Individual subjective forecasts of one's own mortality or survival, referred to as mortality expectations, can affect economic decisions, such as consumption and saving. Such measures may be particularly useful when evaluating social programs serving older populations or otherwise analyzing populations with potentially high mortality rates. In an evaluation of a pension program in Yucatan, researchers found that low literacy, language barriers, and cultural biases hindered older adult respondents' ability to answer mortality-expectation questions and experimented with new ways of asking the questions. The initial version of the questions used a ruler with a numeric scale representing a percentage chance of living five more years. Researchers tried direct and indirect versions of the questions, versions with possible answers being contingent on answers to previous questions, and versions incorporating use of visual aids, and combinations of these variations. Through cognitive interviewing, they studied how respondents understood, processed, and responded to the survey items. They then used what they learned to revise the survey measures. The visual aids tested were the original numeric-scale ruler, a sliding ruler, stones, and stick figures. With their survey population, they had the most success using the stick figures for conditional questions and the stones for direct questions.