Individual differences in Tolerance for Ambiguity (TFA) are important in predicting how people deal with data subject to multiple interpretations (ambiguity) in contexts including therapy and conflict resolution. This study examines models of humor based on bisociation and meta-pattern recognition, using humor variables to predict TFA, and implications/applications for therapy, second-order change, and future research. Using an unrestricted online sample of 691 adult participants it was found that higher levels of sense of humor were significantly associated with higher levels of TFA (β = .30; p < .001), even after controlling for social desirability. Adding humor styles (HS) to the regression model further increased predictive power. TFA was positively associated with Affiliative and Self-enhancing styles, and not associated with the "potentially detrimental" styles (Aggressive and Self-defeating). The largest correlation with TFA occurred with Self-enhancing HS (r = .33, p < .01). Relationships between HS, social dominance, and TFA were also investigated. Constructivist and Gestalt therapists had higher levels of TFA than those with orientations based on rationalist assumptions. Results of a previous study relating humor to HS were replicated in expected directions. (Full text of this dissertation may be available via the University of Florida Libraries web site. Please check http://www.uflib.ufl.edu/etd.html)