Predicting face similarity judgements with a computational model of face space

Current models of face representation involve the notion of a high-dimensional face space. Computational models of face space based on principal components analysis (PCA) have been successfully used to predict human judgements of face sex or race. In this work the capability of PCA-based face spaces to predict human judgements of face similarity is examined. Three different paradigms were used. In Experiment 1 subjects learned face-name associations for 18 faces and identified these faces on tachistoscopic presentation. The number of confusions was used as a measure of face similarity. In Experiment 2 the same subjects subjectively rated the similarity of all 153 possible face pairs. In Experiment 3 reaction time to identify a face in an odd-man-out task was measured as an index of face similarity. These empirical measures were correlated with distance of the faces in PCA-based spaces of different dimensionalities. For Experiments 1 and 2 these correlations were highest for one-or two-dimensional face spaces (r=−0.27 vs. −0.28). For Experiment 3 the correlation was highest for a space consisting of 13 dimensions (r=−0.51). Thus PCA-based spaces seem capable to predict human similarity judgements to some extent. Possible reasons for the differences in predictability between paradigms are discussed.     

Personality perceptions and relationship satisfaction in couples

The present study investigates how perceptions of personality are related to relationship satisfaction in an age-heterogeneous sample of romantic couples. Self- and partner-perceptions as well as perceived similarity and self-other agreement were examined separately for the Big Five personality traits. Results of Actor–Partner-Interdependence Models revealed substantial effects of partner-perceived personality in all Big Five traits on both partners’ relationship satisfaction. In contrast, effects of self-perceived personality on relationship satisfaction were small. Over and above self- and partner-rated personality, perceiving one’s partner as similar to oneself made a small unique contribution to relationship satisfaction in couples. These results emphasize the importance of integrating self- and partner-perceptions of personality for relationships outcomes.     

Social comparison processes in the experience of personal relative deprivation

Drawing on social comparison and equity theories, we investigated the role that perceived similarity of a comparison target plays in how resentful people feel about their relative financial status. In Study 1, participants tended to choose a comparison target who was better off, and they selected a target they perceived to be more similar than dissimilar along dimensions that surrounded their financial outcomes. In Study 1, perceived relative disadvantage was positively associated with resentment regardless of the perceived similarity of the comparison target. The results of Studies 2 to 5b clarified these findings by showing that being both similar and dissimilar to a target can cause resentment depending on the context. Using hypothetical and real social comparisons, we found that people are more dissatisfied with their financial outcomes when their comparative targets have the same background qualifications (i.e., are similar) but are financially better off (Studies 2, 3b, 4, and 5b). However, we also found that when the comparative financial contexts were similar (i.e., equal affluence), participants were more dissatisfied when their target for comparison had lower qualifications (i.e., was dissimilar; Studies 2, 3a, 4, and 5a). In all cases, perceptions of unfairness mediated the effects of social comparison on financial dissatisfaction. Taken together, these studies address some of the ambiguities around what it means to be similar to a target in the context of social comparisons of affluence, and they underscore the importance of perceived unfairness in the link between social comparison and resentment with one's financial status.     

The Role of Surface Similarity in Analogical Retrieval: Bridging the Gap Between the Naturalistic and the Experimental Traditions

Blanchette and Dunbar (2000) have claimed that when participants are allowed to draw on their own source analogs in the service of analogical argumentation, retrieval is less constrained by surface similarity than traditional experiments suggest. In two studies, we adapted this production paradigm to control for the potentially distorting effects of analogy fabrication and uneven availability of close and distant sources in memory. Experiment 1 assessed whether participants were reminded of central episodes from popular movies while generating analogies for superficially similar versus superficially dissimilar target situations. In Experiment 2, we modified this procedure to assess the retrieval of autobiographic memories, more familiar to participants than fictional episodes. Both studies revealed a strong effect of surface similarity on the retrieval of participants' own sources, thus suggesting that the superficial bias typically observed in experimental studies—and simulated by most computational models—does not originate in a lack of ecological validity.     

My Mentor,My Self: Antecedents and Outcomes of Perceived Similarity in Mentoring Relationships

This study examines the antecedents and outcomes of perceived similarity in mentoring relationships in a sample of 82 matched mentor-protégé dyads. Polynomial regression analyses were used to examine the association between actual and perceived similarity. Protégés were more likely to see themselves as similar to their mentors when they and their mentors shared either higher or lower levels of attachment security. However, sharing similar levels of relational self-construal was unrelated to protégés perceptions of similarity with their mentors. Path analyses revealed that protégés reported more organizational and professional commitment when they saw themselves as similar to their mentors, and that these associations were mediated by protégés’ reports of their mentors providing role modeling functions. These findings suggest that the effects of actual mentor-protégé similarity on perceptions of similarity may depend on the shared attribute, and provide further evidence of the importance of perceived similarity to not only mentoring relationships, but also to organizations.     

Deliberate Numerical Discrepancy of Generation in the Genealogy of Matthew

In the genealogy of Matthew, there is a numerical ‘discrepancy’ between the actual totals of generations (fourteen, fourteen, and thirteen) recorded in Matthew 1:12-16 and the stated totals (fourteen, fourteen, and fourteen) in 1:17.

This paper argues that Matthew's genealogy is devised by inserting breaks based on the pattern and concept of Chronicles in order to convey the unique Matthean message. This argument will be defended by showing the similarities between Chronicles and Matthew: (1) Matthew follows the Chronicler by putting the genealogy at the beginning of the whole book as an introduction to convey the Matthean messages; (2) by combining concepts of kingship and priesthood; (3) by mentioning ‘the sons of Israel’ and ‘all Israel’; (4) by including Gentiles; (5) by using retribution theology; (6) by employing intensification in his chiastic genealogical structure; (7) Matthewâ..s omission of several kings is in line with the Chronicler's ideology of cultic-oriented kingship; (8) Matthew shares the Chronicler's point of view that Judah's deportation to Babylon does not invalidate Yahweh's promise for the Davidic kingdom. Thus, the apparent numerical ‘discrepancy’ is deliberately devised by the Matthean author, just as the Chronicler uses the numerical discrepancy to highlight his emphases in the genealogies.     

The role of perceptual similarity of the task environments in children’s perseverative responding

Perseverative responding is often seen in children’s performance in a variety of contexts. One such context is symbolic comprehension in which 2- and 2½-year-olds demonstrate difficulty in appreciating the association of symbols (pictures and scale models) and their referents and show a high proportion of perseverative responding. Representational-level explanations of perseveration were explored in the current studies via examination of the impact of perceptual similarity of visual environments across trials. Across two experiments, children saw either a picture (Experiment 1) or a scale model (Experiment 2) of a hiding location of a room and were then encouraged to recover the toy from an identical room. Manipulating the perceptual similarity of the environments across successive trials affected performance and perseverative responding. These results highlight the critical role played by perceptual information not only in symbolic tasks but also in many other tasks and have important implications for theories of perseveration.     

Trust Development: Testing a New Model in Undergraduate Roommate Relationships

Interpersonal trust is a vital component of social relationships. In this study the roles of parental attachment, perceived similarity of trustee to self, and social exchange processes in trust development were investigated longitudinally with randomly assigned, same-sex undergraduate roommates during emerging adulthood. A total of 214 first-year students completed weekly self-report measures during the first 5 weeks of the fall semester. Perceived similarity measured the second week and social exchange with roommates across the 5 weeks predicted participants’ trust in their roommate, with social exchange mediating the relation between perceived similarity and trust. Results highlight interrelations of social exchange and trust in established relationships.     

Agreement and similarity in self-other perceptions of moral character

We surveyed well-acquainted dyads about two key moral character traits (Honesty–Humility, Guilt Proneness), as well as several other individual differences. We examined self-other agreement, similarity, assumed similarity, and similarity-free agreement (i.e., self-other agreement controlling for similarity and assumed similarity). Participants projected their own level of moral character onto their peers (i.e., moderately high assumed similarity), but were nonetheless able to judge moral character with reasonable accuracy (moderately high self-other agreement and similarity-free agreement), suggesting that moral character traits can be detected by well-acquainted others. Regardless of reporting method, Honesty–Humility and Guilt Proneness were correlated with delinquency, unethical decision making, and counterproductive work behavior, suggesting that unethical behavior is committed disproportionately by people with low levels of these character traits.