Can we predict human life? This study presents a machine learning model for predicting six outcomes of children belonging to disadvantaged, or fragile families. It is motivated by recent work developing models with enhanced predictive power as well as previous attempts from researchers to model fragile families. The model is based on data from the Fragile Families & Child and Wellbeing Study, a data set of over 4,000 families from cities across the U.S. that aims to determine how over 12,000 variables about the children and their parents, schools, and overall environments collected over 15 years can impact life outcomes. The outcomes are child grit, GPA, family eviction, job layoff, material hardship and job training. My approach relies on existing data science techniques -- imputation of missing data, elimination of low variance features, prediction via ridge and logistic regression models -- and proposes and evaluates a new model -- a neural network pre-trained on synthetic data created from Gaussian noise. This approach creates a flexible model with reasonable bias for human behavior and enhanced predictive power. Results revealed that the proposed model outperformed benchmark models for four of the six outcomes by up to 28% and outperformed existing researchers’ models for grit and GPA by up to 40%.

Moral inferences are critical for interpreting and responding effectively to others' behavior. Previous research in moral psychology has explored how people make moral judgments and form inferences about others’ behavior. Within this literature, two key dimensions of moral character inference have been identified: situational vs. dispositional attribution and localized vs. generalized morality. More recently, dispositional attribution and generalized morality have been theorized to contribute to essentialism. However, the nature of this contribution remains unclear. Although all of these frameworks have been studied extensively within their respective areas of psychology, their potential connections have not yet been systematically examined within a unified framework. Through an empirical study, we bridge these research areas by examining the relationship among dispositional attribution, generalized morality, and essentialism. Implications are discussed for law, social psychology, and clinical psychology, where moral inferences impact decision-making, interpersonal evaluation, and treatment strategies. Keywords: Moral inference, moral cognition, dispositional attribution, generalized morality, essentialism

Prior research has examined the effects of peer influence on risk-taking behavior, but substantially less published work has investigated the social influences on health safety behavior during the COVID-19 pandemic. The present study examined Princeton students’ pandemic risk behavior through an exploratory survey of their retrospective and prospective experiences with COVID-19. Statistical analyses of participants’ responses revealed that students relied heavily on their observations of peers to determine their pandemic risk behavior, particularly when such observations were incongruous with institutional guidelines surrounding COVID-19 safety. Such reliance on peers appeared to operate below the level of awareness. In a follow-up pilot study, participants were engaged in a deception experiment in which their COVID-19 safety behavior was observed in the presence of different mask-wearing norms. In both the survey and the pilot experiment, participants who were exposed to people wearing masks were significantly more likely to put on a mask than those who were not, regardless of the institutional guideline surrounding mask-wearing. These findings underscore the importance of taking a social psychological approach to disease transmission. Limitations, directions for future research, and implications for public health guidelines are discussed.

Addiction exhibits pronounced heterogeneity across symptoms, substances, neurobiology, contexts, and treatment outcomes, constituting a persistent challenge for unified theoretical explanation. Addressing this challenge, this work examines how such variability can be systematically understood without reducing addiction to a single social, biological, or psychological mechanism. After outlining the scope of heterogeneity in addiction, the analysis turns to computational models of decision‑making—particularly Redish et al.’s vulnerability framework—as a powerful account of how failures within learning and choice systems generate addictive behavior. It is argued, however, that decision‑making models alone are insufficient unless they account for the contextual conditions in which choices are made. In response, an extended vulnerability framework is developed in which addictive behavior emerges when action selection becomes constrained by contextual and choice‑based vulnerabilities over time. This framework is then linked to precision medicine approaches, which naturally operationalize heterogeneity through individualized diagnosis and treatment. The work concludes by identifying limits of the framework and directions for future research, proposing vulnerability as a unifying lens through which addiction’s complexity can be understood.

A key balance must be achieved by learners of any language: In order to create novel utterances, learners must generate an abstract syntactic system representing the grammar of the language, however in order to ensure that all novel utterances are grammatical, this system must also be appropriately limited to agree with the language’s specific syntactic rules and restrictions. Speakers face two main challenges: They must first construct an appropriately abstract grammar on the basis of concrete instances from the input and then constrain this abstract grammar such that it does not include ungrammatical forms. These are challenging tasks because language input does not come pre-marked with abstract syntactic categories and also provides no direct negative evidence regarding what cannot be said. The present study investigates how learners are able to address these challenges by examining the development of a specific seemingly arbitrary linguistic restriction: the a-adjective restriction. Children between the ages of six and sixteen participated in an elicited production experiment using real and nonce a-adjectives and non-a-adjectives in which response structure choice (attributive or relative clause forms) was the main variable of interest. The results showed a variety of age-related patterns which suggest that (i) children become competent with the attributive structure earlier than the relative clause structure, (ii) children acquire the a-adjective restriction fairly late in development through a process of learning the syntactic structure preferences of different word classes in an initially item-based manner, and (iii) young children’s syntactic representations are not initially abstract but rather develop to become increasingly abstract with age. Implications for theories and potential mechanisms of language acquisition are discussed.

College students are a particularly stressed population, often being cited as having the highest prevalence of mental illness, specifically related to stress and anxiety. This stress can lead to unhealthy eating habits, including both undereating and overeating practices. During the COVID-19 pandemic, student stress was further exacerbated. The purpose of this study is to examine the relationship between academic stress and eating behaviors in an undergraduate population. Participants were recruited from a US college and completed surveys on academic stress and eating behaviors. Perceptions of control and stressor immediacy were measured to identify mediating variables. Results indicated that participants experienced eating behavior changes, either increases or decreases in food intake, in response to stress. Perceptions of stress were not seen to have a significant effect on the relationship between stress and eating. This study is the first to look at college student stress and eating behaviors during COVID-19. The results have implications for how student stress is addressed in college and by psychologists.

There is evidence to suggest that first-generation and continuing-generation college students vary in their experiences navigating institutions of higher education. This study quantitatively examines the relationships between generational status, academic pressure, and anxiety. Results showed no significant difference in the overall academic pressure, and in the academic pressure these groups experience from family, friends, professors, romantic partner(s), athletic coach(es), and themselves. The study showed no statistical significance in the difference in means reported across the six scales of test anxiety and in the total scores for GAD-7. There was an insignificant difference in the strength of the correlations between academic pressure and anxiety in first-generation and continuing-generation college students. Future directions include the creation of more validated and culturally informed measures of academic pressure for college students in the United States to assess if the findings replicate themselves. We also recommend a qualitative approach to understanding why specific sources of academic pressure are cited by first-generation and continuing-generation college students and to assessing differences in the factors contributing to their anxieties. This research is important in informing the kind of support students need to thrive in colleges and universities in the United States.

Confidence and performance are widely examined variables known for their association with one another. This paper investigates the effects that level of confidence, manipulated by two academic tasks, has on expected motor skill performance and actual performance. Results of the study partially confirmed the original hypotheses that when self-confidence is controlled in an academic environment, it has causal effects on expected performances, but no significant effect on actual motor skill performances. However, expected performances did significantly correlate with the actual performances on the motor skill task. Further, the study tests this relationship between undergraduate males and females and discovers an insignificant association. There are several limitations outlined in the discussion section, relating to participant parameters and design of the experiment. Future studies should replicate this experiment using alternate population characteristics and varying academic and motor-task themes.