Psychology in Everyday Life: The Psych Files

Why do your good deeds feel like "still me" while your bad ones feel like "a different person back then"? Research from Felipe De Brigard's lab at Duke University, published in the Journal of Experimental Psychology: General and in Memory and Cognition, shows that people judge their own past lies and hurtful acts more harshly the further back they happened, and that remembering a wrong act makes you feel like you have changed, while remembering a right act makes you feel like the same person you have always been. The effect holds after controlling for how long ago the event occurred, and it disappears when people remember other people's behavior, which points to a self-protective bias rather than a general belief about change. In this episode I explain how this relates to the self-serving bias, walk through how the studies were designed, and look at the limitations. Terms and citations are on the website at thepsychfiles.online.

What do a false ending to the Star Spangled Banner, a food delivery website where nothing ever arrives, and a monkey waiting for a squirt of juice have in common? In this episode I sit down at the piano to demonstrate the deceptive cadence, the musical trick of setting up the "home" chord and then stepping sideways instead. I use it to explain a new trend: "dopamine sites," where people order meals that never come and take smoke breaks with cigarettes that don't exist. The popular story says dopamine is the brain's pleasure chemical, released when you get the reward. The research says otherwise. I walk through Wolfram Schultz's work on reward prediction error, Kent Berridge and Terry Robinson's distinction between wanting and liking, and Valorie Salimpoor and Robert Zatorre's study showing that dopamine is released in the seconds before the musical peak that gives you chills. I also cover George Loewenstein's question about how much you'd pay to kiss a movie star (and when), and a study from Carey Morewedge and colleagues showing that imagining eating 30 M&Ms led people to eat fewer real ones. Along the way, Mr. Spock turns out to have been right about wanting and having. References and definitions of key terms are on the website.

What is embodied cognition, and did it survive the replication crisis? This episode of The Psych Files examines the famous embodied cognition and social priming studies — power posing, the hot coffee study, the Macbeth effect, the facial feedback (pen-in-the-teeth) study, elderly priming, and the action-sentence compatibility effect — and the large-scale replication studies that failed to reproduce them. Learn how small samples, publication bias, researcher degrees of freedom, and experimenter expectancy effects produced findings that later evaporated, and how psychology responded with pre-registration, multi-lab replications, and adversarial collaborations like the Many Smiles project. A clear, sourced guide to the replication crisis in psychology for students, teachers, and anyone who watched the power posing TED talk and wondered what happened next.

This episode looks at why comic timing works, using a 2026 study from Yuxi Ma, Yongqian Peng, and colleagues at Peking University, presented at the Cognitive Science Society. The researchers analyzed 828 professional stand-up sets from Chinese comedy competitions, about 86 hours of material, and measured what actually predicted the laughs: the content of each line, or the delivery. Timing won, and not by a little. The length of a comedian's pauses and how much those pauses varied beat every content measure, and slower delivery helped rather than hurt. The reason comes down to how the brain handles expectations: a setup builds a prediction (assimilation), the punchline breaks it and forces a fast mental revision (accommodation), and the pause gives you time to encode the setup so the twist lands harder, with the payoff running partly on dopamine. The most useful finding is that top comedians stretch the pause before a big laugh far more than weaker ones do, which means timing is a trainable skill, not a mysterious gift. The practical takeaway for anyone telling a story, giving a toast, or teaching a class is simple: set your best line up clearly, then stop, and let the silence sit a beat longer than feels comfortable, because that pause is you loading the spring.

Most of us who use AI tools live with a quiet contradiction: the tools are genuinely useful, and using them raises real ethical problems — work that might have gone to a person, the energy and water that data centers consume, and models trained on creators' content that was never paid for. This episode of The Psych Files sits in that uncomfortable middle and asks what psychology can tell us about it, drawing on Festinger's cognitive dissonance, the trivialization research of Simon, Greenberg, and Brehm, Bandura's moral disengagement, and Kohlberg's stages of moral reasoning to show how we quietly talk ourselves into staying comfortable. Along the way Michael turns the lens on himself — as someone whose own YouTube videos have been scraped to train these systems — weighs whether local AI models actually resolve any of the ethical problems, and lands on the idea that the discomfort may be worth keeping rather than rushing to explain away.