While psychotherapy treatment manuals define the broad structure and targets of a therapy, therapists must decide how to implement treatments with a specific patient. Yet, patients are heterogeneous even within a disorder class, and there is little systematic research to guide a therapist to make principled adaptations. We examined the question of whether individual differences moderate the treatment effects of therapist interventions and both in-session and between-session processes of change, using data from a randomized controlled trial for panic disorder comparing panic-focused psychodynamic psychotherapy (PFPP) to cognitive-behavioral therapy. In Chapter 1, adherence to PFPP (n = 65) was observer rated in Sessions 2 and 10 to predict panic change after the rated session. Panic-specific interpretations predicted improvements, while non-panic-focused interventions did not. Concordant with dynamic theory, patients with more interpersonal problems benefitted especially from heightened focus on the interplay between interpersonal-emotional conflict and panic. In Chapter 2, we examined whether higher levels of observer-rated emotional expression—a marker of therapeutic engagement—across early PFPP sessions predicted subsequent panic improvements. We hypothesized that this relationship would be moderated by certain personality disorder traits related to emotionality: (1) borderline traits, which denote heightened, labile, dysregulated emotionality; and (2) obsessive-compulsive traits, related to muted, constrained emotionality. As predicted, borderline traits attenuated the otherwise positive relationship between emotional expression and symptom improvement, but obsessive-compulsive traits had inconsistent relationships between trial sites. Finally, in Chapter 3, we built on mediational analyses for both CBT and PFPP (n = 138), examining whether the presence of different psychological vulnerabilities moderated the symptomatic impact of improvement in two mediators of panic change: catastrophic, body-focused interpretation style, and panic-specific reflective functioning (PSRF). Patients beginning treatment with a more catastrophic style benefitted more from improvements in either mediator. Personality disorder traits blunted the impact of improvements in PSRF, but patients with no personality disorder evidenced high benefit from PSRF improvements. This set of findings suggests that personalization of psychotherapy can be empirically grounded. Clinical characteristics may inform how a therapist chooses to intervene, what in-session processes they focus on, and what types of psychological changes they aim to encourage.