Bidirectional feature selection with global and local structure preservation for small size samples

Collecting samples is a challenging task for face recognition, especially for some real-world applications such as law enhancement and ID card identification, where there is usually single sample per person (SSPS) used to train a face recognition system. To extract discriminative features from the small size samples, in this paper we propose virtual samples via bidirectional feature selection with global and local structure preservation (VS-BFS-GL) to augment the number of training samples. In VS-BFS-GL, bidirectional feature selection is developed, which introduces L_2,1 norm to explore the face variations from both horizontal and vertical directions. Further, to include more variations in the virtual images, the global structure information and sample-specified local structure information of the SSPP training set are considered. By integrating bidirectional feature selection, global and local structure, the limited training samples are fully utilized and more knowledge are mined. To further improve the effectiveness of VS-BFS-GL, an auxiliary database containing different face variations can be used to explore the local structure information. We extensively evaluated the proposed approach on AR and FERET database. The promising recognition results demonstrate that VS-BFS-GL is robust to expression, pose and partial occlusion variations in the faces.     

Investigating How Test-Takers Change Their Strategies to Handle Difficulty in Taking a Reading Comprehension Test: Implications for Score Validation

This article investigates how test-takers change their strategies to handle increased test difficulty. An adult sample reported their test-taking strategies immediately after completing the tasks in a reading test. Data were analyzed using structural equation modeling specifying a measurement-invariant, ability-moderated, latent transition analysis in Mplus (Muthén & Asparouhov, 2011). It was found that almost half of the test-takers (47%) changed their strategies when encountering increased task-difficulty. The changes were characterized by augmenting comprehending-meaning strategies with score-maximizing and test-wiseness strategies. Moreover, test-takers' ability was the driving influence that facilitated and/or buffered the changes. The test outcomes, when reviewed in light of adjusted test-taking strategies, demonstrated a form of process-based validity evidence.     

Heuristics versus direct calculation,and age-related differences in multiplication: an evidence accumulation account of plausibility decisions in arithmetic

We offer a new theoretical angle for cognitive arithmetic, which is that evidence accumulation may play a role in problem plausibility decisions. We build upon previous studies that have considered such a hypothesis, and here formally evaluate the paradigm. We develop the finding that performance differences, due to variations in strategy use and aging effects, can indeed be reasonably explained through these accumulation-to-bound cognitive models. Results suggest that these models may be effectively used to learn more about the underlying cognitive processes. In this study, we modelled young (18–24) and older (68–82) adults’ solution times in performing arithmetic verification (e.g. whether 8?×?5?=?41 is true/false). The domain-relevant factors in strategy use (problem-verification heuristics) and aging differences (older/younger adult groups) were analyzed by a response process model of the latency data, that is fit by participant and item. Lower thresholds accounted for the faster response times (RTs) for problems solved with heuristics (arithmetic rule-violation checking strategies), as opposed to problems solved by calculation approaches. A more rapid accumulation accounted for faster RTs on problems in which two arithmetic rules were violated (strategy combination) rather than one. Third, higher thresholds (i.e. preferring to have greater certainty before responding) accounted for older adults’ slower speed. These findings are in support of accumulation models being relevant for more complex cognitive tasks, as well as to account for the age-related differences therein.     

Women’s hair as a cue to desired relationship and parenting characteristics

We investigated how women’s hair color (blond, brown, black) and length (short, medium, long) influences males’ judgments about the women’s age, health, physical attractiveness, relationship potential, and parenting capability. Results, which are generally consistent with evolutionary psychology approaches, indicate that hair color and, to a lesser extent, length can affect perceptions of personal characteristics. More specifically, we found that lighter hair (blond and brown) compared to darker hair (black) is generally associated with perceptions of youth, health and attractiveness, and generally leads to more positive perceptions of relationship and parenting potential. Furthermore, the relationships between variables suggest that characteristics directly related to reproductive potential may be inferred from more obvious indirect characteristics. These results suggest that males are able to make complex judgments about women concerning their desirable relationship and parenting potential based on discernable characteristics such as hair color and length.     

Does mediator use contribute to the spacing effect for cued recall? Critical tests of the mediator hypothesis

When study is spaced across sessions (versus massed within a single session), final performance is greater after spacing. This spacing effect may have multiple causes, and according to the mediator hypothesis, part of the effect can be explained by the use of mediator-based strategies. This hypothesis proposes that when study is spaced across sessions, rather than massed within a session, more mediators will be generated that are longer lasting and hence more mediators will be available to support criterion recall. In two experiments, participants were randomly assigned to study paired associates using either a spaced or massed schedule. They reported strategy use for each item during study trials and during the final test. Consistent with the mediator hypothesis, participants who had spaced (as compared to massed) practice reported using more mediators on the final test. This use of effective mediators also statistically accounted for some – but not all of – the spacing effect on final performance.     

Object-Tool-Actor Interaction: Object Information Drives Intended Action

Tool use is typically explored via actor-tool interactions. However, the target-object (that which is being acted on) may influence perceived action possibilities and thereby guide action. Three different tool-target-object pairings were tested (Experiment 1). The hammering action demonstrated the greatest sensitivity and therefore subsequently used to further investigate target-object pairings. The hammer was removed as an option and instructions were provided using pictorial (Experiment 2), written (Experiment 3), and both pictorial and written formats (Experiment 4). The designed tool is chosen when available (Experiment 1) and when removed as a choice (i.e., the hammer), participants perform the same action associated with the designed tool (i.e., hammering) regardless of instruction method (Experiments 2, 3, and 4).     

The Wished‐For Always Wins Until the Winner Was Inevitable All Along: Motivated Reasoning and Belief Bias Regulate Emotion During Elections

How do biases affect political information processing? A variant of the Wason selection task, which tests for confirmation bias, was used to characterize how the dynamics of the recent U.S. presidential election affected how people reasoned about political information. Participants were asked to evaluate pundit‐style conditional claims like “The incumbent always wins in a year when unemployment drops” either immediately before or immediately after the 2012 presidential election. A three‐way interaction between ideology, predicted winner (whether the proposition predicted that Obama or Romney would win), and the time of test indicated complex effects of bias on reasoning. Before the election, there was partial evidence of motivated reasoning—liberals performed especially well at looking for falsifying information when the pundit's claim predicted Romney would win. After the election, once the outcome was known, there was evidence of a belief bias—people sought to falsify claims that were inconsistent with the real‐world outcome rather than their ideology. These results suggest that people seek to implicitly regulate emotion when reasoning about political predictions. Before elections, people like to think their preferred candidate will win. After elections, people like to think the winner was inevitable all along.     

APPLICANT REACTIONS TO PRE-EMPLOYMENT APPLICATION BLANKS: A LEGAL AND PROCEDURAL JUSTICE PERSPECTIVE

The current research investigated applicant reactions derived from common application blanks. Previous work has found applications to be riddled with legally inadvisable items, but limited research has attempted to understand how such items might affect applicants. By using organizational justice theory as a framework, two studies examined applicant reactions from two application blanks: a legally advisable application and a legally problematic application. It was found that applicants completing the legally problematic application had lower perceptions of justice and expressed higher litigation intentions compared to those completing the legally advisable application, especially for those that were rejected without an explanation.

Don't hate me because I'm beautiful: Anti-attractiveness bias in organizational evaluation and decision making

Within organizational judgment and decision making contexts, biases based on an evaluated person's attractiveness are among the most salient and frequently investigated. An enormous amount of research indicates favoritism for attractive people compared to unattractive ones. The current research demonstrates that the nature of this bias depends on whether one is evaluating a member of the same sex or the opposite sex. Experiment 1 (n = 2639) investigated selection of scholarship applicants and demonstrated that a pro-attractiveness bias held only for selection of opposite-sex scholarship applicants; no such bias was observed for highly attractive same-sex applicants. Experiment 2 (n = 622) investigated evaluations of prospective job candidates and demonstrated again that pro-attractiveness bias was observed only for opposite-sex candidates; participants discriminated against highly attractive same-sex candidates. Moreover, this bias was not observed among highly attractive participants; it held only for moderately attractive participants, those for whom highly attractive same-sex individuals can pose especially potent social threats. Findings suggest that attractiveness biases in organizational decision making are rooted partly in the social threats and opportunities afforded by attractive people.