In a laboratory study designed to investigate variables that affect self-regulatory dysfunctioning, 96 college students volunteered to practice solving mathematics problems similar to those found on graduate school admissions tests. They were assigned to groups in which they either self-recorded inaccurate problem solving (negative self-monitoring, NSM), self-recorded accurate problem solving (positive self-monitoring, PSM), did not self-record but received immediate performance feedback, or did not self-record or receive immediate feedback (control). Groups were matched on ability and received either easy or difficult problems. Ss also completed Byrne's Repression-Sensitization Scale. Predictions derived from the closed loop model of self-regulation were supported in that NSM, relative to other conditions, led to lowered self-evaluations, decreased favorableness of self-consequations (self-reinforcement/punishment), and somewhat increased association of anxiety with performance. NSM also led to decreased accuracy in performance, but in comparison to PSM, it facilitated sustained self-monitoring (self-initiated viewing of a videotape of Ss' own problem-solving activity) when the task was relatively simple. All groups decreased sustained self-monitoring when the task was relatively difficult. Theoretical and clinical implications are discussed, and attention is drawn to the role of affect in self-regulation. (24 ref) (PsycINFO Database Record (c) 2006 APA, all rights reserved).