmentation to support human interpretation and improve clinical workflow. However, medical imaging must be curated by trained clinicians, which is time-consuming and expensive. Further, data is often scattered across multiple institutions, with privacy regulations limiting its access. Here, we present FedDis (Federated Disentangled representation learning for unsupervised brain pathology segmentation) to collaboratively train an unsupervised deep convolutional neural network on 1532 healthy MR scans from four different institutions, and evaluate its performance in identifying abnormal brain MRIs including multiple sclerosis (MS) lesions, low-grade tumors (LGG), and high-grade tumors/glioblastoma (HGG/GB) on a total of ~500 scans from 5 different institutions and datasets. FedDis mitigates the statistical heterogeneity given by different scanners by disentangling the parameter space into global, i.e., shape and local, i.e., appearance. We only share the former with the federated clients to leverage common anatomical structure while keeping client-specific contrast information private. We have shown that our collaborative approach, FedDis, improves anomaly segmentation results by 99.74% for MS and 40.45% for tumors over locally trained models without the need for annotations or sharing private local data. We found out that FedDis is especially beneficial for clients that share both healthy and anomaly data coming from the same institute, improving their local anomaly detection performance by up to 227% for MS lesions and 77% for brain tumors.

Never waste a “good” crisis! Priming the economic aspect of crises fosters social capital build-up and prosociality

Do crises make people more prosocial? And what role does communication play in promoting such attitudes and behavior? These answers matter for post-crisis economic recovery as social capital has been linked to growth. We leverage the incidence of Covid-19 --a multifaceted global crisis-- and using a representative panel of US residents, surveyed in April and October 2020, we explore how a) pandemic-induced economic and health anxiety map to prosocial inclinations and behavior, and b) whether communication (and what types) can foster social capital formation. We find that individual exposure to the economic and health consequences of the pandemic had no effect on prosocial inclinations and social capital; but perceived economic vulnerability reduced trust in government and respect for authority and increased preferences for redistribution. Yet information about the aggregate economic consequences of Covid-19 fosters social capital build-up (e.g., altruism, giving, patience) and prosocial preferences. In contrast, information about the health costs of the pandemic has the opposite effect; it greatly reduces interpersonal trust. These information effects also map into policy preferences beyond the Covid-19 crisis. Our findings are consistent with cultural accounts on the determinants of Americans' prosocial inclinations and preferences.

Factors that influence clinical efficacy of live biotherapeutic products

Traditional probiotics are increasingly being used in a medical context. The use of these products as drugs is considerably different from the traditional use as food or food supplements, as, obviously, the target population is different (diseased versus healthy or at risk population). Besides the target population, also the regulatory context is different, mainly with respect to production, administration regime and type of clinical studies required. In this paper we will, besides the regulatory differences, focus on aspects that may impact the efficacy of a live biotherapeutic product (drug), especially in a clinical setting. The impact of the dosage seems to depend on the strain and the application and may follow some rationale. In contrast, information on the impact of the time of administration or diet, is often still lacking. The matrix and the use of protective measures may clearly have an impact on the survival and efficacy of the strain.

Structural Influences on the News Finds Me Perception: Why People Believe They Don’t Have to Actively Seek News Anymore

Using data from a two-wave panel survey among 18 countries worldwide, this study investigates the individual- and country-level antecedents of the “News Finds Me” perception (NFM). Results show that older, more educated, and individuals belonging to the ethnic majority are less prone to develop the NFM. However, social media (news) use, incidental news exposure, discussion frequency, and group affiliations lead to a higher NFM. In contrast, information elaboration as well as news use online were found to weaken the NFM. Testing various country-level factors, only gross domestic product was found to be negatively related to the NFM. The findings form a theoretical and empirical basis for future studies that aim at investigating news use in today’s high-choice media environment.