Posters: Social Psychology

The authors tested a new theory of persuasion as a sequential process (PSP). Drawing on assumptions from both the Heuristic-Systematic-Model and the Parametric Unimodel of Persuasion, PSP predicts assimilation and contrast effects based solely on the sequence in which identical arguments are encountered. Specifically, if an initial argument is immediately followed by an argument of opposite valence (vs. a neutral argument), contrast effects are predicted to occur, despite the fact that the overall information is held constant. In a laboratory experiment, students (N = 216) read four arguments for or against the abolishment of cash. They were assigned randomly to the conditions of a 2 (initial argument: positive vs. negative) x 2 (subsequent three arguments: neutral, positive, very positive vs. neutral, negative, very negative) x 2 (order of subsequent arguments: neutral to extreme vs. extreme to neutral) between-subjects-design. After each argument, they listed one thought related to that argument; thoughts were subsequently rated as favorable versus unfavorable and aggregated into a thought valence index. Participants also reported their overall attitude toward the abolishment of cash. A-priori contrast analysis fully supported the predictions of PSP for the pattern of attitudes. In addition, the effects of experimental conditions on attitudes were fully mediated by thought valence. Theoretical and applied implications will be discussed.

“Fake News” – the term that The American Dialect Society elected Word of the Year 2017, became one of the most discussed phenomena among scholars and the public alike. While there is research on dissemination mechanisms and prevalence (Guess, Nagler, & Tucker, 2019), potential interventions (Lazer et al., 2018) and susceptibility to fake news (Pennycook & Rand, 2019) there are no empirical studies on how exposure to fake news affects individuals on a cognitive level. Previous studies show that corrections of political misperceptions can fail to reduce these misperceptions among the targeted ideological group (Nyhan & Reifler, 2010) and potentially increase ingroup identification when the sender is perceived an ingroup member (McKimmie et al., 2003). We suggest that this principle could also hold true for fake news and that exposure to and subsequent correction of fake news can lead to affective and attitudinal polarization. In an experimental between-subjects setting four fake news stimuli were displayed. The information that the respective stimulus was factually wrong was disclosed either early on or at the end of the study. We find that if the sender of the message is perceived an ingroup member and if identification with that ingroup is high, individuals displayed more extreme policy attitudes when fake news information was disclosed early (Study 1). A similar pattern was true for affective polarization where the distance between ingroup-like and outgroup-dislike was maximized (Study 2). Among other things, these results highlight the importance of social identity processes in political polarization.

The aim of our research was to assess short-term and long-term effects of social comparative social media use on self-esteem and depressive tendencies. For this, we conducted three different studies. First, we conducted an exposure experiment (N =75) including two experimental groups and one control group showing that social comparative internet use decreased participants’ performance-oriented state self-esteem as a short-term effect. Second, we conducted two correlational studies (Study 2 &3, Ns = 809, 145) indicating that passive Facebook use is associated with higher depressive tendencies mediated by a higher ability-related social comparison orientation and lower self-esteem as long-term effect. Third, to obtain more generalizable findings, we transferred the serial multiple mediator model successfully from private (i.e., Facebook) to professional (i.e., XING) SNS use (Study 3).

Studies on person perception showed that stereotypes can be activated by presenting labels associated to particular groups. However, it is not clear whether these semantic information activate stereotypes directly, or via an indirect cognitive pathway leading through brain regions responsible for affective responses. To disentangle the effects of semantic and affective content, first we scrutinized whether the representation of occupational labels is independent of the emotions they evoke. Participants (n=70) were asked to complete two tasks. In the first task they had to arrange 20 labels, randomly chosen from an item pool of 60 labels representing occupations, on a two-dimensional surface along the two axes. Participants were free to use any considerations for the arrangement along the axes which they thought makes a proper differentiation among the labels. With this spatial arrangement method, based on the coordinates of the labels, the relative semantic distance between the occupations were calculated. In the second task the axes’ names were defined a priori. Subjects were asked to arrange the labels according to valence and arousal, which referred to the extent to which the words evoked pleasant or unpleasant feelings, and excitement or calmness, respectively. The relative difference between the coordinates gave the affective distances between each of the labels. Based on the semantic and affective distances, two separate cluster analyses were carried out. The comparison of the two cluster structures revealed significant overlap, suggesting that semantic representation of occupations largely relies on the affective content evoked by the presentation of the labels.

Anticipation of pain triggers defensive responses such as startle potentiation, while its avoidance and termination result in positive sensations such as startle attenuation, called relief. These responses can become associated with and later elicited by other stimuli concurrently present. Importantly, social support reduces aversiveness of pain. We investigated the effect of social influence on relief responses and whether active relief (avoidance) differs from passive relief (pain termination) in 102 healthy females. During acquisition, the active group (N=33) learned to actively oppress a painful stimulation (unconditioned stimulus, US), the social group (N=35) believed that another participant oppressed the US, and the passive group (N=34) had no influence. A visual stimulus (conditioned stimulus, CS+) followed US' termination. Subsequently (test phase), participants heard aversive startle probes presented with the CS+ or a novel visual stimulus (Control). Startle responses and fear ratings were collected as learning indices. After acquisition, all participants rated CS+ as more frightening than Control, suggesting that on the verbal level, the relief-associated stimulus elicited fear in all groups. After test, fear ratings of CS+ further increased but did not differ from Control. The same was evident for startle responses to CS+ and Control in both passive and active group. Thus, relief-physiological responses indicate equal implicit valence for passive and active relief. In comparison, the social group showed overall lower startle responses. However, responses were higher to CS+ than New. In sum, we confirmed that social influence reduces physiological fear response in general, but inhibits implicit relief-learning.

Generalized reciprocity is the phenomenon of reciprocating experienced behavior towards a third party. We investigated if generalized reciprocity occurs in a double dictator game (DDG) and if individual differences therein can be explained through personality traits. In a DDG, participants first receive money from a previous participant and then allocate additional money between themselves and a future participant. We hypothesized that there is evidence for generalized reciprocity, that is, we expected that the amount of money a participant received in the DDG would influence the amount this participant gave to a future participant. In addition, we investigated the relationship of generalized reciprocity with the HEXACO personality factors Agreeableness and Honesty-Humility. Due to its association with direct reciprocity, we predicted Agreeableness would moderate the relationship between the amount received and the amount given so that the relationship would be stronger for individuals low in Agreeableness. Additionaly, we expected that Honesty-Humility would explain the overall contributions of participants (regardless of the behavior of previous participants). In a preregistered online experiment, the amount of money received and Honesty-Humility predicted – as expected – the amount of money a participant gave to future participants. However, Agreeableness did not significantly moderate the relationship between the amount received and the amount given. In sum, the current study provides further evidence and a suitable paradigm for investigating generalized reciprocity in economic games. At the same time, the influence of Agreeableness on generalized reciprocity remains uncertain and requires more attention in future studies.

In this preregistered study, a subtle method was tested to reduce prejudice against one outgroup (Arabs) and generalize its effects to another non-targeted outgroup (Eastern Europeans). Students read a pretested alleged letter addressed to their university insisting on certain changes of the university’s infrastructure. Participants then signed a letter on a page that contained signatures of either only German students or German and Arab students. Combining elements of the contact hypothesis (institutional support, common goal, equal status, and cooperation) without inducing actual vis-á-vis contact, we hypothesized that the German and Arab condition (vs. the German only condition) (H1) would increase positive attitudes toward Arabs and (H2) that this effect would generalize to attitudes toward Eastern Europeans. Dependent variables were explicit evaluations, specific prejudices, and implicit associations toward both target groups. Social dominance orientation (SDO) and authoritarianism (RWA) were assessed as potential moderators. Supporting hypothesis H1, participants explicitly evaluated Arabs more positively in the German and Arab condition. However, opposite to hypothesis H2, participants tended to evaluate Eastern Europeans less positively in that condition, which resulted in an interaction effect of condition and target group. The prejudice measure and the implicit measure showed no effects. SDO and RWA were correlated with the dependent variables but did not moderate the effect of experimental condition. Results will be discussed in light of the contact hypothesis and social identity theory.

Every day millions of people perceive discrimination, yet discrimination is also overlooked or wrongly accused. In our line of research, we are interested in socio-cognitive processes underlying discrimination perception. Since discrimination is defined as the process by which people are treated differently (especially unfair) because of their group membership (e.g. Jary & Jary, 1995), the main question to be answered in discrimination perception is whether the difference in treatment is attributable to differences in group membership rather than a socially accepted feature (SAF; e.g. performance). Hence, to perceive discrimination, people must infer causal priority of the group membership over SAFs. As such, the question of discrimination perception would essentially concern the perception of covariation of group and treatment mediated by the SAF. By measuring the perception of this partialized covariation, one would be able to access participants’ discrimination perception with reduced biases. In our experiment, we investigated participants’ (N=86) sensitivity to covariations between two variables in the presence of a potential mediator or suppressor in dependence of different question formats. The results revealed significant differences between the three formats and indicate that people may be sensitive to trivariat covariations if the right question format is used. With this new perspective, we attempt to shed light on processes underlying discrimination perception through a new paradigm to represent discrimination in a laboratory setting. We designed a new sampling task to demonstrate the precision of the new measurement under various variables with previously observed effects on discrimination perception. Results are still pending.

Empathy, i.e., sharing another’s feelings, and reciprocity, i.e., reciprocating kindness, are two strong motives for prosocial behaviours. However, so far it is unclear which of the two motives is more sustainable, i.e., more likely to elicit prosocial behaviour over a longer period of time, even in the absence of receiving further reinforcement. Here we use a learning theoretical approach to investigate how the empathy and the reciprocity motive develop and decay over time. In a first experimental phase (corresponding to acquisition), we reinforced the respective motive (empathy/reciprocity) with high probability (80%) and in a second phase (corresponding to extinction) with low probability (20%). In a parallel control condition, the motive was randomly reinforced (50%) in both phases. When modelling motive strength over time using a variant of the Rescorla-Wagner model, we observed that the strength of the reciprocity motive closely mirrors the frequencies of reinforcement in the respective phase, i.e., increased in the first phase and decreased in the second phase. In contrast, the strength of the empathy motive increased in the first phase and persevered in the second phase, indicating that the empathy motive might be more sustainable than the reciprocity motive. Together, these preliminary results suggest that (i) motives driving the same observable behaviour can differ decisively in their development over time and perseverance, and (ii) these differences can be well captured and described in terms of a learning model.