Incorporating relevant numerical information into decision-making is a fundamental and important aspect of numeracy. However, the process through which weight is assigned to particular numerical values is not well understood. The central theory proposed in this dissertation is that the weight assigned to numerical information may be conceptualized as a function of a set of discrete and largely independent cognitive and representational cues, each of which makes a relatively unique contribution to the information weighting process. This study focuses on five cues that may frequently be important for assigning weight to numerical information (though this list is not meant to be exhaustive): a) the ease with which information can be processed, b) the extent to which information deviates from expectations, c) the manner in which information is acquired, d) the precision of numerical representations, and e) the perceived uncertainty associated with a piece of information. For ease of processing, existing research was reviewed and implications for the practice of presenting numbers were considered. Experiments were conducted to investigate the roles of the remaining cues in the information weighting process.The first experiment's results indicated that when confidence in expectations was high, information received more weight when expectations were incorrect. However, contrary to predictions, information received more weight to the extent that it conformed to expectations when confidence was low. The second experiment's results showed that the weight assigned to numerical information was greater when it had been actively searched for before its value was learned (as opposed to being passively received). In the third experiment, the precision of numerical representations (as measured by the number of significant figures) was observed to be proportional to the weight assigned to information, while differences in precision across two numbers in a comparison made that comparison less influential. Finally, in the fourth experiment, the weight assigned to the probability level (as opposed to the interval width) was greater for confidence intervals with lower probabilities and narrower ranges, and also for intervals based on subjective judgments rather than empirical data. Implications were discussed for fields including education, journalism, and risk communication, among others.