There are no behavioral models for testing anxiety in amphibians, a group of animals widely used for developmental, ecotoxicological, and genetic research. We aimed to validate two common rodent paradigms, the plus maze and the scototaxis test, for use in the aquatic African clawed frog (Xenopus laevis). We predicted: (a) that frogs would prefer the dark, vs. light, portions of the testing arenas (face validity), (b) that this behavior could be altered with acute administration of anxio-selective drugs (construct validity), and (c) that time spent in the dark portions of the arenas would be positively correlated (predictive validity). Prior to testing, frogs were treated with fluoxetine (selective serotonin reuptake inhibitor [SSRI]), desipramine (serotonin- and norepinephrine-reuptake inhibitor), caffeine (methylxanthine, adenosine receptor antagonist, phosphodiesterase inhibitor), saline, or were left unmanipulated. Each drug was administered acutely (1 h prior to testing; caffeine) or subacutely (24, 3, and 1 h prior to testing; fluoxetine, desipramine) at one of three doses. Plus maze and scototaxis testing were separated by 1 week; each frog completed both behavioral tasks and was treated with the same drug regimen prior to testing. Overall, both tests showed face validity, however, data suggest these paradigms lack both construct and predictive validity.