THIS THESIS DEALS with three studies of methodological approaches to the decision problem concerning provision for retirement. The first study is made of the decision problem during the accumulation period, the working years before retirement. Background investigations are made concerning the institutional and investment environment within which the decisions and investment growth occur. Based on these investigations, a probabalistic simulation model of the accumulation period up to retirement is constructed with various elementssome beyond the decision maker's control and others which are decision variables. Experimental evidence with several subjects is discussed. One major conclusion is that within reasonably broad limits decision makers can select appropriate strategies by direct observation of the probability distributions of consequences, i.e., without more complete decision analysis involving explicit assessment of utility and backward induction. Secondly, it is felt that the summary statistics of aggregate wealth at retirement might be too brief since what the decision maker is really interested in is the standard of living he will enjoy during the retirement period, i.e., consumption levels. The second study investigates the institutional, investment and mortality environments during the period beyond the date of retirement in order to develop a simulation model which replicates the essential elements of the retirement period. Again several subjects are exposed to an interactive computer simulation model and evidence is gathered. From the experimental studies it is found that there is considerable educational value in such a model to the subjects. Individuals tend to prefer a specific annuity option; probably spend too little in the early years of retirement; and should probably spend more out of capital in these early years. The model is found to be cumbersome for analytical purposes but quite useful for educational uses. Suggestions are made for improving the model. As part of the second study utilities for wealth at retirement are assessed for several decision makers. Finally, the third, more theoretical study examines the structure of the utility for consumption during the retirement period when the lifetimes of both the decision maker and spouse, as well as investment returns, are uncertain. The theoretical underpinnings of the structure are developed and assessment techniques are investigated. Since the assessment task can be enormous there is a discussion of means by which the task can be simplified.