Upload your resume. PhdFit's six research agents compare your background with faculty profiles, recent publications, lab focus, and outreach opportunities, then rank professors with evidence you can review.
Resume-aware matchingEvidence from recent researchOutreach drafts included
Advisor Match Brief Resume uploaded
Resume signals5 found
NLPHCIFairnessGraph MLPyTorch
Six-agent analysisRunning
Axel · Candidate AnalystResume parsed
Nova · Professor ResearcherRecent papers
Echo · Match ExplainerEvidence linked
Atlas · Strategy AdvisorFit ranked
Lyra · Outreach WriterDraft queued
Juno · Meeting Prep CompilerNext step ready
SC
Dr. Sarah Chen
Stanford · NLP & interpretability
91
fit
Fairness + NLP projects overlap with her recent work.
Strong methods fit: model evaluation and interpretability.
Clear outreach angle for current trustworthy AI research.
Outreach angle
Ask how her lab is extending interpretability methods into fairness audits for real-world AI systems.
Nic Matthews is an Assistant Professor at the School of Communication at The Ohio State University. He holds a Ph.D. in Mass Communication from Indiana University, an M.A. in Telecommunications from Indiana University, and a B.A. in Telecommunications from the University of Georgia. His research investigates the mechanisms and contexts that bias moral judgments, with a focus on understanding how and why people perceive immoral actions as moral or vice versa. Matthews aims to predict moral judgment biases by intersecting communication theory with moral psychology, grounding his work primarily in mass media entertainment and interpersonal communication theories.
The rise of generative artificial intelligence has prompted claims that large language models (LLMs) can substitute for human participants, particularly in moral judgment tasks where correlations between ChatGPT and humans approach r = 1.00. In response, we conducted a pre-registered study where two LLMs (text-davinci-003 and GPT-4o) predicted human moral judgments of 60 scenarios prior to a large human sample (N = 940) rating them. Despite strong correlations, difference scores revealed substantial, systematic errors: Compared to humans, LLMs provided more extreme morality ratings of moral and neutral scenarios and more extreme immorality ratings of immoral ones. Moreover, ChatGPT differed significantly and with moderate to large effect sizes from human averages on ~ 87% of scenarios. Further, LLM ratings clustered around a restricted number of values, failing to reflect human variability. Re-examination of earlier published data also reflected this clumping. We conclude that broader evaluation criteria are needed for comparing LLM predictions and human responses in moral reasoning tasks.
Despite their popularity, online video games possess pervasive toxicity. However, players do not categorically judge toxic behaviors as wrong. Attribution theories are well suited to disambiguate such judgment variance, but debate exists on the usefulness of motivated versus socially regulated blame perspectives. By exploring a new, potentially toxic behavior called “smurfing,” we innovate on methodological barriers that make experimentally disentangling socially regulated and motivated attribution perspectives difficult. In Study 1, we empirically present, describe, and explore smurfing and its perceived effects as a novel cheating behavior in online gaming. In Study 2, we extracted player-generated reasons for smurfing and manipulated the stakes of games to manipulate transgression salience (a key factor of blame attribution) across a moral continuum. By having participants use a mock crowd-sourced judgment platform, we observed the (in)stability of stakes across a continuum of reasons. We subsequently replicated our findings with a novel sample in Study 3.
Journal of Communication · 2024-08-01 · 2 citations
articleOpen access
Abstract Researchers often use attribution theory to understand how people make sense of messages. Unlike the ephemeral actions typically investigated using attribution frameworks, messages can persist. Our study observed how persistence influences the harmfulness of messages and how people levy blame upon harmful posters and those ostensibly obligated and capable of intervening. Grounded in the path model of blame, a randomized experiment (N = 520) tested whether persistence cues in an online environment (low vs high message persistence) influenced appraisals of the harmfulness of online toxicity (i.e., sexism on Reddit), inferences of mental states for relevant agents (e.g., awareness, capacity), attributions of blame, and motives to respond. Results indicated that greater persistence increased perceived harmfulness and motivated individuals to place greater blame upon inactive/ineffective moderators who were capable of intervening.
Communication Monographs · 2023-07-11 · 1 citations
articleOpen access1st authorCorresponding
Moral differences hinder communication and relationship formation. However, perceptions and reactions to moral dissimilarity varies. Accordingly, we explored how moral adaptiveness (a flexible application of morality) relates to the intent to communicate with and befriend morally dissimilar others by focusing on moral relativism (believing that morals are subjective) and moral tolerance (believing that one should not condemn/change dissimilar others). We observed adaptiveness’ relationships with the willingness to communicate (WTC) with morally dissimilar others and greater moral diversity in close friendship networks. Surveys of convenience samples of adults recruited through MTurk (Nstudy1 = 325; Nstudy2 = 1219) demonstrated partial evidence of a positive link between moral adaptiveness and WTC. By comparison, a consistent relationship between adaptiveness and the moral diversity of friendship networks did not emerge.
Moral differences hinder communication and relationship formation. However, perceptions and reactions to moral dissimilarity varies. Accordingly, we explored how moral adaptiveness (a flexible application of morality) relates to the intent to communicate with and befriend morally dissimilar others by focusing on moral relativism (believing that morals are subjective) and moral tolerance (believing that one should not condemn/change dissimilar others). We observed adaptiveness’ relationships with the willingness to communicate (WTC) with morally dissimilar others and greater moral diversity in close friendship networks. Surveys of convenience samples of adults recruited through MTurk (Nstudy1 = 325; Nstudy2 = 1219) demonstrated partial evidence of a positive link between moral adaptiveness and WTC. By comparison, a consistent relationship between adaptiveness and the moral diversity of friendship networks did not emerge.
Media psychologists commonly study how narrative elements (e.g. characters) influence entertainment and perceptions. Research on the sequencing and structure of these elements (i.e. metanarrative; the shape of the story) is less common. In both areas, morality tends to ground theorizing (e.g. disposition theory). To extend knowledge in these domains, we conceptualize and observe the effects of side-taking (i.e. choosing a side during conflict), a core concept in narratives and moral psychology. Dynamic coordination theory explains that side-taking is fundamental to morality because it signals moral judgment/condemnation. In a preregistered experiment (N = 577), we observed how the direction (i.e. siding with/against the protagonists or taking no side) and timing of side-taking (i.e. early, middle, or late in the story) influenced variables at multiple levels of analysis (i.e. micro-to-macro). Although timing did not produce effects, we found robust evidence that the direction of side-taking affected variables at all levels of analysis.
Journal of Media Psychology Theories Methods and Applications · 2022-11-23 · 6 citations
articleSenior author
Abstract: We examined how morality subcultures moderate judgments of a narrative character’s behavior along the moral continuum. Using the moral continuum procedure (MCP) across two studies ( Matthews, 2019 ), we identify the point along the moral continuum where trait moral salience (i.e., one’s sensitivity to different moral foundations) begins to influence moral judgments. Findings indicate that trait moral salience does not influence judgments of less immoral behaviors (i.e., behaviors that fall along the beginning the continuum). However, trait moral salience does impact judgments of more immoral behaviors (i.e., behaviors that fall on the latter half of the continuum), in a pattern consistent with past research. Our data imply that a moral tipping point along the continuum exists, where moral judgments shift from moral consensus (i.e., a general uniformity in moral judgments) to moral disagreement (i.e., divergence in judgement caused by individual differences in trait moral salience). We posit that dynamic coordination theory’s conceptualization of common knowledge helps explain the observed tipping point. Thus, the current project contributes toward extant media theorizing ( Tamborini et al., 2012 ) by specifying how morality subcultures function along a moral continuum.
Many frameworks exist that explain how people interact with avatars. Our core argument is that the primary theoretical mechanisms of a user-avatar bond (i.e., UAB) rest with the way people engage avatars and, thereby, the broader digital environment. To understand and predict such engagement, we identify a person's skill in handling/engaging the avatar in the digital environment as an ordering parameter (i.e., organizing predictor). Accordingly, we define skill as a person's ability to enact their agency successfully to achieve desired states. To explain how skill orders experience, we ground our theorizing in ecological perception and systems theory. In our explication, we describe how stable action coupling (i.e., the linking of action inputs to perceived outcomes) enables a state of embeddedness (i.e., when the environment facilitates and constrains behaviors) in the digital environment. Then, we explain how embeddedness promotes motivational attunement (i.e., orienting of motivational systems) and what the digital environment affords to users at different levels of skill. Throughout, we consider how our theoretical scaffolding generates tractable contentions regarding how skill influences UABs.
Affective disposition theory (ADT) explains that the moral judgments of character behavior inform dispositions toward characters. These dispositions bias moral judgments of characters’ subsequent behaviors and establish behavioral expectations. We used expectancy violations theory to help specify people’s dispositions toward characters. In study 1, we modified the footbridge dilemma to develop experimental stimuli and predictions. Studies 2 and 3 observed the disposition formation process longitudinally and validated our stimulus: a custom-built visual novel. Study 4 tested our predictions. Studies 2 through 4 used pre-registered hypotheses, sampling, and data analyses. Results demonstrated that the current disposition (positive vs. negative) changes how a novel (im)moral behavior affects that disposition. Schema-violating behaviors provoked larger mean differences in participants’ dispositions toward protagonists compared to antagonists. Specifically, people were hyper-scrutinous of moral paragons and entrenched despised characters in moral skepticism. Additionally, we observed differences in dispositions toward characters who did not act when they could (inaction).
Human Communication Research · 2021 · 41 citations
Psychology
Social psychology
Epistemology
Abstract In two pre-registered studies, we leveraged recent advances to disposition theory to examine whether character judgments are relative. We used a Pilot Study to develop a moral continuum of behaviors for a hypothetical television series. We referenced our established moral continuum to create behavioral sequences that represented two characters descending into immorality. We manipulated whether one or both characters were present in the narrative. The simultaneous presence of both characters polarized participants’ moral evaluations of character behavior, categorization of the characters as heroic/villainous, and character liking. Our findings substantiate the systematic effects that character interdependence has on disposition formation. An improved understanding of narrative context can specify when between- and within-character comparisons occur and what effects character interdependence has on disposition theory’s processes. We discuss how narrative schemas, character schemas, and character networks can serve as the elements for explicating the role of narrative context in disposition theory.