The effects of confirmation bias on juror decision-making have long been a topic of investigation, as this line of questioning allows for academic study of human decision-making while also speaking to an important social reality. However, very few studies have extended this work into the field of memory. Across two experiments, this dissertation addressed this gap in the field by assessing memory retention in a juror decision-making task where participants were subject to an experimentally-manipulated, trial-specific initial bias towards the defendant. The second experiment also addressed the novel area of confirmation bias mitigation by implementing a perceptual disfluency manipulation—a manipulation which has been hypothesized to operate by shifting individuals from a heuristic mode of processing into an analytical mode of processing. The decision-making results showed a dissociation where Juror Bias Scale score (a measure of pre-existing personal biases towards the criminal justice system) predicted verdicts while the experimentally-manipulated initial bias predicted suggested length of sentence—both in a direction that supported an effect of confirmation bias on decision-making. Additionally, results failed to replicate previous findings where disfluency mitigated the effect of confirmation bias on decision-making. Results on memory showed that information that supported a guilty verdict and information that supported a not guilty verdict were both remembered better overall and had steeper slopes of forgetting than neutral information. Overall, this study emphasized the need to consider factors beyond trial-specific bias (such as the pre-trial bias captured by the Juror Bias Scale) when predicting juror decision-making and found that memory outcomes in this type of task are better predicted in light of theories such as the goal relevance hypothesis of memory facilitation (rather than through the posited effects of confirmation bias).