The effect of a pet on psychological consequences of stress (i.e., state and trait anxiety levels) of college students was examined under three test conditions (i.e., reading aloud, reading quietly, and interacting with a friendly but unknown dog). A repeated-measures analysis of variance with three covariates was used to examine the effect of the treatment on each dependent variable (state and trait anxiety). Reading aloud differed from baseline measure under all treatment conditions (p less than .001). Reading quietly and interacting with the dog were slightly below baseline for variables, with more effect seen by reading quietly than by interacting with the dog. Examination of interactions among variables showed no significant differences. Effects upon state anxiety were significant, while trait anxiety levels remained fairly constant throughout the treatments. Baseline differences in trait anxiety scores indicate a potentially greater benefit for pet owners than nonowners. Selected social network and relationship data related to the role of the pet during anxiety-producing times were also analyzed. While interaction with the pet produced a decrease in anxiety level, pet owners did not report the use of their own pet as a social support (i.e., as confidant) significantly more than did previous owners. Results indicated that interacting with a pet for some individuals does affect both physiological and psychological responses by lowering response levels. However, a parallel effect was also seen by reading quietly. Given the effect of pet interaction upon selected social support indicators of health in well college students, these data suggest the importance of examining this treatment with an "at-risk" group in which it is possible to control for ownership characteristics.