Research in consumer financial decision making has received considerable attention in recent years. This thesis contributes to the literature on financial literacy and measuring underlying behavioural characteristics to explain individual choice, as well as the experimental literature on the certainty effect and robustness of experimental results. The thesis is separated into three substantive chapters. The first two substantive chapters use individual level survey data of a representative sample of UK households to investigate the role of financial literacy and behavioural traits in (i) the simultaneous holding of consumer credit and liquid savings, i.e. the `co-holding puzzle', and (ii) mortgage choice. My results show that underlying individual traits are important predictors for consumer choice. Households with self-control issues are significantly more likely to co-hold substantial amounts, consistent with the notion that co-holding is a form of self-control management to limit consumption. My results also show that individuals with low levels of financial literacy and an impulsive present bias for consumption are significantly more likely to hold alternative, non-amortising, mortgage products. This suggests that these mortgages may attract consumers who are less likely to sufficiently understand their features, and who put more weight on present consumption. The third substantive chapter reports and discusses evidence from two experimental studies, motivated by evidence that people may prefer simple and/or certain options disproportionally. The first study investigates the certainty effect using a new laboratory design that goes beyond the pairwise-lottery choices typically used in the literature. The results provide little evidence of a certainty effect in this setting, where subjects can choose from eleven options. In the second study, I attempt to replicate the results of the original experiment that suggests that people are significantly more likely to prefer simple options when faced with larger choice sets. I follow the procedure of the original design, but my results provide no evidence for a disproportionate preference for simplicity. Instead, subjects choose according to their risk attitude.