In eliciting health state valuations, two widely used methods are the standard gamble (SG) and the time trade off (TTO). Both methods make assumptions about individual preferences that are too restrictive to allow them to act as perfect proxies for utility. Therefore, a choice between them might instead be made on empirical grounds. This paper reports on a study which compared a "props" (using specifically-designed boards) and a "no props" (using self-completion booklets) variant of each method. The results suggested that both non props variants might be susceptable to framing effects and that TTP props outperformed SG props.